Submissions on RBNZ Capital Review Paper 4: Optimising Bank Capital Ratios

G Slappendel I'm in favour of increasing the amount of capital banks need to hold. OIA s9(2)(a)

Gareth Vaughan Please see attachment. OIA s9(2)(a)

As a financial journalist I have spent a considerable chunk of my working life over the past decade looking into the banking sector from the outside. It is in this capacity that I make this submission. I support a simple bank regulatory capital regime that helps create a competitive landscape for consumers and strives to minimise risks for taxpayers and bank depositors. The requirements of a competitive market operating for the benefit of consumers means all competitors should be on a level regulatory playing field. Since 2008 this has not been the case in banking with ANZ, ASB, BNZ and Westpac able to hold less capital than their competitors. These four banks have 88% of both banking assets and liabilities, meaning NZ consumers of banking services face an oligopoly. A fundamental building block of giving other banks a better chance to compete, for the benefit of consumers, is by having a level capital playing field. Holding capital costs banks. Recent analysis by interest.co.nz highlighted the competitive advantage available to banks using the Internal Ratings Based approach. This showed ANZ with a risk weight of 19% on residential mortgages. This meant a minimum capital requirement of $1.261 billion against home loans of $76.168 billion. In contrast Kiwibank, using the standardised approach, has risk weights on residential mortgages starting from 35% and a minimum capital requirement of $520 million against $16.956 billion of home loans. It must be remembered that until 2008 all banks had residential mortgage risk weights of 50%. The chart below, from the Reserve Bank’s Dashboard covering the December quarter, shows significantly higher return on equity at the four banks using the Internal Ratings Based approach.

I support the Reserve Bank’s focus on so-called Tier 1 capital such as ordinary shares and retained earnings as part of a desire to see financial engineering opportunities minimised. However any new regime needs to enable small customer-owned banks to raise the required capital. The potential for banks to game the system, and the time and resource required by Reserve Bank staff to police any such behaviour, should be minimised. The Global Financial Crisis, and its CDOs or collateralised debt obligations, taught us the dangers of financial engineering. Use of the Internal Ratings Based approach to calculate regulatory capital requirements by ANZ, ASB, BNZ and Westpac appears to have been challenging for both regulated and regulator. In 2017 Westpac was forced, by the Reserve Bank, to significantly increase its capital after a series of errors in its capital modelling dating back as far as nine years were revealed. Last year BNZ restated historical regulatory capital adequacy ratios after disclosing

a data error dating back to 2003. And ASB was in breach of its banking registration for nine years, albeit in its case capital ratios were higher than they should have been. Then there's ANZ. On deadline day for this submission the Reserve Bank has censured the country's biggest bank and revoked ANZ's accreditation to model its own capital requirements due to "persistent failure" in its controls and director attestation. ANZ is "disappointed" this error occurred. This revelation calls into question whether the Internal Ratings Based approach is fit for purpose. Thus I support the proposal to increase Internal Ratings Based banks' requirements to 90% of standardised bank requirements, from about 76%, and to have them also calculate their requirements using the standardised approach. I won't comment on the specific levels of capital proposed by the Reserve Bank, other than to say I support increases from current regulatory minimum ratios. New Zealand's major banks are certainly not overburdened by their current capital requirements. In their report on the ANZ Banking Group’s recent interim financial results, analysts from investment banking firm UBS put ANZ NZ’s return on regulatory capital at 28.4% for the six months to March 31. Banks' minimum capital requirements are meant to cover all of credit risk, market risk and operational risk. Anti-money laundering, cyber-crime and conduct risk are areas requiring increasing compliance focus by banks and oversight from regulators. Having simple and strong regulatory capital requirements in place should allow more time and resource, from both the regulator and regulated, to be focused on policies that can effectively combat money laundering, cyber-crime and conduct risk. Banks are highly leveraged, as demonstrated by interest.co.nz’s bank leverage page. This shows the number of times the total assets of a bank are larger than the shareholder's investment in the bank. History shows us that banks can lose money and they can fail. And when they do the damage caused can be widespread and costly. Banks get the vast bulk of their money by borrowing it. However when they lend they typically demand borrowers have skin in the game. For someone buying a house this comes in the form of a deposit. In most cases banks would be reluctant to lend to businesses or individuals as highly leveraged as they are themselves. Thus bank owners having more skin in the game means their decisions risk their own money, and hopefully lessens the chances of a taxpayer funded bailout and losses for depositors, which must be a good thing. Gareth Vaughan, OIA s9(2)(a)

Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Level 26 101 Miller Street North Sydney NSW 2060 Australia Tel 1300655 422 Fax 1300 662 228 Genworth.com.au 17 May 2019 Ian Woolford Manager, Financial Policy Prudential Supervision Department Reserve Bank of New Zealand PO Box 2498 Wellington 6140 New Zealand Dear Mr Woolford, RBNZ Review of the Capital Adequacy Framework for locally incorporated financial institutions: Capital Review Paper 4: ‘How much capital is enough?’ Thank you for providing Genworth Financial Mortgage Insurance Pty Limited (Genworth) with the opportunity to participate in your consultation regarding locally incorporated financial institution capital in New Zealand and the question of how much capital is enough. Genworth has focussed on several key themes in our response to the RBNZ Capital Review Paper 4, each of which build on our 19 March 2018 submission. In that submission, we made recommendations to ensure that asset risk weightings appropriately reflected the risk retained on financial institutions’ balance sheets and the benefits of sharing risk more broadly across the local and global financial systems. This continues to be the approach from which we have addressed the questions raised in the latest Consultation Paper. We look forward to further engagement with the RBNZ as you work your way through these issues. Please do not hesitate to contact Alexander Drake, Head of Government and Industry Affairs on alexander.drake@genworth.com with any questions relating to our submission. Yours sincerely, Bradley Dean Head of Strategy and Innovation Genworth Mortgage Insurance Australia Limited

Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Contents

1. Executive Summary ....................................................................................................................................2
2. Introduction..................................................................................................................................................3 Structural risks to financial stability in New Zealand ............................................................................3 LMI in New Zealand..............................................................................................................................4
3. Benefits of LMI.............................................................................................................................................5 Consumer benefits of LMI.....................................................................................................................5 Borrowers in the HLVR category ..........................................................................................................6 Mitigating residential mortgage risks ....................................................................................................7
4. Response to consultation questions ........................................................................................................8
5. References .................................................................................................................................................18 Genworth, Genworth Financial and the Genworth logo are registered service marks of Genworth Financial, Inc and used pursuant to licence © 2019 Genworth Mortgage Insurance Australia Limited.

2 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review

1. Executive Summary New Zealand’s financial system is dominated by four financial institutions that rely on foreign capital to support lending. Sources of capital are therefore concentrated, creating risk within the financial system. Providing efficient alternatives to diversify sources of capital, such as lenders mortgage insurance (LMI), enhances the stability of New Zealand’s major financial institutions and the overall financial system. LMI supports the financial system by providing system capital and risk management to lenders through the economic cycle. Not only does LMI protect lenders from significant loss during an economic downturn, it also provides lenders with comfort to continue lending through downturns, potentially reducing the severity of these events. Additionally, the oversight provided by LMI insurers supports prudent lending policies and underwriting practices through the cycle. While LMI does not directly protect the borrower, it assists borrowers to enter the housing market by enabling financial institutions to provide loans where the depositor is unable to generate a sufficiently large deposit and can put downward pressure on the interest rates paid by borrowers. Genworth broadly supports the proposals in the RBNZ Capital Review Paper 4, ‘How much capital is enough?’. Introducing a Tier 1 capital level (requirement and buffer) of 16% will reduce the probability of a banking failure and systemic crisis and, therefore, strengthens the New Zealand financial system. However, it is important to consider that stability in the financial system is supported both by capital adequacy and access to various sources of capital. For this reason, Genworth suggests closer alignment with Basel III’s final set of reforms by removing the IRB scalar and relying on the clear guidance inherent in the Standardised approach, which we believe provides a greater opportunity to introduce diversified capital into the system. RBNZ’s goal of maintaining a 90% relativity of risk weighted asset (RWA) estimates determined by internal models compared to the Standardised approach could still be achieved by revising the risk weight floor. Within the mortgage portfolio, Genworth recommends that the floor be levied on a per-loan exposure basis, not at the aggregate portfolio level. This is consistent with the granular nature of the Standardised risk schedule and is the most effective means of achieving a level playing field. From an insurance perspective, a per-loan application of the floor helps address adverse selection where the bank may be incentivised to insure the highest risk loans and keep the lower risk loans on their balance sheet. Genworth appreciates the opportunity to participate in your consultation, and we look forward to further engagement with the RBNZ as you work your way through these issues.

3 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review 2. Introduction Strong banking prudential regulation globally and within a country aims to maintain the longer-run stability of the financial system, which can serve the financial needs of individuals, households and businesses through both buoyant and difficult periods while encouraging an appropriate level of competition among those offering financial services. There is a cost to financial safety and it is the task of a prudential regulator to balance the community’s desire for financial stability, its willingness to pay for that stability and the risks associated with certain levels of financial safety, while maintaining access to credit. Regulators achieve this by encouraging prudent risk management by financial institutions whose failure could precipitate a financial crisis. The key mechanism available to achieve this balance is the setting of minimum capital standards for authorised financial institutions. Regulation of capital is required because the community’s risk tolerance is likely to be different to that of bank shareholders, boards and executives. This mismatch is more acute where the banking system is dominated by large systemically important financial institutions that are perceived to have implicit taxpayer subsidies from government. Capital is not a well understood concept in the general population. Many people think of it in a liquidity sense, that of actual deposits held somewhere that can be used to repay depositors or other investors in the event of a crisis. The RBNZ Capital Review Paper 4:’ How much capital is enough?’ makes a valuable contribution to clarifying this issue through its non-technical summary of the regulatory capital proposals and emphasising that capital is best thought of as money the owners of a bank put in to operate the bank. This money (rather than depositor funds) is the first to be lost when bank assets fall in value. The more money owners have invested in a financial institution, the more sensitive they will be to the risks undertaken by that financial institution. Higher capital puts more ‘skin in the game’ for financial institution owners. Structural risks to financial stability in New Zealand Despite New Zealand’s advanced economy and high standard of living, it has characteristics that make it relatively more vulnerable to financial instability. Four factors are:

1. New Zealand is a small country with an economy open to trade and capital flows;
2. All the systemically important financial institutions are foreign owned;
3. The banking system is relatively concentrated, with the New Zealand real economy, therefore, highly dependent on the stability of the largest financial institutions; and
4. The balance sheets of the largest financial institutions are concentrated in particularly sectors of the economy. The New Zealand economy is largely dependent on foreign investment to support growth. From a financial system perspective, the four largest Australian-owned financial institutions dominate the residential lending market in New Zealand, and these lenders rely on capital provided by their Australian parents to support that lending. Sources of capital are therefore concentrated, creating risk within the financial system. Providing efficient alternatives to diversify sources of capital, such as LMI enhances the stability of New Zealand’s major financial institutions and the overall financial system. These four factors pose challenges for prudential regulators. The Reserve Bank of New Zealand (RBNZ) has, to date, managed these risks well and has been ahead of many other global regulators. For example, the RBNZ moved in advance of the Basel Committee on Banking Supervision (BCBS) Basel III proposals to increase the risk sensitivity of the risk weighting system in New Zealand, both for the internal ratings based (IRB) financial institutions and the financial institutions that use the Standardised approach.

4 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Further, the RBNZ has been a leading regulator in terms of facilitating strong disclosure obligations and empowering markets to discipline bank activities. A recent example of this is the adoption of an award-winning RBNZ data dashboard. However, material challenges remain. One area where New Zealand financial institutions are significantly exposed is the residential mortgage market. Approximately 48% of financial institutions’ balance sheets represent loans backed by residential property. During the Global Financial Crisis (GFC), we observed the impact of excess exposure to residential property in different global economies. New Zealand (and Australia) came through that crisis relatively unscathed. Nevertheless, the lessons learned were the importance of capital adequacy, of reducing bank balance sheet leverage and of maintaining robust underwriting standards around residential lending. At the core is the need for better risk management and capital settings. Genworth believes capital requirements should accurately reflect the residual risk to the financial institution’s balance sheet. The existing RBNZ capital adequacy framework contains significant granularity in both IRB and Standardised approaches and is largely aligned with the recent reforms published by the BCBS in December 2017 and Consultation Papers published by national regulators. LMI in New Zealand LMI supports the financial system by providing system capital and risk management solutions to lenders through the economic cycle. Not only does LMI protect lenders from significant loss during an economic downturn, it also provides lenders with comfort to continue high loan to valuation ratio (HLVR) lending through downturns, potentially reducing the severity of these events. Additionally, the oversight provided by LMI insurers supports prudent lending policies and underwriting through the cycle. In the 1990s and early 2000s there was a competitive market for LMI in New Zealand. Financial institutions shared risks associated with higher loan to valuation mortgages with other parties. In this scenario, both the bank and the LMI provider held capital to support housing risk in the market. Shortly after the Basel II framework was implemented in New Zealand in 2008, major financial institutions received IRB accreditation and lost the appropriate capital recognition of the risk mitigation benefits of the LMI cover. This, combined with the inclusion of the LGD floor, resulted in the cost of LMI to the bank becoming uneconomic. Today, the LMI market in New Zealand effectively no longer exists. Both Genworth and QBE withdrew from the market due to the lack of scale relative to costs. New Zealand financial institutions manage higher risk loans by applying a “low equity” fee to borrowers with a higher loan to valuation ratio. In a benign environment, the additional costs of higher defaults may be funded through this fee; in a stressed environment the bank may incur significant loss as the associated risk is retained on the bank’s balance sheet. The result is that the New Zealand market has lost the risk management and capital diversification benefits of LMI. Under our main product, Genworth brings capital to any HLVR loan equivalent to a 22% risk weight, on average. Put another way, our LMI cover contributes approximately A$2 million in capital for every A$100 million of high LVR loans covered by LMI1 . The LMI market has changed significantly in recent years. The industry has recognised that different financial institutions, financial systems and security types may require alternative risk management solutions. In recent years in Australia, Genworth has worked with financial institutions and other lenders to provide a broader range of risk sharing solutions that support loans that lenders have not traditionally insured (such as 70-80% LVR loans). These have included, for example, providing mortgage insurance to “micro markets” – smaller cohorts of loans with homogenous risk characteristics and providing catastrophic cover for financial institutions looking to manage the potential impact of a significant downturn on these active loan portfolios. To support financial institutions in managing their counterparty risk to Genworth, the risks associated with these solutions may be shared with global reinsurers. This expanded product range has provided financial institutions more options to manage their risk profile as their risk appetite has shifted or economic conditions have changed. In this submission, Genworth makes brief comments on questions posed in the RBNZ Capital Review Paper 4: “How much capital is enough?” (Consultation Paper).

1 $100 million x 22%

5 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review 3. Benefits of LMI LMI can provide considerable value to New Zealand borrowers, financial institutions and society in general. LMI is an insurance product that protects a lender (e.g. bank) against losses associated with a borrower failing to make contractual payments on a home loan and subsequently going into default. As an insurance product, the price of LMI depends on risk factors. These include loan to value ratio (LVR), price of the house, location of the house, employment status of the borrower, and commercial arrangements between the LMI provider and the financial institution. While the financial institution is the party being insured against losses, the financial institution typically passes on the cost of LMI to borrowers, along with other origination costs. This can be in the form of a direct fee, or it may be added to the loan amount and repaid over time. Consumer benefits of LMI While LMI does not directly protect the borrower from events such as losing income through retrenchment, death or illness, the availability of the product means consumers (borrowers) benefit in several ways. First, it assists borrowers to enter the housing market by facilitating financial institutions to provide loans where the depositor is unable to generate a sufficiently large deposit. This enables borrowers, particularly first home buyers, to own a home earlier than they would otherwise be able. Second, the availability of LMI can put downward pressure on the interest rates paid by borrowers. The existence of LMI means that any financial institution, regardless of risk tolerance or balance sheet structure, can offer higher risk loans (e.g. HLVR or non-standard loans) and potentially use LMI to manage risk exposure. This increases the number of lenders participating at the riskier end of the market, improves competition, and expands choice. Third, LMI supports stability in the wider New Zealand financial system by shifting risk into international markets. Borrowers benefit from this because the diversification protects the New Zealand financial system and increases the probability that mortgage credit will be available both in times of stability and instability. It helps protect against a process of deleveraging which typically follows a banking crisis.

6 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Borrowers in the HLVR category Lending against residential property is a core exposure of the New Zealand financial system; approximately 48% of New Zealand financial institutions’ balance sheet exposure is to loans backed by residential property. Most financial institutions have a very large appetite to lend to borrowers where the loan is small relative to the value of the underlying property. Conversely, there are a relatively large number of consumers – individuals and families – with an appetite to borrow a larger proportion of the value of the underlying property. These HLVR loans expose the bank, and ultimately the financial system, to higher risk. There are benefits to both consumers and to the broader economy in ensuring that the demand for both low and high LVR lending can be met. Home ownership is important to individuals and families. For most New Zealanders, the family home represents their largest financial asset. Many will need a HLVR loan from a lender to secure a mortgage as an entry point into the housing market or to move to a bigger home as their family grows. Consumers should have access to such loans at a fair and reasonable cost. From the perspective of financial institutions, meeting HLVR demand provides business opportunities, but only if they can a) offset the particular portion above their risk appetite and b) effectively allocate against regulatory capital. Financial institutions should be able to pass on the additional costs of these higher risk loans to consumers.

7 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Mitigating residential mortgage risks Whilst HLVR loans are riskier there are different market mechanisms which allow a bank to offset additional risks associated with these loans while allowing the bank to maintain prudent lending standards, sound capital ratios and robust risk controls. LMI is clearly one of these. In other markets, solutions which fill this gap include government guarantees, securitisation of loans and structured reinsurance arrangements. These mechanisms have the benefit of sharing the risk across the financial system. Globally, accelerated growth in lending/housing markets and concentration of risk on financial institutions’ balance sheets have led to macro-prudential intervention. Regulators have looked to protect and stabilise their markets by tightening underwriting standards around serviceability and verification or placing concentration limits on HLVR loans, investment lending and interest-only segments, etc. However, when risks are shared across multiple market participants, each party holds the others accountable. Underwriting standards are monitored. Policies and procedures are verified, and system risks are reduced. Better mechanisms to share risk, along with recognition of the risk mitigation of these mechanisms through the capital standards applied to financial institutions, will help maintain and promote financial stability and mitigate the need for macro-prudential measures. Currently, most New Zealand financial institutions absorb the risk of HLVR loans on their own balance sheet. Prior to the introduction of Basel II, financial institutions in New Zealand used LMI but have more recently chosen to retain this risk and pass the cost through to the consumer via an annual “low equity” fee. Securitisation in New Zealand has been very limited and confined to non-regulated lenders and lower loan to valuation portfolios. As such, most residential property mortgage risk is being retained by New Zealand financial institutions and the domestic financial system. To help New Zealand financial institutions manage these risks, in addition to traditional LMI, Genworth provides a broader range of risk sharing solutions that target loans which lenders have not traditionally insured (such as 70-80% LVR loans).

11 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Table 2 Funding a $100 loan –Impact on funding costs from CET 1 increase & decrease in wholesale funding costs 8% CET 1 16% CET 1 Funding source Amount ($) Price Cost Amount ($) Price Cost CET 1 $8 12% $0.96 $16 12% $1.92 Deposits $65 1% $0.65 $65 1% $0.65 Wholesale debt $27 4% $1.08 $19 4% $0.76 TOTAL $2.69 $3.33 Difference 64 basis points Savings due to lower debt costs (Modigliani-Miller offset) -32 basis points Estimated cost impact of CET 1 increase to 16% 32 basis points However, while the funding cost increase will be marginal, it may transpire that shareholders of the affected financial institutions seek to at least partly restore the return-on-equity (ROE) ratio that existed before the capital increase. The most obvious means of doing this would be to increase the net interest margin. In practice, even if financial institutions attempted to increase earnings to restore the ROE, the extent to which this would impact on housing and business borrowers is unclear given the role of monetary policy, especially over the five-year implementation period. For example, in 2015 the four major Australian financial institutions raised substantial amounts of tier 1 capital to meet APRA’s adoption of a 25% risk weight floor recommended by the Murray Financial System Inquiry. In response, the financial institutions increased their lending rates marginally as can be seen in Figure 1. However, interest rates have subsequently reduced, due in part to a 50 basis points reduction in the overnight cash rate.

17 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review Genworth proposal Genworth recommends the RBNZ eliminate the scalar with respect to the residential mortgage portfolio and rely solely on the risk weight floor of 85-90% to achieve the policy goal. The scalar may still have some risk differentiation use in the non-residential loan asset categories. Within the mortgage portfolio, Genworth recommends the 85-90% floor be levied on a per-loan exposure basis, not at the aggregate portfolio level. This is justified by the granular nature of the Standardised risk schedule and is the most effective means of achieving a level playing field. From an insurance perspective, a per-loan application of the floor helps address adverse selection where the bank may be incentivised to ensure the higher risk loans and keep the lower risk loans on their balance sheet. One argument against the per-loan application of the floor is that it will materially raise the risk weight on the least risky residential mortgage loans. However, this is where there is the greatest gap between the IRB and Standardised approaches, and the area of most concern for competitive neutrality. Evidence from the United Kingdom (UK) Evidence for this has been born out in UK research. Researchers found that the gap between IRB and Standardised resulted in IRB financial institutions specialising in the safest loan segments: Well before Basel II was implemented, economists predicted that the IRB-SA gap in risk weights would lead firms to specialise. It’s no secret that firms prefer to economise on capital, so if IRB firms get a capital discount on their low-LTV lending, it makes sense that they would prioritise this market segment. Now, eight years after the policy’s introduction, we’ve been able to take the theory to the data and check that this prediction has been borne out. (Bennetton, Eckley, Garbarino, Kirwin, & Latsi, 2017) The researchers concluded that the difference in risk weight methodologies in UK mortgages has implications for the balance of loan type and pricing: These results imply that differences in risk weight methodologies between banks can affect the balance and pricing of mortgage types offered by those banks. Financial stability is also affected, because the distribution of risk across lenders can affect the resilience of the financial sector. Basel II led to smaller standardised banks taking on a higher share of high-LTV lending… …How could policymakers respond? Two options are to revise the standardised approach to make it more risk sensitive, or put a floor under the risk weights available under the IRB approach. From an economic efficiency perspective, it is better to have large, more diversified institutions funding higher risk assets. It is a counterproductive situation whereby the largest financial institutions are incentivised to originate the safest loans (low LVR) residential mortgages. A per-loan floor would help address this issue.

18 | Genworth Financial Mortgage Insurance Pty Limited ACN 106974305/ NZBN:9429035477037 GENWORTH SUBMISSION TO RBNZ | Capital Review 5. References Bennetton, M., Eckley, P., Garbarino, N., Kirwin, L., & Latsi, G. (2017, March 13). Unintended consequences: specialising in risky mortgages under Basel II. Bank Underground. Retrieved from https://bankunderground.co.uk/2017/03/13/unintended-consequences-specialising-in-risky￾mortgages-under-basel-ii/ RBNZ Financial Policy Team. (19 June 2018). Capital Review - Presentation. Reserve Bank of New Zealand. Retrieved from https://www.rbnz.govt.nz/-/media/ReserveBank/Files/regulation-and￾supervision/banks/capital-review/Capital-review-presentation.pdf?la=en&revision=43f65f6f-7064- 4394-ace5-c5f8d18cc1bd Reserve Bank of New Zealand. (2018, December). How much capital is enough? Capital Review Paper 4. Reserve Bank of New Zealand. (2019, April). An outline of the analysis supporting the risk appetite framework. Capital Review Background Paper. Retrieved from https://www.rbnz.govt.nz/- /media/ReserveBank/Files/Publications/Policy-development/Banks/Review-capital-adequacy￾framework-for-registered-banks/Capital-Review-An-outline-of-the-analysis-supporting-the-risk￾appetite-framework.pdf?revision=058df82e-5fc8-4e4

From: Geof Mortlock To: Capital Review Subject: Submission on Reserve Bank bank capital proposals Date: Monday, 4 February 2019 6:27:23 PM Attachments: Article - Capital proposals for banks - February 2019.docx RBNZ, Please find attached my submission on the bank capital proposals. In short, the RBNZ's paper falls well short of the quality that could reasonably be expected of the central bank. The RBNZ has failed to provide adequate justification for its proposals and needs to fundamentally reassess its approach. Regards G J Mortlock Mortlock Consultants Limited OIA s9(2)(a)

Reserve Bank’s bank capital proposals Tane strikes again. From reluctant regulator to rogue regulator. Geof Mortlock1 In January this year, the Reserve Bank released a consultation paper (Capital Review Paper 4: How much capital is enough?), in which it has proposed a major increase in the capital requirements for banks in New Zealand. The proposed increase in capital ratios for banks puts the Reserve Bank’s proposals well out of line with international norms. While there is certainly a need to consider the question of what the appropriate minimum capital ratio requirement should be for banks, the Reserve Bank’s consultation paper falls well short of the quality of analysis that is needed to answer this question. Despite its length (59 pages), the paper contains little substantive argumentation to support the very high capital ratios proposed. It also contains very little cost/benefit analysis. For a central bank that is portrayed by its CEO as a ‘god’ of the ‘financial forest’ (tane mahuta), the quality of analysis is simply not good enough – and certainly falls well short of the mark expected of such a deity. What has the Reserve Bank proposed? Before assessing the Reserve Bank’s capital proposals for banks, let’s recap what they have proposed. In brief, the Reserve Bank proposes to: • set a tier 1 capital requirement (consisting of a minimum requirement of 6 percent and prudential capital buffer of 9-10 percent) equal to 16 percent of risk-weighted assets (RWA) for banks deemed systemically important (the so-called ‘D-SIBs’), and 15 percent for all other banks; and • limit the extent to which capital requirements differ between the Internal Ratings Based (IRB) approach and the Standardised approach, by re-calibrating the IRB approach and applying a floor linked to the Standardised outcomes, raising the average RWA for the four IRB-accredited banks to approximately 90 percent of what would be calculated under the Standardised approach. The Reserve Bank’s proposal represents a very substantial increase in bank capital requirements relative to current requirements. Currently, banks are required to have a total capital ratio of at least 10.5% of RWA (of which 6% must be in the form of tier 1 capital (essentially issued shares plus retained earnings), 2% in tier 2 capital (such as perpetual or term subordinated debt) and 2.5% as a capital conservation buffer. Under the Reserve Bank’s proposals, the large banks (D-SIBs) will have to meet a prudential capital buffer of 10% of RWA in the form of tier 1 capital, lifting their minimum common equity capital ratio to 14.5%

1 Geof Mortlock, of Mortlock Consultants Limited (www.mortlock.co.nz), is an international financial consultant who undertakes extensive assignments for the International Monetary Fund, World Bank and other organisations globally, dealing with a wide range of financial sector policy issues. He formerly worked at senior levels in the Australian Prudential Regulation Authority and Reserve Bank of New Zealand.

2 and total tier 1 capital ratio to at least 16% of RWA. Smaller banks will have their tier 1 capital ratios hiked to at least 15%. How does this compare internationally? These are extremely high tier 1 capital ratios by international standards. The international norm is for tier 1 capital to be no more than around half of these figures, with some countries then imposing additional capital requirements in the form of additional loss-absorbing capital (often in the form of fixed term or perpetual debt instruments that convert to equity upon a defined non-viability trigger). The Reserve Bank’s proposed capital ratios are considerably higher than those recently proposed by the Australian Prudential Regulation Authority (APRA), which would raise the common equity tier 1 ratio for the four major banks in Australia to a minimum of 10.5% of RWA. APRA has proposed no increase in capital for the smaller banks in Australia unless resolution planning reveals a need to do so. The Reserve Bank’s proposal to limit the gap between average risk weights under the IRB framework (which applies currently to the four major banks) and the average risk weight under the standardised framework (which applies to all other banks) is arbitrary and is also penal by reference to some international norms. It will increase the average RWA calculation of IRB banks to 90% of the outcome under the standardised approach, which is considerably higher than the 72.5% RWA floor finalised by the Basel Committee (the international standard-setter for banking supervision) in December 2017. This will have the effect of further increasing the capital burden on the four IRB accredited banks. Is the capital increase justified? The main justification given by the Reserve Bank for increasing the capital ratio requirements for banks is to minimise the damage to the economy and financial system that could result from a bank failure by reducing the probability of bank failure. They have indicated a desire to achieve a probability of bank failure equivalent to around 1 in 200 years. It is certainly appropriate to be seeking to ensure that banks maintain capital ratios that enable them to survive severe economic and financial shocks, such that bank failures are rare events. That is especially important for the systemically important banks (the D-SIBs), given that their failure could have a major impact on the financial system and economy, and would doubtless involve a substantial amount of resolution funding from the taxpayer if the Australian parent banks were not able or willing to provide the capital funding needed. However, what is missing in the Reserve Bank’s analysis is any in-depth assessment of the magnitude of shock needed to cause the major banks to fail. The Reserve Bank’s own stress tests indicate that banks come nowhere near to the point of failure under very severe stress scenarios. That being the case, why is a major increase in capital needed? If they think that the stress tests they have applied are not severe enough, then why do they not recalibrate the tests to even more severe impacts (e.g. through higher and longer contractions in GDP, steeper falls in asset prices and higher and more sustained increases in the rate of unemployment)? Indeed, one would think that, in order to develop capital proposals of the nature contemplated, the Reserve Bank would have undertaken ‘reverse stress tests’. These are used by some foreign supervisory authorities to assess the level of economic and financial shock needed to generate a bank failure. They provide helpful information in assessing how large the economic and financial shock would need to be to cause one or more of the big banks

3 to fail. It would then be possible to assess the probability of such a magnitude of shock actually occurring. Yet there is no substantive analysis of this in the Reserve Bank’s discussion paper. They appear to have done no reverse stress testing at all. Nor, it seems, has the Reserve Bank asked the banks to undertake reverse stress tests. All the Reserve Bank do in their paper is draw on international literature that provides high￾level (and frankly quite challengeable) data on the level of probability of bank failure associated with given levels of bank capital ratios. That is interesting information. But it tells us little about the magnitude of shock needed in the New Zealand economy to generate bank failures at current levels of capital or the probability of such a severe magnitude of shock occurring. Moreover, because the data they refer to covers a span of time that includes pre￾GFC periods, it does not adequately take into account the strengthening of bank risk management frameworks and governance arrangements that have been part of the post-GFC regulatory environment. These developments would arguably reduce the probability of bank failure for any given capital ratio. Another deficiency in the Reserve Bank’s analysis is their failure to adequately differentiate between the systemic and economic consequences of a large bank failure from a small bank failure. The failure of a D-SIB (e.g. any of the four largest banks in New Zealand) would have a much greater impact on the economy and financial system (depending on how it is resolved) than would the failure of a small or medium-sized bank. Yet the Reserve Bank largely ignores this in its capital proposals. It is suggesting that small banks should hold almost as much capital as the D-SIBs, despite the reality that their failure would cause barely a ripple to the financial system (in terms of contagion risk) or economy (in terms of impact on credit channels and adverse wealth effects). In contrast, most other supervisory authorities draw a major distinction between the impacts of D-SIB failure and the impact of small bank failure, and they logically impose considerably lower capital ratio requirements on small banks than on large ones. A further failing in the Reserve Bank’s analysis is that they take no account at all of the means by which bank failure resolution planning can reduce the economic and financial impact of bank failure and reduce the amount of taxpayer funding that might be needed as part of the resolution process. Other countries – including Australia and the United Kingdom – have explicitly taken this into account. They have concluded that, with effective bank resolution planning (something the Reserve Bank has largely ignored to date, aside from its ill-conceived OBR policy), small to medium-sized banks that fail can be resolved without the need for high levels of capitalisation. This can be achieved through various techniques, such as ‘purchase and assumption’, merger and bridge bank solutions. That is why most jurisdictions do not impose capital ratios at as high a level on small banks as they do on the systemic banks. Yet the Reserve Bank has completely ignored this reality. It is largely silent on resolution options and the linkage to capital requirement calibration. The Reserve Bank also ignores the role that deposit insurance plays in a bank failure. Deposit insurance insulates small depositors from loss and provides them with prompt access to their insured deposits. This markedly reduces the adverse impacts of small bank failure. Yet the Reserve Bank paper is silent on the matter. That is probably not surprising given their (largely irrational) opposition to deposit insurance. However, if a sensibly structured and funded deposit insurance scheme were established in New Zealand – which it should be – then this would significantly reduce the adverse impact of small to medium-sized bank failures. All else equal, it would provide an argument for lower capital ratios for small banks than for larger banks, subject of course to bank levies being calibrated on the basis of their risk profile.

4 Another failure in the Reserve Bank’s analysis is the importance of bank risk appetite settings, risk management and governance arrangements in influencing the probability of failure. They ignore it completely. Instead, the Reserve Bank places 100% emphasis on the role of capital in minimising the probability of bank failure. In reality, many factors come into play in influencing a bank’s probability of failure, including capital, liquidity, credit exposure concentration, funding concentration, quality of lending, risk appetite, quality of risk management systems and controls, and governance arrangements. That is why a professional supervisory authority, such as APRA or Canada’s OSFI, put such emphasis on comprehensive requirements for governance and risk management frameworks, risk appetite settings and the like. It is also why these supervisory authorities invest the time, resource and expertise in on￾site bank assessments, benchmarking analysis between banks, and comprehensive preventive and corrective action frameworks. In contrast, the Reserve Bank still keeps its head in the sand on these matters. It is content to be a ‘light touch’ ‘reluctant regulator’ on all of these matters that so heavily influence the probability of bank failure. And yet they now seem to be the ‘rogue and reckless regulator’ when it comes to seeking to put all of the financial system soundness eggs in the capital basket and to increase tier 1 capital ratios to levels unseen in most countries. Where is the balance in this? It seems that the Reserve Bank is happy to impose a very large tax on banks (and bank customers) in order to continue to enable it to continue to be the reluctant regulator and supervisor it is in respect of the details of bank risk assessment. A further piece of the jigsaw puzzle missing in the Reserve Bank’s paper is the absence of any discussion of bank recovery planning. Internationally, most OECD supervisory authorities have, since around 2010/11, required banks to develop, maintain and regularly test recovery plans. These are comprehensive plans that set out the means by which banks would restore themselves to financial soundness in the face of a capital or liquidity shock. They are an essential element in the risk management framework of any bank. All else equal, the more comprehensive and reliable is a bank’s recovery capacity, the lower is the risk of bank failure. Yet, unlike other supervisors, the Reserve Bank has done nothing to require banks to establish, maintain and test comprehensive recovery plans and to link them to their risk management frameworks. All their focus has been on ‘separation plans’ to enable a New Zealand subsidiary of an Australian parent bank to be separated from the parent bank in a failure situation, which merely reflects the continuing irrational paranoia the Reserve Bank has about the Australian authorities in a bank failure scenario. If comprehensive recovery planning requirements were introduced for all banks in New Zealand, that would achieve at least two things. It would shed light on each bank’s capability of restoring themselves to financial soundness from an impaired capital position (and hence the amount of capital needed). And it would strengthen their recovery capability and thereby lower the probability of failure (without necessarily increasing capital ratios). The Reserve Bank’s paper is also very light in its justification for requiring IRB banks to have an average risk weight of not less than 90% of the average risk weight under the Standardised model. This is a significant proposal and one that could impose significant capital costs on the D-SIBs with flow-on effects to borrowers and depositors alike. It is perfectly reasonably for the Reserve Bank to rigorously assess the analysis underpinning banks’ IRB calculations of risk. The Bank should be doing this as part of its supervisory approach. However, imposing an arbitrary floor on IRB risk weights is a costly and inefficient approach to achieving the goal of ensuring that IRB outcomes reflect risk accurately. Much more analysis is needed by the Reserve Bank on this matter, especially if they are proposing a risk weight floor that is more conservative (penal in impact) than in other comparable countries.

5 Finally, the Reserve Bank’s paper is incredibly light on cost/benefit analysis. It notes in a sentence or two that higher capital ratios might lead to higher interest rates and reduced credit availability, but dismisses this as insignificant. Yet there is no quantification of these impacts. They should undertake a substantive cost/benefit assessment of the effect of higher capital ratios on lending rates and deposit rates, on bank risk appetite and on the likely effect on credit availability. This is essential if we are to meaningfully assess the impact of the proposals on the economy and bank customers. Likewise, there is no mention of the impact of the capital proposals on the contestability of the banking system – i.e. whether introducing one of the highest capital ratio requirements in the world would make it less attractive for new entrants (domestic or foreign) to enter the banking system, with reduced competitiveness and efficiency for the financial system and economy. And there is next to no analysis of the impact on competitive neutrality in the proposals – i.e. whether this would impose competitive burdens on small banks relative to large banks (given that the failure of a small bank is systemically insignificant) or the competitive non-neutrality vis a vis non-bank deposit-takers. Given that efficiency of the financial system is one of the implied objectives of the Reserve Bank, one could reasonably expect much more comprehensive analysis of the financial system efficiency issues that arise from the proposals. All in all, the Reserve Bank’s proposals are poorly thought through and unsupported by analysis. The Reserve Bank needs to fundamentally reassess what is being proposed and provide much stronger argumentation and facts to support it, backed by a comprehensive cost/benefit assessment. Treasury needs to step up here too. They need to independently review what the Reserve Bank is proposing, either themselves or by commissioning a report from an independent expert. New Zealanders can reasonably expect a lot more from its banking regulator, in terms of quality of analysis, than it is getting.

Time for an independent assessment of the Reserve Bank’s bank capital proposals Geof Mortlock, International Financial Consultant1 Reserve Bank’s capital proposals for banks In December 2018, the Reserve Bank released proposals to increase the Tier 1 capital required of systemically important banks to 16 percent of risk weighted assets, with 6 percent of this a regulatory minimum and the remaining 10 percent a prudential capital buffer. Under the Reserve Bank’s proposals, the capital requirements for all banks will be much higher than in any comparable country, including Australia, Canada and much of the OECD. As the Reserve Bank acknowledges, its own stress tests of banks (including stress tests done jointly with the International Monetary Fund (IMF) in 2017) reveal that the banking system is currently resilient to very high shocks to the economy; no bank comes even close to failure under any of the stress testing scenarios applied by the Reserve Bank or IMF. Regardless of this, the Reserve Bank is proposing extraordinarily high capital ratios for the banks – ostensibly to achieve a ‘1 in 200 year’ resilience to economic shocks. The aim is to reduce the probability of bank failure to a very low level and thereby reduce the risk of taxpayer costs in any rescue of a bank and to minimise adverse impacts of bank failure on the economy and society. Implications of the proposals The Reserve Bank’s proposals will have major implications for the banks and the New Zealand economy. All else being equal, an increase in capital will lower the probability of bank failure (unless, of course, banks counter this by taking a higher risk appetite to maintain their return on equity – which is a possibility in some cases). However, as every economist knows, there is no such thing as a free lunch. The increase in capital requirements, while potentially achieving a lower probability of bank failure, will almost certainly have adverse impacts on the economy by increasing the cost of borrowing and probably reducing the availability of credit. Some banks (or their parent banks) might decide to reduce their balance sheets or at least slow the rate of growth in assets, rather than put more capital into the banks. For example, it would not surprise me to see the Australian parent banks take this approach. The impact of higher capital requirements is likely to be felt mostly by those who can least afford to bear the cost – i.e. those on low to middle incomes borrowing to buy a house or for other household consumption purposes, and small to medium-sized businesses. They will be faced with higher borrowing costs and/or reduced capacity to borrow. The increase in capital requirements also makes it less attractive for foreign banks to establish and expand a presence in New Zealand, thereby reducing the contestability, competitiveness and efficiency of the banking system, to the detriment of all of us.

1 Geof Mortlock is a Wellington-based international financial consultant who undertakes regular consulting work for the International Monetary Fund, World Bank, Toronto Centre and may other organisations. He draws on a long career in bank regulation and financial stability/central banking, including years at a senior level in the Reserve Bank of New Zealand and Australian Prudential Regulation Authority. He has participated in senior￾level committees of the Financial Stability Board and Basel Committee in Switzerland. OIA s9(2)(a)

2 Cost/benefit assessment One would think that these costs and other adverse impacts would have received a great deal of attention by the Reserve Bank in undertaking a cost/benefit assessment of the proposals. But no. By their own admission, they have not yet undertaken a comprehensive cost/benefit analysis. (Perhaps Adrian Orr was too busy engaging in god-to-god dialogue with Tane Mahuta and the forest fairies to give serious policy matters his attention! Deities are so busy of course.) In a further paper released by the Reserve Bank on 3 April, providing more information on the capital proposals, the Bank acknowledged this by noting that the additional information released is not a cost/benefit assessment. And indeed it is not. The additional information is helpful in providing context for the Reserve Bank’s proposals, but is mainly confined to theoretical assessments of elements underlying the proposals, rather than enabling any meaningful assessment of costs and benefits to be made. The additional information released recently the Bank does not adequately address most of the key issues at stake. For example: • There is no analysis of ‘reverse stress tests’ of the banking system – i.e. assessing the magnitude of economic shock needed to cause any of the systemic banks to fail and then assessing the probability of such a magnitude of shock occurring. This should have been done as part of the calibration of the proposed capital requirement. My hunch is that, if the Reserve Bank conducted reverse stress tests to assess the point of failure for the banks on current capital levels, it would reveal that the level of shock required would be extremely high – much, much higher than New Zealand has ever experienced. It would be interesting to see this analysis and an assessment of the probability of such a severity of shock occurring, based on global experience. • The Reserve Bank’s material contains no analysis of alternative ways of achieving the objective of a very low probability of bank failure, such as strengthening the requirements in relation to bank governance, risk management, risk appetite, risk culture and exposure concentration. Instead, the Reserve Bank focuses solely on increasing the capital requirement. In contrast, in other OECD countries, there is a much more balanced approach to achieving a low probability of bank failure that includes not just capital, but also requirements in relation to bank governance, risk management, quality of lending and exposure concentration. The unbalanced approach by the Reserve Bank is out of step with the rest of the world. It makes no sense. And I suspect that it will prove to be a costly mistake for New Zealand. • The Reserve Bank’s material contains no analysis of alternative loss absorption mechanisms. Its focus is on Tier 1 capital (essentially share capital). In contrast, other countries have assessed and some are implementing alternative loss-absorbing mechanisms to enable banks to be resolved without taxpayer costs or damage to the financial system, including establishing tranches of contractual bail-in debt issued into wholesale markets. Indeed, globally systemically important banks are now required to have this form of loss absorption. Yet no serious analysis of this has been published by the Reserve Bank.

3 • There is no analysis of the means by which banks can restore themselves to financial soundness following an impairment to capital or liquidity, such as through recovery planning. In almost all OECD countries, supervisory agencies have established recovery planning requirements for banks. This has been one of the key responses to the global financial crisis and is designed to enable banks to re-establish capital and liquidity buffers following a severe financial shock. In New Zealand, the Reserve Bank has done nothing at all in this area, in contrast to the very substantial progress made in recovery planning requirements in Australia, UK, EU, US and much of Asia. • There is no analysis of the resolution options that can be used to resolve a failing bank in ways that minimise the adverse impact on the banking system and economy, while minimising taxpayer costs. If reliable mechanisms can be used to manage bank failures in ways that greatly reduce their systemic impact, this reduces the need for very high up-front capital requirements. Yet all the Reserve Bank has done in this space is to implement the requirements for its so-called ‘Open Bank Resolution’ (OBR) form of resolving a failing bank. This policy is half-baked at best. It is actually dangerous because, if OBR were implemented without deposit insurance (and the Reserve Bank has consistently opposed deposit insurance), then it would likely trigger a mass depositor run on the banking system. It is also a failed policy on other grounds because it relies on government guarantees being applied to the post-haircut liabilities of the bank in question. In all likelihood, if any Minister of Finance were stupid enough to implement OBR, it would require a blanket guarantee of all banks’ liabilities for a considerable period after the implementation. The bizarre outcome might be to actually worsen taxpayer risk and increase moral hazard. OBR was a useful mind experiment when we (including me, as one of its early creators) came up with the concept many years ago. But no serious, intelligent resolution authority/central bank would be daft enough to implement it in the form conceived by the Reserve Bank. But then again, no central bank on the planet promotes itself as a god of the financial forest. Only in New Zealand! Hopefully, some adult thinking will emerge in the Reserve Bank to develop realistic resolution policies that minimise government liability and risk to the financial system. One such avenue is a joint resolution arrangement with the Australian authorities in the form of a properly structured contractual bail-in debt arrangement that keeps the New Zealand subsidiaries as part of the global parent banking groups. This would be much lower cost for New Zealand than the Reserve Bank’s oddly conceived plan to separate the subsidiaries from the parent banks in a resolution. Australia has been open to such an idea. The Reserve Bank has not. I hope that the Minister of Finance will knock some heads together to make progress in the direction of some serious joint Australia-New Zealand resolution option. • The Reserve Bank has also failed to provide any analysis of whether the capital proposals will result in significant financial sector activity moving from the regulated part of the financial sector (i.e. the banks) to unregulated non-bank lenders, to the detriment of the efficiency and possibly the stability of the financial system and the economy. The Reserve Bank notes that it will carry out a cost-benefit assessment for a Regulatory Impact Statement to help inform and describe final decisions in the review. This is an extraordinary statement by the Reserve Bank. It has released proposals for a very large increase in capital requirements that will have considerable impacts on the economy and financial system, but has not yet undertaken or published a cost/benefit analysis. When the Bank does – finally – get around to undertaking such an analysis, this will only be once the final decisions have been

4 made. What is the use of that? A meaningful cost/benefit analysis should occur at an early stage in policy formulation. It should have accompanied the release of the capital proposals in December. It should inform the discussions. It should have been undertaken for a number of policy options so that different options could be meaningfully assessed in terms of their respective benefits and costs. Instead, the Reserve Bank will undertake its cost/benefit assessment only once it has finalised its decisions. How objective will this analysis be, you ask? It does not take much cynicism to reach the correct conclusion on that. The cost/benefit analysis that the Reserve Bank finally releases will be formulated to justify its proposal. It will not be objective. It will not be rigorous. It will likely not be vetted by an independent party. And it will probably not be accompanied by cost/benefit analyses of alternative policy options. Sadly, this is the standard we have come to expect of the Reserve Bank and of many government agencies. Cost/benefit analyses for regulatory proposals in New Zealand are typically only done once the agency in question has already made up its mind on the policy to be pursued. They are rarely released at an early stage in the consultation process so as to inform the scrutiny of the proposal and competing options. They are typically engineered to justify the proposal selected. They are poorly developed. And they are subject to inadequate independent scrutiny. Some government agencies do a better job of this than others in preparing cost/benefit assessments. The Reserve Bank has a poor track record in this regard. And that is because it operates as a power unto itself, sheltered by an excessive degree of operational autonomy and lacking in any meaningful external scrutiny. In other countries – including Australia – regulatory proposals are typically required to be released at a much earlier stage in the consultation process and are subject to rigorous external scrutiny. In Australia, this is done by the Office of Best Practice Regulation (OBPR) - a senior￾level government agency that has real teeth and resources, and can reject policy proposals which do not pass scrutiny by them. In New Zealand there is no equivalent. Treasury reviews cost/benefit analyses, but lacks the resources to do the job well. It has none of the powers that the Australian OBPR has. So, what we have from the Reserve Bank is a capital proposal that is not backed by robust analysis and is not accompanied by alternative policy options to assess the best policy/policies to achieve the desired objective of a resilient banking system at low economic and efficiency costs. In short, the Reserve Bank has developed a simplistic policy to substantially increase bank capital requirements instead of exploring a much more sensible, balanced array of policies that could achieve the same outcome at much lower cost. Moreover, it has failed to undertake any meaningful cost/benefit analysis of its proposals, including the potentially very significant adverse impact on longer term economic growth. This is not good enough. Independent review of the capital proposals What is needed is an in-depth independent, professional assessment of the Reserve Bank’s capital proposals. Treasury will no doubt review the Reserve Bank’s cost/benefit assessment once the Regulatory Impact Statement has been prepared. However, that is too late in the policy formulation stage. Moreover, with all due respect to Treasury, it lacks the depth of knowledge needed for a rigorous assessment of the different policy options and the costs and benefits of each. What is needed is for Treasury, at the direction of Grant Robertson, to engage a couple of independent professionals (such as an academic specialised in bank regulation and

5 a recently retired foreign senior bank regulator – e.g. John Laker, former Chairman of APRA) to undertake a comprehensive assessment of the Reserve Bank proposals, plus alternative options. The findings of this assessment should be incorporated into the Treasury’s own review and the results be made public. The independent review would considerably assist the quality of the process of assessing the Reserve Bank’s proposals and may assist in reaching a more sensible outcome. An independent review would sensibly include consideration of: • the magnitude of economic shock needed to cause any of the large banks in New Zealand to fail, and the probability of such a shock occurring; • the level of capital needed to enable banks to survive a plausible range of severe economic shocks; • the composition of capital requirements and other loss absorption facilities that would be suitable to enable banks to survive severe economic shocks, including whether (as in many countries) a substantial proportion could be in the form of debt instruments that convert to equity upon defined breaches of core equity capital ratios; • the impact of the proposed increase in capital ratios on bank lending, interest rates, property prices and economic growth (with implications for government revenue); • the impact of the proposed increase in capital ratios on banking system contestability, competitiveness and financial inclusion (especially for those on low incomes); • the alternative policy options for strengthening the resilience of the banking system, including strengthening Reserve Bank supervision of banks’ governance, risk management practices, lending policies, and recovery planning; • the bank failure resolution options that could be applied to maintain financial stability with minimal taxpayer risk (and hence reduce the need for high capital ratios). The Minister of Finance and Treasury need to give serious attention to these matters. And the sooner the better.

Bank profits are too high for the risk – implied Government Guarantee  NZ due to our size and distance from other markets is a land of oligopolies, few which are cosier than the big four Australian Banks. Throughout my time in banking since 1990 till today the NZ Bank’s returns have consistently well exceeded their WACC. In fact NZ Banks have consistently been the most profitable in the world followed only by their Australian parents.  The discount rate appropriate for an investment with zero risk would be equal to the risk free rate ie -an actual rather than implied Government Guarantee.  Reuters currently reports NAB’s ROE at 11.78% with CBA the highest at 16.12%.  Using RBNZ’s own data for ROE for 2018 calculated as NPAT/ Average Equity – ROEs were ANZ 13.9%, BNZ 13.5%, ASB 15.3%, and Westpac 13.1%.  As a comparison Guru Focus calculate Bank of America’s current WACC at 8.84%.  On the face of it the Bank’s ROE’s seem extraordinarily high for businesses that needed a government guarantee – ie too big to fail post 2008.  Profitability of the 4 Banks also measured by NIM is amongst the highest in the world at 2.2%-2.3%.  Nb:- Totally disregard low ROA the Bank’s use as a defence at they can leverage up 90%.  These statistics suggest NZ consumers, businesses and farmers are being fleeced to support returns to their Parents and international shareholders.  The question then comes how should they be regulated or capped AND WHAT ARE THE IMPLICATIONS? Regulatory Observations

1. Basel 3 requires Tier 1 capital of 8% and total of 10.5%. The NZ banks generally operate with tier 2 of circa 2% and a buffer of 1-2% meaning total capital of 11%-12%. The RBNZ has announced a proposed jump in Tier 1 to 16%. This is one means of lowering their ROEs, but is it the best way?
2. Regulatory risk weighting for residential mortgages at 35% to 100% for commercial and business and more than 100% for some asset types has led to a misallocation of capital away from where it is needed, driving Banks into mortgages which has been negative for the NZ economy and world trade.
3. Government Decision re default Kiwisaver Funds going to conservative = madness even ANZ said it was stupid - result is it has funded the NZ Bank’s balance sheet.
4. AML is a dog. It is well intentioned but misguided. It is a bureaucratic nightmare that simply frustrates both Banks and their customers leading to huge time delays = lower productivity. It is ridiculous that having been a customer for 30 years you now need to provide your passport, drivers license, birth certificate, credit cards etc as proof your not a criminal, despite the bank manager knowing you for many years. Getting a new bank account for a Trust has become next to impossible. It has got so ridiculous that all committee members for some sporting organisations are being asked for their details as part of the sporting organisations approval. Market /My observations NZ banks charge more for almost all products than Banks overseas. You need to benchmark the margin on fixed and floating rate mortgages vs SWAP and BKBM. Examine – fx / interest rate /

option pricing / trade / fx / credit cards / merchant fees/ wealth management / kiwisaver fees / insurance etc. Bank’s operate no differently to securitisation vehicles which were the scourge of the GFC. They borrow short and lend long – historically relying on the time value of money and an upward sloping yield curve. Much to my disbelief at time of the GFC, entire Institutional and Treasury was funded via overnight money. New regulation has forced increased prudence by banks eg the core funding ratio etc , but a mismatch still exists. Bank’s are fortunate they don’t have to mark to market their loan book which was the downfall of many securitisation vehicles (an accounting nonsense in my view)– however those securitisations that held to maturity or bought at 40 cents in the dollar invariably got back 99 cents plus in the dollar. My experience of Credit Risk Scoring tools in Corporate, Property, Agri and Institutionals is they can be gamed but are generally conservative. This is based on my experience in lending recovery from at where with some restructuring we got all our money back safely with the exception of . Many of my work outs were the result of poorly structured lending in the first place that did not help the customer. From when I worked for it was clear was a much leaner organisation in headcount than in the departments I worked in. Given the cuts in staff and service since, I suspect NZ is even leaner. There has been a major loss of experience ( grey hair) across the sector post 2008, and a number of very senior managers I have spoken to don’t understand how their overdraft and loan documents work, risk weightings, financial ratios etc. This shows through in the poor service their departments deliver. Detailed Cost to Income ratios by department would prove this. As an adviser to SMEs since 2012, service levels for SME customers in particular are at an all time low. Call centres are operated by people with no authority or knowledge so unless the request fits “squarely in the box” no one can help. Lending is invariably structured incorrectly, and customer working capital needs are not met. This is why independent brokers/ advisers are needed. Customers need independent financial advice so they get the best deal – which is invariably not offered by their long term bank. However, Customers will pay for a relationship. A number of the conduct recommendations from Australia are simply wrong -eg Banks have cut their sales teams as it is cheaper and more profitable to outsource it to brokers. To make this work they need to pay brokers commission. Likewise getting rid of incentives so sales people perform is stupid – a business will be left with dullards. Unfortunately how some of the incentive schemes have been structured eg to cross sell products people don’t need / or shouldn’t be sold / or flipped into has shown both the limited understanding of the long term business key drivers, and intelligence of Senior Management who have put them in place– who in many cases have been grossly overpaid. Incentives structured correctly, measured, verified and checked are great. IMPLICATIONS / Questions

1. Regulations are likely to lead Banks back to being Building Societies – residential mortgage providers only.

2. Is 12% capital with some other requirements eg a cap on mortgages as a percentage of the loan book a better solution? I don’t think 16% capital is needed.

3. NZ Banks have begun rationing capital to SME’s, Corporates and the Agri-sector (ANZ and ASB remain overexposed). Customers in these sectors will need external advice if their needs don’t fit “square in the box.” Some tough refinancing and pricing discussions will be had.

4. Independent Financial Advisers will be needed more than ever.

5. Cost effective working capital / debtor financing is not being offered by Banks. There is a clear need for GSA lending.

6. Where is the capital going to come from to grow small business and meet the government’s rebalancing growth objectives? It is very difficult to get a loan without a house as security. Going forward Millennials will have 80% LVRs vs their houses. How will working capital for the business that they might own be financed?

7. Lower Bank profits means a lower Government tax take from Banks

8. Is there a different way of redirecting capital from housing eg Change risk weightings?

9. Are the NZ Banks and other foreign banks going to be adversely impacted, therefore further tightening credit.

10. If tier 1 debt is limited are we proposing it is replaced by tier 2 and new equity providers?

11. Does the RBNZ or similar need a Commerce Commission enforcement role with some teeth? Appendix 1 – BNZ’s estimated actual market risk premium I have not used the full WACC = E/V Re + D/V Rd (1-t) analysis to explain this as I know it is understood, rather I have kept it to discount rate that is applied to the after tax cashflows to illustrate the point more simply. Deriving the discount rate using the CAPM where:- r = Rf + Beta (Rm- Rf) r is the discount rate Rf is the risk free rate – generally the market uses the 10 year bond rate – currently 1.58%. Beta is a measure of relative risk to the market as a whole, ( Reuters as of 18/4 report NAB beta at 1.06x with ANZ the highest at 1.41x – International Peers 1.07x), and Rm is the expected market return – most long term US Studies show this be 5-6%, Australian studies up to 8%. The discount rate appropriate for an investment with zero risk would be equal to the risk free rate ie -an actual rather than implied Government Guarantee. Reuters currently reports NAB’s ROE at 11.78% with CBA the highest at 16.12%. My understanding is BNZ’s ROE is currently 14%, but using RBNZ’s own data for ROE for 2018 calculated as NPAT/ Average Equity – ANZ 13.9%, BNZ 13.5%, ASB 15.3%, and Westpac 13.1%. If we use BNZ and NAB’S beta as a proxy The Actual market risk premium achieved is 11.24% = (13.5%- 1.58% ) /1.06.

Graeme Legg Dear Sirs, I am a private individual and a semiretired professional engineer living in Auckland. I wholly support your proposals to raise the minimum capital requirements for the main banks as well as equalizing the requirements between NZ owned & Australian owned banks. At present I understand that in event of a bank running out of capital portions of the depositors funds would be retained to recapitalize the bank. Alternatively (and more likely) there would be substantial political pressure for govt guarantees and recapitalization. To avoid same the one in 200 year risk is the minimum personally acceptable (and given that such failures tend not to follow a bell curve at all, but more likely a break similar to "catastrophe theory" a calculated 1 in 200 year risk is almost certainly an understatement of the real risk. If a bank does not which to run with the higher capitalization & feels that their profitability would be adversely affected there is nothing stopping them from moving out of the regulated banking sector completely and letting the professional market decide how to price their increased risks relative to relative to the highly capitalized regulated banks. The regulated banks are used by the bulk of non professional ordinary customers and are more akin to a utility. Given that the NZ govt does not guarantee the depositors, and due to the necessity for the major banks to continuously function so that the economy does not freeze in event of a major bank failure those organizations which operate in this territory have to accept risk minimization, which should be well under a 1% level. I do not accept that there will be major increased financial costs associated with this – the increased security in a pure capitalist system should result in a lower external costs of funds. Also for equitable functioning in a capitalist system the inherent financial advantage given to Australian owned banks versus New Zealand owned banks cannot be optimal for the New Zealand financial system and this inherent subsidy to foreign organizations should therefore be removed. regards, Graeme Legg BE (Mech), CMEngNZ, CEng (UK). FIMechE

Hall I support the new changes in increase the capital held by banks OIA s9(2)(a)

Harbour Asset Management – April 2019 Page 1 Submission to the RBNZ on the Proposal for Changing Capital Adequacy for New Zealand Banks Harbour1 has written two short research papers on the RBNZ’s proposal to lift capital held in New Zealand by banks (see here and here). These reports highlighted that higher lending rates, lower deposit rates and potentially lower risk weighted asset growth could be expected outcomes that might be associated with lower investment and GDP. There are broader implications for our capital market that are also worth considering. We doubt that investors will accept either a lower cost of debt financing or a lower return on equity reflecting increased equity capital and hence we anticipate higher medium costs. This belief generates an optimal level of capital lower than 16%. We would however allow greater flexibility in the use of the proposed Countercyclical Capital Buffer. Finally, we believe uncertainty in transition costs warrants a longer period for introduction of strong bank capital. At a high level we support the six principles of the review. Banking systems globally have significantly increased capital requirements. The most obvious example is Australia’s move to a 10.5% unquestionably strong level. We note Australia has yet to introduce a Countercyclical Buffer suggesting 10.5% may just be a stopping point to yet higher bank equity capital. New Zealand’s banking system is unique in our reliance on Australian bank shareholders to provide capital; in exposure to the systemic risks from housing (like Australia); and exposure to relatively high levels of household debt. We note that stress tests2 currently signal that individual banks carry a significant capital buffer. The latest 2017 stress test pointed to a potential 6.9% point fall in CET1 capital, with a 6.4% to 7.4% range across the banks. This scenario would eat significantly into current bank capital buffers. Whilst this scenario may have very significant output and unemployment costs, the “banks could absorb material losses while remaining solvent”. These stress tests provide estimates that come with a wide range of uncertain outcomes, but in this scenario it is not obvious that the banking system itself would become an independent cause to exacerbate the output impacts or sustain the sort of banking crisis and permanent output costs outlined as a concern in the capital review. In other words, higher capital, while useful for reducing more extreme events and scenarios, wouldn’t itself be useful to mitigate external shocks. Our perspective Our judgement is that something closer to 14.0% to 14.5% for the bank capital stack seems a more likely median position through the cycle that optimises the trade-offs. This would still be a substantial increase in bank capital when the move to a 90% IRB floor shift is included. We push for a larger (0-3%) countercyclical capital buffer (CCyB), allowing the RBNZ to vary capital stack between 13.0% and 16.0%. We would want to see equalisation across financial institutions, with flexibility reflecting risk assessments of the economy and sectors. We think that significant policy change has uncertain impacts and we recommend taking relatively smaller steps over a longer horizon with a definitive date to pause and re-examine the evidence.

1 Harbour Asset Management manages funds on behalf large wholesale clients and retail advisers, spanning KiwiSaver, superannuation, charitable trusts, IWI, and financial independent advisers. Today our Funds Under Management are a little over $4.7bn. We are a significant investor in both bank debt and bank equity. We also manage money for many banks and bank clients. We therefore have deep relationships across the bank sector. 2 RBNZ 2017 Stress test included a downturn in the Chinese economy, a collapse in the demand for commodities a fall of 35% in house prices, and a dairy payout ratio below $5kgMS.

Harbour Asset Management – April 2019 Page 2 Our key points: • Transition costs could be significant and unpredictable • Slow down the proposed timetable – implement over 10 years not 5 years • Interest rate impacts could be larger, and remain uncertain • Loan growth could be constrained and right now credit is somewhat fragile • Push for a 0-3% CCyB, set at 1.5% now, be flexible • Treat D-SIBs and other banks with the same capital stack • Reconsider the role for Tier 2 capital Harbour is interested in the broad economic impacts of the proposals, the impact on bank shareholders, on bank debt investors, and on the structure and risks in New Zealand’s capital markets. We have conducted an independent analysis and undertaken direct research with Australian banks and shareholders. In addition to responding to some detailed questions we also raise specific issues that we consider merit more inquiry and debate. We agree (support) with more in the proposals than we disagree with, and broadly concur with the direction of the proposals. But as with most submissions there tends to be a focus on controversies and hence a negative tone can permeate responses. We request that you normalise our response accordingly and we look forward to engaging with you on this very important proposal. We think that rapidly adding higher capital requirements for New Zealand banks could carry higher than anticipated transition costs. We found the RBNZ proposal relatively silent on the costs and benefits of transition and the case for the relatively fast (5 year) timetable for higher capital was not made persuasively. We note that some of the literature3 advises against using conventional monetary policy (lower interest rates) to ameliorate or offset transition costs. With rates in New Zealand already near intergenerational lows the scope for monetary policy to offset an unexpected significant tightening in monetary conditions as a result of a rapid adoption of higher bank capital is limited. We note that this is a clear area in which the bank is seeking feedback and that in a recent speech the Governor suggested some flexibility around the timing of adding capital. We welcome a stronger consultation on timing, and the potential for a pause to assess the transition costs would be a useful consideration. This is a key recommendation in this submission, and we make our comments on transition costs at the end of this document. Five years should move to ten years, including a pause to assess the evidence. A second observation is that we suspect that the proposal for a 16% capital requirement may induce significant regulatory arbitrage, encouraging rapid growth in New Zealand’s non-bank financial system. We note already an increase in so called ‘warehouse facilities’ provided for non-banks from Australian and other global banks. Whilst this growth is from a low base, we note that the bank capital review itself is largely silent on the issue of the potential role of non-bank financial institutions funded by non-NZ deposit taking institutions in increasing credit growth. The extent to which leverage in the broad financial system generally exacerbates economic shocks has implications for the cost-benefit analysis used in the RBNZ’s modelling. New Zealand has a clear

3 Cline 2016, “The problem with this line of thinking is that monetary policy should stick to its central mandate—maintaining price stability along with high employment—and not be burdened with extraneous obligations.”

Harbour Asset Management – April 2019 Page 3 track record in using tax-payer funding to bail out non-banks, and over the long-term policies need to evolve to reduce just shifting risk from one sector to another. This may require policy co￾ordination between the RBNZ and APRA and further regulation of the non-bank sector. An underlying theme is that the large Australian owned banks, with circa 88% market share ought to have higher capital reflecting their systemic role. Some global regulatory bodies require systemically important banks to hold higher capital than smaller banks. Many of these banks are Globally Systemically Important Banks. These differ from our D-SIBs. Our view is that the negative impact on society from any bank failure can be significant. The example of Northern Rock comes to mind, where the failure of a relatively small financial institution had a major impact on public confidence. Whilst the capital positions of the non D-SIBs in New Zealand are potentially more constrained given their ownership structures and restrictive shareholding positions 4 , this can be overcome through the capital market and global capital funding. The RBNZ paper seems to make a case that the smaller banks are special and would struggle to gain capital through retained earnings compared to the D￾SIBs5 . We don’t agree that retained earnings is the only path towards stronger equity capital and we think that the output costs and risk appetite for any bank failure are similar enough in a New Zealand context to not differentiate between institutions with the exception of assessing sectoral and specific risk issues. Cost of capital, dividend policies and shareholders’ reactions

• We doubt the Modigliani-Millar model holds in the context of Australian-owned banks. The reaction of global and Australian capital markets to this proposal will largely determine the costs of holding significantly higher capital. We think bank shareholders (and debt holders) are unlikely to accept significantly lower potential returns for holding higher capital (and hence less risk). Our view differs from the assumptions in the Reserve Bank’s consultation papers. We have had numerous discussions with the Australian banks and institutional shareholders and the evidence in capital markets in our opinion is that capital structure matters particularly when the cost of debt is largely fixed. We do not think that debt costs will fall materially for NZ D-SIBs as the marginal cost of debt reflects the financial health of the entire bank entity. In the case of the Australian-owned subsidiary banks in New Zealand, the cost of debt funding is largely set by the credit rating of the overall corporate. Further in terms of dividends and returns to shareholders we observe already the extent to which shareholders prefer high dividends and stock buy backs in Australia, as opposed to retaining earnings for further investment at lower profitability (return on equity). We do not think that increasing equity in NZ banks via retained earnings is an option that will be palatable for shareholders if doing so requires dividends to be cut. In the absence of a bank crisis we would expect significant resistance from Australian shareholders to cut dividends to provide higher capital for the NZ subsidiaries. Instead it seems more likely that banks would engage in a combination of raising fresh capital from shareholders; selling down non-core assets; repricing net interest margins; or cutting risk weighted assets.

4 We are looking forward to further work on tier 1 capital for mutually owned New Zealand banks. 5 RBNZ Capital Review Paper 4 (See paragraph 114 and 115).

Harbour Asset Management – April 2019 Page 4 The evidence that shareholders in Australia like buybacks and dividends has been widely canvassed and reduces the impact of the Modigliani-Miller6 thesis. We agree that the academic and theoretical literature suggests that investors should be indifferent to various tax effective income streams and capital injections. However, in practice this is not what we observe. Both retail investors, and increasingly, institutional investors (through funds that focus on yield) are important providers of risk capital. Our most likely scenario is that shareholders push banks to stem the growth of lending in New Zealand in marginal lending sectors while also seeking to re-price interest margins. A special issue arises for the Australian banks with very large relative businesses in New Zealand. We note that APRA’s proposed APA222 25% subsidiary restriction could have a significant impact on some Australian banks’ capability in terms of increasing capital in New Zealand. This highlights the difficulty of seeing debt-equity substitution across the sector. We cannot find enough evidence in Australia to provide a reasonable estimate of the offset impact on the cost of capital from debt to equity substitution. We note that the RBNZ Capital Review Paper also largely cites global evidence. Cline7 for instance cites evidence for the US over the 2002-2013 period which suggests that less than half the offset has been observed. We disagree with the RBNZ’s assumption of a 50% offset effect. We do not think that debt costs will fall meaningfully, and we do not think that shareholders will accept a substantially lower return on equity. We suspect a reasonable range for an offset impact on the cost of capital for New Zealand D-SIB banks, given the Australian ownership and debt rating structures, is less than 50%. Cline says that higher capital requirements impose increases in lending costs, with associated output costs. We note in work cited by the RBNZ, Cline8 has more recently published work suggesting a 11.7% to 14.1% optimal capital ratio of risk weighted assets for banks. We note that the optimal capital level is sensitive to this key assumption on the offset impact which is difficult to observe ex￾ante. This is another reason why we recommend a longer transition and patience to observe empirical outcomes. In sum, after consultation we are of the view that Australian and global shareholders will move to push for sustaining return on equity, and hence we expect a combination of higher net interest margins (either higher mortgage rates or lower deposit rates) exceeding the RBNZ’s expectation, and/or a targeted reduction in risk weighted assets in specific New Zealand lending sectors. Whilst we do not hold the extreme view in the market provided by UBS estimates, we are closer to the CLSA estimates of a 50-70 basis point re-pricing of net interest margins, plus some degree of risk weighted asset growth constraint in certain sectors.

6 The M&M theorem holds that the average cost of capital to the firm is independent of capital structure, because any reduction in capital cost from switching to higher leverage using lower cost debt is exactly offset by an induced increase in the unit cost of higher-cost equity capital as a consequence of the associated rise in risk. 7 Cline, 2015. Testing the Modigliani-Miller Theorem of Capital Structure Irrelevance for Banks 8 Cline, 2016. Benefits and Costs of Higher Capital Requirements for Banks

Harbour Asset Management – April 2019 Page 5 Consequences of higher bank capital We agree that the RBNZ needs to factor into a cost-benefit analysis the risk that net interest margins are higher. However, we would push for a higher negative impact on capital formation and growth than assumed in the RBNZ’s paper. We also think that risk weighted asset growth could be constrained, especially over a transition period if the capital proposals were introduced in five years. Both these views suggest a lower optimal capital ratio than suggested in the RBNZ’s capital review. In that regard we consider that the second-round effects from this capital review might have a significant impact on GDP and fiscal risks (from the non-bank financial system). Our view is that the Reserve Bank’s model, which captures the costs of higher capital, could be amended to include an estimate for the impact lower risk weighted assets may have on investment and GDP growth. We doubt that the local banks or the non-bank financial system has the near-term capacity to either provide sufficient interest rate market pricing tension, or the capital to fill the void to back higher risk weighted asset growth. We suggest that the RBNZ adopts a longer transition period with more flexibility and the potential for a pause and reassessment of actual transition outcoimes after five years. Capital allocation across the economy Many commentators (and potential submitters) have possibly focussed on the RBNZ’s change to a 1 in 200-year risk assessment from possibly an earlier 1 in 50 or a 1 in 100-year risk assessment. In our view there are many industries and sectors in New Zealand that face risks where further investment now could substantially improve GDP and reduce the volatility of outcomes, and potentially improve well-being. Examples include investing in social housing, the electricity network, highway safety, early adoption of 5G, decarbonisation, investing in pre-school education, child health, and more spending on disease prevention, detection and treatment. We are financial market participants, not public policy experts, but the RBNZ request we submit our risk appetite on the move to a 16% capital level. Our view is that this is higher than an optimal long-term level of capital given competing needs. Calls for $20bn to $30bn more capital in the banking system in New Zealand need to be assessed in a wider context, as capital is scarce and the trade-offs in improving potential GDP (and reducing the volatility of GDP, and outcomes for the most at risk people in society) are very evident. We note that INFINZ has called for the Productivity Commission to provide a cost-benefit analysis. That may be a useful exercise, although our reading of any cost-benefit exercise would still be reliant on assessing many assumptions which are hard to gain strong relevant global comparisons. Again, for Harbour this points towards caution and flexibility in the speed of adopting significantly higher capital in the banking system.

Harbour Asset Management – April 2019 Page 6 Countercyclical Capital Buffer needs more flexibility A final technical point we would like to raise is our view that the proposed 1.5% Countercylical Capital Buffer (CCyB) could be made much larger. We would propose at least a 3.0% buffer and we would take at least 1.5% from the conservation buffer. Some of this is semantics and signalling but most of all we want to see the RBNZ have more tools for dealing proactively with a dynamic economy. We doubt that a sustained 16% is the “optimal” capital level for the systemically important banks. In our view, this doubt is implied in the RBNZ’s own analysis given the 1.5% CCyB (so 14.5% is actually a likely central target expressed by the RBNZ). We think it is more useful to consider 16% as an extreme or maximum level of capital. Our preference is for central banks to have more tools in a crisis. Interest rates are possibly too low to provide enough monetary stimulus to gain traction in a significant downturn. We are all learning that macro-prudential tools are an important adjacency for monetary policy and for the security of the financial system. In the context of an end target of 14.0% to 14.5% we would recommend an early set 1.5% CCyB and provide scope to step this up to 3% overtime. We think that in extreme circumstances the CCyB could be set to 0%. Our judgement is that something closer to 14.0% to 14.5% for the capital stack seems a more likely median position through the cycle that optimises the trade-offs. This view also takes into account that the RBNZ also recommends moving the bar significantly on the IRB capital floor further reducing leverage in the banking system. The 14.0% to 14.5% optimal capital level reflects our assessment that the costs in terms of either higher net interest rates or lower net lending growth will have higher costs in terms of capital formation and lower GDP. Our submission is also led by our reading of the literature on optimal capital (Cline 2016 for instance). We think the modelling conducted by the RBNZ does not take into account transition costs, and the RBNZ has a higher estimate for the Modigliani-Miller effect than is likely in practice. As a result, while we would have a capability of stretching overtime to a 16% capital position with a full CCyB, we would anticipate an over the cycle 14.0% to 14.5% average. Additionally, we observe that banks generally will hold their own buffer above regulatory minimums. As Cline notes: “The reasonable policy approach, then, would seem to be to set the regulatory requirement at the central estimate of the optimal level, in the expectation that in practice the banks would add a cushion…”. We think that central estimate is circa 14.0%-14.5% for NZ banks, still materially above current capital levels. Do not differentiate between D-SIBs and other banks in the first instance

Harbour Asset Management – April 2019 Page 7

• And flex the CCyB or other buffers on a case by case basis We would not differentiate the core capital requirement between the D-SIBs and other banks. This reflects our position that public confidence is as much impacted by the failure of a smaller institution as a large bank and that we often observe less diversification in smaller financial institutions – in geography, in loan concentration - and that funding for smaller institutions can be crowded out in a crisis. We think that the RBNZ could impose CCyB variations across the bank sector – differentiating for size, sector exposures or other perceived risks. We clearly think that the RBNZ may wish to apply a larger buffer when identified risk positions are higher (e.g. extended housing prices, rising unemployment, a terms of trade shock, rising interest rates, fiscal risks). We welcome the RBNZ’s introduction of a framework for flexing CCyB and we think it is good policy to pre-publish many of the reasons why you may flex the CCyB. Another reason for higher CCyB in the capital stack is the potential under Basel III reciprocity for APRA to accept this capital within their Level 1 capital for the Australian banks. This could therefore lessen the actual capital that may have to be raised by the banks. Potentially this could significantly reduce the costs of the additional capital, however, it may depend on how APRA views New Zealand held CCyB. We note that the RBNZ are engaging with APRA on this issue. In addition, we note that APRA’s 10.5% CET1 target is premised on a 0% CCyB. We suspect that changes in the near term with APRA noting that they are analysing “whether a non-zero default level of the countercyclical capital buffer should be implemented” to keep banks in the top quartile of global group 2 banks of a fully phased in Basel III basis. Risk weightings are also moving making the banking sector safer The proposals to significantly alter risk weightings on various lending categories seems to have received little commentary. We understand moving to a 90% Advanced IRB capital floor itself increases capital in the system by an estimated 16%. The trend of increasing capital intensity rose sharply in Australia through 2016 and 2017. Together with a tightening of lending standards (especially the introduction of detailed expenditure analysis), the higher mortgage risk weights had a significant impact on lending growth. The major banks in Australia have seen mortgage loan growth only in the low single digits in the last 18 months as capital standards for the D-SIBs has been tightened significantly. The adjustment process to move to a small gap between IRB and standardised banks approaches seems reasonable, but we wonder how quickly this should be accomplished, again mindful of transition costs. Overall the broad thrust of this proposal could encourage lending in the non-bank sector. Whilst these risks may not be systemic, they may nevertheless involve both fiscal and output costs that have not been discussed in the report. Who fills the potential lending gap if the banks reduce New Zealand risk weighted assets?

Harbour Asset Management – April 2019 Page 8 The probability of regulatory arbitrage seems very evident. There is already significant arbitrage in implied capital requirements across various banking regulators, but a potential gap of 5.5% points between Australia and New Zealand in capital requirements is likely to see Australian wholesale bank funding to non-bank New Zealand lenders9 . Although the proposal may substantially reduce risk in the New Zealand bank sector, it may just translate risk to the non-bank sector. New Zealand has a poor history of non-bank financial risks harming investors and borrowers, and eventually tax￾payers. We are concerned that exceptionally strong bank sector capital may just push risk to the non-bank sector. Our judgement is that the funds management community in New Zealand is not currently well resourced to provide significant capital (say $20bn) to support non-bank financial direct lending. In the long term a deeper local capital market may increase the scope to provide capital for bank and non-bank lending, but today in our opinion there is not sufficient funding from investors to fill a $20bn funding gap. This may change of course as the retirement sector develops through for instance KiwiSaver savings pools. As a result, in the near term we think these proposals increase the probability that foreign banks will step into the funding gap and become more prominent in New Zealand. While this may deepen bank competition, history shows that foreign banks can withdraw support at times of stress, reallocating capital more rapidly between countries. This behaviour could add to stress in a pro-cyclical outcome. Tier 2 capital potentially has a role despite the banks no “gone bank” policy stance We have reservations about excluding Tier 2 capital from the Conservation buffer capital stack. We disagree that accepting small elements (1-2%) of “restricted” Tier 2 capital would set up a financial system for a “gone banks” scenario. In a related issue we consider the capital requirements for the mutually owned banks may need a specific solution. Uncertainty argues for small steps and flexibility in timing In summary we think there are significant risks around introducing large policy changes with uncertain outcomes. We would favour a more gradual introduction of higher bank capital with clear timelines for pausing and more flexibility for the RBNZ to exercise a larger range for the CCyB. When intervening with a patient, doctors are taught to “firstly, do no harm.” The idea is that a desire to do good may fix one problem, but intervention may create other problems or even fail to fix the original one. This idea seems relevant to the implementation of the bank capital proposals. It is widely accepted that there may be anticipated or unanticipated consequences arising from a large shift in capital requirements. Quantifying this or even identifying everything in advance is very challenging. The range of estimates for net interest rate margin changes is only one example of the uncertainty. A broader set of forecasts for the impact on risk weighted asset growth most likely sets the scene for greater uncertainty, especially in transition costs.

9 It is our guess however that APRA may introduce a CCyB overtime that would narrow the gap and the adoption of a 14-14.5% central NZ optimal level would reduce further the scope for regulatory arbitrage.

Harbour Asset Management – April 2019 Page 9 Accordingly, from a risk management perspective, we would suggest a slower, staggered approach to reaching the desired capital levels. We propose a time schedule that might, for instance, seek to reach 13% to 13.5% Tier 1 capital within 5 years and includes, say, a 1.5% CCyB. We then recommend the potential for the RBNZ to wait for circa 2 years to observe empirical outcomes. The pause approach may substantially lessen risks of higher than expected transition costs or even reveal longer-term flaws in the proposed policy. The pause may also give time for the RBNZ (and APRA) to interpret further Basel rule changes and to also assess the growing role of the non-bank financial sector as these changes are implemented. At some more distant stage, a further judgement in lifting the target to 14% or 14.5% could be made (either adding a further 1% CcyCB or more to the conservation buffer). Our overall approach would see a ten year plus timetable and create greater optionality around policy, enable a real-time observation of the impact of the policy change and probably be inherently less destabilising. It could also be more politically appealing and therefore reduce the risk of regulatory intervention through a public backlash should interest rate changes prove sharp, or lending drop off substantially. Good policy needs to be sustainable, as the benefit of these proposed changes need to be measured over perhaps 50 -100 years. The only negative aspects would be if there was an event that stressed the financial system in the near term, or if personnel changes removed the appetite for a longer time frame for implementing the policy. Finally, in the near term, given the softer economic confidence we are experiencing it does not seem like the best time to seek to rapidly increase bank capital, rather a more measured approach could be considered. We welcome further engagement. Andrew Bascand Managing Director Mark Brown Head of Fixed Income, Director Simon Pannett Senior Credit Analyst, Director April 2019

Heartland Bank Limited, 35 Teed Street, PO Box 9919, Newmarket, Auckland 1149 I 0508 432 785 I www.heartland.co.nz 17 May 2019 Reserve Bank of New Zealand PO Box 2498 Wellington 6140 Via email: capitalreview@rbnz.govt.nz Submission to the RBNZ on the Capital Review Paper 4: How much capital is enough? Introduction

1. Heartland Bank Limited (Heartland) welcomes the opportunity to submit on the Reserve Bank of New Zealand’s (Reserve Bank’s) “Capital Review Paper 4: How much capital is enough” proposal (the Proposal). Heartland’s support

2. Heartland supports the Reserve Bank’s goal of promoting a sound and efficient financial system and, on the basis that the Proposal enhances that goal, Heartland supports the Proposal.

3. The current capital framework favors the four larger banks, so Heartland supports any “leveling of the playing field” between the large and small banks. Heartland expects some leveling to occur through the closer alignment of the capital requirements of IRB and standardised banks, and the introduction of the D-SIB buffer, and is therefore supportive of those initiatives.

4. An observation has also been made in the context of the Proposal that, if the burden of increased capital requirements was to materially flow through to lending and/or deposit rates, it could be indicative of insufficient competition in the market. In Heartland’s view, both of the Reserve Bank’s above-mentioned initiatives help to mitigate the competitive advantage that the larger banks have in New Zealand, which is ultimately good for the financial system and New Zealand as a whole.

5. However, Heartland also supports the request that the Reserve Bank undertake a cost-benefit analysis. The benefits of the Proposal ought to be considered in light of the economic costs of the Proposal, and it is important that the playing field is not raised disproportionately high. Specific observations and comments re Tier 2 capital instruments

6. The Reserve Bank is seeking feedback on the role of Tier 2 capital going forward.

7. Heartland does not necessarily accept the Reserve Bank’s position that Tier 2 capital is ‘gone concern’ capital. Additionally, in Heartland’s view, there are some real advantages to Tier 2 capital, being that: (a) It is cheaper than Tier 1 capital. Heartland is a commercial enterprise and the cost of its capital is something Heartland must necessarily consider. (b) It provides Heartland with capital raising options. There may be a stronger market for Tier 2 capital (cf Tier 1 capital) at certain points in time (and vice versa). Having options in the face of a need to raise capital is clearly beneficial.

8. Whilst being supportive of there being an on-going role for Tier 2 capital, Heartland considers that there is no place for Tier 2 capital in the context of the Proposal, as: (a) The Reserve Bank has determined that a 15 / 16 percent capital ratio is the level at which its soundness objective is achieved, and notes that there would be an expected negative impact on output were higher levels to be required. (b) The Reserve Bank’s position appears to be that capital ratio may only be met with Tier 1 capital – meaning that any ‘requirement’ for Tier 2 capital in the Proposal would need to be structured as a requirement for additional capital which could be met with Tier 2 capital. (c) Hence, a ‘requirement’ for 2 percent Tier 2 capital would effectively translate to a 17 / 18 percent capital requirement – where the Reserve Bank’s soundness objective is consider to be satisfied at 15 / 16 percent. [NB, banks will hold buffers over and above the 15 / 16 percent requirement in any event, which exacerbates this issue.] (d) This would come at a cost. Whether that was in the form of reduced returns to shareholders and/or depositors, it would likely impact on Heartland’s ability to attract capital and funding.

9. However, were the ‘requirement’ to be structured as an option to meet the required 15 / 16 percent capital requirement with up to 2 percent Tier 2 capital, Heartland would be supportive. Specific observations and comments re AT1 capital instruments

10. As set out above, Heartland considers that Tier 2 capital has certain benefits. Those same benefits may apply to AT1 capital, namely: (a) It could be cheaper than Tier 1 capital. (b) It could provide Heartland with capital raising options. There may be a stronger market for AT1 capital (c.f. CET1 capital) at certain points in time (and vice versa). Having options in the face of a need to raise capital is clearly beneficial.

11. AT1 capital is something that Heartland is not familiar with. However, if only non￾redeemable, non-contingent, perpetual preference shares are accepted as AT1 instruments, we understand that there will likely be no real investor market for AT1 instruments – and banks’ ability to substitute AT1 for CET1 capital will effectively be illusory.

Page 1 of 3 Ian Woolford Financial System Policy and Analysis Department Reserve Bank of New Zealand PO Box 2498 WELLINGTON 6140 By email: capitalreview@rbnz.govt.nz 3 May 2019 Submission on: Capital Review Paper 4: how much capital is enough? Submitter: Horticulture New Zealand Incorporated Submitted by: Mike Chapman, Chief Executive Contact Details: PO Box 10232, The Terrace, Wellington 6143 Introduction

1. Horticulture New Zealand, Apata Group Ltd, Central Otago Fruit Growers Association, Katikati Fruitgrowers Association, NZ Apples and Pears, NZ Citrus Growers, NZ Feijoa Growers Association, NZ Kiwiberry Growers, NZ Kiwifruit Growers, Process Vegetables NZ, TomatoesNZ, Vegetables NZ, Pukekohe Vegetables Growers Association and Potatoes NZ appreciate the opportunity to make a submission on how much capital is enough? This submission is made following consultation with members of the above listed organisations who make up some of New Zealand’s 5,000 commercial growers of fruit and vegetables.
2. Horticulture employs over 60,000 people, occupies some 120,000 hectares of land and provides critical regional development opportunities in Northland, Auckland, Bay of Plenty, Hawke’s Bay, Marlborough, Nelson, Canterbury, Taranaki, Waikato, Manawatu, Wellington, Central Otago and Southland. Horticulture is valued at $5.8 billion, with $3.61 billion in exports to overseas markets. Horticulture is going through a period of significant growth with forecasts to June 2019 predicting a 17% rise to over $6 billion (MPI, 2019). PO Box 10232 The Terrace Wellington 6143 Level 4 Co-operative Bank House 20 Ballance Street Wellington 6011 Phone: +64 4 472 3795 Fax: +64 4 471 2861 Web: www.hortnz.co.nz Email: info@hortnz.co.nz

Page 2 of 3 3. In making this submission we note that the rationale for the changes is set out on page 5 of the Review Paper as follows: “The Reserve Bank is proposing this change to reduce the chances of banks failing in New Zealand. If banks in New Zealand fail, some of us might lose money and some of us might lose jobs. However, there would also be indirect costs on all society that may be harder to see that would negatively impact the well-being of all New Zealanders. In the end, we would bear the cost of bank failures, in one way or another.” Principles 4. Our preparation of this submission involved consulting with our members and in result we develop the following principles:  Horticulture is an international business with our competitors being growers in other countries; therefore, the cost of capital in New Zealand needs to be competitive relative to the cost of capital for other international producers.  Business infrastructure growth will be sustained through more certainty of long-term rates.  Risk protecting the banking systems needs to be balanced against increasing the cost of capital, slowing infrastructural investment and reducing returns to investors.  Pressure needs to be maintained on the banks themselves to operate prudently.  Competition between banks needs to be enabled. Submission 5. We do not support the proposal as currently formulated. We do support Federated Farmers’ submission and, based on the above stated principles and our discussion below, recommend that any proposed increases in bank capital be revisited to ensure:  A more appropriate level of risk tolerance is chosen;  A robust and independent cost-benefit analysis is undertaken; and  Transitional arrangements allow for a more measured, gradual pace of change, thereby easing costs of getting to desired new levels for bank capital. These three proposals were consulted on and unanimously supported by those HortNZ members and affiliated organisations who responded to the consultation. Discussion 6. For New Zealand horticulture is an international business. Our main earnings are through exports which are underpinned by domestic supply feeding New Zealand. Therefore, our competitiveness in our offshore markets is affected by

Page 3 of 3 the cost of capital in New Zealand. To remain competitive horticulture needs to embrace new technology, robotics and artificial intelligence, and innovate through the development of new varieties and new growing techniques. This all requires significant capital investment. We submit that any tightening of capital in New Zealand will result in capital investment programmes needed to support our international competitiveness being put at risk, resulting in our produce being less competitive and reducing returns. 7. We submit that there needs to be a balance between the costs and benefits of a particular risk appetite. We note that risk can never be completely eliminated, even at great cost, and there will be a point where the added cost of reducing risk begins to outweigh the added benefits of the reduced risk. If the banking system has a margin which is 0.2% - 1.0% higher than the underlying default rate, then all you are doing is adding a tax on the underlying economy. We do not believe that this is an appropriate way to cover a 1 in 200-year event when in practice it is unlikely to be enough in such an event. 8. Banks must be incentivised to operate prudently and to maintain their viability. We submit that any regulatory change should not lessen or weaken the pressure on banks to perform. We submit that a 1 in 200-year “buffer” will not achieve the aims set out in the Review Paper. History shows us that we have, and will repeatedly see, banking crises and that New Zealand will experience bigger ones due our international vulnerability and to the ratio of debt growing. 9. We submit that the perverse effect of strict capital rules will be to protect the large established banks and stifle competition, leading to the possibility that New Zealand would miss out on new innovation in banking systems. 10. We believe that as long as New Zealand has independent monetary policy and the ability to use quantitative easing, future banking crises will be able to be managed. 11. In conclusion we submit that to avoid the taxpayer underwriting lending by banks the following is needed:  Enough real risk capital (equity or Tier 1 debt) that is able to absorb any losses prior to the government having to step in; and  A clear understanding of what happens if and when the government has to get involved. It is this that is usually missing, with the problem only occurring when the government on behalf of the taxpayer steps in to protect bank customers. 12. We support both Federated Farmers’ and Business NZ’s submissions.

From: Iain Parker To: Capital Review Subject: RBNZ Review of Capital Adequacy submission by Iain Parker Date: Monday, 29 April 2019 10:59:30 PM RBNZ Review of Capital Adequacy submission of Iain Parker

1. In opening, I refer the committee to Former Vice President of the World Bank and Nobel prize-winning economist Joseph Stiglitz who published this document January of 2018; MACRO-ECONOMIC MANAGEMENT IN AN ELECTRONIC CREDIT/FINANCIAL SYSTEM https://www8.gsb.columbia.edu/faculty/jstiglitz/sites/jstiglitz/files/Macro￾Economic%20Management%20NBER.pdf?fbclid=IwAR2R￾B9teE59jy1IFq1c0kLxlN_XXvLRO3qk8BzqcdHEcIAab5d22wDqbpQ in which he said; “In a modern economy, banks don’t intermediate between “savers” and “investors,” as claimed in the standard textbook models. Banks effectively create credit out of thin air, backed by general confidence in government, including its ability and willingness to bail out the banks, which is based in part on its power to tax and borrow.” “While the modern financial system based on fiat money doesn’t suffer from the vagaries of gold discoveries, it has sometimes suffered from something else: volatility in the creation of money and credit by the banking system, giving rise to the booms and busts that have characterized the capitalist system.” “Much macro-economic instability is associated with instability in credit creation and in the fraction allocated to newly produced goods and services. The paper also explains how, in an open economy, in a system of electronic money, credit auctions combined with trade chits might enable the control of net exports, again enhancing macro-stability. Finally, we explain how under a system of electronic money, the rents that are currently associated with credit creation and

that arise from bank franchises—that constitute a form of appropriation of the returns from trust in the government and its ability and willingness to bail-out banks in the event of a crisis or bank run— could be appropriated by the government to a greater degree than at present.” End quote This document by Joseph Stiglitz is a very recent paper that supports my long￾held well-documented views on the cause and fix of money system funding structure instability. 2) My submission questions the efficacy of this RBNZ capital adequacy review, along with that of the Open Bank Resolution bail-in plan already on the law books, as a guard against the societal chaos of an old fashioned run on a bank, due to its lack of acknowledgment that the present credit liquidity facilities of New Zealand financial infrastructure is far from old fashioned. This process has ignored the model that has come to dominate banking in the money system funding structures of the Western world in modern times. Which has evolved into a hierarchal chain of institutions who refinance with each other at a cheaper rate of interest than that which they then on-lend. With those at the top bestowed with the extraordinary privilege of credit creation with only those lower down, normally at the domestic rather than cross border or government lending level, actually intermediating loans of customers savings of currency originated by the higher up credit creation. The basis of the model is no longer one of what was referred to as a fractional reserve anchor model but has evolved into one of prudential reserve accrual accounting given the official name of Dynamic Stochastic General Equilibrium, in which all potential physical and human resource capital is valued by actuarial accountants (Valuers) then used as the collateral for the base credit of money. If more base credit is issued than the sustainable natural resource collateral to redeem the base credit (Long-term Equilibrium) physically exists, inflation, asset bubbles, and societal wealth inequality follow.

The base credit becomes counterfeit beyond the fundamentals of money as a system to the advantage of those doing the counterfeiting and the detriment of humanity and the biosphere upon which it depends. The more members a money system has the more complex becomes the accountancy of the senior most balance sheet of currency originating base credit versus long-term sustainable natural resource equilibrium and currency stability. The monitoring of Equilibrium of the present system that has become interconnected throughout the Western world has been appointed to a small group of accountancy firms and financial rating agencies. This balance is even primary to another long-running debate of the ethics of interest being charged upon all base credit no matter if it is being used by a government to develop the resources it has at hand for the greater common good of its citizenship, such as the bare, non-consumer choice, essentials of life or large strategic infrastructure of the economy. I believe most people, including most academics, would be surprised to know it is presently large privately owned institutions that have been bestowed the extraordinary privilege of computer entry base credit, that sit one layer above most Western governments in the discount interest chain, including New Zealand, as evidenced in this document from the RBNZ Nov 2015 Financial Stability Report; Implications of global liquidity developments for New Zealand http://www.rbnz.govt.nz/financial-stability/financial-stability-report/fsr2015- 11/implications-of-global-liquidity-developments-for-new-zealand "There are three key channels through which New Zealand could be affected by declining market liquidity: the impact on New Zealand banks’ funding markets; the impact on short-term interest rates and monetary policy implementation; and the impact on the New Zealand government bond market.

New Zealand banks fund a significant proportion of their balance sheets by accessing offshore wholesale debt markets. They do this by borrowing in foreign currency, then ‘swapping’ this back into NZD. Conditions in global financial markets are therefore an important determinant of New Zealand bank funding. New Zealand banks tend to focus on the primary market (new issues) rather than the secondary market for debt. Hence, funding liquidity is of more immediate importance than market liquidity. Funding liquidity refers to the ability of the banks to raise debt as required at a reasonable cost. Reserve Bank discussions with bank treasurers suggest that funding liquidity conditions have deteriorated somewhat in 2015, owing largely to greater market volatility caused by events such as the Greek crisis mid-year and recent turbulence tied to China. New Zealand banks typically use market makers to help facilitate the foreign currency swap leg involved in borrowing from offshore. Market makers take the other side of the transaction with New Zealand banks (providing NZD in exchange for foreign currency that the banks have raised), while charging a spread. This spread has widened as costs have increased for the institutions providing these market making services for the reasons described above. Overall, the cost increases have been manageable thus far, but this highlights the flow-on effects of changes in market liquidity to New Zealand entities seeking offshore funding." end The privately owned institutions are touted as receiving fair remuneration for their expertise in keeping the money system in long-term equilibrium and cover their business costs. The private bank owners claim governments cannot be trusted to administer the base credit of money. New Zealand predominately borrows the base credit of our money from Western institutions for ease of convertibility into Western currencies that some companies stipulate they will only accept as payment for goods and services along with assurances of the protection of wider Western financial regulatory protection agencies. This is represented in a contract between governments and the privately owned

financial institutions referred to in New Zealand as the Policy Target Agreement, promising that they as their part of the deal will deliver stable inflation and price stability. In New Zealand, it amounts to a public-private partnership of which the major intersections of that partnership happen within the State Services Commission and the New Zealand Debt Management Office, a private institution that operates under the umbrella of New Zealand Treasury. I contend the average farm or residential property price versus average income from the normal course of business ratio has made it clear the private banking sector has not delivered its part of the contract. Assisted by the fact that the land portion of property was removed from the consumer price index in 1999. Which has, in turn, made a mockery of the Interest Rate Inflation Targeting regime as admitted by many such as former IMF Cheif Economist Olivier Blanchard when referring to the failure of the great moderation. 3) Most of the foreign institutions from which the base credit of New Zealand money is borrowed have been fined hundreds of billions of dollars for counterfeit credit based frauds they committed causing the 2008 global financial crisis, with barely any of the perpetrators being jailed. Most of the major accountancy houses and rating agencies were heavily fined for having been taking bribes in return for giving covering legitimacy to what former FBI forensic accountant William B Black termed Accountancy Control Frauds. Which is when the owners and executives of market listed companies use fraud to falsify profits to gain from stock price related remuneration packages. Selling out and departing with their proceeds of crime before the fraud becomes obvious leaving the accountability with the nonhuman entity company shell. In this regard, the massive increase in corporate debt bonds within New Zealand then being used for share buybacks rather than productive investment just screams out with very concerning residual current account balance problems with accountancy control fraud tendencies?

   Back in the 1980s William B Black prosecuted and jailed many hundreds within the private banking sector for committing the same crimes many essentially got away with 2008. William B Black says the 2008 GFC related fines amounted to a pittance of the proceeds of crime the perpetrators gained and the chances of it happening again is more 'when than if' as the necessary meaningful reform of the system has not happened. With that in mind I contend this investigation of safeguards needs to extend far beyond capital adequacy requirements, tenure of share ownership or concerns of old fashioned runs on banks, into an examination of the creditworthiness of credit entering our economy and the origins of the purchasing power of private equity funds now roaming the globe seeking rent seeking hard assets, to ensure they have not been criminally gained. The latest review of this process released by the RBNZ making mention of contingency credit lines being used as collateral for loans and preference share deals, along with a recent decision to not go ahead with the full privatisation of our nations electronic settlement system in the national interest, gives me hope the RBNZ has already begun to broaden its thinking on these matters; Capital Review Paper 4: How much capital is enough? 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? revision=2d386b12-1159-483e-9b53-34ecfebed91a&la=en 4) In recent decades I contend there has been a terrible level of odious, destabilising, wealth inequality inducing, predatory lending of counterfeit credit within the money system funding structures of the world, of which the victims deserve compensatory debt relief that has not been forthcoming and the measures being mentioned in this review will not deliver, let alone protect them from into the future.

I refer the committee back to the January 2018 published Joseph Stiglitz document in full with which I opened and suggest they investigate thoroughly the recent money system funding structure reforms implemented by Japan, with which we still trade and accept their currency, as written about often by the impeccably credentialed Adair Turner, in a hope they may continue to broaden their thinking even further in the face of empirical evidence of the highest credibility. Yours sincerely Iain Parker Concerned citizen

From: Iain Parker To: Capital Review Subject: RBNZ Review of Capital Adequacy submission by Iain Parker Date: Monday, 29 April 2019 10:59:30 PM RBNZ Review of Capital Adequacy submission of Iain Parker

1. In opening, I refer the committee to Former Vice President of the World Bank and Nobel prize-winning economist Joseph Stiglitz who published this document January of 2018; MACRO-ECONOMIC MANAGEMENT IN AN ELECTRONIC CREDIT/FINANCIAL SYSTEM https://www8.gsb.columbia.edu/faculty/jstiglitz/sites/jstiglitz/files/Macro￾Economic%20Management%20NBER.pdf?fbclid=IwAR2R￾B9teE59jy1IFq1c0kLxlN_XXvLRO3qk8BzqcdHEcIAab5d22wDqbpQ in which he said; “In a modern economy, banks don’t intermediate between “savers” and “investors,” as claimed in the standard textbook models. Banks effectively create credit out of thin air, backed by general confidence in government, including its ability and willingness to bail out the banks, which is based in part on its power to tax and borrow.” “While the modern financial system based on fiat money doesn’t suffer from the vagaries of gold discoveries, it has sometimes suffered from something else: volatility in the creation of money and credit by the banking system, giving rise to the booms and busts that have characterized the capitalist system.” “Much macro-economic instability is associated with instability in credit creation and in the fraction allocated to newly produced goods and services. The paper also explains how, in an open economy, in a system of electronic money, credit auctions combined with trade chits might enable the control of net exports, again enhancing macro-stability. Finally, we explain how under a system of electronic money, the rents that are currently associated with credit creation and

that arise from bank franchises—that constitute a form of appropriation of the returns from trust in the government and its ability and willingness to bail-out banks in the event of a crisis or bank run— could be appropriated by the government to a greater degree than at present.” End quote This document by Joseph Stiglitz is a very recent paper that supports my long￾held well-documented views on the cause and fix of money system funding structure instability. 2) My submission questions the efficacy of this RBNZ capital adequacy review, along with that of the Open Bank Resolution bail-in plan already on the law books, as a guard against the societal chaos of an old fashioned run on a bank, due to its lack of acknowledgment that the present credit liquidity facilities of New Zealand financial infrastructure is far from old fashioned. This process has ignored the model that has come to dominate banking in the money system funding structures of the Western world in modern times. Which has evolved into a hierarchal chain of institutions who refinance with each other at a cheaper rate of interest than that which they then on-lend. With those at the top bestowed with the extraordinary privilege of credit creation with only those lower down, normally at the domestic rather than cross border or government lending level, actually intermediating loans of customers savings of currency originated by the higher up credit creation. The basis of the model is no longer one of what was referred to as a fractional reserve anchor model but has evolved into one of prudential reserve accrual accounting given the official name of Dynamic Stochastic General Equilibrium, in which all potential physical and human resource capital is valued by actuarial accountants (Valuers) then used as the collateral for the base credit of money. If more base credit is issued than the sustainable natural resource collateral to redeem the base credit (Long-term Equilibrium) physically exists, inflation, asset bubbles, and societal wealth inequality follow.

The base credit becomes counterfeit beyond the fundamentals of money as a system to the advantage of those doing the counterfeiting and the detriment of humanity and the biosphere upon which it depends. The more members a money system has the more complex becomes the accountancy of the senior most balance sheet of currency originating base credit versus long-term sustainable natural resource equilibrium and currency stability. The monitoring of Equilibrium of the present system that has become interconnected throughout the Western world has been appointed to a small group of accountancy firms and financial rating agencies. This balance is even primary to another long-running debate of the ethics of interest being charged upon all base credit no matter if it is being used by a government to develop the resources it has at hand for the greater common good of its citizenship, such as the bare, non-consumer choice, essentials of life or large strategic infrastructure of the economy. I believe most people, including most academics, would be surprised to know it is presently large privately owned institutions that have been bestowed the extraordinary privilege of computer entry base credit, that sit one layer above most Western governments in the discount interest chain, including New Zealand, as evidenced in this document from the RBNZ Nov 2015 Financial Stability Report; Implications of global liquidity developments for New Zealand http://www.rbnz.govt.nz/financial-stability/financial-stability-report/fsr2015- 11/implications-of-global-liquidity-developments-for-new-zealand "There are three key channels through which New Zealand could be affected by declining market liquidity: the impact on New Zealand banks’ funding markets; the impact on short-term interest rates and monetary policy implementation; and the impact on the New Zealand government bond market.

New Zealand banks fund a significant proportion of their balance sheets by accessing offshore wholesale debt markets. They do this by borrowing in foreign currency, then ‘swapping’ this back into NZD. Conditions in global financial markets are therefore an important determinant of New Zealand bank funding. New Zealand banks tend to focus on the primary market (new issues) rather than the secondary market for debt. Hence, funding liquidity is of more immediate importance than market liquidity. Funding liquidity refers to the ability of the banks to raise debt as required at a reasonable cost. Reserve Bank discussions with bank treasurers suggest that funding liquidity conditions have deteriorated somewhat in 2015, owing largely to greater market volatility caused by events such as the Greek crisis mid-year and recent turbulence tied to China. New Zealand banks typically use market makers to help facilitate the foreign currency swap leg involved in borrowing from offshore. Market makers take the other side of the transaction with New Zealand banks (providing NZD in exchange for foreign currency that the banks have raised), while charging a spread. This spread has widened as costs have increased for the institutions providing these market making services for the reasons described above. Overall, the cost increases have been manageable thus far, but this highlights the flow-on effects of changes in market liquidity to New Zealand entities seeking offshore funding." end The privately owned institutions are touted as receiving fair remuneration for their expertise in keeping the money system in long-term equilibrium and cover their business costs. The private bank owners claim governments cannot be trusted to administer the base credit of money. New Zealand predominately borrows the base credit of our money from Western institutions for ease of convertibility into Western currencies that some companies stipulate they will only accept as payment for goods and services along with assurances of the protection of wider Western financial regulatory protection agencies. This is represented in a contract between governments and the privately owned

financial institutions referred to in New Zealand as the Policy Target Agreement, promising that they as their part of the deal will deliver stable inflation and price stability. In New Zealand, it amounts to a public-private partnership of which the major intersections of that partnership happen within the State Services Commission and the New Zealand Debt Management Office, a private institution that operates under the umbrella of New Zealand Treasury. I contend the average farm or residential property price versus average income from the normal course of business ratio has made it clear the private banking sector has not delivered its part of the contract. Assisted by the fact that the land portion of property was removed from the consumer price index in 1999. Which has, in turn, made a mockery of the Interest Rate Inflation Targeting regime as admitted by many such as former IMF Cheif Economist Olivier Blanchard when referring to the failure of the great moderation. 3) Most of the foreign institutions from which the base credit of New Zealand money is borrowed have been fined hundreds of billions of dollars for counterfeit credit based frauds they committed causing the 2008 global financial crisis, with barely any of the perpetrators being jailed. Most of the major accountancy houses and rating agencies were heavily fined for having been taking bribes in return for giving covering legitimacy to what former FBI forensic accountant William B Black termed Accountancy Control Frauds. Which is when the owners and executives of market listed companies use fraud to falsify profits to gain from stock price related remuneration packages. Selling out and departing with their proceeds of crime before the fraud becomes obvious leaving the accountability with the nonhuman entity company shell. In this regard, the massive increase in corporate debt bonds within New Zealand then being used for share buybacks rather than productive investment just screams out with very concerning residual current account balance problems with accountancy control fraud tendencies?

   Back in the 1980s William B Black prosecuted and jailed many hundreds within the private banking sector for committing the same crimes many essentially got away with 2008. William B Black says the 2008 GFC related fines amounted to a pittance of the proceeds of crime the perpetrators gained and the chances of it happening again is more 'when than if' as the necessary meaningful reform of the system has not happened. With that in mind I contend this investigation of safeguards needs to extend far beyond capital adequacy requirements, tenure of share ownership or concerns of old fashioned runs on banks, into an examination of the creditworthiness of credit entering our economy and the origins of the purchasing power of private equity funds now roaming the globe seeking rent seeking hard assets, to ensure they have not been criminally gained. The latest review of this process released by the RBNZ making mention of contingency credit lines being used as collateral for loans and preference share deals, along with a recent decision to not go ahead with the full privatisation of our nations electronic settlement system in the national interest, gives me hope the RBNZ has already begun to broaden its thinking on these matters; Capital Review Paper 4: How much capital is enough? 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? revision=2d386b12-1159-483e-9b53-34ecfebed91a&la=en 4) In recent decades I contend there has been a terrible level of odious, destabilising, wealth inequality inducing, predatory lending of counterfeit credit within the money system funding structures of the world, of which the victims deserve compensatory debt relief that has not been forthcoming and the measures being mentioned in this review will not deliver, let alone protect them from into the future.

I refer the committee back to the January 2018 published Joseph Stiglitz document in full with which I opened and suggest they investigate thoroughly the recent money system funding structure reforms implemented by Japan, with which we still trade and accept their currency, as written about often by the impeccably credentialed Adair Turner, in a hope they may continue to broaden their thinking even further in the face of empirical evidence of the highest credibility. Yours sincerely Iain Parker Concerned citizen

From: Ian Ross Harrison To: Capital Review Subject: Bank capital review Date: Monday, 13 May 2019 11:32:45 AM Attachments: RBNZ April 4 PDF.pdf How much capital is enough march PDF.docx Attached are three documents that are all part of my submission on the Bank capital review. They are: “Third time Lucky? ‘The 30 billion dollar whim’ ‘Tailrisk Pinocchio Awards' “ Ian Harrison

1 Draft - embargoed Third time lucky? A review of the Reserve Bank of New Zealand’s April 2019 bank capital ratio information paper May 2019

2 About Tailrisk economics Tailrisk economics is a Wellington economics consultancy. It specialises in the economics of low probability, high impact events including financial crises and natural disasters. Tailrisk economics also provides consulting services on:

1. The economics of financial regulation
2. Advanced capital adequacy modelling
3. Stress testing for large and small financial institutions
4. Regulatory compliance for financial institutions 5.General economics. Principal Ian Harrison (B.C.A. Hons. V.U.W., Master of Public Policy SAIS Johns Hopkins) has worked with the Reserve Bank of New Zealand, the World Bank, the International Monetary Fund and the Bank for International Settlements. In his time at the Reserve Bank Ian played a central role in developing an analytical approach to financial system risk issues. Contact: Ian Harrison – Principal Tailrisk economics e-mail:harrisonian52@gmail.com Ph: 022 175 3669

3 Third time lucky? ‘A well-run bank needs no capital. No amount of capital will rescue a badly run bank’ Walter Baghot Part one: Introduction On 3 April 2019 the Reserve Bank of New Zealand released another bank capital review background paper. This was the third1 in a series that offered differing explanations of the analysis behind the Bank’s decision to substantially increase bank capital. We suspect that the latest paper is, in part, a response to the criticisms raised in our paper ‘The 30 billion dollar whim’ (TBW). In particular, the Bank may have been sensitive to the suggestion that the capital increase was decided on a whim. Faced with a choice between a 1:100 probability of a ‘banking crisis’, which wouldn’t have required a capital increase, and 1:200, the Governor, went for 1:200, because it was ‘more conservative’. The Bank now wants to show that there really was more to it than that, even if the thinking and analysis wasn’t written in a formal paper that went to decision makers at the time. At 60 pages the current information paper may be the consultation paper that the Bank wished it had written in the first place. The fact that the Bank has had to produce a third paper, in itself, points to problems with the Bank’s policy making processes. The time from the concept of the ‘risk tolerance’ approach being approved, to a decision on the numbers, was only a matter of weeks and the analysis was never subject to external review. There were serious holes in the analysis which forced the Bank to do a lot of back￾filling, but rather than improving, the quality of the analysis has got worse. The purpose of this paper to assess the quality of the analysis and advice that the Bank is now saying it relied on. Our response is not an easy read. The Bank’s 1 On 14 April 2019 the Bank replaced the 3 April paper with a new version, but did not announce that it had done so. We have not checked how the new version differed from the original other than noting that a sign in one of the equations had been changed, as was explained in a footnote.

4 information paper is lengthy, covers a lot of ground, and is sometimes convoluted and inconsistent. In some places the standard of the documentation is poor. It takes time to untangled what is said, and, importantly, to sometimes comment on what is not said. Some of our analysis will only be accessible to risk wonks, but hopefully a wider audience can understand the gist of what is being argued in most of the paper. The paper is also directed at the Reserve Bank and will be a component of our consultation submission. In its introduction the Bank describes the purposes of the information paper: • To outline the analytical framework that underpins the banks analysis • Show how this framework leads directtly to the policy objective that has been defined by the Reserve bank • Describes the information and analysis that has been considered by the Bank in the context of the recent capital proposal Amongst other things, the new information paper offers a defense of the ‘risk tolerance’ framework; introduces the wider social costs of financial crises as a key driver; tries to buttress the case for the high GDP cost of financial crises; and provides yet another explanation of its inputs into the New Zealand modeling that purports to support the 1:200 target. Importantly it presents, for the first time, the results of a New Zealand optimal capital model. So the Bank is no longer solely reliant on overseas modelling results to support its claim that the 16 percent capital ratio meets the ‘efficiency’ test. However, the Bank still didn’t seriously engage on the following critical issues. • The need to adjust for the difference between New Zealand and foreign capital calculations when using foreign data on the relationship between capital and the probability of a banking crisis. • The need to consider the use of the Open Bank Resolution (OBR) option, which is a partial substitute for capital, as part of the capital review process. • The need to consider the impact of foreign ownership of New Zealand banks on the probability of a crisis. • The need to take into account foreign ownership on the cost of additional capital. The Bank has only considered the impact of interest rate increases on economic output. It has ignored the fact that there will be a transfer to foreign owners because of higher lending rates/or lower deposit rates.

5 • The need to explain the gap between its assessment of the ‘soundness’ of the New Zealand financial system and that implied by the rating agencies’ assessments and the Basel advanced model results. • The need to explain why the Bank now considers the New Zealand financial system is unsound, when it had determined that it was sound in fifteen years of financial stability reviews. This review of the Bank’s paper primarily looks at the following issues. Part three: Why was there a need to review the level of bank capital? Part four: The definition of a banking crisis. Part five: The risk tolerance framework Part six: The social costs of banking crises Part seven: The Bank’s optimal capital model Part eight: The Banks inputs into its optimal capital model Part nine: Capital policy and fiscal risk Part ten: Comparing New Zealand and foreign bank capital ratios Our key conclusions are set out in the next section. Part two: Key Conclusions

1. Capital increases unnecessary The Bank has failed to support its case for a substantial capital increase in the information document. The best evidence and logical analysis shows reasonably strongly that increasing banks’ capital ratios will reduce welfare. We stand by our previous assessment that the costs could be very large. Estimates of the net present value costs in the tens of billions would not be alarmist.

2. Risk tolerance approach a backward step The risk tolerance approach is not an advance in thinking about bank capital ratios. It tends to muddle the issues and can, conceptually, result in suboptimal decision making. Other supervisors have similar mandates to the Reserve Bank’s, but none have attempted to quantify it, and define ‘soundness’ in terms of the probability of a financial or banking crisis. Bank

6 crisis is too subjective a notion to be a useful hard metric for bank capital policy. The Bank is trying to solve ‘a problem’ of its own making. On any reasonable assessment the banking system is sound. We do not need the Reserve Bank to ‘make New Zealand sound again’. 3. Modelling analysis is embarrassingly bad There has been a corrosion of the quality of the Bank’s policy analysis. Some of the analysis of the inputs into the capital model is an embarrassment for New Zealand and a risk to the Bank’s credibility. APRA, which can understand the analytics, must be worried about the quality of the analytics decision making in an institution they may have to work with if there is a financial crisis some time in the future. 4. Bank missed a double counting in the capital requirement The Bank missed the fact that they have already increased bank capital by 20 per cent by requiring advanced bank capital to be 90 percent of that required under the standardised approach. Even if the Bank’s analytical modeling of the optimal capital ratio was robust (which it definitely is not) it should be wound back by about a third to correct for this double counting. 5. Impact of foreign ownership continues to be ignored The Bank has continued to ignore foreign ownership of the New Zealand banking system. It has ignored: the possibility that Australian owned subsidiaries will be sometimes supported by their parents, reducing the probability of a crisis; that there is little point in a subsidiary having a higher capital ratio than its parent; and the cost to New Zealand of increased profits to foreign owners. 6. Economic cost of crisis substantially overstated The direct economic costs of banking crises have been grossly overstated. The Bank’s preferred estimate appears to be 63 percent of GDP. A more realistic assessment of the marginal cost of a banking crisis, for New Zealand as opposed to the underlying economic shock, would be no more than 10 percent of GDP. 7. Misrepresentation of the social costs of crises The Bank has grossly misrepresented the literature it extensively quoted from, on the social costs and longevity of banking crises. The World Bank and the UN did not say that financial crisis have long lasting effects as the Bank claimed. The relevant message from the papers the Bank quoted from

7 is that the social costs in any economic downturn are substantially mitigated in countries, which, like New Zealand, have robust social safety nets. We found no evidence of long lasting ‘wider social costs’ in some relevant New Zealand data. Suicide rates, divorce rates and crime rates did not deteriorate during the GFC recession. 8. Fiscal risks benefits overstated Higher capital will have a limited impact on governments’ fiscal risks, which are already limited and manageable. Higher capital may not reduce governments’ gross fiscal costs at all if a government feels obliged to top up a banks’ capital to the new higher level after a crisis. Anything less could mean the banking system would continue to be ‘unsound’. Part three: Why review the level of bank capital now? One of the questions asked at a public seminar on the TBW was ‘why was the Bank reviewing the level of bank capital’. This was a good question. There didn’t appear to be an obvious problem. • New Zealand had adopted Basel III and bank capital ratios have already increased substantially since the global financial crisis. • There seems to be no obvious move by the international regulatory community towards a further increase in tier one capital ratios. • Supervisors are now using stress test as the primary tool to assess banks’ capital adequacy and implicitly their ‘soundness’, and in New Zealand, banks have ‘passed’ some vigorous stress tests; • The banks’ capital ratios are in line with the ‘optimal’ capital ratio of 13 percent presented in the Reserve Bank’ Regulatory Impact statement that supported the adoption of the Basel III requirements • On a like-for-like comparison New Zealand bank’s capital ratios are, at the least, broadly in line with international comparators. So what is new, and what is the problem now? The Bank’s response on this point has been that they have new information. By that they presumably meant new analytical evidence in the ‘literature’. While the Australian proposal to adopt Basel a ‘conventional’ approach on Total Loss

8 Absorbing Capital (TLAC) is new, the Bank did not seem to be aware of APRA’s proposals and did not engage with the TLAC issue in the capital review. There has been a host of new studies since 2012 (in particular a number of analysts produced optimal capital estimates with widely varying results), but the sum of it does not obviously point to the need for higher capital ratios for New Zealand banks.2 The Bank has focused on a few studies that seemed to support higher capital (on close inspection we found that they didn’t), but equally there were many others studies that pointed in the other direction. In particular there was the post 2012 empirical work that suggested that the Modigliani Millar (MM) offset was much less than complete, and that the cost of higher capital could be much higher than the Bank had previously assumed. The TUATARA model assumed an 85 percent MM offset, and that capital was relatively cheap. However, the ‘cheap capital’ perspective was still in place when the initial consultation document was released in March 2017. The Bank said that its starting point was the ‘big capital’ approach, which was premised of a big MM offset effect. At some subsequent point in the process, however, that view seems to have changed, and the Bank has settled on a 50 percent offset assumption. That could have been the end of the capital ratio part of the review, because it is much harder to justify a large capital increase with that assumption. But by that stage the process had its own momentum and the ambition for much higher capital requirements was not tempered. It is hard at the end of a lengthy review process to come to the conclusion that there is just not enough robust evidence to justify a change. So something had to be done and the mission was to find a rationale for higher capital. Evidence and perspectives suggested a different approach were probably not seriously considered, or brushed aside. Undue evidence was placed on a few academic papers, which focused on cross-country statistical analysis using very simple macro-models, which should have been taken with a grain of salt. The Bank appears to have only a very limited familiarity with the broader ‘literature’ on banks experiences in the GFC, and in earlier banking crises. Only one official report (on HBOS) is referenced in the literature survey. The Bank does not really understand what happened and why.

2 For example the Bank’s review of the literature missed the following paper BANK CAPITAL REDUX: SOLVENCY, LIQUIDITY, AND CRISIS Òscar Jordà Björn Richter Moritz Schularick Alan M. Taylor NBER Working Paper 23287. It concludes Higher capital ratios are unlikely to prevent a financial crisis. This is empirically true both for the entire history of advanced economies between 1870 and 2013 and for the post-WW2 period, and holds both within and between countries. We reach this startling conclusion using newly collected data on the liability side of banks’ balance sheets in 17 countries. A solvency indicator, the capital ratio has no value as a crisis predictor

9 What is remarkable here is that what is now presented as the ‘problem’, that the New Zealand banking system is ‘unsound’, only emerged at the very end of a process that proceeded, albeit in fits and starts, for more than three years. The solution (a 16 percent CET1 capital ratio) was found a matter of weeks later. But this is a ‘problem’ of the Bank’s own making. No one else thought that the New Zealand banking system was ‘unsound’. Another motivator could be the Bank’s ‘light handed’ approach to supervision. If capital rather than ‘hands-on supervision is the Bank’s primary instrument, it may feel needs to be especially conservative with the one instrument that it does control. It is probably time to review this approach. In practice the Bank has become increasing ‘heavy handed’, and our the analysis of the Bank’s capital proposals strongly suggest, the Bank’s ‘more capital good’ approach, can be very expensive, and is probably much less effective than good regulation in preventing crises. The Bank’s approach also impacts negatively on rating agency assessments. Standard and Poor’s has a low assessment of the New Zealand institutional framework (compared to Australia’s), because S&P does not rate the Reserve Bank as a supervisor. A more conventional approach, which would include onsite inspections, need not be costly. APRA already is the substantive regulator in this respect, because it does conduct onsite inspections of the subsidiaries of the Australian banks. It would simply be a matter of upgrading and formalising that relationship, and extending the approach to other banks, having regard to the lower systemic risk posed by smaller banks. And finally there is the general tendency of regulators to overstate the importance of their ‘mission’. Bureaucratic incentives lean to an excessively risk adverse approach. In a relatively small institution like the Reserve Bank, which has almost no external constraints on its decision making, the prejudices and predilections of a few individuals can disproportionately affect outcomes.

Part four: What is a banking crisis?

10 The centrepiece of the Bank’s risk tolerance framework is the concept of a ‘banking crisis’. The policy objective is to reduce the probability that a banking crisis occurs to 1:200. But if a ‘banking crisis’ cannot be defined with some precision then it is not possible to calculate the risk that it will occur. This obviously begs the question, what is a ‘banking crisis’?. Crisis means different things to different people, and researchers have used different definitions over the years. On this point the Bank just says that the term banking crisis is ‘defined in the literature’, citing work by IMF analysts Laaeven and Valcenia (2012) and a paper by Romer and Romer (2015) as sources. The Laaven and Valcenia definition has been widely adopted by researchers and is used, in part, because it comes with a readily accessible list of crises. A stress event, then, becomes a banking crisis if it is on the IMF list. The citation of the Romer and Romer paper gives the impression that it is somehow supportive of the idea that there is an acadenmic consensus on what constitutes a crisis. This is almost the opposite of what that paper has to say. Romer and Romer’s contribution is that reseach using the Laaven definition leads to erroneous results because a crisis either occurs, with significant consequences, or it does not, with no consequences. Using a more finely differnentiated measure of financial stress (a 1-15 index) they demontrate that some of the conclusions of the conventional analysis , for example that the effects of financial crises are long lasting, fall away. The value of the Romer and Romer contribution is that it reminds us that a ‘banking crisis’, is not a discrete, easily idenfified event like a plane crash, but is a mostly subjectively determined point on a continuum of severity. Returning to the Laaven and Valencia definition of a banking crisis, a ‘banking crisis’ occurs if three of the following conditions are met.

1. deposit freezes and/or bank holidays;
2. significant bank nationalisations;
3. bank restructuring fiscal costs (at least 3 percent of GDP);
4. extensive liquidity support (at least 5 percent of deposits and liabilities to nonresidents);
5. significant guarantees put in place; and
6. significant asset purchases (at least 5 percent of GDP); Thus New Zealand did not have a banking crisis in the GFC because only two of the conditions (guarantees and liquidity support) were met. Out first comment here is that there is the obvious one that there is an arbitrariness about the banking crisis tests. For example, if the number of

11 conditions had been set at two, then then both Australia and New Zealand would have had a ‘banking crisis’ in 2008. Second, the bank restructuring fiscal cost test, which is often the determinative trigger, again, has been set somewhat arbitrarily. Applied to New Zealand it means a $9 billion government capital injection, with the prospect of getting most of it back when the shares are eventually sold. This would not be much of a crisis-like event for a New Zealand government starting from a very strong debt/GDP ratio position. If the Laaven fiscal trigger had been set not too much higher, then many identified GFC country crises would not have been crises. Third, several of the triggers are in themselves, not particularly important in economic terms; do not generate significant costs, and are not necessarily reflective of a deep underlying problem. For example, the guarantee of banks by the New Zealand government, and liquidity support from the Reserve Bank in 2008, did not directly cost anything, and were a response to external circumstances, rather than concerns about the solvency of New Zealand banks. Our fourth, and most important, point is that the ‘banking crisis’ metric is different in character, and is calibrated differently to the solvency test that is used in the Basel capital framework. The Basel framework is built around a test of capital adequacy against credit losses of eight percent of (risk weighted) loan exposures with a probability of 1:1000 of it being exceeded. As these things go this go, credit losses are a relatively objective test. The ‘banking crisis’ metric, on the other hand, is based on indicators of official actions. There is a rough link between credit losses and government fiscal injections, but it is only rough, because a government might choose to inject capital for other reasons, than just actual credit losses. It might want to maintain confidence in the financial system, clean up long standing banking sector structural issues, or be seen to be acting decisively in a ‘crisis’ situation. And the calibration of the trigger point is quite different. Government recapitalisations of three percent of GDP, can be a lesser percentage of banking system assets, so the trigger point is substantially lower than the trigger point in the Basel capital model. The upshot is that the Laaven definition is both more subjective (a predictor of official action) and captures a lot more events than the Basel solvency test would. So a banking system can have a 1:1000 probability of credit losses that will exceed its capital, and at the same time a 1:200 probability of having a ‘banking crisis’. Neither is necessarily right or wrong, they are just using different frameworks and trigger points

12 However, the critical implication of the lower threshold of the ‘banking crisis’ metric is that the costs associated with events that triggering the crisis threshold are much lower than those that breach the Basel model threshold. The man in the street tends to think of a ‘banking crisis’ as something horrendous and almost beyond the ability of a government to manage. The case of Ireland comes to mind. The reality is that most of the banking crisis events in high income countries in the GFC, were, on the credit loss facts, (as measured by proxies such as non￾performing loans) not especially fearsome, and the associated avoidable deadweight costs would have been corresponding moderate. This explains why the observed economic losses in the GFC recession was not materially higher in countries that had a banking crisis than in those that did not. In some sense all supervisors wish to reduce the likelihood of financial or banking crises to ‘acceptable’ levels. But because the notion of a banking crisis is subjective, and the triggers as much political as economic, no other supervisor, that we are aware of, has tried to formalise that objective and quantify it. Certainly no other supervisory has used the metric to decide whether their banking system is sound or unsound. To do so would risk being unnecessarily alarmist, and overly mechanistic in the execution of their supervisory duties. Instead supervisors have relied on the stress test as their primary public analytical test of bank ‘soundness’. On that basis of the New Zealand bank stress test, the New Zealand banking system should similarly be regarded as sound.

The key problem with the Bank’s analysis is that the Bank did not understand the difference between the ‘banking crisis’ definition and the Basel capital framework. It has tried to use the Basel model to generate banking crisis results. The Basel model, will deliver a 1:1000 outcome on reasonable input assumptions. But it can only be made to achieve a 1:200 outcome, by making extreme assumptions about the inputs. Unfortunately the Bank has gone down that particular rabbit hole, and has had to resort to some bizarre arguments to make the unworkable work, in an increasingly desperate attempt to shore up some unwise decision making. Part five: The risk tolerance framework This is a lengthy and complex discussion. Some readers might want to just rely on our summary, which reads as follows.

13 The risk tolerance approach does not solve the problems posed by complex, uncertain and often apparently contradictory data and analysis. Mostly it just repackages existing concepts in ways that are difficult, at first sight, to reconcile with more conventional approaches. At the end of the day it is just complicating way of the Bank saying: we intuitively know that more capital is good, and we know what is good for you. The risk tolerance explanation The Bank sets out a lengthy defense (about 13 pages) of its ‘risk tolerance’ framework. It is a much lengthier exposition than that set out in the paper ‘Risk appetite framework used to set capital requirments’ to the Financial system Oversight Committee, which sought approval for the new approach. The argument starts with the observation that there are two strands in the literature. The first is the analytical approached in the Basel model which is calibrated to a 1:1000 standard, at a 8 percent capital ratio (the rating agency assessments which suggest a similar standard are not mentioned). The second strand is foreign statistically based studies that look at of the relationship between financial crises and capital. At face value, at least for the countries reviewed, some of these studies appear to show much higher bank failure rates. The results from the two strands of analysis often depart quite radically from each other. It is thus inevitable that one strand of research will receive priority over the other. Our proposed risk appetite framework is a response to that challenge. We propose to first establish a level of capital below which we won’t go, based on our risk appetite, and then look to see whether there are opportunities to increase output and stability from this minimum capital level. What the Bank appears to be saying here is that there is a such a wide range of outcomes that the only way to cut through the apparently contradictory evidence is to take a radically new approach. Just set an arbitrary probability of crisis target. What the Bank did not do is attempt to reconcile the two strands of the analysis. If it had it would have found that the apparently high rate of banking crises was in part an artifact of the short data periods in many studies, and the particular way the GFC played out, which in many repects was not releavant to New Zealand. And it did not, as discussed above understand that The Basel and banking crisis lierature were using different conceptual frameworks and different calibrations. Nor did the Bank stand back and ask the question. If we don’t appear to have an obvious problem, and there are no obvious answers in the literature and in

14 statistical information, which have direct releance to the New Zealand banking system, then perhaps we should not be doing anything, at least unil we better understand the issues? However, the Bank pressed on. With respect to the banking crisis target the discussion switches from the level of capital to the concept of ‘soundness’. “We believe a reasonable interpretation of ‘soundness’ in the context of capital setting is to cap the probability of a crisis at 1% (or 0.5% if we wish to mirror approaches taken in insurance solvency modelling’. The discussion then moves to a description of the standard approach to optimal capital calculation in the literature, which establishes a relationship between the amount of capital and the probability of a banking crisis. The conventional approach is set out in their figure 1 which shows the optimal level of capital as the point where the marginal cost of capital is equal to the marginal benefit.

15 The Bank then reconfigures figure 1 relationship to produce its preferred depiction of the relationship in their figure 3 above. The level of capital is replaced by the words financial stability. The cost and benefit curve are replaced by a single net benefit curve, which is presented in aggregate rather than marginal terms. In itself this demonstrates nothing new. The Bank could just as easily have presented the issues in terms of the marginal cost and benefit curves. They could have placed a red dot, the current situation, as the Bank claims, to the left of the marginal cost and benefit intersection point, the green dot at the intersection and the blue dot at the right. Nor is the Bank’s picture new. Several studies have presented their results in the same way. However, the Bank puts considerable weight on what it thinks are its ‘advantages’ of its new picture. We have opted to illustrate the policy problem as outlined in Figure 3 for several reasons: · We believe it is more accessible to non-specialist audiences than the conventional marginal-based exposition. We wanted to have a genuine conversation with the public about capital policy, so we can reflect their risk preferences, and this required an accessible description of the policy problem. · This illustration shows a wider range of what believe are potentially justifiable outcomes than the conventional treatment. It becomes clear that having capital beyond the expected output maximising level would deliver more stability and this may ultimately be preferable (it depends on society’s attitude towards risk). How you depict a graphical representation of the issue is not fundamentally important, and largely a matter of taste. However, the Bank’s preferred approach

16 may have the advantage, from its perspective, of downplaying the cost side of the issue. It is netted off, so it is not visible. The claim that the representation is more accessible to non-specialist audience is somewhat fanciful. As is the claim that there will be a ‘genuine conversation with the public so their risk preferences can be reflected in the outcome’. How the Bank’s picture will help illicit the publics risk preferences is beyond us. In reality the general public, will, to the extent they engage with the process at all, be relying on where the Bank has placed the dots, and will be swayed by often misleading Bank rhetoric. We doubt that any non-specialist will truly understand what the Bank is doing. They may, however, be influenced by the way the Bank has packaged and promoted its policy. It is a’ win-win’. We get more stability at no cost. Who could be opposed to that? The problem is that the win-win is pure assertion, and the Bank airbrushes out the very real costs. Risk aversion We then get on to a discussion of risk aversion. The Bank establishes that people are generally risk adverse, which is uncontroversial. So obviously you would want to take account of risk aversion in your model. The foreign optimal capital models that the Bank surveyed and relied on, to some extent, in its earlier analysis, do not take account of risk aversion, though some allude to the issue. The Bank’s solution was the risk tolerance approach: to establish the risk level below which you will not go, to deal with risk, and then deal with the efficiency issue with models that ignores risk. The issue here is that if risk aversion should be included in the analysis, why didn’t the Bank use its TUATARA model that explicitly takes risk aversion into account. That model assumed that risk aversion could be captured by doubling the value of output saved when capital is increased and found that this increased the optimal capital ratio by about 2.5 percentage points. We suspect the reason is that having made a formal decision not to use the model back in 2016 (on the rather spurious grounds that it produced a range of results with different assumptions), the Bank forgot about it. When the it was pointed out in the TBW that there might a better way to address the risk aversion issue, using the TUATARA approach, than arbitrarily selecting a probability of crisis threshold, they had become too invested in their risk tolerance approach, with its attractive,

17 from a public relations perspective, win-win outcome, and didn’t want to resile from their position. To justify this position there is a discussion that seeks to cast doubt on the TUATARA approach. Incorporating risk aversion in the analysis via utility curves requires making assumption about what value people place on avoiding a loss rather than making a gain. The evidence that might inform this assumption can be quite conflicting. In support the Bank cites a paper by O’Donoghoe3 . This paper is just a high level theoretical discussion that argues that you don’t necessarily need an expected utility approach to justify placing a higher weight on adverse outcomes. From a policy perspective we don’t have to be too precious about the theoretical underpinnings of the TUATARA approach. All you need is a rationale for placing a higher value on losses in a downturn (whether your conceptual foundation is expected utility or some model from behavioral economics doesn’t really matter) and some basis for setting the weighting. The TUATARA model used observed behaviour in the life insurance market to double weight downturn losses. The Bank then goes on. However the Reserve Bank is an agency with delegated authority (and obligations) to make such decisions on behalf of the public. We adopted a pragmatic position, acknowledging the importance of risk aversion but not adopting the utility curve approach. On the one hand, the evidence strongly suggests that society is not indifferent to risk, so it seems important to accommodate the possibility that society may prefer capital levels that deliver a great deal of stability even if it means some sacrifice of expected output. On the other hand, we accept the point that any modelled representation of society’s preferences depends on assumptions (the accuracy of which may be impossible to verify ex ante) and the results will be very sensitive to the assumptions made. The approach adopted by the Reserve Bank incorporates society’s risk tolerance in a simple way. We represent the costs and benefits of capital objectively (without any weightings) and incorporate risk aversion by aiming to cap the probability of a crisis at some predetermined level.

3 O’Donoghue, Ted and Jason Somerville (2018) Modelling Risk Aversion in Economics. Journal of Economic Perspectives Vol. 32 No. 2 Spring 2018.

18 Essentially what the Bank is saying is that an arbitrarily chosen target by the Governor is to be preferred to an attempt to incorporate an estimate of the publics preferences. But what we really have here are the Governor’s preferences, and there is no reason to assume that this leads to a reasonable result. The Governor will not be bearing the costs of his decisions, and has every incentive to be overly conservative. High capital ratios possibly buy a quieter life, and in the event something does happen the Governor can say that he took appropriate action in anticipation. And of course one of the highest capital ratios in the world gives the Governor bragging rights (‘mine is bigger than yours’) in supervisory circles. The Bank then summarises the advantages of its approach. In our view the approach is transparent, simple, pragmatic and consistent with the risk aversion literature. It is only transparent in the sense that the arbitrary target is announced. But it is extremely difficult to map the relationship between the target and the policy outcome. It is certainly not really simple. The Bank spends 13 pages just trying to explain the basic concepts. It is pragmatic in the sense that any arbitrary decision is pragmatic. But the approach is not certainly not consistent with the risk aversion literature. No literature supports plucking a number out of the air based on nothing more than a statement that 1:200 is more conservative than 1:100. The Bank then moves on.

The approach is to set a two-part policy goal. The first element of the policy goal is to set bank capital requirements with the aim of achieving a banking system that retains the market’s confidence in the face of large unexpected shocks (delivering ‘soundness’). Here the Bank is setting a very subjective target. The large unexpected shock is not specified, and the requirement that the markets confidence be maintained is a hugely subjective test. It will depend on the surrounding circumstances on how much capital will be ‘enough’. But the Bank then backs off At a minimum, maintaining the market’s confidence means the banking system remains solvent.

19 This test is more objective, but it will definitely not meet the ‘maintaining market confidence’ test. A bank that has lost all but one percentage point of its capital will almost certainly fail. In practice what the Bank has set up is the bank solvency test identical to that in the Basel Capital model. But it then uses the model to test the ‘banking crisis’ trigger, which almost certainly requires a decent buffer. A buffer could be put in, but then the model outputs would be substantially driven by the size of the buffer, and really wouldn’t be data driven. The problem is that how big a buffer will maintain confidence and avoid a ‘banking crisis’ is anyone’s guess. It is not readily amenable to estimation. In the event the Bank decided to take the zero buffer route, and doesn’t explain why it shifted from the two percentage point buffer it used in its January 25 paper. The Bank then moves to the efficiency objective.

The second element of the policy goal relates to ‘efficiency’. If, at the level of bank capital implied by the soundness objective, stability can be increased further with no loss of expected output, then bank capital can be increased beyond what society’s risk tolerance would require. This second ‘leg’ of the policy goal is akin to delivering a constrained maximisation of expected output - expected output is being maximised but this is conditional on achieving the stability objective. The Bank’s constrained optimization approach gets to the nub of the theoretical issue on risk aversion. In general, a constrained optimisation will underperform an unconstrained optimisation, if the constraint is arbitrary and unnecessary. If the ‘soundness’ constraint is set high then it will be higher than the optimal level that takes account of risk aversion. Equally, if the soundness target is set lower, and capital is set by ‘efficient’ level of capital, this capital level could be too low, because the ‘efficient’ level of capital does not take account of risk aversion at all. Finally, the Bank doesn’t seem to apply the risk tolerance model in practice. To explaining why the 1:200 target was switched for the initial 1:100 there is the following statement.4 An additional factor in the decision to adopt 1/200 was the preliminary finding from the early modelling work. This indicated that, for the range of input values being considered,

4 The Bank was obviously embarrassed by the lack of evidence of any substantive analysis supporting the change from a 1:100 to 1:200. The only information was the following footnote in a November 2018 document. ‘We considered two values for the cap on the probability of a crisis – 0.5% and 1% - but in the final analysis opted for the more conservative option 0.5%’ The analysis they refer to may be just an attempt at backfilling.

20 adopting a risk tolerance of 1/100 would typically lead to an inefficient result (i.e. adopting a lower risk tolerance than 1/100 would likely lead to both greater stability and increased expected output). If the Bank had stuck with its framework, the predetermined soundness target would have been left at 1:100, but the capital ratio would still be set at 16 percent on the basis (from their modeling) that it maximized expected output. The capital policy decision would have been the same, but it would just be a ‘win’ result rather than a win-win. The Bank doesn’t seem to understand its own framework. Other supporting arguments for the risk tolerance approach First, there is the use of probability targets in other regulatory models. The general insurers capital framework has a solvency test of 1:200. But this is a solvency target, not an ‘insurance industry crisis target’, whatever that might mean. The Basel model solvency test of 1:1000 is also mentioned. It is difficult to see why this test justifies the need for a conceptually different capital test based on the probability of a banking crisis. Banks already have one risk metric, which if desired can be pressed into service to demonstrate that the system is currently sound. Introducing a second quantitative test is unnecessarily and leads to confusion. And then there are the results of a comparative exercise. It is also important to note that a comparative exercise was undertaken in order to assess the impact of incorporating risk aversion in the policy goal. This was a review of the high level findings in the literature where the policy goal was defined simply in terms of maximising expected output. A summary of the findings is provided in Section 2.4 below. It is difficult to see how this comparative exercise could help. Just looking at a set of models that don’t account for risk aversion, and for the reasons outlining in TMW, can’t be easily applied to New Zealand, says nothing about the impact of incorporating risk aversion into the modelling.

21 Part six: The social costs of banking crises The Bank runs the line that the social cost of financial crises provides an additional and compelling argument for taking a more strongly risk averse approach to bank capital requirements. There is a large literature about the economic and social impacts of deep and prolonged recessions (such as are likely to arise in the event of a banking crises). A common theme in the literature is the harm to mental and physical health, family cohesion and community connectedness caused by the economic stress induced by a severe downturn – through unemployment, falling incomes, reduced savings and/or declining asset values. There is evidence of these impacts in both developed and developing countries although local circumstances can act to mitigate the effects. And It is worth considering examples of the available evidence in some detail, as this can provide insights into why agencies such as the World Health Organisation, the World Bank and the United Nations see the societal impacts of financial crises as being long-lived. The breakup of families, ill-health, reduced spending on healthcare and nutrition and societal unrest can all be expected to have enduring effects on society as a whole. The breakup of families, ill-health, reduced spending on healthcare and nutrition and societal unrest can all be expected to have enduring effects on society as a whole.There is evidence of these impacts in both developed and developing countries. The claim that the UN, the World Health Organisation and the World Bank believe that financial crises have long lasting effects on advanced countries with good social safety nets is highly misleading. The Reserve Bank has relied on just two documents: the UN’s 2011 report on the GFC , and a background paper for the World Bank’s 2014 World Development Report to make its argument. In fact neither report made any judgment about the longevity of the social effects of economic downturns, let alone financial crisis downturns, in high income countries. Indeed one of the key takeouts was that the social effects, in countries with robust social safety nets, is quite limted. For the greater part, the reports were concerned with the effect of the GFC on developing countries which do not have these safety nets. The UN report has over 200 references, but only a handfull refer to social impacts in advanced countries in financial crises.

22 However, the Bank then goes on makes liberal use of quotes from the reports to advance its case for New Zealand, when it was clear that the literature was substantially referring to less developed countries. The Bank also mentions a WHO report for support. Here is what the WHO actually said. Economic downturns result in smaller changes in the mental health of the population in countries with strong social safety nets (Fig. 4) (54). European data indicate that inequality in health does not necessarily widen during a recession in countries with good formal socialprotection. In Finland and Sweden (their banking crises) during a period of deep economic recession and a large increase in unemployment, inequality in health remained broadly unchanged and suicide rates diminished The Bank concludes its discussion with. We believe these impacts are likely to lead society to be relatively intolerant of banking crises. However, one aim of the consultation is to generate a public conversation, and prompt feedback, about this important issue. In the Bank’s long list of documents consulted in appendix A there is another World Bank report5 . The Bank did not cite it in its information paper, but it does claim to have read it. It has the following figure which summarises the impact of the GFC on key development indictors for a range of country groups. The high income industrial country group is on the right of each graph. In each case the indicators either improved or were stable over the GFC. The report concludes. Financial crises affect human development indicators but outcomes have not been uniform across countries. In advanced economies no deterioration has been identified in health and education indicators as a consequence of financial crises (our emphasis). Well￾functioning credit markets, well-established social protection programs, and a low opportunity-cost of attending school could explain this outcome.

5 Otker-Robe, Inci and Anca Maria Podpiera (2014) The social impact of financial crises, evidence from the global financial crisis. The World Bank Policy Research Working Paper 6703.

23

We also looked at four indicators of social wellbeing for New Zealand over the GFC. Although New Zealand did not, on the IMF definition, have a banking crisis, the downturn was just as severe, as it was, on average, for those countries that are identified as having a crisis. Our findings were: Life expectancy: No impact on the upward trend in life expectancy. Suicide rate: No impact. The relevant figure is shown below. Divorce rate : Rate per 1000 of existing marriages 11.27 in 2008, 9.72 in 2011. Crime: Theft and related offenses 141,000 in 2008, 137000 in 2010. While there are other social indicators that we have not considered, on this brief survey there is no obvious evidence that the recession had a serious impact on broader wellbeing measures and certainly none that the social effects of the recession was long lasting.

25 Some results In their figure 5. reproduced below, the Bank sets out some results. But these are not immediately clear, because the horizontal axis is not depicted in terms of the variable of interest, the CET1 capital ratio. Looking at the outputs the black line reaches a maximum (the most efficient level) beyond the 0.995 level (1:200 probability of crisis). We are just left to draw the conclusion that the efficient levelof capital is above the sound level and to assume that a 16 percent capital ratio is consistent with the efficient point.

26 What is clear from the figure is that altering just one of the assumptions (the cost of crisis) has a material impact on the ‘optimal’ probaility of crisis (and hence on the optimal capital ratio,) It changes to around 0.990. If we go to the Bank’s table 2 we see that this is equivalent to a 1.9 percent fall in the required (leverage) capital ratio. Problems with presentation of the results • There are no results for the Bank’s best estimates. As discuused below athe Bank avoids identifying its best estimates. Presenting an array of outcomes makes it difficult to trace the link between the analysis and the policy outcomes. • The results are all presented in terms of the leverage capital ratios, not the policy variable of interest the CET1 ratio. • It seems clear that the analysis was conducted without taking into account the increase in capital due to the requirement that advanced bank capital be 90 percent of the standised model capital. This will increase the starting leverage ratio by around 20 percent and reduce the benefit of marginal increases in capital. The Bank has effectively counted this increase twice • The presentaion of results is selective. In particular, there is no presentation of the results for the scenario with all of the Bank’s lowest estimates. By any reasonable assessment they should be the Bank’s best estimates and of the most interesting to the reader. The Bank of course will say that their approach considered all of the outcomes, so they did ‘take them into account’. This is just unartful obfuscation.

Part eight: The inputs into the Bank’s optimal capital model This part deals with the core of the analytical work that drives the Bank’s conclusions. There are three main inputs into their model

1. The cost of capital We do not have a major issue with the Bank’s core assumption on the MM offset that drives the interest rate increases. We do, however, a fundamental issue with the Bank’s failure to account for the increased

27 profits of foreign banks. This omission means that the cost of capital is undestated by at least half. We do not further discuss cost of capital issues in this paper. The reader should go to the TBW paper for more analysis. 2. The relationship between capital and the probability of a crisis. This is generated by the Basel model which ,in turn, has three inputs: • The probabilty of default (PD) • The loss given default (LGD • The autocorrelation input (R) 3. The cost of a banking crisis The probability of a crisis In any economic model the quality of the outputs depends on the robustness of the mdel inputs. it is a case of garbage in – garbage out. In the TBW we basically concluded that the analysis backing the Bank’s outputs in the 25 january paper were, in effect ,‘garbage’ – just a back filling execise to produce the right result. The Bank has reponded to our criticisms by producing some new inputs, justified by some new arguments, which if anything, are worse than those set out in the January 25 information paper. The Bank explains that it made the following key judgements in setting the ranges for inputs • Basing the ‘PD’ input on NPL ratios and/or impairment rates; • Using a simple average of historical New Zealand bank impairment data, not a value weighted average; • Referencing overseas experience when reviewing possible ‘PD’ input values; and • Reflecting observed relationships between house values and output, and contrasting New Zealand with overseas countries, when setting a range of values for correlation R. There is two other key isues that are not seriously addressed • The decision to ignore advanced modelling bank’s (and regulatory) inputs into the models that generate their capital requirements. The Bank has approved these inputs as being conservative best estimates. What they are now seem to be saying now is that the banks’ estimates and its own previous judgments are wrong, and wrong by very large margins. The Bank needs to explain why.

28 • They do not provide best estimates for the input values. They say:

Our approach was not to derive a single ‘best estimate’ of the relationship between capital and the probability of a crisis – and, in particular, the capital needed to cap the probability of a crisis at 1/200 - but to identify a range of reasonable estimates. This is not really a justification. It is conventional practice to produce and justify a best estimate, and then investigate the sensitity of the result to different input assumptions. The Bank’s approach allows it to fudge the connection between its calculations and conclusions. The conclusions are just ‘informed’ by the analysis, which could mean anything. This is not consistent with the Bank’s professed commitment to transparency and accountability. If and when the Bank gets around to producing a cost benefit analysis they will have to settle on their best estimates. They should have done it already. We now discuss the three model input estimates. Probability of default Here the Bank attempts to justify its use of non-performing and impairment rates rather than actual probability of default estimate. The Bank has good data from the advanced banks on their historical default rates (the average PD is about 1.2 percent in the bank disclosure documents), so why would it use non-performing loans, which as calculated by the Bank, overstate default rates by a factor of more two. Banks’ data does not include crisis events The Bank’s first argument is that the New Zealand banks’ historical PD data does not includes a ‘banking crisis’. So it is necessary to go to foreign data where there have been crises, and the most relevant data that is readily available is nonperforming loan data. This argument doesn’t stand up to scutiny. • The PD data produced by the banks does includes information from the GFC ,and to a varing extent, the 1989- 92 recession. While these recessions are not identified as ‘crises’ on the IMF definition, they were just as severe as the downturns in countries that did experience a crisis. The Cline (2016)6 data shows that the aggregate output loss in New

6 Cline W. (2016) ‘Benefits and Costs of Higher Capital Requirements for Banks.’ Peterson Institute for International Economics Working Paper Series, 16(6).

29 Zealand over the GFC (23 percent) was less than the average for the crisis countries (19 percent). • Including an, as yet unobserved, severe loss event does not make much difference to average longrun PD estimates. To illustrate, consider a New Zealand housing portfolio that has an average PD of 0.50 percent, mainly driven by the GFC recession (default rates in normal times are very low). Now assume that there is some unobserved extreme (say 1:100 ) event that results in a default rate of 10 percentage points. Allowing for this event increases the average default rate to about 0.6 percent. As the banks’ calculations all included a conservative overlay, they already implicitly account for unobserved severe events. Multiple counting of defaults The Bank does attempt to address the criticisms in the TMW that the nonperforming loan approach can count a fault multiple times. We are aware that the NPL ratio may overstate the likelihood a loan will go bad in a given year – because a loan may sit unresolved for more than a year and thus count twice. We responded to this problem by using, at the top of our range of ‘PD’ input values, the average historical NPL or impairment ratio (based on whatever series we were looking at) and, for the bottom of the range, 50% of the historical average (this would be appropriate if all loans classified as non-performing took two years to resolve, rather than one). We believe this is a conservative assumption, because similarly some loans may enter and leave the NPL pool within the year. The Bank’s adjustment simply doesn’t work. Below is the figure for impaired loans taken from the Bank’s previous information paper. It is obvious that the data is dominateded by the BNZ’s experience in the late 1980s, and that the nonperforming loans stayed in the portfolio for several years. Almost all were there for four years because in was not until 1993 that they started to fall. The banking system average of the number of years defaulted loans stayed on the books, for the entire period, would have been probably higher than three – though it is hard to say just by eyeballing the graph. Which is why any researcher would use the banks’actual default rate estimates. While defaulted loans may enter and exit the nonperforming category within one year, the default rates in benign times are very low so this fator will have a limited impact on the historical average. We can roughly assess from the data that non-performing loans overstate the default rate by a factor of around three. The Bank divides by a factor of two for its bottom of the range estimate, but not at all for the top of the range (it doesn’t say what it does with its intermediate estimates). As both the high and

30 low estimates receive the same weight in their analysis, the effective reduction is to divide by 1.5, half the rate required to match the default rate. This is not the conservative treatment the Bank claims. They have, on average probably overstated the default rate by a factor of two. All this is a bit obvious, which probably explains why the Bank dropped the above figure from its April 3 effort, and replaced it with their figure A, reproduced below. Figure A shows the impaired asset ratios for all banks. Of course it suffers from the same multiple counting problem as the earlier effort, but the Bank goes one step further to boost the PD estimate. The estimate is disproportionately affected by the impairment rates of a few small banks in the 1980s, because the data is unweighted. These small banks were new entrants, and made many highly risky loans in an attempt to enter the market. They withdrew after a few years after incurring disastous losses. Looking at the data there are observations at 70-75 percent, more that 60-65 percent, and so on. The great bulk of the observations are quite low – more than 50 percent below 1 percent, reflecting the experiences of the incumbent banks. Even if you believe that the disasterous expeirences of long gone banks are relevant to assessing the future risk profile of New Zealans’s current main banks, you would not weight those experiences as equal. But that is what the Bank has done.

31

The Bank’s defences of its methodology are that:

1. Large samples are better than small samples. This argument does not, of course, apply if the observations are not from the same population. And it is clear that the ‘long gone’ small banks differed fundamantally from today’s main banks in loan composition, risk culture and risk manageent frameworks. While the small banks were given banking licenses at the time, they would not receive a banking license today with the business models they had in the 1980s.
2. The rules will apply to small banks as well as large This is desperate stuff. It does not explain why those long departed banks’ experiences receive an weighting equal to todays main banks
3. If the small banks impairments rates were exposure weighted they would not have a material impact on the overall estimate Precisely. In other words if the Bank used a more respectible statistical technique by weighting the dara their ‘data scam’ would not work. The Bank does set out one of the arguments against its methodology. Given New Zealand is a banking system that has long been dominated by just four or five banks, it might be tempting to confine the historical sample to simply the large banks. Small banks, it could be said, have had very different portfolios than large banks and

32 therefore their loss experience is of little relevance when assessing the capital needs of systemic banks. That is precisely our view. Further the Bank’s approach is fundamantally inconsistent with its modelling approach. The banking system is modelled as a single bank, using aggrgate information about the entire banking system. It necessarily follows that that it is the aggregate weighted non-performing or impaired loans ratios that matter, not an unweighted average of individual bank numbers. And it is inconsistent too, with their use of international data. This is based on national data, which is also a weighted average of the country experience. The Bank’s responses to the challenge it sets is are as follows Firstly, there is the obvious point. We are setting capital policy for the long term and for an unknown future. Portfolio composition among the large banks today is not necessarily a reliable guide to the portfolios that will be held in the distant future. The obvious flaw in this argument is that as a bank’s loan portfolio changes to a riskier asset composition, as the Bank implies, then its average risk weight will change – offsetting that risk. It is not necesary to increase the capital ratio ahead of time. While we do not know the actual portfolio composition of the small banks that incurred the disasterous losses, we do no that they focussed on high risk propoerty development and investment company loans. Today only 0.4 percent of bank lending is in commerical property development, and the risks are carefully limited by much more restrictive loan requirements than were in effect in the 1980s. It is nonsense to suggest that the main banks will suddenly switch their business models. If they do change course in the decades to come the Bank can respond if it believes that the risk weights are not picking up all of the additional risk. Supervision is not a one shot game where the rules have to be set now and never changed regardless of evolving circumstances. Secondly, many factors impact on loan performance, not simply loan type. Governance qualities, macroeconomic vulnerabilities and correlations between borrowers are relevant for all banks. These factors impact on all banks and thus, in order to best capture the effects of these factors, it seems reasonable to include small banks in any sample. The Bank seems to be arguing against inself here. It is precisely because the old small banks were so different that they shouldn’t be included in todays assessment.

33 What the Bank seems to be suggesting here is that the factors that drove the small bank defaults are more or less random, because they cannot be identified ex ante. Hence any bank in the future, and in particular the main banks, can be struck down by the same problems. Without warning and without the supervisor or anyone else noticing they could: • Abandon their risk management systems • Hire inexperienced staff • Be subject to no parental oversight • Makes massive expansions into the most risky lending categoies • Be subject to adverse selection when they attempt to massively expand their lending portfolios. • Have no supervison We think not. Using data from countries that have experienced banking crises The second limb to the Bank’s new PD estimate analysis is the use of foreign data. This is a bit incongurous as the point of the Bank’s modelling exercise was to look at the New Zealand evidence. The Bank explains as follows. The absence of a banking crisis in New Zealand’s history makes the issue of settling on a range of values for the ‘PD’ input particularly problematic. We had to be open to using data drawn from other countries’ histories. While every accounting measure can be assumed to have had varying definitions through time and across jurisdictions the definition of ‘non-performing’ appears to be quite comparable across countries and to have been relatively stable through time. In contrast, the real-world meaning of ‘default’ could potentially vary more widely, making PDs a less reliable loss indicator in a cross￾country context. The Banks produces absolutely no evidence that nonperforming loans are more stable, across time and countries, than actual actual PDs. Indeed the opposite is more likely to be true. PDs are a Basel advanced model input and there are rules for its consistent measurement. In addition the Bank uses impaired assets for its New Zealand assessment, which is a different and more subjective measure than non-performing loans. The Bank then goes on to reference the IMF table it used in the consultaion document. It argues. The peak impairment rates for the BNZ and two stressed Australian banks are included in Figure B for comparison purposes. The impairment experience of these banks was on par with the average experience (indicated by peak NPL ratios) of banks in countries that have experienced crises. The impairment data for New Zealand banks is thus a suitable data set to use for model inputs.

34 This is (yet again) nonsense. Comparing the worst performing bank in New Zealand back in 1990 against system averages for other countries does not establish any equivalance. if there were to be a comparison it should be the relative system performance in the GFC when the economies were subject to similar economic shocks. This obviously shows that New Zealand banks performed much better (a peak NPL ratio of about two percent) than the banking crises countries. The argument that the IMF table somehow justifies the use of non-performing loans simply doesn’t work. And then the Bank adds the ‘precedent’ argument. There are precedents for using cross-country NPL ratios to derive loan loss rates (the NPL ratio is multiplied by an assumed LGD value to arrive at a loss estimate). A recent example is a report published by the IMF in 2016 There are precedents for many things in the thousands of articles or studies on regulatory issues, but that does not make them right. As we argued in TBW, the IMF paper was a lazy piece of analysis with a methodology, which in most cases, will substantially overstate the loss rates it purports to measure. The Bank’s next step is the comparison with some international evidence on non￾perfoming loan rates. Table A below shows historical average NPL ratios for countries with GNI per capita above US$15,000 in 2017 (a Reserve Bank criteria for comparability with New Zealand) that have been identified by the IMF as having experienced banking crises. Omitting Greece and Cyprus, the median NPL ratio sits between 2.9 percent and 3.3 percent and the average is 3.8 percent. The Bank concludes that the international evidence supports their New Zealand analysis, buttressing its case. Oddly the Bank didn’t exclude Canada and Australia, which haven’t had a financial crises, from the table. But they also included several counties (Norway, Sweden and Finland ) that did not have a crisis over the data period (1998-2017) And why did the Bank exclude Greece and Cyprus, if they truly believed in larger samples. The Bank’s answer would probably be that they were not suitable comparators, because they had many economic and banking structual issues that are not relevant to New Zealand. But the same argument could be made for excluding many of the countries in table A. Certainly the former communist countries, and Uruguay. And then there is iceland (an execise in banking madness that was run like New Zealand finance companies prior to the GFC); Italy (it is well known that problems with the Italian

35 legal system means that defaulted loans can sit in bank portfolios for many years, and that any average level of non-performing loans will be a poor proxy for default rates); and Ireland (again a case of identifiable banking madness, and gross, partially politically driven, regulatory failures, not just bad luck). If we take the fuller sample, excluding the non-crisis countries would produce an average non-performing rate of 6.9 percent. On the other hand, if we take a more sensible view of what is a relevant comparator, the average nonperforming loan ratio (of those countries that had a crisis) would be 2 percent. Halving that to take account of the multiple counting of non-performing loans, gives us a default rate of 1 percent, which is about the same as the New Zealand banks’ estimated rate. The main point here is overseas evidence can be manipulated to demonstrate almost anything. They are not a good substitute for actual New Zealand evidence and in particular banks’ PD evidence. Loss Given Default (LGD) The new evidence in this information paper is that “the minimum prescribed (LGD) value (which applies just to high LVR farm loans) is 42.5 percent.” In the Banks view “this analysis is supportive of LGD inputs in the vincinity of 40 percent”. This, yet again, is nonsense. High LVR farm loans account for around 1.3 percent of total lending. The Bank then goes on to say that this estimate is supported by LGD results from the stress tests (31 percent in 2017 and 37 percent in 2014) As the Bank has previously explained, the lower LGD was largely explained by a change in the structure of banks loan portfolio, and is the relevant result. A LGD of 31 percent doesn’t support an estimate of 40 percent, though it does support the lower range estimate of 30 percent. The Bank presents no evidence at all to support the 50 percent assumption. Correlation coefficient R There is a discussion that purports to support a range of inputs (0.2 to 0.4) for the correlation coefficient, R. Almost all of it is wrong, irrelevant or misleading. The starting point should be the Basel model estimates, which are less than 0.2 on average, so the Bank has to make a solid case for using much higher coefficient numbers. The Bank’s discussion is entirely limited to the housing loan portfolio. While housing loans are about 60 percent of bank lending, they account for about one

36 third of the risk. Thus the Bank has nothing to say about the correlation coefficient for two third of banks’ risk. The Banks discussion of the housing loan R coefficent is as follows. Borrowers’ income is, in the Basel III case, assumed to be the determinant of the likelihood of borrower default. However, equally, the value of the borrowers’ assets could be a driver of default. Based on the limited data available, the correlation between asset values and output in New Zealand seems relatively high. The ASRF model allows the capital implications of high correlation between asset values and GDP to be explored. Borrower income is not assumed to be the determinant of default in the Basel III model. This is an emprical issue driven by the banks’ model.s More to the point this inaccurate observation is not relevant to estimates of the correlation coieffcients in the Basel model. R It is a measure of autocorrelation, the extent to which exposures within the portfolio are correlated with each other. While a high correlation between GDP and house prices will increase R, there are many other factors in play. In particular, simple models that attempt to derive a correlation coefficent from the volatitity of business assets don’t work with housing portfolios, because housing defaults are a complex function of both the value of the house and the borrowers income, not just the value of the house. The Bank proceeds as follows. In New Zealand house prices and GDP appear to be highly correlated. Figure C presents a simple plotting of the annual percentage change in GDP (real, relative to trend) against the annual percent change in house prices (relative to trend). The statistical correlation between the two series is 0.63. The statistical correlation does not translate directly into the ‘R’ input needed for an ASRF model, but an approximation to R can be achieved by squaring the statistical correlation (i.e. this data generates an R input value of 0.39). The data used in this calculation is quarterly observations of 12-month changes. Repeating this exercise for other countries, suggest the correlation between asset values and output in New Zealand might be particularly high. Some studies suggest that the correlation between asset values and output (and by inference, borrowers with each other) increases noticeably during crises. Despite a banking crisis in the data (sic), the correlation evident in New Zealand data is higher than in the UK and Ireland, for example, two countries that have experienced banking crises.

37 The impression the Bank is trying to leave here is that this evidence somehow justifies the use of R assumptions that are much higher than those imposed in the Basel models. Its range of R inputs is 0.2, 0.3 and 0.4. The Bank also attempts to lend its methodology an air of authority with a couple of references from the ‘literature’. The first reference7 is an early paper (2003) that discusses various elements of the Basel model ,which was still under development at that time. Asset correlations of assets with S&P credit ratings were estimated (the estimates were under 0.10). There was no discussion of the housing correlation at all. The second8 paper estimates correlation coefficient using Nordea Bank data. The housing correlation coefficient was estimated, by backing in it out of actual loss experiences, but unfortunately the actual results are blanked out in the online version of the paper we were able to access. A third, very recent paper,9 from the Bank of England, was not cited by the Bank, but it is referenced in the Bank’s Appendix A. It reviewed the full literature on the estimation of the housing lending corrrelation. There is no suggestion, anywhere in the literature, that the housing correlation can be read off a GDP/house price correlation. A new methodology, assuming a fat tail default distribution, was

7 Hamerle, Alfred, Thilo Liebig and Daniel Rosch (2003) Credit Risk Factor Modeling and the Basel II IRB Approach. Deutsche Bundesbank Discussion Paper Series 2 No. 02/2003 8 Martin, Lionel (2013). ‘Analysis of the IRB asset correlation coefficient with an application to a credit portfolio’ Uppsala University Project Report 2013:28 9 Neumann, Tobias (2018) Mortgages: estimating default correlation and forecasting default risk. Staff working paper No. 708. Feb 2018

38 developed and applied to the US and the UK. The conclusion is that the results are consistent with the Basel model estimate of 0.15 Issues with the correlation estimates While a strong relationship with house prices and GDP may influence the housing correlation coefficient, the Bank does not make a very convincing case that New Zealand house prices are more sensitive to GDP than house prices in other countries. • The wrong correlation is estimated. The relationship of interest is the correlation between nominal house prices and GDP, not the relationship between deviations of real house prices from their trend and GDP. Banks are at material risk of loss when the nominal house price falls below the value of the loan, not when real prices depart from trend. • The relationship identified is economically inconsequential. A departure of GDP from trend of around 15 percent is associated with real house price deviations of only four to six percent. Table C: Correlation and 'R' estimates for NZ and other countries 1990-2007 1990 to 2017 Full sample Sq root NZ 0.52 0.63 0.63 0.39 UK 0.37 0.62 0.47 (1976- 2018) 0.22 Canada 0.24 0.26 0.31 (1971- 2018) 0.1 Norway 0.40 0.44( (1993- 2018) 0.44 0.15 Ireland 2006-2018 0.44 0.19 Comparion with overseas countries may have been contrived There is no explanation of why the sample of comparator countries was selected. There must be a suspicion of cherry-picking. In addition the data doesn’t really show that New Zealand has a higher housing price correlation, even on the Bank’s flawed measure. Two of the calculations, Norway and Ireland, are not directly comperable. The Norway estimate leaves off the 1990-92 years when house prices were falling. The Irish data only starts from 2006, leaving out the prior run up in prices, which would have affected the results. Over the same time period, the UK correltion is the nearly the same as New Zealand’s

39 The Bank got the math wrong The Bank says that the square root of the New Zealand correltion of 0.64 is 0.39. The square root of 0.64 is actually 0.8. What the Bank may have intended to say is that the square of 0.64 is 0.39. Conclusion The Bank’s analysis of the model input values is, almost from start to finish, nonsense, an increaingly desperate attempt at backfilling. This game has put the Bank’s reputation at risk. Bank risk analysts are appalled at what they are seeing. The Bank may not particularly care, because they are playing to a different audience: politicans, the general public and the media, who don’t understand the subject area. But people who do undestand the analysis will eventually read it and their assessments will leech out, putting New Zealand’s international reputaion at risk. APRA is probaly aghast. They may have to deal with the Reserve Bank in a stress situation.

Broader implications If the Bank really believes it is right on the numbers, then it should share their conclusions with the Australians, and take them through their analysis. Australia has not had a financial crisis, so the logic of the Bank’s analysis, and some of the numbers, should apply to Australian banks. In particular banks’ PD estimates should be replaced by estimates derived from impaired loan ratios that include unweighted data from the now defunct Australian state banks. The economic cost of crises Although the Bank doesn’t like to be pinned to a point estimate of any of the variables in their model, it is fairly clear that there is a strong preference for an economic impact effect of 63 percent of GDP. This just happened to be the median of the estimates in the studies reviewed by the Basel Committe in their original 2010 cost benefit analysis. The Bank picked it up because it was the easiest thing to do. The Bank also uses assumptions of 20 and 40 percent of GDP, but there is no discussion on how those figures were derived. In appears that they were just made up. The Bank identifies what it sees as the key issues in estimating the cost of a financial crisis. Whether the output effect is assumed to be temporary or permanent The temporary approach looks at the shortfall of GDP from a non-crisis growth path counterfactual, from the start of the crisis to the finish.

40 The permanent approach compares the absolute difference in post crisis output with the pre-crisis growth path. If there is a difference it is assumed to be permanent. With this model the effects of tulipmania, for example, are still being felt by the Dutch today. The permanent impact assumption can generate very high GDP losses. A measured reduction in output from its potential, of say, 3 percentage points and a discount rate of 3 percent generates a GDP loss of 100 percent of GDP. Conceptually this approach is a bit of a nonsense. Economies do get over shocks eventually. . But it appealed to some researchers who were on a mission to justify bank capital increases post the GFC. The base the loss is measured from The Bank agrees that the methodologes used in many of the earlier studies overstated the losses. If it is the most recent level of output, and that was unsustainable (for example, fuelled by under-priced credit), the loss may be over-estimated. And the Bank’s solution is: This is acknowledged in our framing of the policy problem as we have chosen potential output (which we interpret to mean non-inflationary steady state output) given current interest rates (not actual output) as the benchmark against which to measure the output impacts of capital. The Bank may have acknowleged the problem but it does not address it when selecting its prefered GDP estimate. Its preferred estimate of 63 percent of GDP is mainly based on ealier studies that used inflated pre-crisis growth ratesthat boosted their estimates. Some of these studies also included low income countries, which are not releavnt to New Zealand. The marginal impact of the financial crisis A third issue relates to what, if any, of the output loss would have occurred even in the absence of a financial crisis. And again there appears to be a solution. This issue is addressed in the literature by using statistical techniques to separate out the effects of many factors on realised output. Here the Bank cites Romer and Romer (2015) as a supporting source. How they could do that is with a straight face is beyond us. The thrust of that paper, as

41 discussed above, was that the conventional way of thinking about crises as something that either ocurrs or it doesn’t, substantally overstates the impact of financial crises. The truth is that it is almost impossible to reliably disentangle the effect of the banking crisis from the underlying shock. Cline addressed the issue and concluded that the ‘problem’ had not been solved. Very few of the GDP loss estimates in the literature make any attempt to account for underlying drivers of the recessions associated with banking crises. They just assume that all the GDP losses are due to the banking component of the GDP downturn. Thus all of these estimates are biased upwards. The best evidence on the matter is probably Cline’s emperical evidence on the GDP shock in countries that had a financial crisis in the GFC and those that didn’t. As noted above, there was no major difference If one believes that the effects of a crisis dissappate over five years or so (which is likely except for the most extreme of crises). If they are longlasting, then the difference according to Cline is about 10 percent of GDP. This accords with common sense, for countries like New Zealand. Financial crises have underlying causes which will have real consequences that will be felt regardless of whether the IMF crisis definition is triggered or not. A property boom that generates over-building will naturally be followed by a slump in GDP as the excess stock of property is absorbed. The demand and supply for loans will naturally fall as borrowers become less optimistic and bankers become more cautious. More capital is not some kind of magic that makes those effects go away. The Bank gives the impression that all of the technical problems have been resolved in the literature and that it is relying on some robust estimates. This is misleading. The Bank then it throws in its lot with the ‘longlasting impacts’ school, citing a passage from the Firestone (2017) the lead author of the Federal Reserve’s 2017 optimal capital paper. We share the view of the authors of the following quote, namely that the empirical results appear to suggest the output effects are long-lasting. “Other studies leave the duration of a crises’ effects open as an empirical question, and generally find support for long lasting effects. Furceri and Mourougane (2012), analyze OECD countries and compare actual output after a crisis with a measure of potential input. They estimate autoregressive equations and the implied impulse response functions, finding an average permanent reduction in GDP of 2 percent. Cerra and Saxena (2008),

42 analyzing data from over 120 countries, find evidence that effects of a financial crisis on GDP are barely reduced by one percentage point after ten years, remaining at a level of six percent. These studies provide evidence for robust long-lasting effects. We assume that financial crises have persistent effects in the rest of the analysis.” Firestone (2017) The Bank makes several references to the Firestone paper for support at different points in their discussion, and It may be one of the few papers that they actually read. However, Firestone did not do any original research on the long-run effects of a crises. Instead the long run effects analysis relied entirely on the Mourourange paper cited above. This paper, which effectively relied on a sample of just seven crises (their data period did not include the effects of the GFC), found a permanent long run impact of about 2 percent of GDP. It also analysed the ‘big five’ crises of Spain (1977 crisis), Japan Finland, Sweden and Norway, and found an permanent impact of 4 percent. Notably Korea was left out of the analysis. Korea had the most severe crisis, in terms of banking system losses, of all countries considered, but it experierned a very rapid recovery. After three years it was back on its precrisis growth path. If it had been included in the data set (it was an OECD country at the time), Mourourange’s conclusions probably would have collapsed. In any event a sample of just seven crisis events, all with different recovery paths, and a range of other things going on, which could not have been captured by the model (a simple VAR model of GDP and its laggs), provides a very weak basis to draw general conclusions about how crises really work. The other study cited by Firestone mainly related to lower income countries and would have been dominated by the likes of Argentina. They did produce an estimate for industrial countries and found a significant effect, but that was based on a sample of just two countries. Even ‘respectable’ analysts can be open to massaging the numbers to get a decent output cost effect. The following figure, which is is taken from the 2015 Bank of England cost benefit study, shows the crisis output path, for a group of banking crisis countries, over the 7 years from the start of the crisis. It is clear from the figure that recovery was almost complete, and would be have been complete by years eight or nine. However, it was assumed that there would be a permanant effect of over 1 percent of GDP. With a some low discount rate assumptions, the effect of this assumption was to more than double the estimated GDP costs of a banking crisis from about 20 percent to 43 percent.

43 The discount rate Overseas studies each have to apply a discount rate. These tend to be lower than what is currently required of public projects (other than accommodation and office buildings) by the NZ Treasury, for example. This suggests that the output cost of a crisis reported in these studies would be less if the costs had been discounted using the rate currently prescribed by the NZ Treasury. Again having recognised the issue the Bank does nothing about it. Conclusion The Bank presents three estimates of GDP costs: 20 percent, 40 percent and 63 percent, with a clear preferenc for the latter. Only the 20 percent estimate has any credibility in the New Zealand situation, and a lower figure, say 10 percent, could be the best estimate. Higher estimates are more credible in overseas studies, when the issue is how much capital a globally significant bank should have. Stress to those banks can have global external effects. That is not a consideration when thinking about capital for New Zealand banks. In international terms they are small players

44 Summing up the results. The following is a summary of our assessment of the inputs into the capital model. Cost of capital The cost of capital should be at least doubled to account for the increased profits of foreign banks. Likelihood of a ‘crisis’ PD The Bank’s mid range assessment of 2.25 percent is based on nonsensical arguments.The banks’ basel model estimates of 1.2 percent should be used. This should be further reduced by an assessment of the impact of parental support. LGD The Bank’s mid-range estimate of 40 percent is too high. A 30 percent estimate based on the stress test results and banks’ Basel model estimates is more appropriate. R The Bank has not established any basis for using a correlation coefficient that is any higher than the Basel model of around 0.2. Costs of Crisis A reasonable estimate of the costs of downturns that can be mitigated by higher capital is quite low. An estimate of 10 percent would be appropriate. Taken together these will generate a low optimal capital ratio. Part nine: The fiscal cost of crises In TBW we drew attention to the fact that the fiscal costs of banking crises have not been, on average, particularly high. The average gross cost of historical crises in high-income countries was about 11 percent of GDP, and the net less than 5 percent. The median costs are lower, at 6 and 3 percent respectively. With today’s higher capital ratios the net costs would have been significantly lower. In

45 addition, New Zealand banks have good franchise values, which will be reflected in an eventual sales price, so it is likely that the government will recover all of their capital injection. Nevertheless governments and the Treasury may be more concerned about the gross fiscal costs, as those costs will come when government borrowing is increasing due to the economic downturn. The problem here is that an increased capital ratio may not help. This might seem counterintuitive, from a perspective that sees a financial crisis as a one-off game. Consider a banking system which has $100 in loans and $7 in capital, and the bank losses $15. The Government pitches in $8 so the depositors don’t lose. If the capital is $11 then the Government only has to contribute $4. There is a fiscal savings of $4. But this assumes that the banking system will be closed down and the assets sold off. This is not an option. What will happen is that the banking system, or most of it, will be recapitalised so that it is ‘sound’. Prior to the release of the Bank’s consultation document, everyone (rating agencies, depositors, bankers etc.) thought the system was sound with capital of $7. So the government would have to contribute $15 to restore capital to $7. Under the Bank’s proposals, however, it will still have to contribute $15 to restore system capital to $11. If capital were only restored to $7 the system would be ‘unsound’. Having allowed the Reserve bank to drawn its line in the sand on soundness, it would be difficult for a government to just meet the $7 target, arguing perhaps the Bank’s 2018 soundness assessment was just rhetoric designed to put more capital into the system. That is unlikely to work, so the effect of the Bank’s proposals on the Governments fiscal risk, in this scenario, is nil. Part ten: Banks’ capital ratios in an international context The Reserve Bank did not discuss the issue of comperability of New Zealand and foreign bank capital ratios in its information release paper. It is a sensitive point. If the Bank were to admit that the foreign sourced data should be adjusted before applied to New Zealand, then many of the conclusions it has drawn from that research would have to be withdrawn.

46 However, the Bank has released, under the OIA a document titled Media Resource, where it set out the arguments for not making any adjustment. Because these document has not been in the public arena so far, we repeat the relevant parts of the ‘Introductory and guidance’ section, which presents the Bank’s arguments. The Bank’s arguments for not making adjustments The first five points in the ‘introduction and guidance’ are some basic information about a risk weighted capital regime. 6. In each jurisdiction the risk weights applying to a subclass of assets reflect the regulator’s views about the potential losses the sub-class could generate for banks. This is mostly misleading. The risk weights are mostly driven by banks’ advanced risk models. With the advanced models regulators did not impose their view on the risk weights as a matter of course. The risk weights used for a particular subclass of assets can vary from regulator to regulator because the circumstances in each country vary (objective factors) and because regulators vary in terms of how they view and respond to these risks (subjective factors). In order to make accurate comparisons of the ability of banks in different jurisdictions to withstand shocks – their relative capitalisation in other words – it is necessary to remove the subjective element from the RWA calculated in each jurisdiction. This is inherently very difficult and the results of any such attempt cannot be relied upon with any confidence. What the Bank appears to be saying here is that there is ‘objective’ information about the risk of a loan, which is unknowable. In the advanced Basel model banks attempt to calculate the true risk, but this is just an estimate. The supervisor then comes along and sometimes adjusts the banks’ estimates upwards, because there is something about the economy, that they know, that the Basel model and the banks’ modeling has not picked up. To illustrate, assume that the true ‘objective’ risk weight is 50 percent in country A and 30 percent in country B, but the Basel model produces a risk weight of 30 percent in both countries. Supervisor A spots the problem and ‘subjectively’ imposes a risk weight of 60 percent. If we were to compare just the risk weights we would say that country A is requiring twice the capital for the same level of risk. However, if you look at the objective element the capital requirement is overstated by just 20 percent (a risk weight of 60 percent compared to the ‘objective’ 50 percent). The problem with this story (particularly with respect to the large residential

47 mortgage portfolio) is that it does not describe what happened in New Zealand when the banks were first accredited10 on their advanced models. The Reserve Bank required banks to use higher model inputs to boost the risk weights. This did not reflect an assessment that New Zealand was objectively intrinsically more risky than foreign jurisdictions in a way that should impact on risk weights. Rather it was due to a lack of confidence in the Basel model and in the very low risk weights it was producing. Other supervisors, at least initially, went along with the model outputs. The result was that New Zealand had much higher housing risk weights, for what was the same levels of risk than most overseas jurisdictions. In recent years the Bank has run the argument that the higher New Zealand risk weights reflects New Zealand specific risk factors but it has never supported these assertions with any analytical work. We note that the Bank does not respond here to the fact that APRA did not see the ‘objective/ subjective distinction’ as a sufficient impediment to conducting a comparative exercise. The Bank then downgrades the relevance of the Basel risk weights and capital ratios. Reflecting the difficulties in separating the objective and subjective factors leading to a given RWA value in any country, ratings agency S&P has developed bespoke capital ratios that draw on bank balance sheet data (and other measures) to calculate risk measures (in contrasts to the official RWA values and Because of the inherent difficulty in separating objective and subjective factors impacting on official RWA values we do not actively monitor other countries official capital ratios. However we do monitor the relative position of NZ banks in ratings agencies studies. This lack of confidence in the Basel capital ratios does not sit well with the Bank’s use of the overseas literature to support their case. Most of the evidence relies on Basel definitions of capital. If the Basel measure is unreliable then that evidence is also unreliable. On the S&P capital ratios we are told.

10 The author of this paper was deeply involved in this process and knows what happened. The Bank’s current analysts are unlikely to know.

48 ‘For example, our interpretation of the most recent S&P findings is that NZ banks are at the median of their peers. But on the detail. We acknowledge there is a genuine interest in comparing our proposal to the capital position in other countries, an interest which cannot easily be met because the S&P findings are copy- write protected. While copyright protection apparently stopped the Bank from being transparent about its assessment that New Zealand is at the median of its peers, it did not stop the Deputy Governor releasing, in a recent speech, a set of S&P capital ratios, by individual bank, that showed New Zealand towards the bottom of the pack. The Deputy Governor also said that the S&P ratios did not play much of a role in the capital review, which is probably true, but this does not really square with the statement that the RBNZ is now just actively monitoring S&P capital ratios. The statement that New Zealand is at the median of its peers is significant. If the S&P capital ratio is a reasonably robust measure of relative risk and we are at the median, and if, according to the Bank the New Zealand banking system is ‘unsound’, then it follows that either half of the comparators are unsound, or the Bank is wrong on its soundness assessment. Understanding S&P’s capital model Given the weight the Bank says (at least sometimes) it is now placing on the S&P capital model it is useful to understand how it works. Basically it is like the Basel standardised model, with some additional categories for some loan classes (such as property development loans), and fewer in others. It does not, for example, distinguish the risk weights of residential mortgages by LVR, whereas the Basel standardised model now does. The distinguishing feature of the S&P model is that its risk weights are multiplied by a BICRA (banking industry country risk assessment) ratio. This ratio is meant to capture the country economic, industry and institutional factors that affect the relative riskiest of loans. To illustrate, the residential risk weight, with a BICRA of 1 (the best possible) is 20 percent. With BICRA’s of 2, 3, and 4 the risk weights are 23, 29, and 37 percent

49 respectively. This assessment will not have regard to the LVR structure of the book11 or of the actual loss performance of the loan book over time. A country with a BICRA of 2, will have a higher capital ratio, say 16 percent, than New Zealand, which with a BICRA of 4 would have a capital ratio of 10 percent With a BICRA of 4 New Zealand is in pretty poor company12 On the economic aspect of the BICRA we do even less well scoring a 5. So why are we rated so poorly?

The first point to note here is that the S&P capital model is a simple (perhaps in places simplistic) one-size-fits all model that is designed to apply to a wide range of countries. For individual countries, it is a paint-by-numbers approach, which in some cases doesn’t work very well. New Zealand appears to be one of those countries. In terms of the BICRA scoring regime there are a number of drivers of our low score. On economic imbalances New Zealand first gets hit on the private sector debt to GDP ratio. We are just over the trigger point of 150 percent of GDP. Then we get negative points on the growth of real house prices and private sector credit. This is a ‘point-in-time’ score, based on a four year rolling averages of those changes. This means that New Zealand ‘s ratings will improve as slowing credit and house price growth numbers feed though the S&P assessment. We will probably revert to a BICRA of 3, which is where we started when the framework was introduced. That alone would increase the capital ratio, from the illustrative 10 percent presented above, to about 13 percent. The other drivers are on the industry side. We score relatively poorly on the institutional environment, compared to Australia, primarily because S&P does not rate the Reserve Bank as a supervisor. The rest of the New Zealand’s institutional environment would rate as very favourable from a risk perspective. We also rate very unfavorably on systemwide funding because of banks’ relatively high level foreign bank funding. For a country with its own floating currency and an independent central bank to deal with liquidity issues, this is not a material credit risk. But given the historical experiences of countries that did not share these

11 The LVR structure by country does enter, somewhat subjectively into the BICRA assessment but unless it clicks the country over to the next BICRA grade it will have no effect 12 We are rated as a 4 with Estonia, Iceland Israel, Kuwait, Malaysia, Mexico, Saudi Arabia and Taiwan. Most of what we think of as comparators rate as a 2. Australia is 3 but will revert to a 2 as the recent house price falls work through the model.

50 characteristics, and the number of S&P clients that do not have their own currency, S&P is overly sensitive to the bald New Zealand balance sheet numbers. The other point to note here is that even with the capital ratio produced by the BICRA rating of 4, New Zealand banks pass the implicit stress test that was used to calibrate the risk weighting model, for an A rating. This means that the implied probability that all of the bank’s capital will be exhausted in that test is about 1:750. And this is without taking account of the implicit support from the Australian parent, which contributes to the final rating. The capital ratio only applies to the stand-alone bank assessment. New Zealand and Australia BICRA risk assessments New Zealand Australia Economic resilience Very low Very low Economic imbalances Very high Very high Credit risk in the economy Intermediate low Institutional framework intermediate low Competitive dynamics low low Systemwide funding high intermediate .

51 .

1 The 30 billion dollar whim A review of the Reserve Bank consultation paper: ‘How much capital is enough’

2 About Tailrisk economics Tailrisk economics is a Wellington economics consultancy. It specialises in the economics of low probability, high impact events including financial crises and natural disasters. Tailrisk economics also provides consulting services on: The economics of financial regulation Advanced capital adequacy modelling Stress testing for large and small financial institutions Regulatory compliance for financial institutions General economics. Principal Ian Harrison (B.C.A. Hons. V.U.W., Master of Public Policy SAIS Johns Hopkins) has worked with the Reserve Bank of New Zealand, the World Bank, the International Monetary Fund and the Bank for International Settlements. In his time at the Reserve Bank Ian played a central role in developing an analytical approach to financial system risk issues. Contact:Ian Harrison – Principal Tailrisk economics e-mail:harrisonian52@gmail.com Ph: 022 175 3669

3 The 30 billion dollar whim A review of the Reserve Bank consultation paper ‘How much capital is enough’ “We are going to build a great big beautiful capital wall - and Australia is going to pay for it” Part one: Introduction On 14 December 20181 the Reserve Bank of New Zealand released a discussion document ‘How much capital is enough’, which proposes an increase in the minimum common equity tier one (CET1) capital ratio to 16 percent. As banks will need to hold a buffer over that minimum, the practical effect will be to increase the banks’ average capital ratio from the current level of around 11 percent to around 18 percent, a seven percentage point increase2. There will also be an increase in capital calculation floors and adjustment ratios that will further increase required capital. The justification for the higher capital ratio is based on what is represented as a new approach to setting the regulatory capital ratios - ‘the risk appetite framework’. The Reserve Bank says that this follows from a consideration of the words in part 68 of the Reserve Bank Act, which reads: The powers conferred on the Governor-General, the Minister, and the Bank by this Part shall be exercised for the purposes of— (a) promoting the maintenance of a sound and efficient financial system The Bank argues that the financial system is ‘sound’ if ‘there is enough capital in the system as a whole to cover losses that are so large that they might only occur very

1 On 25 January 2019 the Reserve Bank issued an updated version of the consultation paper but there do not appear to be any changes that would affect the analysis in this paper. 2 The bank may be able to reduce some of their current tier one capital that no longer counts for regulatory purposes but some of it will be ‘stranded’. We have assumed that half will recalled so the net effect of the policy changes will be a 6 percentage point capital increase.

4 infrequently’. Infrequently is quantified as a one in 200 probability. It is claimed that regulatory capital of 16 percent is required to reduce the probability to this level. The capital implication of word ‘efficient’ in the Act is considered separately. It is considered in more conventional terms, requiring a balancing of the costs of the higher capital and the benefits from having higher capital. If the ‘soundness’ and ‘efficiency’ analyses generate different answers, then the soundness criteria rules, and some inefficiency should be tolerated for the sake of an appropriate level of soundness. But, happily, according to the Bank, there is no such conflict – there are improvements from both a soundness and efficiency perspective, so there is a ‘win￾win’ situation. The costs of the policy receive little attention. It is admitted that the higher capital requirements could make it more expensive for New Zealanders to borrow, but the Bank claims that the impact will be ‘minimal’ and that they have taken it into account. However, even on the Bank’s own assessment of 8.2 basis points3 for each percentage point increase in the capital ratio, the cost to New Zealand will not be minimal. It is likely to cost around $1.5 billion per year, and possibly more. The present value of the cost of the policy could reasonably be assessed at $30 billion. A medium size business with a loan of $5 million could be paying $50,000 additional interest a year. A homeowner with a $400,000 mortgage could be paying an additional $1,000 or more a year. The Reserve Bank has not taken borrowers direct costs into account. The ‘soundness’ test which is driving the policy explicitly excludes any consideration of cost. The foreign cost benefit analyses, which the bank seems to have relied on in its ‘efficiency’ assessment, also ignore borrowers’ increased costs. $30 billion is a lot of money, and Australia is not going to pay for it. It could fund for, example: huge improvements in the national roading system; $10,000 subsidies for three million electric cars; or, if you are quite detemined to waste the money, a 1000 km wall on the southern US border. Given competing claims on scarce resources the capital increase should only proceed if you are reasonably sure that the benefits will well exceed the costs. The burden of that proof should sit with the Reserve Bank.

3 This is the number that appears in the decision document. A figure of 6 basis points appears in the consultation document but there is no explanation for the difference. Assuming a 5-percentage point increase and $430 billion of bank lending the annual cost is $1763 million.

5 The central question that is addressed in this paper is whether the benefits, (‘being more resilient to economic shocks’) are worth more than $30 billion. Our assessment is that very clearly, they are not. New Zealand could secure nearly all of the benefits of higher capital by increasing tier two capital, as the Australians are proposing to do, at about one fifth of the cost of the Reserve Bank’s proposal. The Reserve Bank does not appear to have considered this option. The key driver behind the 16 percent capital ratio proposal is the Bank’s assessment that the banking system is ‘unsound’ because the risk of failure is worse than 1:200, which is at odds with rating agencies and informed analysts’ assessments. However, the Bank now says it has alternative facts. The Bank trumpets the breadth and dept of their analysis supporting their decision. One of the purposes of this paper is to ‘fact check’ the Reserve Bank’s claims. We have made ‘Pinocchio’ 4 assessments of the Bank’s more public statements on the capital adequacy proposals in an accompanying paper. This paper proceeds as follows. Part two sets out the key conclusions from our analysis. Part three looks at the logic of the ‘risk appetite’ approach, and considers the evidence presented by the Bank that a 16 percent minimum capital ratio is necessary to reduce the probability of a ‘crisis’ to one in two hundred years. Part four examines the evidence on the ‘efficient’ level of capital, and in particular it revisits the Reserve Bank’s analysis presented in its 2013 Regulatory Impact Statement on the application of the Basel III rules to New Zealand. This analysis suggested that the optimal (tier one) capital ratio was 13 percent, which is about where the banking system is now. Part five looks at the Reserve Bank’s discussions of its stress test results, and what this tells us about the soundness of the New Zealand banking system. The next part takes a closer look at the benefits, in the New Zealand context, of increasing CET1 capital from 11 to 18 percent, focusing on the costs of financial crises. Part seven briefly considers the Australian approach of increasing tier two capital. The Reserve Bank also says that it is inclined to introduce a new minimum capital requirement based on the leverage ratio. Here the Bank has made a 180 degree turn from its initial position. When a leverage ratio was proposed by the Basel Committee just after the GFC, the Bank made a submission opposing its introduction. It then declined to apply it to New Zealand banks when it became part of the Basel framework. The Reserve Bank was supported in this (privately) by APRA, who also

4 The Pinocchio assessments broadly follow the Washington Post fact Checker methodology. The assessment document is available at tailrisk.co.nz/documents

6 opposed the leverage ratio, but had to go along because they were a member of the Basel Committee. Put briefly and bluntly, the leverage ratio is a silly idea. It will achieve nothing, but might confuse members of the public and analysts if they were to look at it. We will set out the arguments in a subsequent paper. Part two: Key conclusions

1. The ‘risk tolerance’ approach is a backward step that ignores a consideration of both the costs and benefits of the policy The soundness test is based on an arbitrarily chosen probability of bank failure that ignores the cost of meeting the target. The Bank has ignored its own cost benefit model which did take the probability of bank failure, the costs of a failure, the interest rate costs of higher capital and societal risk aversion into account.
2. Bank decision based on fabricated evidence The Banks’s decision to pursue a 1:200 failure target was purportedly based on evidence from a version of the Basel advanced model. It was manipulated to produce the right answer. Initially, a 1:100 target was proposed, but when this couldn’t generate a capital increase, the target was switched to 1:200 at the last minute. The Bank’s model inputs were not credible. It was assumed that all loans were higher risk business loans and that the probability of loan default, a key model input, was more than two and a half times the estimates the Reserve Bank has approved banks to use in their capital modelling. The Bank’s analysis was embarrassingly bad, so it attempted to cover this up with a subsequent information paper that was written after the decision was made, and after the Consultation paper was released. It reached the same conclusion on the required level of capital, but only by assuming a 1:333 failure probability, and by using model inputs that were still not credible.
3. A 1:200 target can be met with a capital ratio of around 8 percent If the Basel model were rerun using credible inputs if would probably show that a 1:200 failure rate can be met with a capital ratio of around 8 percent.

7 4. The policy will be costly The Bank has down played the interest rate impact of the policy, saying any increases will be ‘minimal’. Based on its own assessment of the interest rate impact, the annual cost will be about $1.5-2 billion a year. The present value of the cost of the policy could be in excess of $30 billion. A homeowner with a $400,000 mortgage could be paying an additional $1,000 a year. A business with a $5 million loan could be paying an additional $50,000. 5. The Bank’s assessment that the banking system is currently unsound is at odds with rating agency assessments and borders on the irresponsible The rating agancies’ assessment of the four major banks is AA-, suggesting a failure rate of 1:1250. The Bank is now saying that, at current capital ratios, the banking system is ‘unsound’ because the failure rate is worse than 1:200. Or in other words the New Zealand banking system is not too far from ‘junk’ status. The international evidence does not support the Bank’s contention that the probability of a crisis is worse than 1:200. The Bank has ignored the fact that banks will need to hold an operating margin over the regulatory minimum, and has not adjusted New Zealand capital ratios to international standards to make a fair like-for-like comparison. 6. The Bank‘s analysis ignores the fact that the banking system is mostly foreign owned Foreign ownership increases the cost of higher capital because the borrowing cost increases flow to foreign owners. Foreign owners will support their subsidiaries in certain circumstances, which reduces the probability of a bank failure. There is little point in having a higher CET1 ratio than Australia, because if a parent fails then it is highly likely that the subsidiary will also fail, because of the contagion effect. A New Zealand subsidiary might still appear to have plenty capital, but depositors will run and the Reserve Bank and government will have to intervene. 7. The Australian option of increasing tier two capital has been ignored APRA is proposing to increase bank capital by five percentage points, but will allow banks to use tier two capital to meet the higher target. This provides the same benefits, in a crisis, as CET1 capital, but at about one fifth of the cost. New Zealanders will be required to spend an additional $1.2 billion a year in interest costs for almost no benefit in terms of more resilience to a severe crisis.

8 8. The benefits of higher capital are modest Most of the costs of a banking failure are due to borrowing decisions made before the downturn. This will impose costs regardless of the amount of capital held. With current levels of bank capital failures will be rare, with the main cost likely to be a government capital injection. The experience with most banking crises, in countries most like New Zealand, is that governments have recovered most of their costs when the bank shares are subsequently sold. 9. The Bank is mis-selling insurance The Bank is selling a form of insurance to the New Zealand public, but it vague about the premium costs and has exaggerated the benefits. The premium is the $1.5-2 billion. The benefit would be around a 10 percent reduction in the economic cost of a financial crisis, with an expected return of a few tens of millions. An informed, rational public would not buy this policy. 10. New Zealand banks already well capitalised compared to international norms A recent PricewaterhouseCoopers report argued that if New Zealand bank capital ratios were calculated using international measurement standards they would be 6 percentage points higher, placing New Zealand in the upper ranks of well capitalised banking systems. The Reserve Bank critised some details in the report, but has not produced is own assessment as Australia’s APRA has done. 11. The Bank has forgotten about the OBR The Open Bank Resolution (OBR) bank failure mechanism, was originally conceived as a substitute for higher capital to reduce fiscal risk, and to reduce the costs of a bank failure. While banks are been required to spend almost $1 billion on outsourcing policies to supportthe OBR, it does not appear in the capital review at all - despite the Governor’s arguments that the main justification for capital increases is to reduce fiscal risk.

9 Part three: The risk appetite framework The soundness test The distinctive feature of the soundness test in the risk appetite framework is that it does not require a consideration of the costs and benefits of bank capital. A probability of crisis is chosen, and as long as it can be reasonably described as ‘sound’, it will do. The Bank just happen to have choosen 1:200, but they could just as reasonably have chosen, say, 1:100, or a higher number. Different numbers will delivering very different minimum capital results. The word ‘sound’ does not naturally link to any particular probability. The arbitrariness of the Bank’s choice of 1:200 is obvious. In the risk appetite paper5 that went to the Banking Steering Group for decision, a ratio of 1:100 was suggested, with 1:200 appearing as an alternative, but just in a footnote. In the event 1:200 was chosen6 apparently on the Governor’ whim, because it was more ‘conservative’ than 1:100. This is obviously true, but that does not get us very far. A probability of 1:1000 or 1;1000,000 are both more conservative than 1:100 but that does not mean that they should be adopted. There is no evidence of any analysis on the relative costs and benefits of the 1:100 and 1:200 options. With a 1:100 target the tier one capital ratio, even on the Bank’s modelling, would have been 12-13 percent, suggesting no change in policy settings. Which is probably why the 1:100 target was dropped at the last minute. The efficiency test Under the ‘efficiency’ test, however, the Bank should consider all of the risks, costs and benefits (which involves a consideration of the probability and costs of avoided crises) of different capital levels, which, in principle, should lead it to the right answer. The separate ‘soundness’ test adds nothing to the exercise, because the probability of a financial crisis is already considered under efficiency. Soundness trumps efficiency However, because the soundness test ‘trumps’ the efficiency test if the latter delivers a lower result, it is only the soundness test that matters. Efficiency only comes into play if it delivers a higher ratio, and the Bank is not making that claim.

5 Susan Guthrie ‘Risk Appetite framework used to set capital requirements’ Paper to FSG. 30 October 2018 6 Susan Guthrie ‘Capital ratio calibration’ Memorandum for Financial Oversight Committee

10 The Reserve Bank’s ‘risk appetite’ approach simply doesn’t make sense. Why would anyone make a decision based on a balancing ot cost, benefits and risk, then ignore all that in favor of an arbitrarily chosen probability of having a crisis? What is interesting here is just how late the ‘risk appetite’ framework was in making an appearance. After more than two years of consideration of the regulatory capital ratio, it first appears on 7 November 20187, a little more than a month before the consultation paper was released. As for the paper, we won’t get into the detail here, but we found it a combination of muddle, unsubstantiated assertions and vacuity. The following statement illustrates the point. “Soundness” and “efficiency” have a reasonably clear meaning in everyday language – if it was applied to the family car it would mean it doesn’t break down and filling the tank fits easily within the weekly budget A family car that doesn’t break down too often and doesn’t use to much petrol could cost anywhere from $5,000 to $60,000. The ‘sound and efficient’ metric doesn’t help in making the purchase decision. It ignores the vital cost element – as in practice does the Reserve Bank’s risk tolerance framework. It appears that the risk tolerance framework is someones, late, ‘bright idea’, and that the staff have then scrambled to wrap some ‘intellectual substance’ and ‘analytical and empirical’ support around it. New Zealand banking system is currently unsound? If we accept the Bank’s assertion that a regulatory capital ratio of 16 percent, and an actual ratio of closer to 18 percent8 is required to deliver ‘soundness’, then it follows that the New Zealand banking system, with a CET1 ratio of around 11 percent is currently unsound. As recently as November 2018, in its Financial Stabilty report the Bank said (repeating its public assessments in all of its RSR reviews) that the system was sound. But now, apparently, on the basis of the analysis presented in the consultation document the Bank has come to a different view. The system is unsound. The Bank does not say what the current probability of a crisis is with the current level of capital, but assuming that it is, say, 1:125, this would equate to a credit rating of somewhere between BBB and BBB-9. However, based on the AA- credit

7 ‘Risk appetite framework used to set capital requirements’ paper to FSO. 8 We have assumed that banks will maintain an operating margin over the minimum of 2 percentage points. 9 Widdowson D. Wood A. 2008 ‘A users guide to credit ratings’

11 ratings of the four large banks, New Zealand had a AA- system with an implied failure rate of 1:1250. Of course the banks’ stand alone ratings are lower, and it would be legitimate to ‘deduct’ the implied support from the New Zealand government, but the implied support from the Australian parents is obviously relevant to an assessment of financial strength of the system. Stripping out the New Zealand government support might give us an A rating, with a implied failure rate of 1:750, stil well ahead of the Bank’s assessment. The Reserve Bank Governor attemped to address this issue in a recent speech10 Banks also hold more capital than their regulatory minimums, to achieve a credit rating to do business. The ratings agencies are fallible however, given they operate with as much ‘art’ as ‘science’. Bank failures also happen more often and can be more devastating than bank owners – and credit ratings agencies – tend to remember. The inference here is that rating agencies’ default probability assessments are understated because they have forgotten about historic defaults is simply untrue. Say what you will about the agencies’ rating methodologies, but they are assiduous about long-term data collection. They haven’t forgotten past history. The question is, who is right. Is the New Zealand banking system uncomfortably close to ‘junk’ status, as the Bank seems to be suggesting, or are the rating agencies more or less right in their assessments? The Bank provided two sets of evidence to support its new perspective on banking system risk. The first is evidence from the international ‘literature’ on the probability of financial crises. The second is, evidence from its ‘portfolio modelling’ of New Zealand bank risk. The international evidence The Bank concluded that on the basis of the international studies it has reviewed that: ‘there is a consistency in the findings that suggest that Tier 1 capital equal to or exceeding 16 percent of RWA is needed to limit the probability of a crisis to 1 in 200 years or thereabouts’. RBNZ Bulletin Vol. 71 No. 3 10 Address to Business NZ CEO forum 30 November 2018

12 We have reviewed these studies in detail, first making two adjustments to the measurement of capital in our assessments. Actual capital not regulatory minimum All of the studies cited consider the actual level of bank capital, not the regulatory minimum, so we have to adjust the proposed minimum to the actual amount of capital banks are likely to hold. If we assume that banks will hold a 2 percentage point operating buffer over the minimum, then this implies an 18 percent actual ratio is required to meet the 1:200 standard. Adjust New Zealand capital to international standards We have to compare like-with-like when applying the international default evidence to New Zealand. So we need to adjust the reported New Zealand capital ratios to international capital measurement standards when looking at foreign studies based on historical capital ratios. A recent comprehensive study by PricewaterhouseCoopers (PWC)11 argued that if New Zealand banks’ capital ratios were restated using international measurement conventions, they would be 6 percentage points higher. The Reserve Bank12 went to considerable effort to discredit the PWC analysis, on matters of detail, but, unlike APRA, they did not produce their own ‘official’ version to put the technical arguments to rest. One of the points the Bank made in their review was that PWC conducted a comparitive exercise for Australia that produced an adjustment figure that was one percentage point higher than the APRA estimate. To account for possible upward bias in the PWC adjustment estimate we have reduced it by two percentage point to four percentage points. The Reserve Bank also argued that some of the higher capital accounts higher risk characteristics of the New Zealand financial system that are not captured by the Basel risk measurement framework. The Bank has often made this claim but has never substantiated it with any analysis. On the empirical and analytical evidence it does not appear to be true.

11 PricewaterhouseCoopers “International comparability of the capital ratios of New Zealand’s major banks’ October 2017 12 ‘International comparison of capital ratios (PwC report)’. Information paper 6 December 2017. ‘2017PwC (NZ) study’. Information paper with more details. 8 May 2018.

13 Taking the two adjustments together the appropriate test of the 1:200 standard against the international evidence, is a 22 percent capital ratio. The Basel Committee on Banking supervision (BCBS) evidence The first set of evidence the Bank used is from the 2010 BCBS report, later adjusted by the Bank of England to put the numbers on a tier one capital basis. The adjusted BOE numbers are presented in table one. The estimates do not go up to the 22 percent ratio, but if we extrapolate the downward trend, it is reasonably clear that the probability at 22 percent is well below the 0.5 percent level needed to support the Bank’s argument. Table 1 : BCBS crisis probability estimates Tier one capital Probability of crisis % 8 8.3 11 2.9 14 1.2 16 0.8 A second point is that the numbers presented by the BOE are not actually Basel Committee estimates. Rather, they an average of seven different studies. We went back to the original studies and adjusted them, individually, to a tier one basis. The highest reported capital ratios were 15 percent, which translates to 18 percent of tier one capital. The results were as follows. Table 2: Individual studies reported in BCBS 2010 appendix. 18 percent capital ratio Study Prob. Crisis percent FSA 0.5 Bank of Japan (1) 0 Bank Japan (2) 0.1 “bottom up” 0 BOE 0 BIS 0.7 Bank of Canada 0

14 The average at the18 percent capital ratio is less than 0.2. At a 22 percent ratio, the number would be lower again, say, less than 0.1 percent. The failure rates are roughly in line with the implied rates from the credit ratings of the New Zealand banking system. Our conclusion here is that using the appropriate test, and looking at the actual studies, the results from the seven models are not oven close to 0.5 percent failure rates that would support the Bank’s claims. The Bank of England (BOE) modelling The BOE’s used a bottom-up model 13(using individual bank loss data) and a top - down model (which estimates the relationship between financial crises and country banking system capital ratios) Their results are as follows. On the BOE’s reported numbers they clearly do not provide support for a 0.5 crisis rate at a 22 percent capital ratio. Table 3: BOE probability of a crisis Tier 1 capital ratio Bottom up prob % Top-down prob % Topdown leverage ratio 8 0.5 1.8 3 11 0.4 0.6 3 14 0.3 0.5 5 16 0.3 0.4 6 And there are issues with the analysis. Due to data constraints the estimation period for the bottom-up model was very short (from 1993 to about 2013), a period that captured the most turbulent period in post war banking. If the data period was stretched back to, say, 1970 and forward to 2018, the measured probability of a crisis would likely fall quite sharply. Similarly, for the top-down model (see the discussion of the Firestone model below). The top-down model was estimated using the leverge ratio and then converted to a tier one ratio using the UK average tier one/leverge ratio ratio of 2.63. If the New Zealand ratio of about 2 had been applied to the results, tier one ratio corresponding to a leverage ratio of 6 percent would have been 12 percent, a difference of 4 percentage points from the BOE’s 16

13 Financial Stability Paper No. 35 – December 2015 Measuring the macroeconomic costs and benefits of higher UK bank capital requirements Martin Brooke, Oliver Bush, Robert Edwards, Jas Ellis, Bill Francis, Rashmi Harimohan, Katharine Neiss and Caspar Sieger

15 percent. This demonstrates that our 4 percentage point adjustment for measurement differences was almost exactly right in this case. There is clearly no support for increasing capital ratios in the BOE analysis. It suggests that New Zealand banks are already solidly meeting a 1:200 target. The IMF analysis14 The IMF paper does not make an assessment of the relationship between the probability of a crisis and capital ratios, so it is not altogether clear why the Bank presented it. Presumably it was an opportunity to present the IMF table on non￾performing loans together with some additional high nonperforming loan rates, which we have reproduced below. All the IMF analysis is, is a listing of the peak non-performing ratios for countries that have experienced finanicial crises. It assumed that the loss rate on non-performing loans is 50 percent, which is a crude estimate which will overstate loss rates in most crises. For example, we checked the Italian non-performing loan rate 15. It was 9.5 percent in 2009, not the 18 percent presented by the IMF. The Italian banking system was modestly profitable in 2009 and did not record a disasterous loss implied by the IMF methodology. The IMF then arbitrarily assumed that the 85th percentile has some special significance and that the ‘optimal’ capital ratio is therefore 15 percent. Figure 1 Non performing loan rates in financial crises

14 Benefits and Costs of Bank Capital Jihad Dagher, Giovanni Dell’Ariccia, Luc Laeven, Lev Ratnovski, and Hui Tong SDN 16/04 15 Banco d’Italia Financial Stability Report No. 1 2010

16 The implied assumption behind the IMF analysis is that because the banks in some countries had a high level of losses at a particular point in their history, this outcome is relevant to banking systems everywhere, at all times. This doesn’t make much sense. Take the case of Korea, which sits near the 85th percentile. Korean banks did incur heavy losses in the Asian crisis because: • Korea had a crony-capitalist system, rife with corruption. • Banks were lending to business groups (Chaboels) with debt to equity ratios of 500 percent. • Many loans were in US dollars which were taken out when the Korean currency was fixed to the dollar at an overvalued exchange rate (sharply increasing the value of loans when the currency eventually devalued). None of these circumstances are relevant to New Zealand. The Korean experience (and most of the other high loss experiences) is not necessarily some kind of relevant cautionary tale of what could happen to New Zealand, that justify a regulatory capital ratio of 16 percent. Similarly the Icelandic, Greek, Cypriot and Irish experiences16 generally have little or no relevance to present day New Zealand.

16 The Irish banking disaster was primarily driven by very loose commercial property lending. The loss on this lending was ¢40 billion on a portfolio of ¢70 billion, compared to Irish housing loan losses that peaked at ¢10 billion before subsequent recoveries. The report ‘The Irish Banking Crisis Regulatory and Financial Stability Policy 2003-2008 A Report to the Minister for Finance by the Governor of the Central Bank’ sets out some of the main reasons fo the losses: High LVR limits (80 percent), which were frequently exceeded; statements of net worth accepted at face value; no assessment of impact of lening on propoerty supply; a political imperative to expand the banking system. These factors were not present in New ealand in 2008 so the losses were moderate. They are not present today.

17 The Bank also took the opportunity to present some evidence that is closer to home, inserting the maximum nonperforming loan ratios for two Australian banks (the ANZ at about 10 percent and Westpac at 15 percent) and the BNZ at around 25 percent (all in the early 1990s), in the IMF table. The inference again, presumably, is that big losses are possible, and that if it happened once in recent times, then it can happen again. However, the actual Australian loss outcomes were less scary and show that simply assuming that 50 percent of peak non-performing loans translates into losses is a poor way to estimate bank losses. Marianne Gizycki and Philip Lowe17 report that total losses by privately owned Australian banks through the recession amounted to 16 percent of bank capital at the starting point in 1989, when capital ratios were substantially lower than they are now. The Reserve Bank’s claim that the BNZ’s non-performing loans ratio reached 25 percent appears to be overstated. According to the BNZ 1991 annual report the number is more like 17 percent. Over half of those bad loans were in Australia. The New Zealand peak bad loan rate was more like 10 percent. The BNZ story is well known to bankers, if not to the Reserve Bank. The BNZ had a heavy exposure to property and investment company lending in the midst of a building boom that resulted in a massive oversupply of commercial properrty space and a subsequent collapse in prices. The BNZ rushed into Australia and got involved with a large number of bad deals to boost their lending. There was little or nothing in the way of robust underwriting standards and effective credit controls and reporting. In its 1989 annual report the BNZ Chair reported “little regard was paid to effective monitoring of the exposures to each sector, a situation compouned by suspect judgments and a serious lapse in security administration” and “the rapid growth of the BNZ in recent years has unfortunately been accompanied by some undisciplined corporate lending practices” This is not the situation in the New Zealand banking sector today. Commercial property lending is only about 8 percent of total lending; there is no evidence of an commercial property overbuilding boom; and New Zealand banks are not flinging their money at very marginal Australian lending propositions. The Australian banks’ subsidiaries are subject to parental and APRA oversight. It is worth noting too that bank capital ratios were much lower than they are today. After its first capital injection the BNZ proudly stated that its tier one risk adjusted capital ratio was 5 percent.

17 ‘The Australian Financial System in the 1990s’ Marianne Gizycki and Philip Lowe17

18 The relatively moderate losses in the GFC attests to the effective change in New Zealand underwriting standards and risk culture. Total nonperforming loans peaked at 2 per cent of lending and no bank recorded a loss in any year. Over the two peak years of the recession total credit losses amounted to 1.2 percent of lending, one of the lowest rates in the OECD. Federal Reserve Board (Firestone et al., 2017)18 The Bank describe the substance of Firestone bottom up approach as follows. The authors concluded that when Tier 1 capital was 8 percent of RWA the probability of a financial crisis was 3.8 percent. The authors estimated that, in order to reduce the probability of a crisis to 1 percent or less, Tier 1 capital would need to be 17 percent of RWA or more. The first part of the statement is not accurate. The authors did not conclude that the probability of a crisis, with risk weight of 8 percent was 3.8 percent. They actually assumed this result to line up with an estimate they claim was reported by Laeven and Valcenia 19, and used it to calibrate their model. All of the results for the higher risk weights follow from this calibration. Laeven and Valcenia did not calculate a crisis rate of 3.8 percent. Rather, in the cited paper, they listed all identified financial crises over 1970-2011. Firestone calculated the default rate using only the period from 1989 to 2011, leaving out the earlier more benign years. We calculated a crisis rate of around 1.5 percent for advanced economies, based on the data in the latest, 2018, Laeven paper. We think that the longer data period is more appropriate guide to future failure rates as it captures a more conservative banking period as well as the higher risk strategies that lead to high loss rates in the GFC. If a 1.5 percent financial crisis probability calibration had been used in the Firestone model the calculated numbers for different capital ratios would have been lower. It is not possible to say by how much without a rerun of their model, but the difference would probably have been quite significant given the difference between a 3.8 percent and a 1.5 percent calibration.

18 ‘An Empirical Economic Assessment of the Costs and Benefits of Bank Capital in the US’ Simon Firestone, Amy Lorenc and Ben Ranish 2017. US Federal Reserve Board 19 Laeven L., and F. Valencia (2012), “Systemic banking crises database: An update”, International Monetary Fund Working Paper No. 12/163.

19 The second step in the Firestone model is to adjust for enhancements in liquidity management regulations and improved resolvability post the GFC. These estimates are the most appropriate to the New Zealand system where similar improvements have been made. The results are set out in table four below which shows the probabilities from both the bottom up and top down models. The top-down approach is a standard regression on the relationship between financial crisises, capital and other risk drivers. Two models were specified but in specification 1, capital was not statistically significant and the results should be ignored. Capital was (weakly) statistically significant in specification 2. Table 4: Firestone crisis probabilities Tier 1 RWA Bottom-up model % Top-down Spec 2 % 8 2.6 4.7 11 1.3 2.3 14 1.0 1.1 17 0.7 0.5 21 0.5 0.2 25 0.5 0.1 The relevant capital ratio for the 1:200 test is between 21 and 25 percent. The reported bottom-up result was 0.5 percent but as discussed above it would have been materially lower, using a default probability from a longer historical data period. The top-down result, at between 0.1 and 0.2 percent, clearly didn’t support the 0.5 percent test. Summary of historical experience cited by the Bank. 1.BCBS Only one of seven studies support the 0.5 percent crisis rate. 2. Bank of England Clearly doesn’t support 0.5 percent crisis rate. 3. IMF No information on the crisis rate. 4. Firestone

20 The bottom-up estimates were exaggerated by the short data period and do not provide credible support for a 0.5 percent crisis rate. The top-down approach clearly does not provide support. New Zealand evidence We should not lose sight of the fact that we are dealing with New Zealand banks here, and that it is New Zealand evidence that it is most relevant. Taking a 100 year perspective, New Zealand has experienced many economic shocks that would have lead to bank failures in many other jurisdictions. Reddell and Sleeman discussed20 six such events in a 2008 Bulletin article and if we add the GFC that makes seven. Through that time there was one bank failure, the BNZ, (which actually failed twice, but to the same event). It is difficult to count the number of banks (there were twelve trustee savings banks though much of the 100 years) in New Zealand times the number of years they have been operating, but if we assume an average of say 16 that gives us 1600 bank years, or a failure rate of 1:1600. Of course, caution should be exercised when applying this number to a forward looking risk assessment of the New Zealand banking system. There have been many economic, institutional, legal and behavioral changes over 100 years. But equally, caution should be exercised when applying the international evidence to New Zealand, where those differences from modern day New Zealand are just as pronounced, and probably more so. Several of the countries that whose experience is captured in the data are poor comparators for New Zealand and there is always a risk in using foreign outcomes, when the history and institutional environment is poorly understood by the Bank’s analysts. Portfolio modelling of New Zealand bank risk The Bank says: In order to incorporate the New Zealand context in our analysis, and as a complement to our review of overseas findings, we used a portfolio risk model. We used the model to explore what level of Tier 1 capital might be sufficient to ensure the sector retained the confidence of the market after a large shock. But on the results, we are just told Based on our range of input values for PD, LGD and correlation R, our conclusion from our risk analysis is that Tier 1 capital equal to 16 percent of RWA is sufficient to cover credit-

20 Some perspectives on past recessions Michael Reddell and Cath Sleeman RBNZ Bulletin Vol. 71. No. 2. June 2008

21 related losses (after taking account of provisions) and operational and market trading￾related risk. In other words, a Tier 1 capital ratio of 16 percent of RWA is needed to ensure our banking sector retains creditor confidence after enduring an extreme shock This is a vague statement, which could mean anything. It depends on what they mean by “ensuring our banking system retains creditor confidence after an extreme shock”, and, of course, what inputs were fed into the model. The consultation document was also very vague about the structure and outputs of the model. • No working paper is presented in the bibliography. • They do not explain what model input values were used to generate the 16 percent capital ratio requirement. • There is no information on where the critical correlation coefficient inputs came from. The Bank has belatedly disclosed a relevant technical paper21. It is dated 25 January 2019, so it did not exist when the consultation document was released. A reader might think that this was the writeup of the analysis the Bank relied on to make its decision on the 1:200 target and the related capital requirement. But this is not so. The analysis that really provided the basis for the 16 percent capital ratio decision was presented in a decision paper to the Financial Oversight Committee dated 13 November 2018. The results, which seemed to support the 16 percent capital ratio, were clearly fabricated. Mostly implausible assumptions were fed into what was probably the wrong Basel IRB model to get the ‘right’ result. In particular, the average probability of default was assumed to be 2.8 per cent, when the average PD used by banks in their capital models is 1.1 percent. The Reserve Bank has accepted those estimates as appropriately conservative. However, to the Bank’s decision making committee, which is not close to the numbers, and knows little about bank risk, the analysis would have looked sophisticated and plausible. They would have had no idea that they were being duped. On reflection the Bank obviously thought that the original analysis was embarrassingly bad and that they would need something better to put out for public consumption, when the inevitable OIA requests came. Hence the later paper. The hope, presumably, was that no one would spot the real analysis amongst the host of papers the Bank released.

21 Explanatory note on portfolio risk modelling in the New Zealand Context 25 January 2019

22 As the 25 January paper is the Bank’s best shot at explaining themselves, we have analysed it here. In the paper there is a table, reproduced below, that sets out some results. The 16 percent capital ratio was derived from the column on the left – Initial setting. Table 5: Model inputs and outputs There are some obvious problems with the Bank’s analysis. IRB Model appears to have been inappropriately selected to generate higher capital requirements The Bank initially said that they used a variant of the Basel IRB model, but did not provide more detail in the Consultation paper. In the Explanatory paper we are told that they used the corporate loan model. But on the Bank’s data presented in table 5 these loans account for only 11 percent of total lending. In makes little sense to assume all bank loans are corporate loans. One of the advantages, we presume, from the Bank’s perspective, is that the corporate model has a couple of moving parts that can be inappropriately tweaked, if the objective is to cheat in order increase the measured capital requirement. 1:200 event not modelled The confidence interval reported in their table 4 is 99.7%, which means that the Bank was modelling a 1:333 year event, not a 1:200 event. If a 1:200 event had been

23 modelled all of the calculated capital ratios would have been materially lower. The Bank’s excuse is that the outcomes are uncertain and that using a 1:333 probability somehow accounts fthis. The uncertainty is analysed through their sensivity analysis. The use of a higher confidence interval is not legitimate. Note that in the analysis presented to FSO they used the 1:200 calibration with no mention of a different calibration being necessary to account for ‘uncertainty’. Single representative asset approach not appropriate The Bank uses a single representative asset approach that does not suit the Basel modelling framework. There are several models depending on the risk characteristics of the subportfolios. The correlation coefficients range from 0.04 to 0.24. It is difficult then to assess the appropriateness of the Bank’s aggregate numbers without breaking them down into the component estimates. Further, the aggregated approach does not work when estimating the probability of default (PD) from historical data because there have been such large changes in portfolios structures over their data period. To illustrate the point assume that historically a bank had 85 percent of its portfolio in high risk commercial property and business loans, with a measured probability of default of 5 percent, and 15 percent in low risk housing loans with a default probability of 0.4 percent. Its average PD is 4.31 percent. The bank then changes its portfolio structure to 60 perecent housing loans, and 40 percent high risk loans. Its average default rate falls to 2.24 percent. The bank would not describe its current average PD as 4.31 percent based on its historical portfolio experience. Operating margin forgotten Again the bank has forgotten to allow for an operating margin above the regulatory minimum. The model works on actual capital, not on the regulatory minimum. Assuming an operating margin of two percentage point would, by itself, reduce the capital requirement by that amount. The following are more specific comments on the model inputs. Probability of default: Mean estimate 1.5 percent The Bank could have used the advanced bank PDs, which, as noted above, it has already approved as being reasonable and conservative representations of long run default rates (these are set out in their table 3 reproduced below). All it would have to do is adjust the reported housing PD of 0.89 percent, which is not a bank driven estimate and which substantially overstates the ‘true’ long run PD. A number more like 0.4 percent, which would still account for some as yet unobserved large housing shock, would be more appropriate. A reasonable portfolio average would be 0.8-0.9

24 percent. The reported advanced bank portfolio average PD of 1.1 percent was one of the inputs they ‘considered’, but this is diluted by their consideration of four other sets of ‘information’. All of them are poor proxies for a long run PD estimate. They are:

1. Impaired assets to total loans - 2.0% estimate. • It significantly overstates PD rates because nonperforming assets were held in portfolio for years in the early 1990s. • It is dominated by the BNZ 1989-93 experienced, which, as discussed above has limited relevance for a forward looking assessment. • Average default experiences over 1989-93 will not be applicable now because portfolio structures have changed radically from that period. For example housing loans BNZ’s were about 15 percent of the loan portfolio in 1991, but about 60 percent in 2018.
2. Impaired asset expenses: 0.3 to 0.4%. • Again, this estimate is dominated by BNZ experience. • There is an inference that the impaired asset expense can be converted to a PD estimate by dividing it by an assumed LGD of 20 percent to give a PD of 1.5 to 2 percent. But the reported impaired asset expenses are dominated by commercial lending with significantly higher LGDs, so the implied default rates would be much lower (more like 0.8-1percnt) than the Bank implies.
3. Impaired and past due assets to total assets: 1.4 percent This doesn’t provide any more information than 1. It just uses a different denominator.
4. Stress tests 2.7-2.8 percent An average default rate over a stress test scenario is not a long run average PD estimate, which will also include default rates from benign periods. The stress test number appears to be there just to introduce a high number into the mix. Note that a PD of 2.8 percent was used in the modelling that was presented to the FSO. This was clearly a gross overstatement and one of the embarrassments that the Bank attempted to coverup by reworking the analysis. LGD: mean estimate 35 percent There are two data sources. • The weighted average advanced bank reported LGD of 29 percent. • The stress test outcomes for 2014 and 2017 of 37 and 31 percent respectively. The 2017 result is more relevant as it accounts for a more up to

25 date portfolio structure, but even taking the straight average this gives an estimate of only 32 percent. A mean estimate of 30 percent is more appropriate. The FSO modelling assumed a 40 percent LGD based on the same information. Correlation coefficient: Mean estimate 0.28 From the January paper we are told that the correlation was derived from the Bank’s TUATARA22 23model which was built in 2011 as a cost benefit model that would have particular regard to key structural features of the New Zealand banking system (in particular its substantial foreign ownership). The correlation coefficient was a ‘made￾up’ number designed to fatten the tail (by a factor of 10) of the Basel model, which assumes that default rates follow a normal distribution. While it was appropriate to make some adjustment to the Basel model, the TUATARA estimate did not have any empirical or analytical support and is overstated. While this might be acceptable in a the original ‘proof of concept’ cost benefit model, we would expect the Bank’s current modelling, which is driving a 30 billion dollar plus decision to be backed by much more substantive analysis, and properly documented. From the Bank’s consulation paper, and the information provided to the FSO, it is clear that the modeller(s) did not understand what the Basel correlation coefficient is. They thought that it was (presumably in the context of estimating housing loan capital) an estimate of the statistical correlation between GDP and house prices. It is not. It is a measure of how defaults within a housing portfolio are correlated. The higher this correlation the more loans will fail together and the higher the capital requirement. It is difficult to estimate empirically24, but it cannot be read off a GDP/house price correlation. However, on the basis of an GDP/house price correlation presented to them, the FSO Committee was probably lead to believe that the Basel calibrations were nowhere near conservative enough for New Zealand conditions. Table 6 : RBNZ IRB bank estimate table

22 TUATARA is an acronym that stood for ‘Teaching Uninformed Australians To Assess Risk Appropriately’. This is unfair to Australians. The reader might wish to substitute, according to circumstance and taste: Americans, Authorities, Amateurs, Assess or Adrian. 23 Disclosure: the author of this paper built The TUATARA model. 24 There was an empirical estimate of .025 before the Basel housing model was calibrated, but this was based on pre-GFC data. A figure of 0.15 was adopted on the basis of some reverse engineering of Federal Reserve Board housing loss modeling with a margin for conservatism thrown in.

26 Re-estimating the model We do not have access to the full suite of Basel IRB models that would allow us to rerun the model using reasonable input estimates. However, the Bank did a sensitivity analysis that provides some guidance. The lowest set of inputs (PD =1.0 percent; LGD =30 percent; correlation = 0.24), might be a reasonable best estimate. These assumptions generate a capital ratio of 10-11 percent at a probability of 1:333, suggesting a ratio of around 8 percent might be about right for a 1:200 probability. It could be lower again if the Bank manipulted the corporate model to produce higher results. Sensitivity analysis The information paper reported on a sensitivity analysis which showed how the capital requirement respond to changes in the model inputs. The technical issue here is that they used a uniform distribution to estimate the ‘true’ distributions of the inputs. This is an extreme assumption, which assumes that the highest and lowest inputs have the same probability as the best estimate. The uniform distribution assumption has the effect of blowing out the tails of the distribution, making it appear that there is a relatively high risk that the ‘true’ capital requirement could be significantly higher than the mean estimate. A distribution that is more centred around the best estimates should have been used. A back filling exercise The Bank’s modelling exercise looks to be an obvious ‘backfilling’ exercise. A decision had been made to have 16 percent CET1 regulatory minimum, and this had to be consistent with a 1:200 failure rate. But there was no New Zealand evidence to

27 support it, so they ignored what relevant New Zealand evidence that was available, and trumped up some to suit their story. We understand that The Reserve Bank has briefed bank risk managers on their modelling, and have said they will be providing more information in March. The bankers were told that the Reserve Bank was not interested in the the banks’ views on the model inputs because the Bank was going to use inputs which are ‘good for New Zealand’. This is admitting that they will ignore any facts or analytical evidence that doesn’t suit. Apparently, the Reserve Bank is operating in a post-fact world.Ifthe bank says an inpit is ‘good for New Zealand ‘ then it is a fact. The portfolio risk modelling ‘analysis’ should not be taken seriously and should be withdrawn by the Bank, before they further embarrass themselves. The exercise modelling exercise was a sham and they might as well admit that the 16 percent ratio is mostly based on gut feel, or, perhaps, a revelation from the tree god. Part four: Stress testing evidence On the face of it the Reserve Bank’s stress testing results do not support higher capital ratios. The banks get through a pretty severe stress test with a substantial capital buffer in hand. The Bank’s response, in the consultation paper and elsewhere, is to downplay these results, suggesting that the outcomes are really too optimistic, and that losses could be much worse. Our view is that the Bank’s concerns are overstated, and that they have artificially boosted some outcomes to paint a more negative picture. A fair assessment of the 2017 stress tests is that the more likely outcomes are actually significantly more favorable than the results presented in the consultation document. Our assessment is based on the description of the stress test in two 2018 RBNZ Bulletin articles 25. Some of the more obvious issues are as follows. Dairy lending losses overstated It is assumed that the dairy payout falls to $4.90. No information is given on the magnitude of the assumed commodity price and exchange rate falls that should have driven this result. However, we are told that the exchange rate fall, that is used to generate losses from a counterparty failure, is ‘substantial’. On plausible assumptions on the size of the dairy commodity price and exchange rate shock, it is likely that the dairy payout shock would be much less than assumed,

25 Dunstan and Lilly

28 or even be positive. The Bank probably did the numbers, which would have shown these kinds of results. Which is probably why the key commodity and exchange rate assumptions were not presented. Bank mitigating actions not taken into account In the 2017 stress tests banks estimated that mitigating actions would improve their capital positions by 1.1 percent. While the Bank presented this information in the stress test write-up, it was not part of the ‘headline’ results that are presented in the consultation document. Most of the banks’ actions are plausible and should have been used, at least in part, to give a balanced picture of the stress test outcomes. Increase in risk weights partially unnecessary. An increase in risk weights decreased capital by 1.5 percent. The main reason is that banks using the advanced capital models are marking their collateral to market, which pushes loans into higher risk weight classes as asset prices fall. This ‘procyclicality’ is an undesirable and mostly unnecessary characteristic of a capital adequacy regime. It appears to be partially mitigated in the Basel framework by the simple expedient of applying a minimum advanced/standardized capital ratio of 72.5 percent. If the measured ratio is, say, 50 percent and this increases to 65 percent in a cyclical downturn then this has no impact on the actual capital requirement, which is still determined by 72.5 percent.26The New Zealand regime could, and should, be set up to reduce the systemic price effect. It already does so, in part, on the upside of the cycle. Banks cannot mark the value of their housing collateral to market as prices increase. Operational risk loss is implausible It is assumed that banks will be subject to a successful class action suit (costing 0.6 percent of capital) because of inappropriate lending, and that this will impact on capital in the third year of the stress test. The prospect of a successful suit in New Zealand, which does not have a class action friendly legal environment, is a stretch (what for – advising home buyers that house prices never fall or that they will never become unemployed?), but the likelihood that it will impact in the third year must be close to nil. A case would take years to get organised and court action would drag on for more years. Housing default rate is 12 percent This looks to be implausible given the severity of the house price and unemployment shocks, which are the main drivers of housing mortgage defaults. This was forced on the stress test to increase the loss rate and is not a fair mean estimate.

26 There will still be a cyclical element in the Standardised model

29 Capital deductions Capital deductions reduce capital by 0.8 percentage points. This is a material number, but there is no explanation in the Bank’s bulletin article of what this deduction is for, and how it is related to the stress test scenario. Taking these first five factors together could reduce the measured impact on the capital ratio by nearly four percentage points. Dairy lending (say) 1.0 Capital ratio measurement (say) 0.5 Bank mitigations (say) 0.8 Operational risk 0.6 Housing 1.0 Total 3.9 In any event the stress testers did not think their results supported a 16 percent regulatory capital ratio. In a report to FSO dated 28 August 2017 on the implications of the stress tests for capital requirements, they advised an increase in the CET1 capital ratio to 12-13.5 percent (a 12.75 percent midpoint). And that was before they were aware of the 11 percent increase in capital that the new capital floors would impose. That would bring their required rate back to about the level banks will hold after the floors take effect. Losses in other jurisdictions In the stress test section the Bank sets out a figure, reproduced below, with the cumulative bad debt charges in several other jurisdictions, in particular the countries affected by the Nordic crisis in the early 1990s. Again, the figure is there to give the impression that in reality things could turn out much worse than the stress tests suggest. This figure is misleading. • It excludes several countries, which according to the IMF definition had a ‘financial crisis’, but did not experience such high bad debt charges. • Most importantly, the comparison does not account for the difference in portfolio composition between the Nordic countries that feature prominently in the table, and current New Zealand bank portfolios. The Nordic banks’ lending was much more heavily concentrated in sectors affected by the steep fall in commercial property prices. They also had poor risk management systems. Despite sharp falls in residential property prices (equivalent or worse than the New Zealand 2017 stress test assumption)

30 mortgage losses were low – 1 to 2 percent of loans. The losses were concentrated in the commercial sector. The Bank’s response to this is argued in a recent bulletin article27 on the stress test results. Historical experience and insights from stress tests suggest three key areas of upside risk to losses. First, residential mortgage losses could be higher than expected. Although banks’ loss estimates are consistent with historical evidence prior to the GFC (Kragh-Sorenson and Solheim, 2014), household debt levels have increased significantly in recent decades and some countries experienced much higher loss rates on household loans during the GFC. The most relevant country experience here is Denmark. House prices fell by nearly 30 percent in the GFC, and there was a sharp downturn in economic activity. But the losses on residential mortgage lending came to only about 0.8 percent over 4-5 years.28 Notably, Denmark had the highest household debt to income ratio in the world at that time, and still does now. At around 250 percent it is substantially higher than New Zealand’s current ratio of 160 percent. Two of the key reasons for Denmark’s low loss rates: relatively high buyer equity; and creditor friendly laws, are shared by New Zealand. There have been heavy losses on residential mortgages in a few other developed countries, but for reasons not shared by New Zealand. The US has a number of institutional and behavioral features that predisposed them to higher loss rates, particularly in the subprime sector. Icelandic banks incurred heavy losses in part

27 Dunstan A. 2018 ‘The Reserve Bank’s philosophy and approach to stress testing’ Vol. 81 No.8 28 Jesper Berg Morten Bækmand Nielsen James Vickery ‘Peas in a Pod? Comparing the U.S. and Danish Mortgage Finance Systems’ Federal Reserve Bank of New York Staff Report No. 848

31 because homeowners were borrowing in foreign currencies and got caught when the currency was sharply devalued. Then there is the Irish experience, which is a frequent driver of the Reserve Bank’s concerns about the New Zealand mortgage market. Our review of the provisioning experience by the major Irish mortgage lending banks showed that the accumulated losses over 5 years from 2008 were about 8 percent of the starting loan balances, but this fell to 6 percent with write backs as the property market recovered. The Irish losses can be explained by: • A 55 percent fall in property prices. The losses would have been substantially lower with the 35 percent fall assumed in the New Zealand stress tests as the Irish banks’ provisioning was driven by the fall in the housing price index. The 55 percent fall, was in part, collateral damage from the speculative property boom. New housing builds outstripped demand and helped drive house prices down. • Underwriting standards were very low, with borrowers being able to access 100 percent plus mortgages in some cases. • A court decision meant that Irish banks could not foreclose on defaulting homeowners. This drove up default rates as, and losses as borrowers, who were underwater on their investments, did the rational thing and stopped paying. They could live mortgage free and had a free option on a house price recovery. None of these risk factors apply in the New Zealand market. The Australian experience is also reported, and again it has limited relevance. It is affected by heavy losses by the State banks, but privately owned banks had much lower loss rates. The banks’ portfolios at the time had a limited exposure to housing, which did not incur material losses; and even the private banks had weak risk management systems by current standards.

Part five: The optimal capital literature The Bank’s analysis of the ‘efficient’ level of bank capital, which considers both the benefits and costs of higher capital, is cursory, in part because it does not really matter. The capital ratio is set under the soundness test, so the costs and benefits of

32 the policy becomes irrelevant. The only purpose of the analysis, apparently, was to claim another ‘win’ so the Reserve Bank would have its ‘win-win’ policy. The discussion starts by drawing attention to the fact that the optimal capital models in the international literature does not account for risk aversion, with the inference that the optimal capital ratios would be higher if risk aversion were taken into account. It is true that the foreign models do not account for risk aversion, but this makes it even more difficult to understand why the Bank ignored its own TUATARA model, which explicitly accounts for risk aversion in a structured manner. The discussion then turns to some of the key inputs in the models including the cost of a crisis, and the increase in borrowing costs and their subsequent impact on economic output. The increase in borrowing costs is put at 6 basis point per percentage point increase in capital, but this inconsistent with the 8.2 percent increase presented in a more detailed discussion in the FSO decision paper. The cost of a crisis is put at 63 percent of GDP. This was the BCBS 2010 mid-estimate from all studies, which was boosted in some by the inclusion of high developing country costs, and to an extent by the permanent shock assumption. There was no discussion of why the BCBS midpoint was selected over other later analyses, which suggest that the earlier estimates presented an overblown account of the cost of crises. However, this was all just padding. The preferred inputs did not really go anywhere. The Bank simply presented the results of an array of optimal captal studies without any attempt to assess or amend them on the basis of their key inputs, still less adjust for the differences in capital measurement between New Zealand and foreign countries or allowing for an operating margin as discussed above. But they then go on to conclude ..the results of our modelling work, and our read of the international literature, suggest that the Tier 1 capital ratio of 16 percent that we think will meet our desired soundness objective is within the bounds of an optimal capital ratio, taking into account New Zealand conditions. In our view, at a Tier 1 capital ratio of 16 percent there would be little room to increase stability further without some impact on expected output. This is not saying very much because the bounds they presented were very broad (7 to 26 percent). Even then it appears not even one of the studies lends much support for a 16 percent CETI capital ratio. If we look at the nine model results and assess them against a 22 percent capital ratio test, to take account of the operating margin and setting the New Zealand rate

33 to the international measurement standard, eight of the models do not support the Bank’s proposal. This leaves just the Firestone results which ranged from 13 to 26 percent. However, the higher result relied on a cost of crisis estimate of 100 percent, substantially above the Bank’s preferred number of 63 percent. Switching to the latter ratio would likely have generated a capital ratio below 20 percent, so the Firestone support for the Reserve Bank’s policy falls away. The Bank also says that they complemented the international studies by conducting their own modelling exercise. This was the TUATARA model which was used to support the Bank’s Basel III decisions with an estimated optimal tier I capital ratio of 13 percent. The Bank went to some effort to review and formalise the model in 2016, but then they did nothing with it. It is played no part in their decision making on the ‘efficient’ level of capital. They claimed that this was because of uncertainty due to the ‘wide range of plausible outcomes depending on model assumptions and inputs’. This is disengenuous. All models will produce a range of outputs if their assumptions are changed. The international models reviewed, and used by the Bank also present a range of outcomes with different assumptions. It appears that the real reason the model wasn’t used was that certain element in the Bank were on a mission to increase bank capital and they thought that the TUATARA model would not be helpful to that end. Table 7 : Bank’s Summary of Optimal capital results

34 The TUATARA model Here it is useful to briefly look at the TUATARA model. The model follows the same basic framework of the other models considered by the Bank. The cost of additional capital is a function of the effect of the additional capital on lending rates. The benefits are a function of the reduction in the expected cost of financial crises. This in turn, is a function of the probability that there will be a financial crisis and the cost of a crisis, should it occur. In this framework the capital ratio is successively increased to the point where the marginal benefits from the lower cost of capital equals the cost of capital. This is the optimal capital ratio. The following figure taken from the Reserve Bank’s Regulatory Impact Statement shows some of the key TUATARA outputs. Figure 2 TUATARA model outputs

35 The red line shows the benefits of additional capital, and the purple line the costs. The optimal level of capital is where they intersect at about a 13 percent tier one capital ratio. The shape of the benefit curve is typical of all the models. They start off high and then tail off quite sharply. While the optimal capital level is 13 percent in the TUATARA model there is little additional benefit (the difference between the red and purple lines) in moving from 11 to 13 percent. Similarly, if the red line was positioned higher (say by increasing the assumption on the cost of a crisis), the optimal capital ratio would increase but there would be limited benefit. The different shapes of the benefit costs and benefit lines have something to say about decision making under uncertainty. If we underestimate the benefits and set the capital ratio too low, then the losses are small. If, however, we assume that the costs are low because there is little interest rate effect (as this particular model setting does), but it turns out that the interest rate effect is much higher, then the cost of the mistake is much greater than a benefit mistake. The TUATARA model differed from the foreign models the Bank has examined in the following key respects. Explicitly takes account of risk aversion It does this by increasing the value of the avoided costs of crises. The bigger the crises, the bigger this affect. The impact of the risk aversion affect can be seen in the figure above. The blue line shows the benefits without the risk aversion effect and the red line with it. The impact is to increase the optimal capital ratio from 10.5 to 13 percent. The explanation for the risk aversion calibration in the RIS is as follows.

36 To account for this effect (risk aversion), we have multiplied the expected dollar benefits of the GDP and fiscal savings by a factor that is intended to proxy society’s risk aversion. The factor varies from just over one for small expected shocks, to more than two for the large shocks that enhanced capital ratios are designed to provide protection against. The ratios were calculated using an expected utility model that assumes a relatively high degree of risk aversion reflecting society’s wish to reduce the risk of socially and economically disruptive events. The higher ratios are also consistent with the revealed preference in the New Zealand life insurance market, which covers low probability but high impact events. Average premiums are over twice expected payouts to policyholders. Accounts for the direct cost of capital increases on lenders In the foreign models the increase in the lending rate is not a cost directly. Rather it has a negative impact on GDP, which is then counted as a cost. This is the output effect of a capital increase. The reason why the higher lending cost to borrowers is not counted is that the banks are assumed to be domestically owned, so the cost to borrowers is offset by increased bank profits. From a national perspective the higher interest cost is a transfer, there is no net cost. This assumption does not hold in New Zealand, where most of the banking system is foreign owned. The transfer effect should be counted as a cost. The effect is significant because the interest rate transfer effect is higher than the output effect in most models. Typically, the output and transfer affects together are two and a half times the output effect. This means that capital is much more expensive in New Zealand than in foreign jurisdictions and, other things being equal, the optimal capital level will be lower. However, by looking at just the foreign models in their ‘efficiency’ assessment the Bank has completely missed the transfer effect. Around $1.5 billion in increased interest costs to borrowers has been ignored. If the foreign models were adjusted for the transfer effect to be fit for purpose for New Zealand then their optimal capital estimates would have been materially lower than the reported numbers. Modigliani Miller effect strong The model was calibrated with a strong, 85 percent, Modigliani Miller (MM) offset effect (the MM offset accounts for a fall in the required return on capital because banks with a higher capital ratio are safer). This meant that there was only a small interest rate effect and the costs were corresponding low. It is now generally agreed that this was too optimistic. The Reserve Bank’s view, based only on the international evidence is that the offset is around 50 percent. However, that figure could be lower given New Zealand specific circumstances.

37 A recent (December 2016) Centre for International Finance and Regulation paper29 (which the Bank did not considered) estimated the MM offset for Australia. Using the leverage ratio measure of capital adequacy, no MM offset was found. Taking a market value approach to measure capital, it found that the MM offset was 25 to 30 percent. The study did not offer any explanation as to why the offset effect was lower than the international average, and so much lower than the theoretical effect. A possible reasons is that the Australian system is dominated by four very similar banks who believe in the ‘fixed cost of capital’ theory, and who have the market power to enforce it. A further test of the MM offset effect in Australia was the ‘natural experiment’ of the increase in residential mortgage capital. The Reserve Bank of Australia noted, in its October 2016 Financial Stability Report, that housing loan interest rates had increased by 15-20 basis points in response to the capital increase, but that there had been some subsequent discounting. If there was no MM-offset then the expected rate increase would have been around 12-13 basis points, suggesting that the initial increases were beyond those required to maintain a target return on capital, and that the subsequent discounting may have been a return to a zero MM offset effect. If the offset effect is weak in Australia, the same outcome would apply in New Zealand, but more so. The system is dominated by subsidiaries of the Australian big four banks, which from a capital pricing perspective, are just operating divisions of larger organisations. They are distant from the market forces that, theoretically, could drive down the required rate of return on capital. The big four banks already face limited competition from New Zealand banks and small international competitors. That competition will be further reduced by the Bank’s capital increases, as the smaller banks are mostly capital constrained It would be reasonable to apply a lower MM offset to New Zealand than the international evidence would suggest. A twenty-five percent offset could be appropriate. The upshot, whether a 50 percent or 25 percent offset is used, is that using the TUATARA model, as currently calibrated, the optimal capital ratio would fall to under 9 percent.

29 Linh Nguyen & James Cummings December 2016’Impact of Higher Capital requirements on bank funding costs: Australian evidence’ CIFR WP No. 132.

38 The TUATARA model was obviously not the last word in optimal capital modeling for New Zealand. It was built, rather hurriedly, over seven years ago. The model inputs could be recalibrated, and the model itself could be enhanced. But, however that was done it is unlikely that it could provide very convincing evidence to support that the large increase in capital as the Reserve Bank is contemplating. Which is probably why the Bank ignored it. Part six: What difference will higher capital make to the cost and frequency of financial crises in New Zealand? In this part we consider, in more detail, the evidence on the cost of financial crises and look at question of what difference the increase in bank capital will make in a severe New Zealand crisis. How costly are financial crises? The Reserve Bank’s position on this question is that financial crises are very costly, with long lasting effects, so it is almost self-evident that reducing the probability of crises with more capital is a good thing. First a word on what is a financial crisis. It is a fuzzy concept, but a convention for their identification has emerged based the definition developed by the IMF’s Laeven and Valencia. A systemic crisis is defined as a national event that meets at least 3 of 6 criteria. Some of these are determined somewhat subjectively, and some have a quantitative trigger. The most important criteria are: widespread liquidity support; government guarantees, and fiscal support. The trigger point for the latter is 3 per cent of GDP. So a crisis is not necessarily the massive economic shock experienced by Ireland. It is helpful to keep this in mind when think about the cost of crises.

39 There have been a very wide range of estimates of the cost of crises (from under 10 to 400 percent of GDP) reported by various analysts. The key points of difference in the estimates, are first, whether the costs of the crisis are temporary, relating to the loss in GDP over a defined period of the recession associated with the crisis, or whether financial crises, leave a permanent scar that reduces the level of GDP for all time, or for a very long time. The second issue is the extent to which the GDP losses in the recessions associated with the crisis can be attributed to the banking crisis as such, and the extent they relate to underlying economic events. Permanent versus temporary effect The difference between the permanent and temporary effect of a financial crisis is illustrated by the following stylised diagrams that were presented in the BCBS report. Example 1 in their Graph 1 shows the cost of a temporary shock. Output eventually catches up with the pre-crisis trend and the GDP loss is the marked area under the trend GDP line. In example 2, the rate of growth returns to its pre-crisis level, but the new GDP trend is below the previous trend growth. This difference is treated as permanent, and the permanent cost of the crisis is the present value of the difference. This assumption can generate some big cost of crisis numbers. For example, if the shortfall is 5 percent of GDP, and a discount rate of 4 percent is used the present value of the permanent effect of the crisis is 125 percent of GDP. . The permanent loss approach is open to some obvious criticisms. First, it is highly dependent on the assumed pre-crisis trend growth rate. However, that trend will have been biased upwards by the pre-crisis ‘excesses’ that will have been building for a number of years. The observed lower level of post crisis GDP might be a return to something closer to more normal levels. Second, while might be reasonable to assume that some ‘scarring’ effects from a

40 crisis, there is no compelling reason to expect that the scars will last for ever. If that were the case, a county’s history of financial crises might be the dominant cause of its current level of GDP. US GDP would still be bearing the scars of the depression and of multiple crises back in the 19th and early 20th century. The Dutch economy would still be bearing the scars of tulip mania. There is also a tendency to overstate the temporary effect of a financial crisis. Prior to the GFC many OECD countries, had a credit and property price boom, and when they busted, or petered out, it was inevitable that there would be some impact on GDP, regardless of the banks’ capital ratios, or whether or not there was a domestic systemic financial crisis. A recent paper 30, carefully considered some of the issues in identifying the effects of crises on GDP, adjusting for some of the methodological problems that tended to overstate their effects in earlier studies. Cline then compared GDP losses, over 5 years, for the countries that were identified as having experienced a systemic banking crisis, with a group that (including New Zealand) that did not. The former had a GDP loss of 23 percent. The latter 21 percent. There was just a just a two percentage point difference. The net fiscal costs of crises The Governor’s speech, cited above, focused on the fiscal cost of crises, arguing that they could be very large and that the increased capital would help mitigate the government’s fiscal risk. There have been some large gross fiscal costs in some banking crises but for the most part they have been recovered. The following table on the gross and net fiscal costs is taken from Laeven and Valencia IMF crisis database. The peak nonperforming loan data gives a sense of the scale of the crises. The average gross fiscal cost was 11.3 percent of GDP, but the average net cost was 5.8 percent. Most of the cost was driven by a few outliers. Excluding Greece, Ireland and Korea the net average comes to 2.7 percent31. The best comparators for New Zealand would be the Scandinavian duo of Sweden and Norway, who nationalised their main banks in their 1990’s crises32. These banks had

30 William R. Cline 2016 ‘Benefits and Costs of Higher Capital Requirements for Banks’ Petersen Institute WP 16- 2016 31 The UK cost was presented as 3.8 percent but the cost was probably lower. Mor “Bank Rescues of 2007-09: outcomes and cost” House of Commons Library 2018, estimated the cost to be about 1 percent of GDP. 32 In Finland’s case the losses were concentrated in the savings bank sector, which had transformed itself into something of an investment banking role, amongst other things making risky foreign loans and taking over

41 strong underlying franchises, which were eventually reflected in the selling prices, limiting the governments’ net losses to close to zero. If the government was forced to nationalise one or more of New Zealand’s big four banks then it may well come out ahead once the economy recovered. The New Zealand banks have strong underlying profitability, as have their parents. Table 8: Gross and net fiscal costs of banking crises Peak Nonperforming loans % Gross fiscal cost %GDP Net fiscal cost %GDP Austria 2008 4.1 5.3 1.6 Belgium 2008 NA 6.5 0.5 Denmark 2008 6.0 5.9 2.4 Finland 1991 13.0 12.8 11.1 France 2008 4.5 1.3 1.1 Germany 2008 3.7 2.7 0.7 Greece 2008 37.1 28.7 17.1 Iceland 2008 61.3 37.6 3.3 Ireland 2008 25.7 37.6 26.8 Italy 2008 18.0 0.7 0.7 Japan 1997 35.0 8.5 8.5 Korea 1997 35.0 31.3 23.2 Luxembourg 2008 1.7 7.2 5.0 Netherlands 2008 3.2 14.3 5.1 Norway 1991 16.4 2.7 0.6 Portugal 2008 12.9 11.1 7.6 Spain 2008 9.4 5.4 4.8 Sweden 1991 13.0 3.6 0.2 Sweden 2008 2.0 0.2 0 UK 2007 4.0 8.8 3.8 US 2008 5.0 4.5 0.6 Average 11.3 5.9 domestic companies. They were very bad at it, and after their rescue, these operations were closed down. There was no franchise to sell to offset the losses.

42 The other fact to remember is that the Norwegian and Swedish banks were, by current standards, very weakly capitised prior to their crises. Englund33 reported that the Swedish CET1 leverage ratio was 2 percent. The main reason that commericial property loans with an LVR of less than 75 percent had a 25 percent risk weight. The capital backing a commercial property only had to be two percent of the loan value. If the Swedish and Norwegian banks were as well capitaised as New Zealand banks are now, then there would have been no net fiscal cost. Of the three big fiscal cost outliers Korea is noteworthy because of its rapid recovery, after the crisis. After less than three years the economy was back on its previous growth path and the increase in public debt was very manageable. Greece, of course, has a host of problems not relevant to the New Zealand situation (we have a sound fiscal position and a floating exchange rate), which again leaves just Ireland, with a net fiscal cost of 26.8 percent of GDP. As discussed above Ireland’s banking problems where, in large part where due to incompetent bankers and supervisors and a permissive political environment. There is no evidence that New Zealand bankers are exhibiting the kinds of behaviours that drove the Irish disaster. New Zealand also has the advantage that our large banks are effectively supervised by APRA, which is competent. What difference would an additional 5 percentage points of capital make in a severe financial crisis? The point we want to reiterate here that large financial crisis do not occur randomly, rather they are mostly initiated and driven by real economic processes. The costs of those processes will be incurred regardless of the amount of capital the banking system has. More capital is not some magic pill that makes all of the pain go away. The argument for higher capital, therefore, is that it reduces some of the follow-on effects that can exacerbate the initial shock, and it reduces the government’s fiscal risk. There are a number of follow-on effect channels. The first is a credit crunch, caused by a change in banks’ credit policies. There are two main drivers here. First, banks will be experiencing credit losses which are outside their experience and will

33 The Swedish 1990s banking crisis A revisit in the light of recent experience’ Peter Englund. Paper presented to the Riksbank Macroprudential Conference, Stockholm 23-24 June, 2015

43 naturally tend to pull back from advancing any more loans to problem areas. There will also be a natural shock on the demand side. Property developments will not proceed, and home buyers will become more cautious. It is unlikely that more capital will play much of a role in mitigating these effects. The second is related to the prospect of breaching regulatory capital requirements. Banks will be more reluctant to lend if they see this as a risk, as their capital is being eroded, or they think it might be further eroded through future credit losses. The higher the level of capital the weaker this effect. The Bank’s proposals are designed to be more permissive in this respect. The capital is intended to be used, so there could be a positive effect on the economy through this channel. The third channel is some kind of general impact on business and consumer confidence. The higher the level of capital, the argument goes, the less overly cautious consumers and businesses will be. However as the Cline analysis discussed above suggests the positive effect through these channels is unlikely to be very large. Non-crisis countries, with sound banking systems did not fare much better than crisis countries through the GFC. How would a real crisis play out? We assume here that New Zealand banks have an 18 percent CET1 capital ratio while Australian banks have 13 percent and an additional 5 percent of loss absorbing tier 2 capital. There are three main possibilities

1. A New Zealand specific crisis: 30 percent chance It is assumed that the economic shock is largely specific to New Zealand (say a complete and sudden collapse of the dairy industry that sparks a very large fall in house prices), but Australia is mostly unscathed. Here there is a high probability that Australian parents will provide their subsidiaries with more capital. As discussed above, the subsidiaries have solid franchises (probably worth as much as their accounting equity) worth preserving, and for reputational reasons an Australian bank would be reluctant to walk away from its subsidiary. In this scenario it is not necessary to require the capital ahead of time, paying for the cost of capital, when the support would be provided just when it is needed.
2. A trans-Tasman crisis: 65 percent chance

44 In this scenario banks in both countries are under stress. At first sight the New Zealand subsidiary banks should be more robust against failure because they have more CET1 capital, but in a situation where the course of the recession and future losses are highly uncertain, they will also be highly dependent on the fate of the parent banks. If the parents fail there will be a strong contagion effect on their New Zealand subsidiaries and the additional capital will be of limited value. If the Australian banks require government intervention, then so will the New Zealand banks. So, it is likely that there will be a coordinated rescue effort. In our scenario the banks are guaranteed and are recapitalised by their respective governments. The Australian loss absorbing capital will be ‘bailed-in’ so the effective capital position in both countries will be the same. The New Zealand government contribution will be less than it would have been without the proposed capital increase, but so will the Australian government’s contribution. There is a fiscal savings in both cases. The key takeouts here are: first if the objective is to reduce the probability of government intervention, then additional CET1 capital in New Zealand is of limited value. The fate of the New Zealand subsidiaries will largely be determined by the fate of their parents. Second, tier two capital is just as effective in mitigating fiscal risk as CET1 capital. But it comes at around one fifth of the price. A $1.2 billion a year savings. 3. New Zealand goes it alone with the OBR: 5- 10 percent chance? The bank(s) would be placed in statutory management and would be effectively recapitalised by depositors. This option is unlikely to be used. Australia is likely to (fiercely?) resist a New Zealand go it alone response, and the OBR is a high-risk strategy. It has never been tested and the mechanism for dealing with complex wholesale instruments in an OBR has never been built. There is a high probability that a government will go with the trans-Tasman government recapitalisation option, weighing the overall trans-Tasman relationship, and the likelihood the net cost of the recapitaisation will not be too high, the risk of the OBR alternative, against the immediate fiscal savings. In essence, then, the decision to increase CET1 capital comes down to whether it is worthwhile to hold capital now that will reduces the possible cost to government in the event there is an acute crisis. There is a very limited reduction in the other costs of a crisis, which we put at 30 percent of GDP or around $100 billion.

45 The cost of the capital increase is $1.5 billion a year. The benefit is a maximum reduction in government’s gross borrowing of under $20 billion (the amount of the capital increase). Let us assume that half of this is recovered, for a net cost of $10 billion. So the capital increase reduces the cost of the crisis by about 10 percent. If we assume (generously) that the probability of a very severe trans-Tasman crisis is, say, 1:300 the expected benefit is $33 million. So if the Bank’s policy can be characterized as an insurance policy, the purchaser pays an annual premium of $1.5 billion for a policy that reduces the cost of the bad outcome by 10 percent and has an expected payoff of $33 million. Would any informed rational New Zealander buy such a policy? And would the Reserve Bank run afoul of conduct rules by selling it to the public?

Part seven: The Australian Option The obvious omission from the discussion paper is any consideration of the Total Loss Absorbing Capital (TLAC) approach taken by APRA34 in their recently released discussion paper. APRA proposes a five percentage point increase in capital, which can be met from tier two capital. The increase only applies to domestically significant banks, which means that small banks are not affected. By contrast the Reserve Bank proposes that small New Zealand banks get only a one percentage point reduction from the large bank requirement. The question is why hasn’t the RBNZ taken the APRA route, or even discussed it. It would align the capital frameworks, which would probably make it easier to jointly manage a crisis, should one ever arise. At one-fifth the price, it would mean a savings of $1200 million a year.

34 APRA November 2018Increasing the loss-absorbing capacity of ADIs to support orderly resolution APRA November 2018

46 There is a general discussion of the TLAC approach in one of the policy documents released on 25 January.35 The discussion came late very late, (11 October 2018) when the Reserve Bank should have been known what APRA was proposing, but there is no reference to it in the paper. Instead it was explained that they hadn’t got around to considering TLAC, and they were, at this very late stage, considering it for the first time. There is a discussion of the UK, and Swiss TLAC requirements, and then the discussion gets bogged down by conflating two logically separable issues: the arguments for requiring bail-in capital; and how failing subsidiaries in a banking group should be managed. The Bank seems to think that tier 2 TLAC capital can not be issued inless the failure co-ordination issue with Australia is resolved, which is obviously not the case. APRA has had no problems with proposing a TLAC requirement without ‘solving’ the trans-Tasman failure issue, or possibly even talking to the Reserve Bank about it. An important consideration in the Reserve Bank’s aversion to a TLC approah, is that they wold have to use internal resoures to evaluate the capital instruments. A saving of $1.2 billion should be sufficent to ensure that the Bank could hire at least one competent analyst, or outsource the evaluation to an experienced consulatant. The paper ends up by recommending that the Reserve Bank keeps a ‘watching brief’ on TLAC. There doesn’t appear to be any understanding that TLAC, in the trans￾Tasman, context is a very close substitute for CET1 capital, but is much less expensive, and that it should have been an integral part of the capital review at an early stage.

35 Susan Gutherie ‘Total loss absorbing capacity’ 11 October 2018 Memorandum to Financial Sector oversight Committee

47 Bibliography Alfonso, G, Santos, J, and Traina, J (2014), ‘Do ‘Too-Big-to-Fail’ banks take on more risk?’, Federal Reserve Bank of New York Economic Policy Review, available at: Altunbas, Yener, Simone Manganelli, and David Marques-Ibanez. 2012. ‘Bank Risk during the Financial Crisis: Do business models matter?’ Bangor Business School, Prifysgol Bangor University (Cymru / Wales) Working Papers 12003. Antoniasdes. A. 2015 ‘Commercial bank failures during the Great Recession: the real (estate) story’ ECB working paper series no. 1779 APRA 2018 ‘Increasing the loss-absorbing capacity of ADIs to support orderly resolution’ Baker, M and Wurgler, J, 2013, ‘Do strict capital requirements raise the cost of capital? Bank regulation and the low risk anomaly’, National Bureau of Economic Research Working Paper No. 19018, May. Barrell, R, Davis, E P, Karim, D and Liadze, I, 2010, ‘The impact of global imbalances: does the current account balance help to predict banking crises in OECD countries?’, NIESR Discussion Paper no. 351. Basel Committee on Banking Supervision (2010a), ‘An assessment of the long-term economic impact of stronger capital and liquidity requirements’, Basel Committee on Banking Supervision (2010b), ‘Assessing the macroeconomic impact of the transition to stronger capital and liquidity requirements’ Beltratti A., and Ren M. Stulz. 2012. ‘The credit crisis around the globe: Why did some banks perform better?’ Journal of Financial Economics, 105(1): 1–17. Berg J. Nielsen M., Vickery J. 2018 ‘Peas in a Pod? Comparing the U.S. and Danish Mortgage Finance Systems’ Federal Reserve Bank of New York Staff Report No. 848 Berger, Allen N., and Christa H.S. Bouwman. 2013. ‘How does capital affect bank performance during financial crises?’ Journal of Financial Economics, 109(1): 146–176. Barro R (2009) ‘Rare disasters, asset prices, and welfare costs’, American Economic review, 99(1), pp. 243-264. Brooke M, O Bush, R Edwards, J Ellis, B Francis, R Harimohan, K Neiss, and C Siegert, (2015, December) ‘Measuring the macroeconomic costs and benefits of higher UK bank capital requirements.’ Bank of England Financial Stability Paper(35). Calomiris C., and Mason J.2003. ‘Consequences of Bank Distress During the Great Depression.’ American Economic Review, 93(3): 937–947. Campello, Murillo. 2002. ‘Internal capital markets in financial conglomerates: Evidence from small bank responses to monetary policy.’ Journal of Finance, 57(6): 2773–2805.

48 Casimano, T F and Hakura, D S (2011), ‘Bank behaviour in response to Basel III: A cross￾country analysis’, International Monetary Fund Working Paper 11/119 Cecchetti, S, Kohler, M and Upper, C (2009), ‘Financial crises and economic activity’, National Bureau of Economic Research, Working Paper 15379 Clark, B, Jones, J and Malmquist, D (2015), ‘Leverage and the cost of capital: Testing the relevance of the Modigliani-Miller theorem for US banks’, Office of the Comptroller of the Currency Working Paper. Cole, Rebel Allen, and George W. Fenn. 2008. ‘The role of commercial real estate investments in the banking crisis of 1985-92.’ University Library of Munich, Germany MPRA Paper. Cole, Rebel Allen, and Lawrence White. 2012. ‘D´ej`a Vu All Over Again: The Causes of U.S. Commercial Bank Failures This Time Around.’ Journal of Financial Services Research, 42(1): 5–29. Cline W.R. 2016 ‘Benefits and Costs of Higher Capital Requirements for Banks’ Petersen Institute WP 16- 2016 Dagher J, G Dell’Aricca, L Luaeven, L Ratnovski, and H Tong, (2016) ‘Benefits and Costs of Bank Capital.’ IMF Staff Discussion Note 16/04. March 2016. Demsetz, Rebecca S., Marc R. Saidenberg, and Philip E. Strahan. 1996. ‘Banks with something to lose: the disciplinary role of franchise value.’ Economic Policy Review, , (Oct): 1–14. DeYoung, Robert, and Gkhan Torna. 2013. ‘Nontraditional banking activities and bank failures during the financial crisis.’ Journal of Financial Intermediation, 22(3): 397–421. Dunstan A. 2018 ‘The Reserve Bank’s philosophy and approach to stress testing’ RBNZ Bulletin Vol. 81 No.8 Englund P ‘The Swedish 1990s banking crisis A revisit in the light of recent experience’. Paper presented to the Riksbank Macroprudential Conference, Stockholm 23-24 June, 2015 Gutherie S. 2018 ‘Total loss absorbing capacity’ Memorandum to Financial Sector Oversight Committee RBNZ Guthrie S. 2018 ‘Risk Appetite framework used to set capital requirements’ Paper to FSG. 30 October 2018 RBNZ Guthrie S.2018 ‘Capital ratio calibration’ Memorandum for Financial Oversight Committee RBNZ Ivashina, Victoria, and David S. Scharfstein. 2010. “Bank Lending During the Financial Crisis of 2008.” Journal of Financial Economics, 97(3): 319–338.

49 Jordà, Òscar, Moritz Schularick, and Alan M. Taylor. 2013. ‘When Credit Bites Back.’ Journal of Money, Credit and Banking 45 (s2): 3–28 Junge, G, and Kugler. P (2012), ‘Quantifying the effects of higher capital requirements on the Swiss economy,” University of Basel, Working Paper, July, available at: Kashyap A, J Stein, and S Hanson (2010, May). ‘An analysis of the impact of ‘substantially heightened’ capital requirements on large financial institutions.’ Mimeo. Laeven L, and F Valencia, 2012 ‘Systemic banking crises database; An update,’ IMF Working Papers 12/163, International Monetary Fund. Laeven L, and F Valencia F., 2018. ‘Systemic banking crises revisited,’ IMF Working Papers 18/206, International Monetary Fund. 47 Laeven and Fabian Valencia1 2012 ‘Resolution of Banking Crises: The Good, the Bad, and the Ugly’ IMF Lilly C, (2018). ‘Outcomes from the 2017 stress test of major banks.’ Reserve Bank of New Zealand Bulletin 81(9), July 2018. Linh Nguyen & James Cummings December 2016 ’Impact of Higher Capital requirements on bank funding costs: Australian evidence’ CIFR WP No. 132. Miles, D, Yang, J and Marcheggiano, G (2013), ‘Optimal bank capital’, Economic Journal, Vol. 123, pages 1-37. Modigliani, F and Miller, M (1958), ‘The cost of capital, corporation finance and the theory of investment’, The American Economic Review, 48(3), pages 261-297. Morgan, D (2002), ‘Rating banks: risk and uncertainty in an opaque industry’, The American Economic Review, 92(3), pages 874-888. 33 Financial Stability Paper December 2015 Rochet, J C (2014), ‘The extra cost of Swiss banking regulation’, Swiss Finance Institute White Paper, Mendicino C, K Nokolov, J Suarez and D Supera (2016) ‘Welfare analysis of implementable macroprudential policy rules: heterogeneity and trade-offs.’ ECB Macroprudential Bulletin, No. 1, March 2016. Mor ‘Bank Rescues of 2007-09, 2018: outcomes and cost” House of Commons Library Obstfeld, Maurice. 2013 ‘On Keeping Your Powder Dry: Fiscal Foundations of Financial and Price Stability’ CEPR Discussion Paper No. DP9563 Otker-Robe I, and Podpiera A M (2013) ‘The social impact of financial crises; Evidence from the Global Financial Crisis’, Policy Research Working Paper No,WPS no. 6703 World Bank, Washington, DC.

50 de-Ramon, S, Iscenko, Z, Osborne, M, Straughan, M and Andrews, P (2012), ‘Measuring the impact of prudential policy on the macroeconomy’, Financial Services Authority Occasional Paper 42 de Ramon S, W Francis, and Q Harris (2016) ‘Bank capital requirements and balance sheet management practices: has the relationship changed after the crisis?’, Bank of England Staff Working Paper No.635 Ratnovski, Lev, and Rocco Huang. 2009. ‘Why are Canadian Banks More Resilient?’ IMF Working Paper, 1–19. RBNZ 2012 ‘Regulatory impact assessment of Basel III capital requirements in New Zealand’ RBNZ 2018 ‘Capital ratio decision: memorandum to FSO’ RBNZ 2019 ‘ Capital review paper 4: How much capital is enough’ RBNZ 2019 ‘Explanatory note on portfolio risk modelling in the New Zealand Context ‘. Reddell M. and Sleeman C. ‘Some perspectives on past recessions’ RBNZ Bulletin Vol. 71. No. 2. June 2008 Romer, C and Romer, H (2015), ‘New evidence on the impact of financial crises in advanced countries’, NBER Working Paper 21021. Romer, Christina D., and David H. Romer. 2017. ‘New Evidence on the Aftermath of Financial Crises in Advanced Countries.’ American Economic Review, 107 (10): 3072-3118. Toader, O. (2015, September). ‘Estimating the impact of higher capital requirements on the cost of equity: an empirical study of European banks.’ International Economics and Economic Policy, pp. 411-436. Widdowson D. Wood A. 2008 ‘A users guide to credit ratings’ RBNZ Bulletin Vol. 71 No. 3 Yan M, M Hall, and P Turner 2012,‘A cost–benefit analysis of Basel III: Some evidence from the UK.’ International Review of Financial Analysis, Vol 25 issue C pp. 73-82.

51

.

.

1 Tailrisk Economics Pinocchio awards Reserve Bank of New Zealand’s proposed bank capital increases March 2019 Contact Ian Harrison Ph. 022-175-3669 e-mail harrisonian52@gmail.com

2 Pinocchio Assessment methodology The assessments are adapted from the Washington Post’s Fact checking methodology.  One Pinocchio: "Some shading of the facts. Selective telling of the truth. Some omissions and exaggerations, but no outright falsehoods.  Two Pinocchios: "Significant omissions and/or exaggerations. Some factual error may be involved but not necessarily. A public office holder can create a false, misleading impression by playing with words and using technical language that means little to ordinary people."  Three Pinocchios: "Significant factual error and/or obvious contradictions."  Four Pinocchios: "Whoppers." We also make an overall assessment of each document or paper by awarding ‘Trump’ assessments based on the number of Pinocchios awarded. Moderately Trumpesque 5 - 9 Pinocchios Trumpian 10 - 16 Pinocchios Full blown Trump 17 plus Pinocchios

3 Reserve Bank of New Zealand’s proposed bank capital increases Our overall assessments are: Governor’s speech: Full blown Trump Non-technical summary: Full blown Trump Interest.co.nz interview Trumpian Deputy Governor’s speech Full blown trump Governor’s Speech Notes from an address delivered to Business NZ CEO Forum that argues for higher capital requirements 30 November 2018. The Reserve Bank needs to ensure there is sufficient capital in the banking “system” to match the public’s “risk tolerance”. This is because it is the New Zealand public - both current and future citizens - who would bear the social brunt of a banking mess. We know one thing for sure, the public’s risk tolerance will be less than bank owners’ risk tolerance The Bank doesn’t know for sure that the public’s risk tolerance will be less than the bank owners’. New Zealand bank owner have substantial franchises to protect, which makes them more cautious than they would otherwise be. The Bank has made no effort to assess the public’s risk tolerance. Two Pinocchios

4 How do we know this? Surely the more capital a bank has the safer it is and the more it can lend. This does not necessarily follow. The more capital required, the less a bank can lend in the first instance. If a bank is capital constrained, it must reduce its lending if capital is increased. Two Pinocchios And, the most a bank owner can lose is their capital. The wider public loses a lot more (see Figure 2). Banks can lose more than just their capital. They lose their franchise, which in New Zealand, is particularly valuable as evidenced by their high ongoing profitability. The franchise value is probably worth as much, or more, than the regulatory capital. Figure 2 (not reproduced here) depicts the fiscal cost of bank failure as a percent of banking system assets. The better measure of the burden on the public is the net cost to GDP. With a few exceptions, this cost in advanced countries that have experienced banking crises has been moderate – a few percent of GDP. The Bank has made no assessment of the relative cost to governments and bank shareholders. In the GFC, the fall in banks’ market capitalisations dwarfed governments’ capital contributions. Two Pinocchios Hence, we need to impose capital standards on banks that matches the public’s risk tolerance. We have been reassessing the capital level in the banking sector that minimises the cost to society of a bank failure, while ensuring the banking system remains profitable. The policy doesn’t minimise the cost to society of a bank failure – it just reduces it. The Reserve Bank’s policy doesn’t ensure the system remains profitable – that is market determined. The Bank has made no effort to assess the public’s risk tolerance. One Pinocchio The stylised diagram in Figure 3 highlights where we have got to. Our assessment is that we can improve the soundness of the New Zealand banking system with additional capital with no trade-off to efficiency.

5 This makes it appear that more capital is costless - when the cost could be about $1.5 billion a year. Four Pinocchios In making this assessment, our recent work makes the explicit assumption that New Zealand is not prepared to tolerate a system-wide banking crisis more than once every 200 years. We have calibrated our ‘sweet spot’ thinking about economic ‘output’ and financial stability benefits. The Bank settled on a 1:200 target, when the Governor decided, on a whim, to go for a 1:200 target. The original target was 1:100, which would not have required a capital increase. There was no consideration of output and financial stability benefits in the target decision. Four Pinocchios Banks also hold more capital than their regulatory minimums, to achieve a credit rating to do business. The ratings agencies are fallible however, given they operate with as much ‘art’ as ‘science’. Bank failures also happen more often and can be more devastating than bank owners – and credit ratings agencies – tend to remember. The costs are spread across the public and through time. The inference here is that rating agencies are poor at recording bank failure events. They are not. Taking a long historical perspective, bank failures in New Zealand have been very rare. Three Pinocchios Total Pinocchios: 18

6 The Bank’s non-technical summary Part of the consultation document on the capital increases. Bank Capital: What is it? Banks get their money from two places – their owners (often referred to as ‘shareholders’) and people they borrow from, including depositors (often referred to as ‘creditors’). The money that banks get from their owners is referred to as ‘capital’. This omits capital contributions from non-owners. The failure to say anything about other forms of capital is a material omission in this context. One Pinocchio The Capital Review It is important that the Reserve Bank’s banking regulations are up to date. There is also increasing evidence that the costs of bank failures – both economic and social (well-being) costs – are higher than previously understood. This is why we’re reviewing the capital rules for banks. The latest evidence is that the cost of bank failures is lower than earlier studies. Two Pinocchios Banks currently get the vast majority of their money by borrowing it (usually over 90 percent), with the rest coming from owners (usually less than 10 percent). The Reserve Bank is proposing to change this balance by requiring banks to use more of their own money. This proposal is consistent with steps taken by other banking regulators after the Global Financial Crisis. New Zealand has already increased capital along with other regulators following the financial crisis and New Zealand’s tier one capital, measured on a comparable basis, is higher than international norms. Other regulators, including Australia are now increasing capital using relatively inexpensive tier two capital to reduce fiscal risk. Australian approach imposes much lower costs on borrowers and depositors alike.) Three Pinocchios The Reserve Bank is proposing this change to reduce the chances of banks failing in New Zealand. If banks in New Zealand fail, some of us might lose money and some of us might lose jobs. However, there would also be indirect costs on all of society that

7 may be harder to see that would negatively impact the well-being of all New Zealanders. In the end, we would all bear the cost of bank failures, in one way or another. The increase in capital will impact on the fiscal cost of a failure, but a bank failure, in itself, is not the biggest cause of job and money losses. Bank failures are usually the result of a significant economic shock. More capital can mitigate this only to a limited degree, as New Zealand banks will still be dependent on the health of the economy and the financial strength of their parents. Three Pinocchios This is why we want to make the chances of this happening very small – so small that a banking crisis in New Zealand shouldn’t happen more than once every two hundred years. The best evidence is that New Zealand banks already meet the 1:200 test Two Pinocchios Extent of changes We expect only a minor impact on borrowing rates for customers. The impact on borrowers will be significant. Using the RBNZ’s own estimates a typical first home borrower will pay another $1000 per year initially which will add somewhere in the order of $20,000 more over the life of their loan. That is not a minor impact. Two Pinocchios Possible Impacts If banks increase their capital, they will be more resilient to economic shocks and downturns, which will strengthen New Zealand’s banking system and economy. The impact on resilience will be minor. New Zealand main banks will still be dependent on the health of their parents. One Pinocchio

8 What’s the downside? Because the level of a bank’s capital can have an impact on the interest rate it charges on its loans, it is possible that higher capital requirements could make it more expensive for New Zealanders to borrow money from a bank. While we certainly take this into account, we think this impact should be minimal. The Reserve Bank did not take the increased cost of borrowing into account at all. Their ‘soundness’ test does not take account of borrowing costs. The efficiency assessment rather loosely references some foreign cost benefit analyses, which do not take the direct impact of increased borrowing cost on the borrower into account. A cost of $1.5 billion a year is not minimal. Four Pinocchios Total Pinocchios: 18 Interview with the Governor Interest.co February 18, 2019 Banks' financial rewards have been greater than the risks they've taken, and the Reserve Bank's proposals for banks to hold more capital would bring the reward down to something that better justifies their risk, Reserve Bank Governor Adrian Orr says. What the Governor appears to be saying is that banks are really low risk, but they are earning excessive returns relative to that risk. So, if their capital is increased, their returns will come down. This presumes that banks don’t increase their margins (more on that below). If banks are earning excess returns, this is market efficiency not a soundness issue. It makes no sense to reduce competition by increasing the capital ratios of capital constrained New Zealand banks. One Pinocchio

9 "For every risk there should be a reward and when you look at the banking system globally that reward has been greater than the risk taken. And that happens because banks can take the returns, they don't have to wear all of the losses because the losses are socialised through bank bailouts and failure management and all of the things that we're still struggling to recover from in the GFC [Global Financial Crisis] globally," Banks are limited liability companies like most businesses, so they don’t wear all of the costs of a failure. That is how the economy works generally. What is distinctive about banking is that generally depositors do not take the losses of a failure. Depositors are meant to be taling a small risk, but governments bail them out. This is the market failure. It is a myth that bank shareholders did not bear significant losses in the GFC. In the depth of the crisis shareholders lost more than 80 percent of their investments, a multi-trillion-dollar loss, that dwarfed government capital injections Generally, investors were not bailed out. Two Pinocchios "So that reward has been in excess of the risk. We're saying 'let's better balance risk. So we're doing two things at once, we're lowering the risk through more capital, and if investors are struggling to say 'while there's a lower risk,' they should expect to get lower return." "When I was running the NZ Super Fund we wanted to be rewarded for the risk taken. If we could find excess reward that was fantastic, let's go there. But for the banking system as a whole it's too critical. The fear of failure would create failure so we're saying 'let's reduce the risk,' and hence expected returns on equity would probably decline." But will the expected returns decline, that is the multi-billion dollar question? In terms of what's being proposed by the Reserve Bank, Orr says the central bank and prudential regulator is working from the perspective of society's risk appetite, not any particular bank's risk appetite. "We're trying to calibrate it to society's risk appetite rather than the risk appetite of a particular financial institution," says Orr. There is no evidence that the Bank actually examined what society’s risk preferences might be in the capital review. Did the Bank actually ask any first home buyer if they were prepared to pay $20,000 extra over the term of their mortgage to reduce the

10 risk of their bank failing from 1 in 100 years to 1 in 200 years? Did they ask borrowers if they we prepared to accept lower deposit rates? Two Pinocchios "We know that society's risk appetite is less than a bank's risk appetite because banks get to privatise their earnings but get to socialise their losses. If they fail it's us, it's taxpayers, current and all future citizens, who have to pay for that bank failure. So we're trying to get a better balance. More capital means a safer bank. It means a lower risk bank, which means that they will have a better credit rating themselves, and it means that society is in a better, safer and more efficient position." The argument that society’s risk appetite is lower than the banks’ doesn’t necessarily follow. Again, banks don’t get to socialise their losses much, depositors do. The Bank has introduced OBR in an attempt to internalise this externality. The Bank seems to have entirely forgotten about the OBR in its analysis, or they are conceding that it will never be used in practice. Safer banks don’t automatically mean more efficient banks. Banks’ credit ratings probably won’t change, because the parent rating will set a ceiling and New Zealand capital requirements are unlikely to affect that (the New Zealand subsidiaries’ standalone ratings might improve though). One Pinocchio Asked about criticism of a lack of cost-benefit analysis in the Reserve Bank consultation paper Capital Review Paper 4: How much capital is enough?, Orr said the fact the Reserve Bank has talked about potential for higher mortgage rates, lower deposit rates, credit rationing by banks and an impact on Gross Domestic Product shows it is thinking about the benefit-cost analysis. Thinking about, and then mostly ignoring the cost evidence, is not a cost benefit analysis. Two Pinocchios "What we're saying is that we believe more capital, and better quality capital, could cost a bank something. And we're not hiding from that. We think around 20 to 40 basis points additional premium between the costs at which they borrow and lend will be passed on because of them having to raise equity rather than just leveraging debt. So we're not hiding from that," Orr says. The Bank might not be hiding from the higher costs, but it has tried to minimize them. They have never tried to explain those additional cost in terms of what that will mean to many New Zealanders or calculated the aggregate costs. In the non-

11 technical summary, they were described as minimal. Cost played no part in its decision making. Two Pinocchios "But the wider benefit is that their total risk is lower, their ability to attract equity will be cheaper because the investors will have a lower expected rate of return because they are a safer institution, and then beyond that the cost to the economy because we have more capital, safer institutions, our whole credit rating will be better." The key question is whether the Australian bank owners will lower their expected return on their New Zealand investments. The Governor is just hoping that they will. Standard and Poor’s have already said that the main New Zealand banks’ credit rating is unlikely to improve. (Two Pinocchios There's no shortage of people wanting to grow their balance sheets. Likewise the bigger end of town could issue their own debt. It doesn't always have to go through the banking system. So there are so many other things that could happen. It's not a 'hold everything else this will happen'." Yes, many things could happen. The point is what is most likley to happen. The existing New Zealand domestic bank’s might want to growth their balance sheets, but they won’t be able to because they will have to meet a higher capital requirement. "We want to all agree where we want to get to and then we can talk about what makes sense getting there. The interesting thing of course is that markets are forward looking ...we want to work with you, we want a sound and efficient and profitable financial system just as you do. Let's work through this properly. The current situation is sub-optimal, let's get to an optimal situation in a way that doesn't kill us on the way there." The Governor is just asserting that the current situation is sub-optimal. Working through things properly, includes genuine consultation and openness to other ideas and evidence. The Governor appears to have a closed mind on the endpoint. Over the course of the Capital Framework Review, hundreds of New Zealanders will have contributed to the many submissions made on the papers published to date. These are New Zealanders with deep knowledge of the local financial markets and passionate about what is needed for the New Zealand economy to be successful. To date not one point differing from the RBNZ’s proposals has been accepted.) Two Pinocchios

12 Total Pinocchios: 14 A speech delivered to the Institute for Governance and Policy Studies, Victoria University in Wellington, New Zealand by the Deputy Governor Today I’d like to talk about how we are proposing to improve the lives of all New Zealanders by making New Zealand’s banking system safer, and how the recent proposals by the Reserve Bank with respect to minimum capital requirements will do just that. The lives of the New Zealanders who have to pay higher mortgage interest or receive a lower deposit interest rate will not be improved. Two Pinocchios Since the late 19th century, New Zealand has only experienced two banking crises, one in the late 1880s and early 1890s, which resulted from a credit-fuelled rural land boom in the 1870s, while the other occurred in the late 1980s and resulted in the Bank of New Zealand having to be recapitalised by the government and its shareholders.2 Perhaps one banking crisis per lifetime is one too many? 3 The failure of the BNZ was not a banking crisis as commonly defined. It was the failure of a single largely government owned bank, which engaged in some madcap lending with a capital ratio of just 5 percent One Pinocchio The Potential Impact The Reserve Bank has several options for resolving a failed bank. Under any of these options, there would likely be at least some level of disruption to the banking system as depositors would lose access – at least temporarily – to some or all of their

13 deposits. It is also possible that depositors may never recover the full value of their original deposits. 5 With a direct Government intervention there would be no loss of access to deposits. If the OBR was used with a main bank, a haircut of 10 percent (equivalent to a risk weighted capital injection of about twenty percent) would be sufficient The OBR is designed so the failing bank would open for business the next day. Depositors would then have access to their remaining deposits. One Pinocchio We would expect that the impact of bank failures would be broader and harsher the larger and more intertwined the failed banks are with New Zealand’s economy ….This could mean that it may be more difficult for individuals to borrow to buy a car or home or pursue further education, and businesses may have trouble borrowing to meet their short-term cash flow needs. The decrease in available credit could have disastrous impacts on New Zealand’s economy. Most of this will happen anyway in a stress event, regardless of whether a bank has a 13 or an 18 percent capital ratio. The ‘disastrous impacts’ claim is a gross exaggeration. Three Pinocchios While it is likely that the Reserve Bank would intervene to resolve a failing bank before it is in a negative capital position (i.e., the Reserve Bank would begin the resolution process when the value of assets exceeds liabilities), it is difficult to determine exactly how the bank’s remaining value (and losses) would be allocated among creditors (including depositors) and shareholders. However, what is known with greater certainty is that creditors (including depositors) will realise more value in the failed bank if the bank has a greater proportion of shareholder capital. The last statement is literally true if holders of tier two capital (subordinated debt) are counted as creditors. But they provide equal protection to depositors and/or the government as shareholder capital in a resolution, so the statement is misleading. One Pinocchio

14 Social Costs of Bank Failures International agencies like the World Bank, the World Health Organization, and the United Nations have investigated the economic and social impacts of financial crises. None of these agencies had anything to say about the marginal effect of higher capital on the social costs of economic downturns. All downturns have wider social costs. The question is how does higher capital help. One Pinocchio Since the 1970s, there have been more than 140 banking crises around the world. The great bulk were undeveloped countries and former communist transition economies. There have been a much smaller number of crises in developed countries which are good comparators for New Zealand. One Pinocchio I want to spend some time on the impacts of banking crises on wellbeing. If you ask someone who’s lived through a banking crisis, they’ll likely tell you that the impacts were not only significant, but lasting. Perhaps the person you talk to may have lost their job as a result of the crisis, and if not, it might have been their spouse, a friend, or a neighbour. Maybe you speak to a young couple that had purchased their first home just prior to the crisis, only to see its value decline by 30% in the months following the crisis, forever altering their outlook on the economy and their willingness to make another significant investment. Banking crises did not cause the downturn in house prices in the GFC. Rather, in the US, the fall in house prices was a cause of the crisis. The housing market turned in 2006 as the subprime lending ‘scams’ started to unwind. In no other country (with the possible exception of Iceland) did house prices declines make a decisive contribution to bank losses. In Ireland housing loan losses were dwarfed by commercial losses. Ian Harrison did talk to two homeowners who lost their house in the GFC. It was at a wedding in Mexico in late 2009. They had just mailed in their keys on a house (they had purchased at the peak of the boom without a deposit) to the bank. But immediately before doing so, (while their credit record was still intact) they had purchased a similar, but now cheaper, house (with just 5 percent down – funded from their credit cards). California has a no-recourse law so only the bank lost on the first house. The couple were staying in a luxury hotel in Mexico City and were planning a trip to New Zealand - again all on the credit card. They seemed happy

15 enough at the time and given their (to me) high-risk, but rational, behaviour they are probably sitting on a Californian real estate fortune now. Two Pinocchios .. maybe you speak to someone who just graduated from university prior to the crisis, only to enter a depressed labour market, and forced to accept work well below their educational qualifications and abilities, forever altering their desired career path. New Zealand has had many recessions where this has been the case, where there was no banking crisis. We doubt if the career paths of many of those affected were forever altered One Pinocchio For those that lived through the recession we experienced here in the early 1990s, you will recall that some industries were decimated, and a generation of workers lost. Many of these workers were not able to re-enter the workforce easily and lost valuable skills while trying to find suitable employment. And while recessions sometimes occur in the absence of a banking crisis, it is common for banking crises to ultimately result in recessions. The 1990s recession was not caused by the failure of the BNZ. Both were the fall out of the irrational exuberance of the 80’s, with the recession being exacerbated by the restructuring of the economy and the RBNZ’s tight monetary policy Four Pinocchios The average New Zealand bank gets around 92% of its money by borrowing it. Compare this with the average business in New Zealand, for which this figure is about 55% It is not clear to me why this discrepancy between banks and other businesses is so large, but perhaps at least part of it can be explained by the fact that, historically, governments have been much more reluctant to allow a bank to fail than other types of businesses, which may lead banks to operate closer to the edge. The Deputy Governor does not appear to understand that banks are financial intermediaries, not widget manufacturers. An important economic role is to take people’s money and lend it to businesses and households more efficiently and safely than individuals can do it themselves. A widget manufacturer’s leverage ratio provides no guidance on what is socially optimal for financial intermediaries.

16 Three Pinocchios What is ‘skin in the game’ and why is it important? Put simply, skin in the game refers to the concept of having to bear the consequences of one’s own decision-making. By having minimum ‘skin in the game’ requirements, banking regulators ensure that the owners of a bank have something at stake, something to lose. I would like to note that banks themselves lend on these very same ‘skin in the game’ principles, for example, by requiring mortgage borrowers to provide a deposit. The regulatory measure of capital understates the amount of ‘skin’ banks already have in the game. They are also putting their franchise value, which they will lose if they fail, at risk. The franchise value of the major banks represents the excess of market determined shareholders equity over net assets. All of the main banks trade substantially above book value showing their franchise is worth much more than their regulatory capital. Two Pinocchios We believe that more ‘skin in the game’ for banks will result in: Society being less at risk from banking crises. Because more capital means banks are more likely to survive large unexpected shocks, society is also less at risk from the economic and social fallout that usually accompany bank failures. This ignores the fact that main banks are subsidiaries that will still be dependent on the financial health of their parents, which will be little improved by an increase in New Zealand bank capital. S&P have already said that the big four New Zealand banks’ credit ratings are unlikely to change because of this effect. Two Pinocchios Reduced fiscal risk. When the probability of a banking crises is reduced, so is fiscal risk. Fiscal risk can be reduced by subordinated debt, at a fraction of the cost of shareholder capital Two Pinocchios

17 Bank shareholders and management being less inclined to take excessive risks. Beware the law of unintended consequences. Banks might take more risk to meet their current rate of return targets. Two Pinocchios There were other threads of analysis that supported our proposals. A comprehensive review of the international literature told us that we were within the range of capital ratios that were appropriate. This is misleading. On a like-for-like comparison the weight of evidence from international studies considered is that the optimal capital ratio is significantly lower than the RBNZ’s proposal. In addition, the international studies all leave out the direct cost to borrowers and depositors, so should not have been used without adjusting for this omission. Two Pinocchios Furthermore, building on analysis that began in 2012, we looked at the implications of representing society’s risk appetite differently. This analysis took into account New Zealand specific factors and captured household risk-aversion. 18 As with findings from similar types of analysis in the international literature, there is a wide range of plausible calibrations. However, this analysis ultimately showed that a capital ratio of 18% (or 17%) is in the mid-point of our range of estimates for the ‘optimal capital ratio’ for New Zealand. This gives the impression that the Bank actually used its New Zealand models in this analysis. That is not true. They got the NZ capital model going again in 2016, but it played no part in their decision making. The Regulatory Impact Statement showed a model-based optimal capital ratio of 13 percent. The later FSO report showed a ratio of 17-18 percent, but the difference was not explained. The original model understated the cost of capital by a large margin, and using the Bank’s current cost estimate, the optimal capital estimate would have been materially lower. Three Pinocchios Comparisons of capital ratios The Basel Committee has published tables of capital ratios for a group of large banks operating in its member jurisdictions. Credit rating agencies are another source. For

18 example, Standard & Poor’s calculates its own risk-adjusted capital ratios for many banks around the world, using a methodology that attempts to reduce the influence of differing national applications of the Basel framework while still taking into account the different risk profiles of the countries in which each bank operates. While these cross country studies have not played much role in our analysis of determining what we think is the appropriate capital calibration for New Zealand, they do demonstrate that our proposals are by no means extreme. Instead, they move us towards our goal, expressed back in 2017, of capital requirements that are conservative relative to our peers. This evades the real issue, which is whether the regulatory capital ratios are higher in New Zealand if calculated using international standards, as argued in the PWC report. The Reserve Bank has had plenty of time to produce its own comparison, as APRA has done, so the absence of a Bank report is telling. Listing the Basel Committee’ table does not address this issue at all. Presenting some, potentially cherry picked, S&P capital ratios does not really help. The S&P ratios have a point-in-time element, as they are affected by short-run movements in house prices and credit growth. If these unwind, New Zealand banks’ measured capital ratios will improve. The S&P capital model also tends to a ‘paint-by-numbers’ approach, designed to work across a broad range of countries, but which does not appear to be work well in New Zealand. It has a set of risk weights, similar to, but generally higher, than the Basel standardised model. The risk weights have been calibrated off a generic stress test outcome. When we compared S&P’s implied stress test outcomes with even the most conservative New Zealand stress tests, the S&P losses were a multiple higher. This tells us that the S&P risk weights applied to New Zealand banks are too high and the resulting capital ratios too low. For what it is worth, the banks in the two countries with the highest S&P capital ratios had the same credit ratings as the main New Zealand banks. Two Pinocchios Total Pinocchios:35

19

17 May 2019 Ian Woolford Financial System Policy and Analysis Department Reserve Bank of New Zealand Wellington By Email: CapitalReview@rbnz.govt.nz Dear Ian, INFINZ – Submission on the Regulatory Capital Review Thank you for the opportunity to provide submissions on the Reserve Bank’s Bank Capital Review: Capital Review Paper 4 (“Capital Review Paper”). About INFINZ The Institute of Finance Professionals in New Zealand Inc (INFINZ) is the pre-eminent industry body for finance and capital markets professionals in New Zealand. INFINZ is a voluntary organisation, with membership of more than 1,650 individuals drawn from right across the sector, including treasury professionals, investment analysts, fund managers, bankers, lawyers, academics and students. One of the key missions of INFINZ is advocacy for stronger capital markets in New Zealand. Vibrant capital markets underpin the key matters of productivity, well-being and investment , where there is an increasing focus on addressing New Zealand’s infrastructure deficit and decarbonising our economy. This submission has been prepared primarily with those issues in mind. A note about the preparation of this submission The resource for INFINZ’s advocacy function is drawn solely from members of its board. This submission was prepared by a sub-committee, comprising a corporate treasurer, an economist, an M&A and governance consultant, and a finance/capital markets lawyer, with some input from a finance academic and prudential supervision expert. It also reflects input from:  a range of communication, including forum discussions, with members who are CFOs or corporate treasurers of a number of New Zealand’s leading companies and other institutions; and  specialist treasury advisers, who are regularly in touch with borrowers across the full range of large corporates to small-to-medium sized companies (SMEs). We are grateful for the input from our members on what is a very difficult subject. We would like to acknowledge that, while we have consulted as broadly as we could, the nature of our membership is such that it is not possible to canvass all of them for their views. We have appreciated the time taken by members of the Prudential Supervision Department to meet with us to discuss the Capital Review and the open engagement brought by that group. We look forward to continuing this dialogue as the proposals are developed.

2 Introduction – key themes of our submission The purpose of capital requirements and other tools is to achieve a “sound and efficient financial system” (s 68 of the Reserve Bank of New Zealand Act 1989 (“RBNZ Act”)). The Reserve Bank also has the purpose, in its prudential and other functions, to “promote the prosperity and well-being of New Zealanders, and contribute to a sustainable and productive economy” (s 1A RBNZ Act, as amended recently following Phase 1 of the Reserve Bank Review). These principles form the basis for the Reserve Bank’s prudential mandate. Given the importance of the consultation process, it is essential that submitters and other stakeholders have the information they need in order to assess whether the capital proposal presented by the Reserve Bank achieves the objectives underlying that mandate. INFINZ acknowledges the importance of the Capital Review, since financial stability underpins New Zealand’s broader economic performance. We support the Reserve Bank in wanting a sound, stable and efficient banking sector. This is especially significant given the key place that bank financing has in the New Zealand market, particularly for those (such as SMEs or rural borrowers) who do not have the scale or credit rating to access the capital markets. Clearly measures taken to achieve stability (such as capital requirements) have material economic consequences that need to be assessed, both from the perspective of efficiency and soundness. Although these criteria involve inherent trade-offs, the degree of their interdependence and complementarity should not be neglected – particularly in circumstances where similar stability results can be achieved by alternative, more efficient, means. It is essential that we devote the time and resources to getting this right. Although the Reserve Bank has rightly focused on our acknowledged vulnerabilities (e.g. elevated housing-related debt), the same factors mean that our economy has a high degree of sensitivity to the cost and availability of debt financing. These flow through to our levels of investment and therefore productivity, which are already well below the OECD median, with widespread consequential impacts on our prosperity and wellbeing. Our poor domestic savings record and constrained access to foreign direct investment mean that we are also likely to remain heavily reliant on debt to finance the significant investments required to address our infrastructure deficit and related housing affordability issues, adjust to increasingly rapid technological change, and achieve the Zero Carbon target. In the latter regard, it is encouraging to see the Reserve Bank’s recent commitment to a climate change strategy. Together, these factors make it particularly important that the optimal balance of soundness and efficiency is achieved. To date, the debate has centred on the capital proposal’s funding cost impacts – whether the marginal benefit of higher capital exceeds the marginal cost of the capital requirement. This is important, and in our view should be subject to additional scrutiny,1 but it is not the only issue for consideration. Empirical studies indicate that factors other than capital ratios (for example, controls addressing funding structure and credit growth) have at least as much of a role in achieving stability as the level

1 The UBS calculation of an 80 to 125 basis point margin increase is by now quite well known, and we are aware of other calculations prepared by credible investment analysts that are also around or above a 100 bp increase (+/-), which is two and a half times the high end of the Reserve Bank’s projected range. Other independent analysts have arrived at numbers somewhere in the middle – it is not a precise science. We are not able to test these projections ourselves, but consider that it would be prudent to for a sensitivity analysis to be undertaken.

3 of loss-absorbing capital/leverage. For this reason, a narrow focus on capital is unlikely to be optimal in achieving either soundness or efficiency, without a closer look at the contribution of other options available in the prudential toolkit (summarised in Appendix 1). These complementary tools should be explored in a cost-benefit analysis and should be factored into decisions on the ultimate level of prudential capital, consistent with the approach taken by the Bank of England (Brooks et al 2015). As noted below, the Reserve Bank has implemented policies addressing these matters and thus contributing to stability, but it is not clear that the resulting lower systemic risk probability has been factored into the calculation of the minimum capital requirement. This is more than a question of cost of capital, although that is very important in light of New Zealand’s Net International Investment Position and reliance on debt intermediated by the banking system as the dominant source of capital. There is a broader range of potential impacts from the changed policy, which are not given in-depth consideration in the Capital Review Paper. These could include transitional or lasting changes in lending composition, credit growth, and deposit rates, as well as second order impacts on investment and productivity. Some of these effects may be less prevalent if alternative tools are used (including alternative loss-absorbing capital instruments), but in any event they should be factored into an analysis of the proposal’s contribution to soundness and efficiency. As things stand, we think that insufficient material and analysis has been put forward to enable a confident assessment that the capital proposal is the best solution to achieve optimal financial stability settings for New Zealand, either viewed alone or when considered alongside other prudential options. This is all the more important where the recommended option differs materially from the approach recently taken in other jurisdictions – most notably in Australia – in relation to both the level of capital required and how it is composed. There is little exploration of these alternative approaches in the Capital Review Paper or accompanying material. A cross-country comparison to Australia would be particularly relevant given the home-host relationship in relation to our four largest banks. Our key submission is that the capital proposal warrants a level of scrutiny proportionate to the magnitude and breadth of its potential impacts on the wider economy. Such an assessment can only be made based on a robust cost-benefit analysis. This includes assessing alternative options and consideration of the contribution to stability provided by other existing or available prudential tools, and of the broader range of potential impacts the capital settings may have. That is, what are the optimal settings across each of the prudential tools available that, in combination, deliver the optimal balance of stability and cost? Such an assessment would provide confidence to all stakeholders that both soundness and efficiency will be optimised. Two key questions Before addressing the detail, there are two key questions to consider: (1) what problem are we trying to solve; and (2) why is the position in New Zealand so different to other jurisdictions as to require significantly higher minimum levels of regulatory capital, and regulatory settings so heavily tilted toward the common equity tier 1 (CET1) component of the prudential toolkit? At first sight, New Zealand stands out in the comparative strength of its banking system, not its fragility.2 But we rank low on many efficiency and productivity measures, sometimes among the lowest

2 For example Fitch, which recently conducted its own stress tests on the New Zealand banking system, found it resilient to shocks, consistent with the Reserve Bank’s own regular stress tests. Refer also the assessment in the IMF FSAP (May 2017), pgs 22 to 24, that New Zealand banks are “resilient to a severe global economic downturn”.

4 in the OECD, due in large part to deficient non-residential investment. In these circumstances, it is particularly important to give due emphasis to both soundness and efficiency. In addition to these factors, the determination that greater conservatism is required in New Zealand’s prudential settings is based on a ‘risk appetite framework’ and the related regulatory approach in which the efficiency criterion is analysed as a second level criterion, after requisite soundness is achieved (each of these points is canvassed in more detail below). The policy settings are also influenced to some extent by resourcing considerations and by a regulatory philosophy having a preference for simplicity and for a light-handed approach. A number of these matters were highlighted by the IMF in its recent FSAP assessment. Similar factors that have led to the conservatism bias in the level of capital have also influenced the restriction in its composition to CET1, missing the opportunity to access the complementary qualities of bail-in instruments in creating a risk-averse constituency with ‘skin in the game’ to counteract the incentives of shareholders and to yield pricing signals more responsive to emerging credit issues. In our submission, each of these premises would warrant further evaluation, particularly in light of the FSAP recommendations and the concurrent Phase 2 Review. Taking the time to get it right The New Zealand financial system is consistently described by the Reserve Bank, credit rating agencies, and international agencies as stable and well-capitalised, albeit subject to the familiar vulnerabilities of housing-related and rural debt. Overall economic conditions are sound, credit growth is muted, and there is no sign of the emergence of the sorts of exuberance or financial exotica associated with the GFC (although even then, with only limited penetration to New Zealand). In its most recent Financial Stability Report (published in November 2018, just prior to the capital proposal), the Reserve Bank noted:  Household sector indebtedness remains a key vulnerability, but risk is reducing (thanks, at least in part, to the LVR “speed limits” imposed by the Reserve Bank).  The banking system is sound, with large capital buffers.  Banks have strengthened their funding profiles, through higher deposits, lower market exposure, and weaker credit growth. The percentage of offshore funding has reduced substantially. The banks have passed rigorous and regular stress tests, in which the banking system was able to absorb losses – including in a scenario of a “severe” fall in house prices – and retain sufficient capital to meet minimum capital requirements. Over the longer run, New Zealand is one of only six countries whose banking systems have never suffered a financial crisis (including the similarly structured Australian and Canadian banking systems, despite the relatively high levels of credit in each). 3 This is acknowledged in the Capital Review Paper (para 51 on pg 22), which noted that “the absence of any banking crisis here in the post-war era,4

3 Refer Laeven and Valencia (2018), Huang and Ratnovski (2009) and Reinhart and Rogoff (2009). 4 The reference here is likely to a bank failure in 1895 requiring recapitalisation by its shareholders and the government – refer Chris Hunt “Banking crises in New Zealand – an historical perspective” RBNZ Bulletin, Vol. 72, No. 4, December 2009. Neither that, nor a similar occurrence following the 1987 crash, were systemic in nature.

5 poses some pragmatic constraints on the range of analytical tools that are available to us” and that “there has been no large loan loss events in New Zealand in recent decades” (para 63, pg 25). The IMF in its recent Financial System Assessment conducted stress tests which found that major New Zealand banks are “resilient to a severe global economic downturn”.5 The OECD, in its most recent survey of New Zealand (June 2017),6 similarly found little cause for alarm in relation to macro-economic or financial system stability – each of which has been improved since the GFC, at the same time as the banking system has significantly bolstered its capital adequacy, liquidity and funding resilience, and new tools have been added to the Reserve Bank’s arsenal. Key OECD findings were:  Macro-financial vulnerabilities are generally lower than at the end of the last expansion in 2007, with high house prices and associated high levels of household debt remaining the major sources.  The external position remains a risk, with relatively weak cost and price competitiveness and export performance, but all have improved since 2007, as has the current account balance and net international investment position.  Financial stability has shown the greatest improvement, as tightened regulation has reduced external bank debt and leverage. The changes since 2007, and remaining key vulnerabilities, are shown in the following diagrams: None of this leaves room for complacency, but it does provide good conditions in which to ensure that the review of capital settings is robust and comprehensive. This is not a situation in which, in exercise of its operational independence, the Reserve Bank needs to move quickly, or alone.

5 IMF “New Zealand: Financial System Stability Assessment” (IMF Country Report 17/110, May 2017), pgs 22 to 24. 6 www.oecd.org/economy/surveys/NewZealand-2017-oecd-economic-survey-boosting-productivity-and-adapting-to￾the-changing-labour-market.pptx

6 It is also an ideal time to fundamentally review the central banking arrangements in New Zealand, to ensure that they serve us as well for the next three decades as they have in the previous three, since the RBNZ Act 1989 came into effect. We commend the Government for the current review, the first phase of which – concerning monetary policy – has now become law. Concurrence of the Phase 2 Review of prudential policy-making arrangements Phase 2 of the review concerns the equally important, but less conspicuous, topic of prudential supervision. The terms of reference for Phase 2 were agreed on 7 June 2018, addressing the financial policy provisions of the RBNZ Act, and the broader governance and accountability arrangements for the Reserve Bank, particularly as relate to prudential regulation and supervision. On 1 November 2018, the Phase 2 paper was released for consultation, six weeks before the Capital Review Paper. The Phase 2 paper looks at changes which may be made to the structure and accountability arrangements for prudential policy and supervision. While it is not necessarily the case that the Capital Review should be put on pause pending the outcome of the Phase 2 Review, it equally would be a missed opportunity if one of the most significant decisions in the 30-year history of the RBNZ Act were to take place without at least some acknowledgement of the issues that motivated the Phase 2 Review. In this regard, we think that there is a gap in the arrangements under the existing RBNZ Act for Ministerial and broader government agency engagement and input in relation to prudential policy￾making, by comparison to:  monetary policy, which is subject to detailed provisions as to Ministerial input in sections 9 to 15 of the RBNZ Act, recently bolstered by new institutional arrangements (including the establishment of a Monetary Policy Committee) established under the Phase 1 reforms; and  macro-prudential policy, which is the subject of a Memorandum of Understanding (MOU) between the Minister of Finance and the Reserve Bank,7 including that the Reserve Bank will keep the Minister and the Treasury regularly informed on its thinking on significant policy developments and will consult with them where macro-prudential intervention is under active consideration.8 The former likely reflects the primacy of the Reserve Bank’s monetary function9 and the latter the real economy impacts of policies such as loan-to-value ratios (LVR) limits. Because the macro-prudential MOU arose extra-legislatively and under the same statutory power (in s 74 RBNZ Act), it appears likely that the lack of any equivalent procedures for prudential policy reflects a course of conduct, rather than having a particular statutory basis. There is no question about Reserve Bank’s powers to undertake the capital review or set standards, or its operational independence in doing so. The question is about the degree of engagement and scrutiny that should be brought to a proposal of such broad economic significance.

7 Refer https://www.rbnz.govt.nz/financial-stability/macro-prudential-policy/mou-between-minister-of-finance-and￾governor-of-rbnz (13 May 2013). 8 It is notable, for example, that the government has requested the Reserve Bank to undertake a full cost-benefit analysis if it wished to use debt-to-income controls as a complement to high LVR restrictions (refer OECD (2017)). 9 This is borne out by Hansard reports on the RBNZ Bill, where the debate in the House was overwhelmingly dominated by monetary policy considerations and barely mentions the prudential function.

7 Engagement by the Minister of Finance and agencies outside the Reserve Bank, such as the Treasury, would be consistent with the evident intent of the accountability provisions in Part 6 of the RBNZ Act. Robust governance and accountability arrangements are also a key component of international best practice for the operational independence of prudential supervisors, 10 and have been described as an “indispensable complement to independence” (Masciandaro/IMF (2011)). The capital proposal is a prime example of a prudential decision that would have broad economic significance. As such, pending legislation implementing Phase 2 determinations, it would be appropriate to apply a process or approach of the nature described in the macro-prudential MOU, which provides for Government engagement and input while carefully preserving operational independence. This would reflect the policy direction of the Government in the Phase 2 Review. Ensuring that the scrutiny is proportionate to the proposal’s impacts The consultation process is important and is welcome, as is the transparency that the Financial System Policy and Analysis Department has brought to the consultation process. But the capital proposal is not subject to any Parliamentary oversight process or other checks and balances, such as the disallowance procedure for secondary legislation. Exclusive reliance on public consultation to fill that gap poses a challenge, due to the complexity of modern prudential frameworks and vastness of the related empirical and theoretical literature. These make it a daunting, resource-intensive subject for potential submitters who are non-specialists, creating significant barriers to engagement. Because the impacts of the proposal will flow far beyond the community of regulated banks, it is vital that such dramatic changes to regulatory settings are subject to rigorous and independent scrutiny. The broadening economic impact of prudential policy, and corresponding questions about the accountability and oversight arrangements in its formulation, are a key focus of the Phase 2 Report. Advice to officials in connection with the Phase 2 review notes: 11 “Over recent years there has been a significant growth in the breadth and complexity of prudential regulation and with this comes a wider range of impacts on the wider economy. This raises the question as to whether there are appropriate safeguards in place as to the setting of regulatory requirements. While it is desirable to delegate matters of technical detail to regulatory agencies, the question is whether the prudential requirements are subject to too little scrutiny, particularly in comparison with the IPS Act, the NBDT Act and the financial markets regime.” (Emphasis added.) The capital proposal is a banner example of a policy with a wide range of impacts on the wider economy. We submit that the process for the formulation and implementation of a proposal this significant should include the scrutiny and appropriate safeguards noted above. In this context we note that the Phase 2 review team is due to report back in the next few weeks. As such, it is imperative that time is taken to get it right, meaning – at a minimum – to move forward only on the basis of a comprehensive and robust cost-benefit analysis. Ideally, this process would include engagement with, and input from, broader agencies within the government. The confidence of stakeholders and the public would also be enhanced by some degree of oversight by, or ‘second opinion’ from, an agency that is independent from both the Reserve Bank, as proponent, and the community of regulated banks.

10 Refer Principle 2 of the Basel Core Principles for Effective Banking Supervision. 11 Dr James Every-Palmer QC Reserve Bank Prudential Regulation of Banks (August 2017).

8 Costs and benefits The 3 April Capital Review Background Paper (on pg 1) says that: This paper does not provide a cost-benefit assessment of the proposal to increase bank capital, but has been prepared in order to provide further information to those who wish to provide feedback during the consultation process which ends on 17 May 2019. The Reserve Bank will carry out a full cost-benefit assessment for a Regulatory Impact Statement to help inform and describe final decisions in the review. (Emphasis added.) We submit that a rigorous cost-benefit analysis needs to be undertaken before decisions are made, not just in the Regulatory Impact Statement required to be prepared following the decision. In particular, it is a prerequisite to informed participation in the consultation process, particularly for non￾specialist stakeholders, including small businesses, industry bodies and the public at large. Government requirements for and guidance on cost-benefit analysis Government guidance requires that a cost benefit analysis “evaluates different policy options” to improve decision-making.12 The Treasury guidance notes that a Multi-Criteria Analysis (MCA) may be required, as a complement to the cost-benefit analysis, to determine all factors which contribute to achieving to the relevant objective and to evaluate them against the relevant criterion. The Guide to Cabinet’s Impact Analysis Requirements states that:13 “Before a substantive regulatory change is formally proposed, the government expects regulatory agencies to provide advice or assurance on the robustness of the proposed change, including by: …

• undertaking systematic impact and risk analysis, including assessing alternative legislative and non-legislative policy options, and how the proposed change might interact or align with existing domestic and international requirements within this or related regulatory systems;
• making genuine effort to identify, understand, and estimate the various categories of cost and benefit associated with the options for change; …”. (Emphasis in original.) The Guide to Social Cost Benefit Analysis notes that a more comprehensive cost-benefit analysis “where the importance of the decision requires it” – i.e. the degree of analysis should be proportionate to the potential impacts.14 Approach to considering costs and benefits in the Capital Review Paper After setting out the risk appetite framework, the analysis underlying the recommended CET1 requirement is contained in paragraphs 29 to 78 of the Capital Review Paper. It comprises an assessment of international econometric studies on optimal capital (Brooks, LEI, Dagher, and Firestone), modelling of risks in a New Zealand context, stress tests, and impacts of capital on output. The Reserve Bank last conducted an exercise of this nature in September 2012, when it assessed the costs and benefits of moving to the Basel III framework, and found that the currently applicable

12 Refer https://treasury.govt.nz/information-and-services/regulation/information-releases/regulatory-review￾programme/cost-benefit 13 Refer https://treasury.govt.nz/publications/guide/guide-cabinets-impact-analysis-requirements-html and see also section 162AB(1) of the RBNZ Act, which requires the Reserve Bank to “assess the expected regulatory impacts of any policy that it intends to adopt under Part 5” (which includes the prudential supervision mandate). 14 Refer https://treasury.govt.nz/sites/default/files/2015-07/cba-guide-jul15.pdf

9 minimum regulatory capital level would be optimal in New Zealand. In that paper, the Reserve Bank took the approach of calculating benefits of higher capital by estimating the expected fall in the probability of a financial crisis, and of assessing the likely cost in terms of the probable fall in economic activity. The capital ratio in which the marginal GDP costs and benefits lines intersect was determined to be optimal. This appears to be a reasonable approach in light of the soundness and efficiency mandate. The departure in the current capital review from the approach taken in 2012 appears to result primarily from the adoption of the “risk appetite framework”. In this approach, a provisional capital level is determined as that required to achieve the degree of soundness implied by the risk tolerance, and the efficiency analysis is then applied to see whether it likely that capital could be increased from the provisional amount without loss of expected output.15 This approach places very significant weight in both the analysis and the recommended capital level on the risk appetite framework and, by extension, on the strength, reliability and applicability of the analysis underlying that framework. As noted further below, there are reasons to doubt it has those qualities. More importantly, introducing a heavy conservatism bias into an optimality assessment is not assured to deliver the increase in soundness that the hypothetical risk averse citizen desires. Based on empirical literature and experience (refer below), there are substantial reasons to doubt that much higher levels of common equity capital actually deliver the lower probability of financial crisis inferred from overseas empirical studies (and underlying the 1-in-200 year probability). In addition, existing vulnerabilities arising from persistently weak levels of investment and productivity would be exacerbated by any increase in the cost of capital (the debate is only “how much?”), with consequent impacts on output, wellbeing, and ultimately soundness. Questions arising from literature on the costs and benefits of high capital requirements The assumptions in the Capital Review Paper which form the basis for the capital proposal raise a number of questions that, in our submission, warrant further analysis:  The benefits comprise the marginal benefit that the additional capital provides in reducing the probability of financial crises, with the quantum depending on how costly these are to the economy, compared with ‘ordinary recessions’. Yet this rests on a critical assumption about the efficacy of higher capital in preventing financial crises and on the difficult and debatable question of what costs are attributable to banking crises, and do not have an independent cause. In each case, recent empirical studies suggest that these assumptions should be tested more fully.  The costs depend on the higher borrowing margins engendered by the increased capital requirement, and also would include other potential impacts such as credit rationing, changes in lending composition and sectorial impacts, as well as second order effects, for example on investment and saving. There is significant divergence of views on the former, with a number of independent analysts suggesting an increase in borrowing costs between double and triple the Reserve Bank’s estimates. Potential broader or second order impacts are not evaluated. We address these questions in more detail below. Some further analytical queries in relation to the Capital Review Paper are set out in Appendix 3.

15 Capital Review Paper, paragraphs 24 and 25.

10 Benefits – the efficacy of capital in lowering the probability of financial crises A key assumption in the Reserve Bank’s analysis is that a substantially higher capital requirement will achieve soundness by materially decreasing the probability of financial crises, providing benefits in fiscal, output and other costs avoided. Although this matches intuition, there are a number of important caveats to this arising from both macro-prudential literature and experience. These call into doubt the approach of placing such a high degree of reliance on a very high capital ratio, in the form of CET1, as the core prudential tool, as well as the assessment of the benefit of doing so:  Empirical evidence: The efficacy of significantly higher capital requirements in reducing the probability of financial crisis has been challenged in recent empirical studies. For example, Jordà et al (March 2017) pg 2, find that: “there is no statistical evidence of a relationship between higher capital ratios and lower risk of systemic financial crisis. If anything, higher capital is associated with higher risk of financial crisis.” (Emphasis added.) Factors that were found to be far more significant were funding structure (in particular the loan￾to-deposit ratio) and high credit growth. Each of these is addressed by prudential tools other than capital (e.g. Net Stable Funding Ratio, monetary policy, macro-prudential tools and counter-cyclical buffers). The impact of funding structure is also borne out from the change in the composition of bank balance sheets in the lead-up to the GFC:  Theoretical underpinning: The empirical conclusions about the relative efficacy of capital compared with other prudential tools mesh with studies on the causes of financial crises. The vulnerability arises from banks’ intermediation role in (a) creating ‘riskless’ debt, funded by deposits, so that people and firms have a way to transact, and (b) transforming those riskless deposits into real investments in the economy (such home and business loans), which carry risk. This is critical to the functioning of the economy, but inherently exposes banks to the risk of self-fulfilling panic-based runs: when there are concerns about an upcoming recession, bank money becomes “information-sensitive” because depositors are concerned some banks might fail. Because they often don’t know which ones, there is an incentive to withdraw money

12 quickly rebounds and returns to its pre-crisis path. When measured using GDP, output does not bounce back, but this pattern is driven entirely by the experience of Japan. … Using conventional regression techniques with previous chronologies of post-war financial crises in advanced countries also does not provide strong support for the view that the aftermaths of crises are persistently grim.” (40-41) In addition, as pointed out by the Reserve Bank, the impacts of financial crises are very difficult to disentangle from credit-fuelled asset bubbles that commonly precede them. As a result, the economic and societal impact attributed in some studies to financial crises may to a large extent be due to the underlying factors that caused the financial instability, rather than to the financial stability itself (i.e. the financial instability may be a symptom rather than the cause). This is consistent with the Reserve Bank’s own work, which indicates that recessions are more severe and protracted if they follow house price bubbles, whether or not accompanied by a financial crisis.16

It also matches up to the experience of the GFC, which saw large housing booms and associated accumulation of household debt. Mian and Sufi (2014) argue that these developments, rather than the financial crisis itself, were the main drivers of the downturn and, especially, of the weak recovery. The Reserve Bank’s stress tests also point to the likelihood that adverse economic events of the magnitude that would threaten financial stability – for example very high levels of unemployment combined with a very dramatic collapse in property prices – would themselves carry material and long￾run output costs, independently of any financial crisis impacts. Implications None of this is to suggest that capital ratios are not important – they clearly are, and that is why in the same exercise undertaken by the Reserve Bank in 2012 it recommended capital levels higher than the Basel III minima. It does, however, raise questions about the basis for the risk appetite framework and the associated high degree of conservatism built into the capital proposal, and more broadly about whether the cost￾benefit assessment has been sufficiently developed and stress-tested. Most significantly, it calls into question the methodology of subordinating the efficiency criterion to soundness. The skew in the cost-benefit calculus created by this approach creates a real risk that the proposal to substantially increase the level of regulatory capital, and to fulfil that requirement through CET1 alone, could have an adverse impact on the key objective in section 1A of the RBNZ Act of contributing to a sustainable and productive economy. This risk is heightened by New Zealand’s persistent weakness in investment, and hence productivity, even in the current ‘steady state’. Other implications that arise from the foregoing analysis and from broader prudential literature include:  Focus on other factors and tools: Financial crises reflect a range of underlying imbalances (such as excessive credit growth and asset bubbles) and corresponding vulnerabilities, and resolving those requires a range of tools. As such, it is important that the broader toolkit is appropriately factored into the capital proposal at its formulation stage and is taken into account in a cost-benefit analysis. In this regard, it is also notable that a number of matters which are the subject of the Phase 2 review – for example, depositor protection and crisis management – potentially will have an important bearing on the capital adequacy tool, as will

16 Reserve Bank “Financial stability from housing market cycles” (2016).

13 future determinations to be made by the Reserve Bank (for example, as to leverage ratios, TLAC, and further policy development in relation to the countercyclical buffer).  Incentives and market signals: Since the GFC revealed some inadequacies in excessive reliance on minimum capital requirements, a number of prudential regulators and economists have sought alternative solutions that can complement CET1 minimum requirements. These studies focus in particular on the potential of alternative bail-in instruments to create:  a risk-averse constituency with powerful incentives to monitor banks, because the holders of those instruments (a) are subordinated to depositors and other senior creditors (so that bail-in instruments are more risk-sensitive) and (b) do not receive upside beyond the interest return, as holders of common equity do;  credit-sensitive pricing signals, that (being based on market trading) are more responsive, because they are based on market participants’ current perceptions of credit issues and asset impairments, than capital ratios, which are based on reported financial statements – this effect was particularly notable in the GFC, when some banks in the U.S. which were bailed out with TARP funds had very healthy reported capital ratios on the eve of that bailout, but had market-implied ratios close to the insolvency zone; and  prompts, through those pricing signals and resultant bondholder pressure, to timely corrective action at bank level, by (for example) replacing management responsible for the issues and recapitalisation or other balance sheet repair (for example asset sales) – for this reason it would not be entirely accurate to characterise these instruments as ‘gone concern’, rather than ‘going concern’, capital. These features align well with the Reserve Bank’s regulatory philosophy of emphasising the market discipline pillar. It also interacts well with the open bank resolution (OBR) tool, since these instruments bail-in by explicit contractual agreement. This is particularly relevant given the evidence from new empirical studies – that markets for wholesale bank debt are no longer pricing in a high probability of government bail-out – that the consistent messages from supervisors and legislators about moral hazard and “too big to fail” are starting to get through.17

There is a large literature examining the part that could be played by contingent convertible or other bail-in interests (refer for example Coffee (2011), Calorimis and Herring (2011), and Flannery (2009), and a large and increasingly liquid market in them has developed. Recent experiences with contingent capital in Europe (e.g. Deutsche Bank) suggest that the potential benefits set out above are starting to come to fruition, after some false starts in jurisdictions which do enjoy have strong institutions (emphasising the centrality of that factor to stability).

17 Antje Berndt, Darrell Duffie and Yichao Shu “The Decline of Too Big to Fail” (March 2019). In this regard it is notable that the primary objection by ‘big capital’ advocates Admati and Hellwig, that these instruments would not perform because of regulatory pressure and moral hazard, may have been too pessimistic – refer Anat R Admati and Martin Hellwig The Bankers’ New Clothes: What’s wrong with banking and what to do about it (Princeton University Press, 2013).

14 These factors suggest that questions of the “quality” of capital have more dimensions than simply loss-absorption and that the Reserve Bank should take a wider lens to its evaluation of composition of capital at whatever minimum it ultimately opts for.  Mitigating the costs of crises: The empirical study by Jordà et al (March 2017) concluded that the key impact of capital is not so much in preventing financial crises, but in reducing their severity and the persistence of their impacts on output and trend GDP growth. The same is also achieved by bail-in instruments or other total loss-absorbing capital (TLAC), which carry a far lower cost of capital than common equity. These have been implemented in Australia and a number of other jurisdictions.  Stress-testing of assumptions: A key lesson from analysis by Jordà, Gorton and others is that econometric studies should not be taken in isolation and should be tested against empirical research and observed experience. These models employ simplifying assumptions to analyse the resilience of a ‘representative bank balance sheet’ in absorbing losses of a given magnitude, by reference to historical experience of non-performing loans. Those assessments are made ex post and with perfect information, which does not match the conditions applying on the eve of a potential financial crisis. Those conditions create an equilibrium in which it is rational for deposit-holders, or short term wholesale lenders, to run or to freeze credit (Diamond and Dybvig, 1983). The extent of the efficacy of an enlarged CET1 requirement, and the complementarity of other capital and prudential measures, are matters for debate, but are cornerstones of a rigorous cost-benefit analysis. Such an analysis is crucial in forming a sound basis for stakeholders’ engagement in the consultation and for determining the optimal prudential settings to achieve a “sound and efficient” financial system. The significance and potential impacts of the capital proposal require that this is done at the stage of formulation, not only as part of an ex post Regulatory Impact Statement – particularly given the absence of any Parliamentary process to give effect to the proposal. Costs As with benefits, there are a number of components to the question of what costs higher capital requirements may carry and how they should be factored into the assessment. The focus in the Capital Review Paper, and in much of the ensuing debate, is on the impact on borrowing costs that arise as a result of higher bank capital ratios, and the degree to which they are dampened by the Reserve Bank’s Modigliani Miller assumption. In this regard, there are inherent difficulties in accurately determining the level of the ‘MM offset’ that should be assumed (refer for example Cline, 2015), which are magnified in a New Zealand context by the fact that D-SIBs price debt off their group’s AA- credit rating, not their own far lower stand-alone ratings, and the issue of disentangling equity cost of capital when the capital is at the subsidiary level. While cost of capital is very important, particularly given New Zealand’s relatively high level of private debt, it is not the only potential cost to consider. Other potential consequences of the proposal may include:  A reduction in the availability of credit or in the trajectory of credit growth. For example, Dagher et al (2016), pg 9, find that “one would expect that any rapid increase in mandatory capital ratios would take place at least partially through an adjustment of bank assets, with potentially large negative effects on credit and macroeconomic performance”. Other studies identifying this effect include Aiyar et al (2016), Cohen (2013) and Gropp (2018).  Sectorial/distributional impacts – for example, the potential for the new measures to have a disproportionate impact on the cost or availability of credit to particular sectors with higher

15 risk weights, for example SMEs and agri-borrowers which are heavily reliant on domestic bank funding for working capital and term investment needs and few, if any, alternative sources of capital available to them.  A decline in deposit rates in order to maintain margins or in response to any reduction in the official cash rate implemented to mitigate the effect of the capital requirement. This in turn may put pressure on those (such as retirees) reliant on savings returns, drive savers to higher yielding investments which are more complex and risky (information-sensitive) than deposits, and operate as a further disincentive to saving – already a significant policy issue (Savings Working Group, 2011).  Transitional costs as new lending is re-priced or banks seek to re-price existing loans under customary “increased cost” clauses. In this regard, although a 5-year transition is proposed, this may not have the mitigating effects intended – for example, Dagher et al (2016), pg 4, note that “markets tend to anticipate full compliance with new standards ahead of phase-in periods.” Any such effects will not necessarily be restricted to new lending, as loan agreements almost invariably give banks the ability to reprice, including specifically as a result of the impacts of increased capital requirements on their cost of funding or profitability.  Impacts on capital markets efficiency, for example through an increase in the cost of hedging or reduced willingness of banks to incur the capital costs associated with market￾making, which supports secondary market liquidity and is vital to the performance and sustainability of the debt capital market.  Changes in the composition of lending – for example lending may be tilted toward asset classes with low risk weights, such as housing, at the expense of lending to the productive economy, which carries higher risk weighting (see sectorial impacts below). This factor may operate to exacerbate existing imbalances, undermining financial stability. Another effect for which there is empirical evidence, and which can run counter to the response just mentioned or accompany it, is the incentive to undertake riskier lending in order to maintain profitability (Dautović, 2019).  Changes in financial system structure, through an impetus to move lending to off-balance sheet vehicles or to less regulated parts of the financial system (shadow banking). This is not always a negative phenomenon, particularly in a concentrated banking system such as New Zealand, but it creates risks if credit growth in this sector outpaces risk management procedures, underwriting practices are weak, significant related party lending occurs, or irrational exuberance comes into play. Each of these phenomena were evident in the failure of finance companies in New Zealand in the mid- to late-2000s.  To the extent that monetary policy is employed to counteract any effects of an increase in banks’ cost of funding, this will reduce the Reserve Bank’s ability to respond to weakening economic conditions by lowering the policy rate. Romer & Romer point to this as potentially a significant factor contributing to Japan’s large and protracted slowdown in economic growth, in the “lost decade” following the 1987 stock market crash (and in turn heavily skewing the data on the impact of financial crises). This is potentially a significant issue given the already historically low official cash rate. The unprecedented nature of the capital proposal makes it difficult to predict the likelihood or scale of such impacts. Empirical evidence is mixed, but all have been found to exist in at least some circumstances, and they are assessed in many of the overseas studies reviewed by the Reserve Bank

16 (albeit by reference to the different circumstances applying in those jurisdictions). These are not addressed in the consultation document, but should be considered in evaluating of the merits of the proposal, or at least stress-tested, including by reference to whether alternative options are less likely to give rise to such effects or to do so to the same extent. Funding and sectorial impact One issue identified in prudential studies, but not considered in the Capital Review Paper, is the impact that may be borne by high capital usage sectors critical to the economy (for example, small to medium sized enterprises (SMEs), rural borrowers, and the construction sector). This is particularly important because SMEs and the non-corporate agri sector do not enjoy the access that large companies have to alternative sources of funding, such as the local and international debt capital markets, and so are heavily reliant on bank lending. As a recent OECD report on SME funding internationally highlights, the cost of capital for SMEs is already high in New Zealand and is not reducing at the rate evident in other OECD countries: Because SMEs play such a crucial role in our economy, including in sustaining employment and contributing to productivity, any impact on the availability or cost of bank funding is a real concern, and should be a matter specifically assessed in a cost-benefit analysis of the capital proposal (including, ideally, by reference to the in-depth studies by the OECD into this issue and potential responses). The issues with rural lending are similar but may be intensified by the broad range of issues farmers are currently confronting, including managing emissions, water quality, labour capacity, pressure to deleverage, etc. The rural sector is a vital cog in our economy and concern about how to appropriately manage issues associated with rural debt are evident in the recent Farm Debt Mediation Bill. It is important in this context that the capital proposal is given effect in a joined up way to broader Government initiatives, and particularly that the Capital Proposal does not precipitate a rapid or disorderly response to issues of rural debt, or increase the cost of capital to such an extent as to undermine the significant investments that will be required to respond to the various challenges in this sector.

17 Decarbonisation is another issue that is both linked to efficient capital (as a key input) and wellbeing (as an outcome). Very substantial investments will be required to sources of generation in order to meet zero carbon targets and to transition to EVs, among many others. More broadly, the infrastructure deficit and related housing affordability issues are rightly the focus of a range of Government initiatives and require a very significant investment response. These matters should also be factored into the approach taken to achieving the optimal balance of soundness and efficiency. The part played by other factors, and prudential tools corresponding to them Empirical studies have found that there are a number of factors of as much or more significance as capital adequacy requirements to the incidence or probability of banking crises. We set some of these out below, with the policy response contained in the Basel III framework, and our assessment of the New Zealand position (both in brief, non-comprehensive summary):  Funding structure: In particular the degree to which banks’ loan books are funded by deposits (i.e. the loan-to-deposit ratio (LtD)) is a key indicator. Funding structure, particularly through diminished or low LtD is significantly correlated with financial crises. Basel response: Net Stable Funding Ratio. NZ assessment: Strong: implemented by way of the Core Funding Ratio. A high percentage of the loan book is now funded by either ‘sticky’ deposits or long tenor wholesale or retail debt.  Credit growth / asset bubbles: Financial crises are very commonly preceded by rapid asset price appreciation and a corresponding increase in credit growth and debt levels. Basel response: Countercyclical buffer (in addition to the dampening effect provided by conventional monetary policy). NZ assessment: Strong: Macro-prudential policy enhances the broader framework of prudential regulation by actively varying prudential instruments over time to help reduce the potentially damaging effects of asset and credit booms. This tool, in the form of the LVR speed limits, has been applied and performed broadly according to expectations (Dunstan, 2014). The countercyclical buffer (which the Reserve Bank is currently formulating more detailed policies on) is available. Monetary policy can also have a positive impact on financial stability through its dampening effect on the credit cycle.  Liquidity: Sufficient cash or high quality liquid assets to cover short run liquidity requirements. Basel response: Liquidity Coverage Ratio (LCR). NZ assessment: Strong: LCR applies.  Strength of institutions: Rule of law, independent judiciary, strong regulators, lack of corruption, property rights, audit rules, and similar factors indicating the strength of a country’s institutions help stave off financial crises. NZ assessment: Very strong: New Zealand ranks very highly in this criterion by general acclamation (including in rating agency reports and assessment by international agencies such as the World Bank and OECD).

18 There are also factors which can contribute significantly to the likelihood of financial crises, including the following (again accompanied by an assessment of New Zealand’s position):  Financial liberalisation: Similar to excess credit growth, rapid financial liberalisation has been identified as a common factor in many financial crises (for example Reinhart & Rogoff, Gorton (2012)). NZ assessment: Stable: The financial liberalisation process was concluded in the late 1980s and New Zealand’s banking system is described by the Reserve Bank as relatively conservative and vanilla.  Currency pegs/foreign liabilities: Financial crises can be caused or contributed to by attempts to defend a currency peg (e.g. Sweden ~1992, Asian financial crisis ~1998) or by incurring substantial unhedged foreign liabilities. NZ assessment: Strong: Floating exchange rate removes the first risk, and most borrowing is either done in New Zealand dollars or hedged via an economic or derivative position.  Culture and conduct: Nick Le Pan, a former head of Canada’s top banking regulator, the Office of the Superintendent of Financial Institutions (OSFI), is reported as saying: “What’s more important is how the business is conducted, whether it be the capital markets investment banking business, or whether it be the retail business. Are they conducted with the right culture, and customer focus, and prudence, and sustainability? That’s much more important for safety and soundness”. We agree with that and note that it is a key focus of both the Reserve Bank and the Financial Markets Authority, who have recently reported back with recommendations following an investigation into these issues. NZ assessment: Strengthening: The Reserve Bank/FMA study did not reveal issues of the magnitude that have arisen in other jurisdictions, including under the Hayne inquiry, but nonetheless called for more focus in this area and has backed this up with a number of recommendations that are currently being considered. Finally, there are some factors – such as “regulatory intensity” – that have been shown not to have a statistically significant influence on the probability of a financial crisis, but nonetheless are commonly pursued by prudential authorities. The Reserve Bank currently takes a different approach to this, preferring (for example) not to undertake site visits. To sum up, as noted by the IMF in its recent FSAP assessment, New Zealand’s prudential settings are strong and as a result its financial system is stable. But the issues discussed below (under Broader economic impacts) indicate that we are falling short in underpinning the level of investment necessary to improve our productivity and wellbeing and to adjust to the impacts of rapid technological change and global warming (among others). This has been a significant focus of Government initiatives in recent years, leading to the establishment of the Productivity Commission. As that Commission’s work on housing affordability, infrastructure, carbon adjustment show, the need for a renewed focus on investment has become significantly more pressing. This is the context in which the approach to the soundness and efficiency criteria must be assessed.

19 The efficiency criterion The Reserve Bank has done a lot of work on what efficiency means.18 In its May 2014 Financial Stability Report, the Reserve Bank noted: “An efficient financial system is one that enables economic resources to be allocated to their best use across time and space without imposing unnecessary costs (or ‘rents’) on households and businesses. … [A]n inefficient financial system can hamper economic prosperity by imposing unnecessary costs on households and businesses … and misallocating resources.” Deputy Governor Geoff Bascand recently gave the following interpretation:19 “The efficiency goal means different things in different contexts: we minimise compliance costs; we support innovation and operate a regime that is open to new entrants; we avoid creating unnecessary frictions in the supply of credit to the economy; and we ensure that financial resources are allocated in a productive (and not harmful) way to maximise long term economic growth.” We agree with that. In our opinion, in the context of setting prudential policies which can have broad economic effects, efficiency must be interpreted as extending to the broader concepts referred to above. This also accords with the new objective in the RBNZ Act of contributing to a “sustainable and productive economy”. A key premise of the Reserve Bank in its capital review is that New Zealand’s high levels of household and farm debt, and the fact that we are a small open economy, require an overlay of conservatism in our prudential settings. This is the ‘risk aversion framework’ underlying the new capital requirements,20 and has translated into a proposed capital level that would be among the highest in the world, once the underlying conservatism in risk weights and other inputs are factored in.21

The same factors underlying the recommended risk aversion, however, make us heavily reliant on debt funding and thus particularly sensitive to the cost of debt capital – particularly when combined with our muted equity investment potential resulting from inadequate domestic savings, and the constraints on foreign direct investment owing to the Overseas Investment regime. Another factor cited by the Reserve Bank in setting a very conservative risk tolerance is the concentrated nature of the banking system, but the effects of this are ambiguous as comparative studies have suggested that such systems may be more stable and resilient to shocks – for example the Canadian banking system, like the Australasian banking system, came through the GFC largely unscathed.22 Similarly, Gorton (2012) notes that “countries with highly concentrated banking systems are less likely to have crises”.

18 Refer for example Chris Bloor and Chris Hunt “Understanding financial system efficiency in New Zealand’, Reserve Bank of New Zealand Bulletin, Volume 74(2), June 2011. 19 From a speech entitled “Financial stability – risky, safe, or just right?” delivered to the UBS Australasia Conference in Sydney (13 November 2018). 20 Refer the 3 April Background Paper and the discussion below. 21 An internal memorandum (7 August 2016) notes that the Reserve Bank approach to capital adequacy is more conservative than that taken by either Basel or APRA. It was noted that the effects of this conservatism are difficult to quantify, but “are roughly equivalent to banks holding an additional 1-2 percentage points of CET1 when measured on a Basel basis”. 22 Huang and Ratnovski (2009). See also Laeven and Valencia (2018).

20 The soundness criterion and the risk appetite framework The Capital Review Paper (pg 14) notes that the Reserve Bank’s “risk appetite framework” (intended to reflect society’s tolerance for risk) “plays a central role in both the policy goals and our decision￾making process”. The 3 April background paper sets out further detail on the analytical basis for the capital proposal and the risk appetite framework. Each of these papers indicate that the Reserve Bank has chosen to formulate prudential policy to achieve the soundness limb on the basis of the ‘risk appetite framework’ which is informed by risk aversion literature. The Reserve Bank notes that this does not reflect the conventional approach to prudential regulation (at pg 12): “In the majority of the capital policy studies we reviewed the policy goal was defined solely in terms of maximising expected output, with no role for risk aversion.” And that: “any modelled representation of society’s preferences depends on assumptions (the accuracy of which may be impossible to verify ex ante) and the results will be very sensitive to the assumptions made.” The risk aversion literature seeks to identify and quantify cognitive or behavioural biases to explain deviations from results achieved under standard economic models, which are predicated on rational actors. A number of questions arise in relation to the application of risk aversion principles to prudential policy:  The O’Donoghue paper cited by the Reserve Bank23 does not contemplate the application of risk aversion models to prudential regulation, stressing the models are context-specific and require further development. O’Donoghue also cautions a “model that describes people’s behaviour might not be the metric we ought to use for welfare analysis”.24  The literature suggests a very broad range of conclusions can be drawn, with a high degree of complexity in particular when results are abstracted from a ‘representative person’ to broader society: “measuring attitudes to risk is a difficult task, if not impossible, at a macro-level”. 25  It is unclear how the risk appetite framework interacts with the “efficiency” component of the prudential mandate, which on its face suggests that the goal should be to achieve the optimal level of capital, in line with the approach normally taken in capital policy studies. This comes back to the discussion above about the balance between soundness and efficiency, including the integration of this with broader economic policy goals. Ultimately the question in relation to any prudential policy is the extent to which it contributes to “soundness and efficiency”, and care must be taken in introducing a behavioural bias into the analysis. 26

23 Ted O’Donoghue and Jason Somerville “Modelling Risk Aversion in Economics” (Journal of Economics Perspectives, 32:2, 91-114 (2018)). 24 O’Donoghue, pgs 98, 106, and 111. Perhaps just as salient is the observation, on pg 104, that loss-overweighting generates “risk-aversion and thus a willingness to pay for insurance that is larger than the actuarially fair price”. 25 For example, J François Outreville “Risk Aversion, Risk Behaviour, and Demand for Insurance: A Survey” (Journal of Insurance Issues, 37(2), 2014, 158 at pg 170). 26 Douglas W Blackburn and Anrey D Ukhov “Individual vs. Aggregate Preferences: The Case of a Small Fish in a Big Pond” (May 2008).

22 A degree of interdependence is also evident in the statutory criteria, in the sense that soundness is required to underpin long-term investment and to avoid costly boom-bust cycles, and productivity and wealth creation (i.e. increasing output) are essential to servicing and sustaining our high level of private debt. The feedback loops between the criteria are another reason to keep them in proper balance. Broader economic impacts – investment, productivity and wellbeing As noted previously, the RBNZ Act was recently amended to redefined the purposes of the Reserve Bank in its prudential and other functions, to “promote the prosperity and well-being of New Zealanders, and contribute to a sustainable and productive economy” (s 1A RBNZ Act). The reference to a ‘sustainable and productive’ economy echoes the ‘soundness and efficiency’ criteria, and recognises the connection between prudential policy-making and broader economic outcomes. It also aligns well with the IMF’s definition of “financial stability” noted above, which is closely linked to efficiency and emphasises the contribution to be made by the financial system to the real economy, including through efficient asset allocation, “and ultimately prosperity”. It also implies shifting at least some of the prudential focus to imbalances that have built up in the economy – a significant issue in New Zealand, particularly in relation to housing. Here, for example, that might involve looking at the potential impact on those imbalances of much higher capital requirements. A key issue identified by the OECD in its 2017 survey of New Zealand is that “Productivity remains well below that of leading OECD countries”, restraining living standards and well-being. The Survey goes on to find that productivity is held back by persistently weak investment, such that non-residential capital formation per person in the labour force is less than 75% of the OECD average. The same factors mean that GDP per capita is below the OECD mean, and that “improving productivity growth is a major long-term challenge for improving inclusiveness and living standards”. The fact that New Zealand’s productivity has lagged behind that in most other OECD countries over the past two decades, despite generally productivity-friendly policies, is sometimes referred to as ‘the New Zealand enigma’. This has a number of potential causes, some of which – such as distance from markets and lack of scale – we can’t do much about (beyond, perhaps, investing in technology). However, one of the key factors – cost of capital – lies at least partially in our control. The OECD notes (pgs 34-35): “A higher cost of capital than in most other advanced economies contributes to low capital investment. As national saving has persistently fallen short of investment, New Zealand has accumulated substantial foreign liabilities, and international investors may require a premium to invest there (Rose, 2009; McDermott 2013). Also, owing in part to its small size, New Zealand has thin venture capital, stock and bond markets. Low rates of capital investment depress wages, with negative consequences for income distribution and inclusiveness.” These problems are familiar ones for New Zealand officials. For example, the 2008 Treasury Productivity Paper, to which the Reserve Bank contributed, notes that: “New Zealand faces a big challenge to overcome its productivity shortfall”, as it sits 22nd out of 30 in the OECD productivity league table, generating 30% less output per employee than Australia. Similarly, the Savings Working Group convened in 2011 commented that growth in productivity, incomes and living standards is much too slow, and described that situation as “pretty shocking really”. The Productivity Paper similarly notes that investment is one of the key productivity drivers, and that “New Zealand firms face a somewhat elevated cost of capital compared to other OECD countries…”, with real interest rates showing a persistent premium. Factors identified as contributing to the interest rate premium and high cost of capital were low rates of saving, financial market development,

23 exchange rate volatility, tax and high levels of debt. Wrapping up the link to investment, the Productivity Paper cites a Reserve Bank study, which concluded that “a high domestic cost of capital is almost certain to be holding back the total level of real business investment in New Zealand”. The Savings Working Group also noted that we have very high net foreign liabilities, 90% of which is in the form of debt – emphasising our high level of sensitivity to the cost of debt capital, which will persist unless the approach to foreign direct investment undergoes a radical change (as recommended by the OECD). In relation to the imbalances in our economy, the Savings Working Group goes on to note that asset￾price inflation has been accompanied by a large increase in debt, more than half of which is housing loans: “In the last 15 years, household debt has doubled relative to incomes, most of it in the form of mortgages to buy increasingly expensive houses. … We need to close the GDP gap on the OECD. The gap is the result of low productivity – a lot of hours worked for modest reward. The growth has been driven by the wrong things, an increase in consumption and house prices. These have imbalanced the economy.” Emphasising the interconnected nature of these issues, the Savings Working Group observed that a number of incentives affect saving, including low interest rates – monetary policy means savers lose out and that households could borrow more than previously, further contributing to asset price inflation. The result is an over-investment in property assets that have a low productivity return. That, in turn, has generated what everyone, including the Reserve Bank, agrees is our most significant financial stability vulnerability, as well as having broader impacts on wellbeing through housing unaffordability. Prudential rules did not cause this problem, nor by themselves can they solve it, but there is at least a possibility that an increase in the cost of capital engendered by the capital proposal might make matters worse, by creating the conditions for further suppression of real economy investment, reducing our already very low deposit rates, or creating incentives to direct lending away from the productive – but highly risk-weighted – parts of the economy. Concluding comments The Capital Review comes at an important time, as the Government prepares the world’s first Wellbeing Budget. Amendments to the objectives of in the RBNZ Act made following Phase 1 of the Reserve Bank review draw a clear connection between the Reserve Bank’s prudential role and the goal of promoting the prosperity and wellbeing of New Zealanders. This connection is also evident from the direct line drawn in studies noted above between cost of capital, investment, productivity and wellbeing. It is an important reason for striving for the optimal balance of soundness and efficiency in our financial system. The Capital Review is also happening at the same time as a comprehensive review of New Zealand’s prudential supervision arrangements, which consider important questions relating to the connection between the Reserve Bank and the Government in relation to the prudential policy-making function. For the same reasons as just given, we think it is vital that those are more integrated. These factors also point to the overwhelming case to be made for the Reserve Bank to have substantially more resources to undertake its prudential function, as recommended by the IMF in its recent FSAP, and which is being considered in the next part of the Phase 2 review. In this regard, we note that the Reserve Bank’s 5-year funding agreement can be varied at any time by agreement between the Minister of Finance and the Governor (s 159(3) RBNZ Act).

24 We have not made any submission on what is the “right” level of capital in New Zealand. Rather, we have called for an evaluation of the capital proposal that is proportionate to its importance and which achieves the optimal balance of soundness and efficiency, to underpin the investment required for prosperity and wellbeing. We thank you for the opportunity to submit on these important questions. INFINZ has no objection to any publication of this submission. Yours sincerely, Jim McElwain, Chief Executive Louise Tong, Chair Ross Pennington, Chair, Advocacy Clyde de Souza, Board Member Institute of Finance Professionals New Zealand Inc. Appendices

1. OECD investment and productivity data (New Zealand).
2. Prudential policy toolkit, in New Zealand and internationally.
3. Queries in relation to the analysis in the Capital Review Paper.

25 References Reserve Bank of New Zealand “Capital Review Paper 4: How Much Capital is Enough?” (December 2018) (“Capital Review Paper”).

• “Review of the Capital Adequacy Framework for Locally Incorporated Banks” (May 2017) (“May 2017 Capital Review Paper”).
• “Capital Review Background Paper: An outline of the analysis supporting the risk appetite framework” (3 April 2019) (“3 April Background Paper”).
• “Regulatory impact assessment of Basel III capital requirements in New Zealand” (September 2012). Reserve Bank and Treasury Phase 2 paper “Safeguarding the future of our financial system: the role of the Reserve Bank and how it should be governed” (1 November 2018) (“Phase 2 Paper”). Shekhar Aiyar, Charles W Calorimis, and Tomasz Wieladek “How Does Credit Supply Respond to Monetary Policy and Bank Minimum Capital Requirements” (European Economic Review 82, 2016). Malcolm Baker and Jeffrey Wurgler “Would Stricter Capital Requirements Raise the Cost of Capital? Bank Capital and the Low Risk Anomaly” (unpublished working paper, 2013). Bank of England (Martin Brooks et al) “Measuring the Macroeconomic Costs and Benefits of Higher UK Bank Capital Requirements” (Financial Stability Paper No. 35, December 2015). Charles W. Calomiris and Richard J. Herring “Why and How to Design a Contingent Convertible Debt Requirement” (November 2011), Wharton School Working Paper. William Cline “Testing the Modigliani-Miller Theorem of Capital Structure Irrelevance for Banks” (unpublished working paper, 2015). John C Coffee Jr “Systemic Risk after Dodd-Frank: Contingent Capital and the Need for Regulatory Strategies Beyond Oversight” (2011) 111 Columbia Law Review 795. Benjamin Cohen “How Have Banks Adjusted to Higher Capital Requirements?” (BIS Quarterly Review, September 2013). Jihad Dagher, Giovanni Dell’Ariccia, Luc Laeven, Lev Ratnovski, and Hui Tong “Benefits and Costs of Bank Capital” (IMF Staff Discussion Note 16/04, March 2016). Harry d’Angelo and René Stulz “Liquid Claim Production, Risk Management, and Bank Capital Structure: Why High Leverage is Optimal for Banks” (2015) 116 Journal of Financial Economics 219. Ernest Dautović “Has regulatory capital made banks safer? Skin in the game vs moral hazard” (European Systemic Risk Board Working Paper No 91, May 2019). Douglas W Diamond and Philip H Dybvig “Bank Runs, Deposit Insurance, and Liquidity” (Journal of Political Economy, Vol 91, No. 3, 1983). Ashley Dunstan “The Interaction between Monetary and Macro-prudential policy” (RBNZ Bulletin Vol 77 No. 2, June 2014).

26 Douglas Elliott, Suzanne Salloy, and André Oliveira Santos) “Assessing the Cost of Financial Regulation” (IMF Working Paper WP /12/233, September 2012). Simon Firestone, Amy Lorenc, and Ben Ranish “An Empirical Economic Assessment of the Costs and Benefits of Bank Capital in the US” (Board of Governors of the Federal Reserve System, Finance and Economics Discussion Series, 31 March 2017). Mark J Flannery “Stabilizing Large Financial Institutions with Contingent Capital Certificates” (October 2009), SSRN Working Paper. Gary B. Gorton Misunderstanding Financial Crises: Why we don’t see them coming (Oxford University Press, 2012). Reint Gropp, Thomas Mosk, Steven Ongena, and Carlo Wix “Bank Response to Higher Capital Requirements: Evidence from a Quasi-Natural Experiment” (SAFE Working Paper 156, January 2018). Rocco Huang and Lev Ratnovski “Why Are Canadian Banks More Resilient?” (IMF Working Paper No. 09/152, 2009. IMF “New Zealand: Financial System Stability Assessment” (IMF Country Report 17/110, May 2017). Òscar Jordà, Björn Richter, Moritz Schularick, and Alan M Taylor “Bank Capital Redux: Solvency, Liquidity, and Crisis” (NBER Working Paper 23287, March 2017). Luc Laeven and Fabian Valencia “Systemic Banking Crises Revisited” (IMF Working Paper, WP/18/206, September 2018). Donato Masciandaro, Rosaria Vega Pansini and Marc Quintyn “The Economic Crisis: Did Financial Supervision Matter?” (IMF Working Paper WP/11/261, November 2011). Natalya Martynova “Effect of Bank Capital Requirements on Economic Growth: A Survey” (De Nederlandsche Bank Working Paper No. 467, March 2015). Atif Mian and Amir Sufi House of Debt: How They (and You) Caused the Great Recession, and How We Can Prevent It from Happening Again (Chicago: University of Chicago Press, 2014). OECD Economic Surveys New Zealand (June 2017). Carmen Reinhart and Kenneth Rogoff This Time Is Different: Eight Centuries of Financial Folly (Princeton University Press, 2009). Christina D Romer and David H Romer “New Evidence of the Impact of Financial Crises in Advanced Countries” (NBER working paper 21021, March 2015). Treasury “Investment, Productivity and the Cost of Capital: Understanding New Zealand’s Capital Shallowness” (Productivity Paper TPRP, 08/03, 17 April 2008). Treasury “Saving New Zealand: Reducing Vulnerabilities to Growth and Prosperity: Final Report to the Minister of Finance” (Savings Working Group Paper, 1 February 2011).

27 APPENDIX 1 – OECD INVESTMENT AND PRODUCTIVITY DATA (NEW ZEALAND) International obligations (NIIP) As national saving has persistently fallen short of investment, and New Zealand has run nearly continuous current account deficits, has accumulated substantial foreign liabilities.

Persistent house price appreciation, which has accelerated over the past two decades, has led to a singificant affordability issue, and exposure to interest rate risk. We have low labour productivity, suppressed in particular by low investment levels.

29 Basel treatment Basel capital requirements (in % of risk weighted assets): CET1: 4.5%. CET1+AT1: 6%. CET1+AT1+Tier 2: 8%. Capital conservation buffer: 2.5% CET1. Total requirement: 10.5% of RWA, with at least 7% CET1. Discretionary counter-cyclical buffer: 0-2.5% of CET1 during periods of high credit growth. Up to 1.5% of capital requirement may be fulfilled with AT1 capital. Up to 2% of capital requirement may be fulfilled with Tier 2 capital. New requirement. Driven by FSB (Financial Stability Board) work and currently being implemented in major jurisdictions. Requirement of 16% RWA, 6% LR since January 2019 (rising to 18% RWA and 6.75% LR in January 2022) for G-SIB. Net stable funding ratio (ASF/RSF) =/> 1. Available Stable Funding (ASF): portion of capital/long term liabilities with tenor > 1 year, taking account of ‘stickiness’. Required Stable Funding (RSF): amount of stable funding required to be held. It is a function of the liquidity characteristics and residual maturities of assets/off-balance sheet commitments. Liquidity coverage ratio: Stock of high quality liquid assets (HQLA) must exceed total net cash outflow amount over a 30 calendar day period. NZ position Bespoke. NZ has proposed minimum CET1 levels that substantially exceed the Basel minima (in particular increasing the mandatory capital conservation buffer to 7.5%). Otherwise generally consistent with Basel framework, but with divergences in RWA etc. Disfavoured. The RBNZ strongly discourages use of AT1 and Tier 2 instruments, regarding them as complex, lower quality, and uncertain in how they would perform under stress. Non-conforming. NZ OBR regime takes a different approach, by contemplating bail-in of both wholesale funding and retail deposits (subject to ‘de minimis’ for the latter). The RBNZ is against introducing TLAC policy, referring to complexity and doubts about feasibility of SPE (single point of entry) bail-in approach. Bespoke. Similar in substance but called ‘Core Funding Ratio’ under BS13 rather than Basel. Bespoke. Similar in substance but under BS13 regime rather than Basel.

31 introduced in major jurisdictions. (e.g. G20 Pittsburgh Accord/EMIR) APRA proposes LR of 3.5% for IRB ADI (3% for standardised ADI). where state aid can be provided. NZ position Partial. NZ contemplates bail￾in under OBR, but the RBNZ doesn’t have a statutory bail-in power. Limited development/ planning of OBR may reduce effectiveness in time of crisis, e.g. due to:  Affecting only a limited number of creditors;  Absence of statutory bail￾in order;  Identity of post-OBR shareholders. Non-conforming. NZ has not implemented laws to give effect to the G20 Pittsburgh Accord. NZ has adopted a “wait and see” approach, filling gaps where required for NZ banks to meet international requirements. Non-conforming. NZ has not implemented a leverage ratio as RBNZ has not considered it appropriate for NZ market conditions. RBNZ proposes tightening of IRB approach for calculating RWA, suggesting that it sees issues in current RWA calculations. Yes. NZ carries out stress tests. Yes, bespoke. OBR underlines that there should be no bail-out, but NZ is unusual in extending bail-in to retail depositors while not having (formalised) depositor protection. Limited development/ questions on effectiveness of OBR may put pressure on government to conduct bail-out.

32 QUERIES IN RELATION TO THE ANALYSIS IN THE CAPITAL REVIEW PAPER  Sensitivity analysis: The impacts shown in relevant international studies are very broad in range. As a result, it would be helpful for submitters to have more in-depth stress testing to help understand what results are the most likely in the context of the New Zealand financial system and the degree of risk that outcomes will be greater than predicted. In addition, there is no investigation of which of the studies is likely to be most relevant to New Zealand conditions, in terms of methodology, underlying assumptions, factors taken into account, and sample set.  Economic impacts: The assessment of impact on output is very brief, given the significance of the proposal and the key objective in the RBNZ Act that the prudential policy promotes the prosperity and well-being of New Zealanders, and contributes to a sustainable and productive economy. In particular, there is no assessment of the potential impact on key imbalances affecting and issues in a New Zealand context (most notably the low level of investment and impact of that on productivity, growth and wellbeing – each of which is influenced significantly by cost of capital (Treasury Productivity Paper (2008), OECD (2017)) or of second order effects, such as tightening of lending conditions.  Factoring in the impact of other prudential tools: In contrast to the approach taken in the 2015 Brooks/Bank of England study, which took into account in the optimality assessment the contribution of other prudential tools, such as crisis management and standards for additional loss￾absorbing capacity, there is no evaluation of the contribution or effect of other elements of the prudential toolkit (such as, in a New Zealand context, Open Bank Resolution). See also IMF (2012), pg 42-55. These studies suggest that a robust cost-benefit analysis must include a consideration of the contributions made to financial stability by other components of the prudential toolkit.  Adjustment to reflect underlying conservatism: The analysis does not take into account the in-built conservatism in the calculation of New Zealand capital levels – which is assessed by the Reserve Bank as adding approximately 200bp on a ‘like for like’ basis – or the buffer that is held by banks above the regulatory minimum.  Other potential impacts/costs: The broader range of potential bank responses to changed requirements – such as reducing credit to bank-dependent borrowers such as SMEs, balance sheet reduction (deleveraging), or balance sheet adjustment (shifting toward lower risk-weighted assets such as mortgage loans) – is not assessed (contrast Brookes et al (2015), pg 22) and IMF (2012) pg 23-24).  Composition: There is no analysis of other available options, such as fulfilling some of the additional capital requirement through alternative loss-absorbing instruments – the approach recently taken in Australia. In this regard, the study by Dagher et al (2016) notes capital requirements may be met by other bail-in-able instruments, such that that their results “may be reinterpreted as applying to other TLAC instruments”.

Crombie Lockwood Tower Level 16, 191 Queen Street PO Box 7244 Wellesley Street Auckland 1141 New Zealand Phone: +64 9 377 5570 Email: info@infrastructure.org.nz 17 May 2019 Reserve Bank of New Zealand 2 The Terrace PO Box 2498 Wellington 6140 New Zealand Per email: capitalreview@rbnz.govt.nz Infrastructure New Zealand is the peak industry body for the infrastructure sector and promotes best practice in national infrastructure development through research, advocacy and public and private sector collaboration. Infrastructure New Zealand members come from diverse sectors across New Zealand and include infrastructure service providers, investors and operators. This submission represents the views of Infrastructure New Zealand as a collective whole and may not necessarily represent the views of individual member organisations. Infrastructure New Zealand feedback on the review of the capital adequacy framework for registered banks Summary Infrastructure New Zealand supports the Reserve Bank’s efforts to ensure the stability of the New Zealand economy. However, we are concerned that the proposed changes to the capital adequacy framework for registered banks may impose costs on investment capital. Our concern is that a reduction in infrastructure investment may not be sufficiently offset by the potential benefits of the proposed changes. Consequently, we support the recommendation from BusinessNZ and others that the Reserve Bank conduct a comprehensive cost-benefit analysis of the proposed changes prior to taking any action. We support the submission made to the Reserve Bank on this topic by BusinessNZ.

2 Increasing capital requirements may reduce investment capital and hamper infrastructure investment If the proposed changes go ahead, banks will need to raise more capital to meet the requirements. We are concerned that this will change the quantity and cost of lending to customers. To raise capital banks may choose to lend less or raise interest rates on loans. Less lending and higher interest rates would both pull capital out of the market that would otherwise have been invested in the economy, including, most notably, infrastructure development. This is a critical period of infrastructure investment in New Zealand. The Government has committed over $42 billion in infrastructure investment over the next four years in an effort to rebuild New Zealand’s depleted capital stock. This is a major increase to existing expenditure and delivery is already proving challenging. We are concerned that increasing capital requirements over the next five years will remove investment capital from the New Zealand market at the precise moment when New Zealand needs it most. In addition to reducing lending, or making it less affordable, banks may also adjust their lending to customers who are more profitable or less risky. This may shift lending away from developers, sub￾contractors and other productive companies and towards safer and more profitable sectors such as home mortgage lending. It is unclear how banks will respond to changes in capital requirements, but we have observed uncertainty among infrastructure lenders as to what the net impacts will be on lending in the infrastructure sector. In light of this, we strongly recommend that the Reserve Bank undertakes a comprehensive cost-benefit analysis of the proposed changes. The impacts of infrastructure investment are long-term, widespread, and affect all aspects of life In considering the costs and benefits of these changes, it is important to understand the full impacts that infrastructure has on the New Zealand economy and society in general. Infrastructure is an enabler of economic growth, productivity, innovation, growth and wider environmental and social outcomes. Infrastructure supports economic activity, saving costs for businesses, enabling New Zealand to better compete internationally. Lower costs reduce barriers to entry for New Zealand markets, driving greater entrepreneurship and innovation.

3 Infrastructure is essential in the delivery of more accessible and cost-effective housing, facilitating economic and social mobility and greater social equity. Increasing capital requirements may have a long-term and widespread multiplier effect Given the critical role that infrastructure plays in lifting living standards in New Zealand, we are concerned that the Reserve Bank’s proposed changes to capital adequacy framework will have knock-on effects throughout the New Zealand economy. The infrastructure sector relies on an adequate supply of investment capital to facilitate development. This includes both businesses involved in the construction and operation of infrastructure as well as private and public sector asset owners considering the use of debt to finance new projects. Available funding is unlikely to fully meet the Government’s infrastructure objectives. Access to private finance provides a means to accelerate the provision of public infrastructure. However, we are concerned that changes to the capital adequacy framework may significantly affect infrastructure investment even as the country is attempting to correct decades of underfunding. As infrastructure underpins and supports so much of the country’s long-term economic growth and prosperity, as well as improved environmental and social outcomes, a capital shortfall at this time could have considerable consequences for New Zealand’s long-term competitiveness, social equity, and quality of life. We ask that the Reserve Bank undertakes a comprehensive cost-benefit analysis to fully understand the consequences of this proposal not only on lending costs, but on infrastructure investment and all the impacts infrastructure has on New Zealand’s economy and society. We thank the Reserve Bank for the opportunity to provide feedback on this important proposal.