Reserve Bank of New Zealand Capital Requirements Consultation Reports and External Updates

1 Reserve Bank NZ - Capital Requirements Consultation Social NGO Focus Group Report Independent Report by Travis Wallach from Continuum Consulting Group Report Date 16 May 2019 Background As part of the consultation period into the Reserve Bank of NZ’s proposal to increase capital requirements for NZ banks to improve banking stability and reduce possible future banking crises, a focus group was held to seek feedback from a number of social non-government organisations (NGOs). The aims of the focus group were to engage more with key members of the social representative community on policy formation, and to better understand the key social impacts that may require further consideration. The focus group was an opportunity for input by the attending socially engaged NGOs into the formal capital requirements consultation. The focus group was facilitated by Travis Wallach, an independent consultant, on Friday 10th of May 2019, at The Reserve Bank in Wellington. Participants Seven representatives from a range of socially engaged NGOs, representing a diverse group of NZ society, selected to attend the capital requirements consultation focus group. In addition three Reserve Bank officials were in attendance at the beginning and end of the focus group, but were not present during a period of facilitated open free discussion between the NGO representatives. The full list of participants included: • Simone Robbers – Assistant Governor, Manager of Governance, Reserve Bank • Susan Guthrie – Policy Advisor, Reserve Bank • Ian Woolford – Manager Financial System Policy, Reserve Bank • Travis Wallach (facilitator) – Continuum Consulting Group • Jessica Wilson – Head of Research, Consumer NZ • Paul Barber – Policy Advisor, NZ Council of Christian Social Services • Rebecca Adams – Child Poverty Action Group • David Hanna – CEO, Wesley Community Action • Dr Prudence Stone – Public Health Association • Dr Louise Delaney – Otago University, The Public Health Association • Joanne Reid – Policy Advisor, Age Concern New Zealand Focus Group Agenda Overview The focus group began with a welcome and introduction from Simone Robbers, Assistant Governor and Manager of Governance at The Reserve Bank. The group representatives then each introduced themselves and outlined their key interests in the capital requirements proposal. This introduction period was followed by an overview and general discussion of the capital requirements proposal lead by The Reserve Bank’s Susan Guthrie and Ian Woolford. Following the capital requirements overview discussion The Reserve Bank’s officials left the room to allow for a facilitated open discussion with the social NGO representatives around their support for the proposal and any further comments, concerns, or recommendations they wished to make. The session concluded with Susan Guthrie from The Reserve Bank coming back into the room to hear feedback from the group.

2 Key Supported Outcomes The consensus from the socially engaged NGO focus group representatives was overall support for The Reserve Bank’s proposal to increase NZ bank capital requirements to order to provide greater banking stability and reduce possible future banking crises. There was overall recognition from the group that the catastrophic impacts of a potential banking crises, including reduced access to funds, increased job loses, business failures, reduced family and home security, as well as negative impacts to societal health and well-being; justify an increase in the capital requirements for NZ banks. The group did note that a small rise in interest rates may occur as a result of increasing NZ bank capital requirements, if banks decide to pass on this increase to their customers. This interest rate increase may impact those members of society who are already vulnerable to financial risk. While the group did highlight this impact, the majority of the group believed that the overall benefit to the stability of NZ society of increasing banking capital requirements (mentioned above) is of greater importance. The group did however provide some additional recommendations for minimising negative impacts that may result from interest rate rises. Additional Recommendations In addition to the above supported outcomes the NGO representatives provided some further recommendations to support the capital requirements proposal. The first additional recommendation is that a more integrated approach to policy formation is needed. The group suggested that increasing stability, needs to be looked at and communicated through the lens of a broader systematic policy approach. This broader systematic policy approach needs to include related impacted areas such as social welfare, business and employment, and health. Viewing policy change, such as increasing bank capital requirements, through a singular lens may not optimise the supportive benefits of a more integrated ‘all of government’ approach. To build on the first recommendation the group would like to see a risk assessment completed on the possible impact to vulnerable members of society that may feel any negative consequences resulting from a potential interest rate increase. The group would like to see The Reserve Bank working with the various government agencies to first identify and then help reduce the risk to vulnerable groups. Finally the group recommended that greater policy engagement and communication with different national and local community groups is needed to strengthen the narrative around the broader benefits of increasing bank capital requirements. The group is concerned that the current narrative, mainly controlled by the banks, is focused on the negative impacts of an interest rate rise to everyday New Zealanders; but largely discounts the broader societal benefits of greater stability. Broader, clearer and more positive communications, around the benefits are required; that could include more regular policy focus groups, community hall events, and greater direct engagement with key impacted parties.

• End of Report -

Bank capital hui: 5 June 2019, Holiday Inn Auckland Airport, 5th June 2019 from 9.30am to midday. Report prepared by Susan Guthrie (RBNZ) Participant organisations: Te Pou Matakana (Whanau Ora Commissioning Agency) Te Whānau o Waipareira Trust Nga Tangata Microfinance Muaupoko Tribal Authority, Levin Ngāti Whātua Ōrākei Participating from the Reserve Bank: o Adrian Orr o Ngarangi Haerewa o Ian Woolford o Susan Guthrie Background In December 2018 the Reserve Bank released a policy for public consultation that includes requiring bank owners to invest more equity in their New Zealand banks. The purpose of the hui was to seek feedback about the bank capital proposals, with attendees looking at the pros and cons of the proposal through the values, aspirations and experiences of the communities they represent. The purpose of this report is to provide a brief record of the issues that were discussed during the hui and the messages participants provided to the Reserve Bank for consideration. What matters

1. There was a broad discussion around Māori values and how Māori communities take a long term view, focusing on the intergenerational impacts of policies.

2. There was a discussion about the values underpinning the Reserve Bank’s vision (as expressed in Tane Mahuta) – placing a stronger focus on kaitiakitanga was endorsed by all participants.

3. The attendees discussed what matters in people’s lives. There was a consensus that this should be thought of in multidimensional terms that include financial success, home￾ownership, educational & health outcomes, and community connectedness (for example, access to transport options, facilities, appropriate services and other opportunities to strengthen communities).

4. Attendees said we need to think about gathering qualitative info about the trade-off between safety and cost, not just quantitative (what matters is more than just GDP, but multi-dimensional, e.g. financial inclusion, well-being, home ownership).

5. There was a discussion around the impacts of unemployment at the individual level as well on whānau and wider community wellbeing (for example, on self-esteem and mental health) and how this translates into enduring inter-generational stress at a society level (experiences in NZ’s deepest recession - 1991/92 – were discussed). There was a discussion of where the factual evidence for these impacts might lie.

6. The view of the group was that this policy is extremely important for the communities represented at the hui. These include disenfranchised youth, a growing proportion of the elderly, Māori, Pacific Island and other migrant groups, as well as other vulnerable sectors of society who face structural disadvantages which limit resilience to economic shocks.

7. There was a discussion of the importance of breaking down barriers to financial inclusion, giving people more financial options (eg not confined to pay day lenders) and fostering a savings culture. The impacts of financial crises

8. There was a discussion of the interaction between bank failure, demands on government funding and austerity programmes. The scale of support provided by the NZ government guarantee of deposits during the GFC was considered in the context of how governments would have had to respond if this guarantee was called on. The government guarantee covered $100bn in deposits, compared to present day public debt of $60bn. Overseas when governments have faced large unexpected expenditures related to bank failure they have responded by instituting austerity programmes that dramatically impact social services at a time when need for greater support has sharply increased.

9. There was general endorsement for protecting the most vulnerable communities from the fallout of bank failures, but attendees said the impact on vulnerable people of the protection itself must be considered – will they pay more on their borrowings, and most importantly will they have less access to cheap borrowing options than they did before?

10. There was a more general discussion about the difficulties vulnerable people face getting access to mainstream finance – they are often excluded and this led to pay day borrowing at crippling interest rates. The impacts of higher bank capital

11. The issue of who would bear the costs for the additional protection was discussed at length.

12. Attendees wanted to know what the effect of the proposal would be on third tier lenders – this sector is funded by banks, will costs rise and be passed onto vulnerable borrowers? Will lenders be even less willing to lend to vulnerable borrowers?

13. The potential impact on home ownership, small businesses and employment was discussed. Attendees asked the Reserve Bank to take great care to consider these impacts in advance.

14. Attendees said that there are many in the community who don’t have access to affordable borrowing options, and recent policies (e.g. the failure of the Credit Contracts Act reforms to introduce a ceiling on interest rates) aren’t helping. Attendees asked how does this policy interacts with other reforms that are occurring.

15. The group said the Reserve Bank must exercise great care to make sure the vulnerable are not worse off as a result of changes to interest rates and access to credit following the increase in bank capital.

16. There was support for joined up policy across the government sector. It would be important, for example, for MSD to be aware of what we are doing and the possible impacts.

17. Attendees pointed to the importance of deposit interest for many people and said they would be concerned if deposit rates fell further.

18. There was a discussion of whether bank capital instruments could be useful for people relying on income from savings.

Bank culture, conduct and competition 19. Attendees said there appear to be few consequences arising from poor conduct, and discussed whether it could be time for a different approach. 20. Attendees said there is a changing sentiment amongst the general public and this has prompted customers to think more deeply about who they bank with, to consider the effect of their bank on them and others. 21. Attendees questioned the basis for the level of very high profits earned by the four large banks. 22. Attendees said having disrupter institutions that challenge the dominant players is very important (by this they meant not only price competition but institutions that have a more socially responsible attitude too). Attendees discussed whether we had one disrupter bank already and whether or not it was having any impact. 23. There was a discussion of the time and support needed to foster the growth of market disrupters. 24. Attendees said small cooperative banking institutions are important and very valuable, the cooperative/sharing values underpinning these institutions align with Māori values, but these institutions need increased support as well as time to grow. 25. Attendees said a bigger gap is needed between what the Reserve Bank proposes for the big four versus the smaller, domestic banks. Attendees were of the view that domestic banks should be favoured in some way in the capital regime, so that they can grow and compete. 26. Attendees asked who is fixing the lack of financial market options in NZ – what is happening to better match small businesses with equity not simply bank debt? Bank behaviour 27. Attendees said banks are in a social contract with the rest of NZ, they have a social licence to operate. There was a view that the language around banking needs to change in New Zealand to reflect this. 28. Banks need to consider the social good as well as profit – there needs to be a rebalancing. Attendees asked “What incentive is there for banks to change their behaviour – what is going to make them be more socially responsible?” 29. Attendees said their communities would value information about how banks make loans and other windows into how banks behave. Māori communities are more inclined to assess banks based on bank behaviour – on how socially responsible they are – not just simply on price. 30. There was a view that front line staff do not always know or understand what is expected of them. If more information was available publicly about bank behaviour and what is expected, it would be easier for frontline staff to meet society’s expectations. Messages for the RBNZ from participants Attendees expressed support for: o Levelling the playing field, saying that in fact the Reserve Bank should go further in terms of oversight and regulation o Greater transparency around how banks behave o Recognising the importance of disrupters in the market (i.e. banks with new models) and what they need to thrive

o Increasing protection from financial crises, but be aware of and very careful of the impacts of higher capital. o Being joined up with other government agencies so any harm caused by higher capital is not borne by the most vulnerable and as is as small as possible in general. Attendees did not support: o The risk weights assigned to social housing being the same as those for speculative property development. This is preventing the expansion of social, cooperative and community housing, as well as emergent pathways to home ownership such as shared equity, intergenerational borrowing and building on land in multiple ownership. Attendees expressed a desire to see a parallel non speculative market set up for social housing that is focussed on community ownership and the positive externalities which can be achieved by framing access to good quality housing as a human right.

MEMORANDUM FOR FSC FROM Financial Policy (Author: Charles Lilly) MEETING DATE 3 July 2019 SUBJECT Recent external developments on bank capital FOR YOUR Information Purpose

1. This paper provides the Committee with a brief update on recent external developments that are relevant to our Capital Review: a. Findings of the 2019 OECD Economic Survey of New Zealand. b. The concluding statement of the 2019 IMF Article IV Consultation. c. An update to the Basel Committee’s 2010 assessment of the costs and benefits of bank capital. d. A research paper from the Bank of England on the impact of banks’ funding mixture on their funding cost.
2. The key messages we want to convey to the Committee are as follows: a. Both the OECD and the IMF are supportive of our proposed direction for capital requirements, with the IMF endorsing the proposed level for DSIBs as commensurate with New Zealand’s risk profile. The OECD is more cautious on the level, saying the end point should be guided by our CBA. b. Recent research papers from the Basel Committee Research Task Force and the Bank of England appear to be supportive of our analysis on the net benefits of higher capital: the benefits of capital exceed the costs over a wide range of calibrations (higher than the Basel Committee estimated in 2010), and empirical evidence shows investors (debt and equity) will adjust their return expectations downwards in response to more capital. 2019 OECD Economic Survey of New Zealand
3. On 25 June the OECD released its most recent Economic Survey of New Zealand, which commented on the Capital Review among other issues.1 The OECD concludes that higher capital is likely to have net benefits but does not quite endorse our proposed end point (16% Tier 1 for DSIBs). Instead, the OECD says that the end point should be guided by our cost benefit analysis: The Reserve Bank in late 2018 proposed a large increase in Tier 1 capital requirements for banks to reduce the risk of failure in a crisis. While international comparability is complicated by differences in asset risk weightings used and different economic context, this would take Tier 1 capital requirements beyond those applying in other OECD countries. The increase in actual capital held will not be as large as the increase in regulatory capital requirements insofar as banks choose to hold smaller discretionary buffers. Higher capital requirements reduce the probability of banks becoming insolvent, lowering the likelihood and expected costs of banking crises. They can also reduce economic activity by increasing lending rates due to a higher cost of funding for banks (albeit offset by the reduced risk of bank equity). The Bank has set its proposed capital requirements by first limiting the probability of a

1 http://www.oecd.org/economy/new-zealand-economic-snapshot/

2 financial crisis to 1 in 200 years, then considering the trade-off between economic activity and the expected costs from financial crises for any level of capital beyond this. There is a strong case for higher capital requirements, but considerable uncertainty around the appropriate end point. As New Zealand is a small open economy, the output cost of financial crises is likely to be greater than in larger countries. High levels of household debt also exacerbate risks. Such local contextual factors have been incorporated in quantitative analysis by the RBNZ, which shows substantial uncertainty around “optimal” capital ratios but clear gains from increasing capital to around 12 to 13% of risk-weighted assets (7.5% to 8% of unweighted assets). The Bank’s choice of a 1 in 200 year threshold drives its specific proposal for capital requirements, which is well above what it had previously advocated. The Bank, which completed public consultation on its proposal in May 2019, should proceed with higher capital requirements as warranted by its forthcoming cost-benefit analysis. The effect on lending spreads, bank credit availability and credit activity pushed outside the banking sector (to credit unions, for example) needs to be carefully monitored. 2019 IMF Article IV Concluding Statement 4. The IMF published its Staff Concluding Statement for the Article IV mission on 25 June 2019.2 Financial Policy presented on the Capital Review at the session on Financial Stability, and the IMF team showed a keen interest. The concluding statement effectively endorses our proposed calibration as being commensurate with New Zealand’s risk profile, but suggests that a larger DSIB buffer is warranted (meaning 16% Tier 1 for DSIBs but a somewhat lower requirement for smaller banks than the 15% we have proposed) and that the transition path should be gradual: The proposed increase in bank capital adequacy requirements would provide for a welcome increase in the resilience of the banking system. The new requirements would increase bank capital to levels that are commensurate with the systemic financial risks emanating from the dominance of the four large banks with similar concentrated exposure to mortgages, business models and funding structures. The new buffer for domestic systemically important banks (D-SIB buffer) could be higher, with a corresponding decrease in the capital conservation buffer for all banks. The phase-in period should be sufficiently flexible to ease the transition. A stronger bank supervision regime would still be needed, to complement the higher capital requirements. Basel Committee on Banking Supervision research paper: update on the costs and benefits of bank capital 5. Researchers for the Basel Committee on Banking Supervision (BCBS) published an update to the Committee’s 2010 assessment of the long term impact of stronger capital requirements on 24 June 2019.3 The BCBS’s 2010 paper formed part of the evidence base for the calibration of Basel III, as well as our own analysis for the ratio paper. 6. The 2019 study does not undertake new research to any great extent; rather, it considers the earlier BCBS work in light of the range of studies by academics and central banks on optimal bank capital that have emerged in the subsequent nine years. Given the range of methodologies and design choices in estimating the costs and benefits of bank capital, and its optimal setting, the authors do not attempt to fully harmonise and/or reconcile each paper they surveyed with the original 2010 work. Rather, they work through each of the components of an optimal capital study (impact of capital on the probability and cost of crisis, impact of capital on interest

2 https://www.imf.org/en/News/Articles/2019/06/25/mcs062519-new-zealand-staff-concluding￾statement-of-the-2019-article-iv-consultation-mission 3 “The costs and benefits of bank capital – a review of the literature”, https://www.bis.org/bcbs/publ/wp37.htm.

3 rates and economic activity etc.) to assess the reasonableness of the findings of the 2010 paper. 7. While cautious about making direct comparisons in light of the methodological differences between studies, the authors conclude that the findings of the 2010 study (an optimal capital ratio in the range of 10-15%4 ) are likely to be at the lower end of optimal capital (i.e. optimal capital is probably higher): Consistent with the Basel Committee’s original assessment, this paper finds that the net macroeconomic benefits of capital requirements are positive over a wide range of capital levels. Under certain assumptions, the literature finds that the net benefits of higher capital requirements may have been understated in the original Committee assessment. Put differently, the range of estimates for the theoretically-optimal level of capital requirements – where marginal benefits equal marginal costs – is likely either similar or higher than was originally estimated by the Basel Committee. Optimal capital studies completed after BCBS (2010) generally estimate [optimal capital] higher than 10% (the lower bound estimate for BCBS (2010)). This finding does not appear to be the result of just one design choice, or one component of net marginal benefit. … Overall, the range of estimates for the theoretically-optimal level of capital requirements – where marginal benefits equal marginal costs – is likely either similar or higher than was originally estimated by the Basel Committee. As such, subsequent optimal capital studies confirm the finding in BCBS (2010) that the benefits exceed the costs of capital over a wide range of capital ratios. Bank of England working paper: Bank funding costs and capital structure 8. Researchers at the Bank of England recently published an empirical study of the implications of different mixes of bank funding structures (equity, subordinated, senior unsecured and senior secured) on banks’ funding costs.5 The study uses data from 2005q1 to 2017q4 on a panel of large (>£15b assets) banks across a number of countries (3051 observations). 9. This paper is most useful for us in providing further robust, empirical evidence on the Modigliani-Miller effects of banks’ funding structure – that is, the extent to which having more equity in a bank’s balance sheet (as we are proposing) reduces the risk premium associated with that bank’s funding, negating the impact of changes in banks’ capital ratios on weighted average cost of funding and the lending rates they subsequently charge their borrowers. 10. We are happy to provide more technical detail on the paper’s findings if Committee members would like, but our key takeouts from the paper as follows: a. Under the authors’ baseline specification, a bank’s equity risk premium falls by 24 basis points (and subordinated debt risk premium by 154 basis points) when a bank employs 1 percentage point more equity and 1 percentage point less subordinated debt, as a % of assets. b. 1 percentage point more equity and 1 percentage point less subordinated debt (as a % of assets) increases the weighted average cost of funding by 1.65 basis points in the baseline specification. This suggests our proposed increase in high quality capital (c. 3-5% in terms of assets) would only need lending rates to increase by perhaps 5-8 basis points to compensate banks’ owners for their higher equity stake.

4 10-15% is based on Tangible Common Equity as the numerator and RWA calculated using Basel II rules as the denominator. Tangible Common Equity is roughly comparable to CET1 capital. 5 https://www.bankofengland.co.uk/working-paper/2019/bank-funding-costs-and-capital-structure

4 c. Banks’ submissions on the ratio paper have disputed our assumptions about the extent to which the MM effect would hold in New Zealand. This is based on an view that, because the New Zealand major banks’ debt is priced based on the credit rating of the parent, with no change in the parent’s rating they would see no benefit (in terms of debt funding cost) from operating with more equity. The results of this paper however show that the bulk (70%) of the MM effect in their specification is driven by a reduced equity risk premium; the empirical evidence shows that bank shareholders are happy to accept a lower return if that return is less volatile (as would be the case with higher capital levels). Therefore, even if we were to accept that banks’ debt funding costs wouldn’t change as a result of our proposals, a substantial MM effect can still be justified based on equity risk premium effects alone. This supports our case that the impact of the proposals on borrowers’ interest rates are likely to be modest, and less than suggested by some banks’ analyses.

www.thinkSapere.com i Contents

1. Introduction..........................................................................................................................................................................1
2. Australian evidence of impact on bank funding costs ........................................................................................3 2.1 Further illustration from updating parameter estimates in Professor Mile’s study for New Zealand conditions ........................................................................................................................................6
3. A comment on tax rules...................................................................................................................................................8 3.1 Imputation credits .......................................................................................................................................8 3.2 Holding companies and thin capital rules .........................................................................................9
4. References ...........................................................................................................................................................................12 About Us ....................................................................................................................................................................13

www.thinkSapere.com 1

1. Introduction
2. During a discussion on 26 June 2019,1 Dr James Cummings provided a copy of a recent study he undertook with Linh Nguyen. That study estimated the impact on bank funding costs in Australia of the Basel III capital reforms (Cummings & Nguyen, 2019). Dr Cummings queried whether his empirical work may have implications for our analysis of the Reserve Bank proposals to increase the amount of capital required to be held by New Zealand banks. Our report (Scott, Boyle, Lubberink, & Murray, 2019), was attached to the New Zealand Bankers Association submission to the Reserve Bank on its proposed changes.2 We estimated that the economic costs of the Reserve Bank’s proposals would exceed the economic benefits by about $1.8 billion per annum.
3. Dr Cummings observed that his estimate of the impact of additional capital on bank funding costs in Australia was much lower, for each additional percentage point of capital, than the increase in the cost of bank credit assumed in our analysis. Our report had taken, unamended, the Reserve Bank’s central estimate of an 8.1 basis point increase in the cost of bank credit for each percentage point increase in the ratio of Tier 1 capital to unweighted assets (Scott et al, p. 19). Dr Cummings referred to his estimate that each additional percentage point in the ratio of capital to risk-adjusted assets had raised funding costs in Australia by about 2 basis points, and that his result was similar to that predicted by Miles, Yang, & Marchegiano (2012). 3 We show that these estimates are not directly comparable.
4. Understanding the differences in these estimates is important for any assessment of the costs of the Reserve Bank proposals. An increase in the capital required to be held by the banks will raise real interest rates, leading to higher costs to borrowers and reduced economic activity. Clearly, if real interest rates would rise by significantly less than estimated by the Reserve Bank, that would substantially reduce the estimated economic costs of the proposals.
5. In the analysis that follows, we show that if Dr Cummings’ methodology were applied to the Reserve Bank’s proposals, with New Zealand (rather than Australian) parameter inputs, it would produce a similar estimate to the Reserve Bank’s central estimate—that is, an increase in the cost of bank credit of about 8 basis points for each additional percentage point in capital. This

1 Dr Cummings was meeting with Roger Beaumont and Anthony Buick-Constable from the New Zealand Bankers Association, and Graham Scott and Kieran Murray from Sapere, to discuss submissions to the Reserve Bank on its Tier 1 capital proposals. Dr Cummings is one of three experts commissioned by the Reserve Bank to review its analysis underpinning its proposals to increase bank capital. The terms of reference for this review are available here: https://www.rbnz.govt.nz/-/media/ReserveBank/Files/regulation-and-supervision/banks/capital￾review/Terms-of-Reference-Capital-Review-External-Experts.pdf?revision=b1121938-7649-449e-af7e￾fa26958197ef&la=en 2 A copy of the New Zealand Bankers Association submission and the Scott, G., et al, report are available here: https://www.rbnz.govt.nz/-/media/ReserveBank/Files/regulation-and-supervision/banks/capital￾review/submissions/N-to-O-Capital-Ratio.pdf#page=6 3 Professor David Miles has also been appointed as one of three experts to review the Reserve Bank proposals.

2 www.thinkSapere.com finding reinforces a central theme of our report—that it is dangerous for the Reserve Bank to be overly reliant on estimates of impacts obtained from overseas studies. The Reserve Bank should assess the costs and benefits, as would be incurred, were the proposals to be applied in New Zealand. 5. As a further illustration of the importance of understanding local conditions, Dr Cummings observed in his study that although he had found the Modigliani-Miller offset to be weaker than reported by Miles et al, 4 the cost of reform was reduced in Australia because of its imputation tax arrangements (Cummings & Nguyen, p. 31). The Modigliani-Miller (MM) offset refers to the extent to which an increase in bank costs resulting from an increase in equity capital would be offset by a reduction in risk. As we explain below, the New Zealand imputation tax provisions would not materially reduce the cost of increasing bank capital in New Zealand, and the existing ‘thin capital rules’ would mean that little additional tax would be paid by the banking sector overall, though the tax effects would impact more on New Zealand owned banks than overseas owned banks. 6. Taking account of the local tax laws, as Dr Cummings did for his Australian estimates, reinforces our conclusion that the additional cost to New Zealand borrowers of the Reserve Bank proposals—of about $1.6 billion per annum—would largely be an economic welfare loss; there would be no material transfer, in the form of additional tax payments, to the government. There would however be different tax effects for different organisational forms—as we observed in our report, the Reserve Bank does not evaluate whether its proposals would tilt the field for competition, and if so what effect that might have on the efficiency of the banking services offered to New Zealand.

4 Some care is needed in interpreting the Cummings & Nguyen claim (pp. 3, 31) that their estimate of the Modigliani and Miller offset of 25% is “weaker" than the effect estimated by Miles et al. The Miles et al, estimate of the offset of 45% - 75% is expressed in terms of impact on the Weighted Average Cost of Capital (WACC), whereas the Cummings & Nguyen estimate is expressed in terms of the equity risk premium.

www.thinkSapere.com 3 2. Australian evidence of impact on bank funding costs 7. Cummings & Nguyen examine the long-run cost implications for Australian banks that increased their equity capital to meet the requirements of the Basel III reform package. They proceed as follows: • Using Australian data, they estimate the model:5 ERPit = α + βLi,t-1 + ϵit (1) where ERPit is the estimated (using one of two asset pricing models) equity risk premium for bank i at date t and Lit is the ratio of total assets to equity capital for bank i at date t. • The idea underlying the above equation is that, according to the Modigliani and Miller (1958) capital structure irrelevance theorems, the equity risk premium is directly proportional to L; that is, a doubling of L results in a doubling of ERP. Estimating equations such as (1) therefore enable an estimate of the empirical MM offset. • Using various data samples and different estimates of ERP, the results from estimating equation (1) are reported in Cummings & Nguyen, Tables 2 and 3. Based on the results appearing in column 1 of Table 2, they conclude the MM offset is approximately 25%; that is, about 1/4 of the full MM effect operates in practice for Australian banks. • Cummings & Nguyen use this result to estimate that a 5 percentage point increase in capital requirements (defined as CET1 capital as a percentage of risk-adjusted assets; equivalently, given their assumptions, a 2.2 percentage point increase in CET1 capital as a percentage of unweighted assets) would increase the WACC, for a bank with average risk-weighted assets and average book-to-market ratio, by 10.1 basis points (their Table 4). 8. A superfcial reading might suggest that the result obtained by Cummings & Nguyen conflicts with the approximately 35 basis points increase in the cost of bank credit cost we used in estimating the economic cost of the Reserve Bank proposals (Scott et al, Appendix B). However, such a conclusion would be based on an apples-and-oranges comparison—the Cummings & Nguyen analysis differs from the New Zealand situation we analysed in several fundamental ways:

5 Clyne (2015) estimates a similar model but uses a different dependent variable (cost of equity rather than equity risk premium) and a different measure of leverage (ratio of debt capital to equity capital rather than ratio of total assets to equity capital).

4 www.thinkSapere.com i. Cummings & Nguyen assume as their counter-factual (where the Basel III accords are assumed to be reversed), an increase in L from 9.2 to 11.5. Obviously, this corresponds to a fall in the ratio of equity capital to total assets from 0.109 (1/9.2) to 0.087 (1/11.5); that is, a fall of 2.2 percentage points. However, this is not the same as the Reserve Bank’s proposals: as shown in Appendix A of our report, the New Zealand-specific counter-factual would entail a change of 4.3 percentage points. ii. Cummings & Nguyen assume an “imputation gamma" equal to 0.35, based on an earlier estimate by Cummings and Wright (2016). However, due to the non￾transferability of imputation credits across countries, this number, if correct for Australian banks, will be too high for New Zealand banks. An approximate (and very conservative, due to the home bias phenomenon) adjustment is to multiply 0.35 by the ratio of New Zealand and Australian populations, which is slightly less than 1/5, which in turn implies γ = 0.07.6 We elaborate on this adjustment in discussing tax considerations further below. iii. The New Zealand corporate tax rate is 28 per cent—Cummings & Nguyen assume the Australian corporate rate of 30 per cent. iv. In generating their Table 4 numbers, Cummings & Nguyen use only the MM offset results from Table 2, Panel A, column (1) of their paper, which represents just one of the 20 models they report in their Tables 2 and 3. It also happens to be one of the highest-offset models; that is, most of the other models imply a smaller offset (and hence a bigger WACC increase). v. Finally, Cummings & Nguyen estimate the change in WACC which is not the same as the Reserve Bank estimate of change in bank credit costs, as the Reserve Bank assumes that the higher bank funding costs are recovered only from loans and not total assets; the Reserve Bank arrived at its estimate of an 8.1 basis point increase in the cost of bank credit by dividing its estimate of the increase in WACC of 6.5 basis points by the ratio of loans to total assets (0.065/0.8=0.81). 7 Hence, for an increase in capital requirements of 4.3 percentage points, the increase in WACC would be (4.3 x 6.5) 28 basis points (leading to a 35 basis point increase in the cost of bank credit). 9. The differences between the Cummings & Nguyen calibration for Australian banks and the appropriate calibration for New Zealand are summarised in Table 1. 8

6 This adjustment can be justified by noting that, in a frictionless world, the weight of Australian banks in the average Australian optimal portfolio would be the same as their weight in the average New Zealander’s optimal portfolio. As there are five times as many Australian investors, it follows that total Australian investment in Australian banks should be about five times that of New Zealand. 7 These assumptions are set out in the Reserve Bank Capital Review Background Paper: An outline of the analysis supporting the risk appetite framework, April 2019, pages 36-37. 8 For New Zealand post-L, Scott et al Appendix A implies 1/L = 32.5/565 = 0.0575, and so post-L = 1/0.0575 = 17.39; for New Zealand pre-L, the corresponding calculation is 1/L = (55.5 + 58.3)/2(565) = 0.1007 and so pre-L = 1/0.1007 = 9.93.

6 www.thinkSapere.com effect of Basel III on Australian banks, but applying their methodology to New Zealand conditions and the Reserve Bank proposals provides quite different estimates for New Zealand banks. Updating the Cummings & Nguyen, methodology for New Zealand conditions arrives at a cost estimate range that supports the estimates of the costs of the Reserve Bank proposals provided in our report. 2.1 Further illustration from updating parameter estimates in Professor Miles’ study for New Zealand conditions 13. In a widely cited study, Miles et al, estimated the long-run costs and benefits of having banks fund more of their assets with equity. They claim that (Miles et al, p. 1): The amount of equity capital that is likely to be desirable for banks to use is very much larger than banks have used in recent years and also higher than targets agreed under the Basel III framework. 14. In arriving at this conclusion, Miles et al, compared estimates of the expected benefits from increased capital (that is, a reduced chance of banking crises) with estimates of the expected economic costs (from higher interest rates). Reading the Miles et al, study in the light of the Reserve Bank proposals provides a further illustration of the importance of understanding local conditions when interpreting results from overseas studies. 15. For example, Miles et al estimate an increase in the cost of bank funding of 18 basis points if capital relative to assets were to double for the United Kingdom banks in their sample, resulting in a leverage ratio—defined as assets to CET1—falling from 50 to 25. In New Zealand, the equivalent leverage ratio is currently about 17, and would fall to under 10 should the Reserve Bank proposals proceed—that is, the New Zealand economy is already positioned at a very different point on the cost benefit curves than the position considered by Miles et al.10

16. In estimating the impact of higher bank funding costs on economic activity, Miles et al, assume that firms rely for about 1/3rd of their funding on bank loans, and hence the overall cost of capital would increase by about 6 basis points (for their assumed increase in capital requirements). At the time of writing their article, Miles et al determined that it was reasonable to assume the initital cost of capital to firms was 10 per cent. Hence, a 6 basis point increase would represent an increase of 0.6 per cent (0.06%/10%) increase in the cost of capital. In the current New Zealand environment, the cost of capital to firms is likely closer to 5 per cent (in our report, we adopted the New Zealand Treasury discount rate of 6 per cent). With a 5 per cent cost of capital, a 6 basis point increase would represent a 1.2 per cent (0.06%/5%) increase

10 For comparison, the leverage of New Zealand banks measured on the same basis of CET1 to total assets is approximately 17 (565 bn / 32.5 bn) and would reduce to about 9.7 (565bn / 58.3 bn) under the Reserve Bank proposals, assuming that the banks hold a buffer equivalent to 2 percentage points of risk weighted assets. See Scott et al, Appendix A.

www.thinkSapere.com 7 in the cost of capital—in short, the cost curve would be about twice as steep as estimated by Miles et al. 17. In Appendix C of Scott et al, we applied the Miles et al methodology for estimating the cost of an increase in bank capital requirements to New Zealand conditions. We show that applying New Zealand specific parameters produce much higher estimates of the reduction in steady￾state GDP from a permanent increase in real interest rates than the estimate of 8.8 basis points assumed by the Reserve Bank (the Reserve Bank did not make its own estimates, but had uplifed the 8.8 basis points estimate from a United States Federal Reserve study of the impact of increases in bank funding costs on the United States economy—an economy with significant structural differences to the New Zealand economy). 18. Using alternative scenarios applicable to New Zealand conditions, we estimate that the loss in steady state GDP for each percentage point increase in capital ranges from 17 to 40 basis points (Scott et al, Table 3, Appendix C). The range of estimates arrived at from applying New Zealand parameters, and their sensitivity to the dataset they are estimated from indicate these estimates are subject to considerable uncertainty. However, all of the estimates derived from New Zealand parameters are significiantly higher than—at least double—the values assumed by the Reserve Bank from applying elasticities from overseas studies with no recognition of New Zealand conditions.

8 www.thinkSapere.com 3. A comment on tax rules 19. The effect of tax rules must always be a consideration in policies that are intended to alter financial arrangements. This is because taxes can drive a wedge between amounts paid and amounts received, and hence alter incentives and organisational form. Tax considerations are especially relevant to estimating the economic costs and benefits of the Reserve Bank proposals to increase capital ratios because: • the Modigliani-Miller theorem of capital structure irrelevance assumes an absence of tax, and hence any real-world assessment of likely costs of a regulatory change to leverage ratios needs to account for the specifics of tax policy • the prospect of additional tax payments to the government may mean that some of the additional costs of borrowing reflect a transfer payment and not a welfare loss—Cummings & Wright p 50, suggest that in principle the extra tax receipts could be used to offset the impact on the wider economy of an increase in bank funding costs • the interaction of the change in regulatory requirements with tax laws may create incentives for firms to adopt different organisational forms and business models. 20. The Reserve Bank did not consider the effects of tax in its consultation document discussings its proposal to increase capital requirements. However, in our discussion Dr Cummings highlighted the importance of tax effects. In the comments below, we elaborate on the adjustments necessary to the Cummings & Nguyen methodology to reflect the influence of New Zealand imputation credits. We also explain why little additional tax would be paid by the banking sector overall, and hence why the argument that the additional amounts paid by borrowers is a transfer is misconceived. 3.1 Imputation credits 21. In most countries, taxation policy makes equity financing more expensive than debt financing as firms can deduct interest payments as an expense to set against their corporate tax payments. Both Australia and New Zealand tax regimes operate an imputation system intended to mitigate the double taxation of corporate profits by allowing the company to transfer a tax credit (in Australia, called a franking credit, in New Zealand, an imputation credit) to the shareholder for tax paid by the company. 22. Cummings & Nguyen estimate the proportion of company tax paid by Australian banks that will be rebated against personal income in the hands of Australian shareholders as the product of (i) the portion of after tax profits typically paid out as dividends (70%) and (ii) an estimate of the

www.thinkSapere.com 9 value to shareholders of franking credits of 50%; so: 70% x 50% = 35%.11 This estimate is referred to by Cummings & Nguyen as the “imputation gamma". 23. However, New Zealand dividend imputation credits can only be utilised by New Zealand residents and have no value to foreign residents. The imputation rules also limit the maximum imputation ratio to 28 cents for each dollar of gross dividend. Hence, two adjustments are necessary to adapt the imputation gamma for dividends paid by foreign parent companies to New Zealand residents: 12 • an adjustment to reflect the proportion of the parent company dividends that would be derived from the New Zealand operations, and therefore eligible to have imputation credits attached—for example, as the New Zealand economy is about 15 per cent of the Australian economy, an indicative assumption might be that the New Zealand operations contribute about 15 per cent of the profit to Australian parent entities • an adjustment to reflect the portion of shareholders of Australian banks that are New Zealand resident and therefore able to utilise the imputation credit—as noted above, the ratio of the New Zealand to the Australian population may provide a reasonable assumption, which is slightly less than 1/5th . 24. On these assumptions, the adjusted formula would be: 70% x 50% x 15% x 20% = 1%. In the analysis presented above, we made only one adjustment—to reflect the portion of shareholders of Australian banks that are New Zealand resident. Because the WACC impact is a decreasing and concave function of gamma, making the further adjustment would not have much effect on the estimate of WACC. For example, in the Cummings & Nguyen base case, the WACC effect is 10 basis points when γ = 0.35, 24 basis points when γ = 0.07, and 25 basis points when γ = 0.01. 25. Imputation credits are also unlikely to be used by New Zealand banks owned under a cooperative structure. Cooperative ventures are less likely to use imputation credits as it is more tax efficient for a cooperative to return any surplus via reductions in fees etc, than to pay tax, declare a dividend and attach imputation credits (as the individual would pay the fees using after tax income). 3.2 Holding companies and thin capital rules 26. In considering the tax implications of the Reserve Bank proposals, it is important to understand that the Reserve Bank proposals apply to the operating entity of New Zealand banks. These operating entities are (or can be structured as) a subsidiary of a New Zealand holding company,

11 Australian franking credits can be cashed out by non-tax payers; New Zealand imputation credits cannot be cashed out and hence would likely have a lower value, even if all else were the same. 12 The New Zealand Inland Revenue Department provides a fuller explanation of the rules for attaching imputation credits to dividends available at: https://www.ird.govt.nz/topics/income-tax/imputation.

10 www.thinkSapere.com that is in turn a subsidiary of an offshore parent (for New Zealand banks that are wholly owned subsidaries of offshore parents). 27. When assessing tax liabilities, the Inland Revenue Department looks at the whole New Zealand entity—that is, the operating entity and the holding company. Specific, ‘thin capital’, rules determine the extent to which interest is deductible to the New Zealand business of a foreign￾owned bank as part of calculating its New Zealand income for tax purposes. 13 The Reserve Bank proposals do not involve a change to the tax laws. Hence, the proposals would not change the amount of tax paid by the New Zealand businesses of foreign-owned banks.14

28. These tax considerations raise an obvious question: if the tax liabilities are independent of their actual capital structure, why does the Modigliani and Miller irrelevance theorem not hold exactly? That is, why is the MM offset not equal to 100% for the New Zealand businesses of foreign-owned banks? Several factors make this unlikely.
29. First, the corporate tax shield is not the only example of a ‘leakage’ that favours debt financing and hence invalidates the pure version of MM. A particularly salient example for banks is the presence of under-priced deposit insurance (the implicit subsidy because governments cannot credibly commit not to intervene to protect depositors in event of bank failure). This provides a subsidy on deposit-funded debt financing and hence allows banks to lower their WACC by substituting deposit financing for equity. Conversely, raising equity requirements increases WACC.
30. New Zealand is unusual in not having an explicit deposit insurance scheme. However, past experience suggests that an implicit scheme exists. At the onset of the Global Financial Crisis (GFC), before any overt signs of bank distress, the New Zealand government introduced a depositor protection scheme covering all retail deposits; New Zealand banks were also offered a loan guarantee plan. Thus, both deposit and non-deposit debt financing were fully insured for as long as these schemes remained in operation. Although participation in these schemes incurred charges, these had obviously been avoided in the years leading up to the GFC and so the protection was effectively under-priced.
31. Depositor and non-depositor sources of debt financing are likely to (rationally) anticipate similar protection in future crises. As a result, the cost to banks of debt from these sources is subsidised and there is an advantage to debt financing. Requiring greater equity (less debt) will reduce the amount of this subsidy and increase WACC.15

13 An Inland Revenue Department description of the purpose of the thin capital rules is available here (note, the specifics of the rules have since been amended): https://www.classic.ird.govt.nz/technical￾tax/legislation/2005/2005-79/leg-2005-79-policy-issues/thin-capitalisation/leg-2005-79-thin-capitalisation.html 14 As discussed above, an Australian entity would prefer to pay tax in Australian than New Zealand because franking credits have value to Australian resident shareholders whereas little benefit is obtained from New Zealand imputation credits. 15 The efficiency effects of reducing any implicit subsidy are unclear. The policy intervention would attempt to address the implicit subsidy to those who lend to the bank by raising the costs to those who borrow from the bank. It’s not obvious that this would lead to better outcomes, especially relative to other potential policies.

www.thinkSapere.com 11 32. Similarly, New Zealand is also likely to be seen by credit markets as operating a Too-Big-To-Fail (TBTF) policy. In 1990, the BNZ, at the time the largest bank in New Zealand, failed and had to be bailed out. No depositors lost money. It seems likely that political considerations would force similar action today, the Reserve Bank’s OBR policy notwithstanding. To the extent such behaviour is indeed anticipated, the big-4 banks receive an effective TBTF subsidy that provides an advantage to debt financing. Again, requiring greater equity (less debt) will reduce the amount of this subsidy and increase WACC. 33. Second, the MM irrelevance theorem is an arbitrage relationship that applies to frictionless markets. However, as Shleifer and Vishny (1997) eloquently point out, frictions that result in “limits to arbitrage” can mean that financial market prices in the real world behave quite differently to those in a frictionless world. 34. One example of such a phenomenon is the divergence between the rates at which banks and investors can borrow. The MM theorem, based as it is on the concept of “homemade leverage” requires these two rates to be the same. For many companies, this is not an overly-strong assumption, but it seems much less likely to hold for banks. Banks are firms whose business model involves charging investors who borrow from them a margin over what they themselves pay to borrow from depositors or wholesale markets. But if investors cannot borrow at the same rate as banks, then they will be willing to pay a premium to banks to do the borrowing for them; that is, the MM offset is less than 100%. 35. As Titman (2002) discusses, there are also many supply-side frictions that inhibit the arbitraging behaviour underlying MM. A further complicating factor in this case is the absence of traded claims on the New Zealand businesses of foreign-owned banks, which renders the replicating arguments in MM at least partly infeasible. Again, since investors cannot perfectly replicate the leverage decisions of the banks, they may be willing to pay a premium to have it done for them. 36. Of course, none of these eliminate the basic logic of MM: that cutting a pizza up into different￾sized slices has, in the absence of wastage, no effect on its overall size. But they do suggest that, even in the absence of tax implications, the MM offset is likely to be less than 100%. The relevant question is by how much. 37. Unfortunately, the Reserve Bank has not undertaken a study examining this issue for New Zealand, but other studies from around the world typically suggest the offset is likely to be considerably less than 100%. The Reserve Bank has calculated that the estimate implied by these studies is a 6.5 basis point increase in WACC for every percentage point increase in the ratio of capital to total assets. As we have seen above, nothing in the Australian evidence of Cummings and Nguyen contradicts that position 38. As explained in Scott et al pp 18-19, the outcome is that the Reserve Bank proposals would result in an additional cost of bank credit to New Zealand borrowers of about $1.6 billion per annum. There may be some tax effects—additional tax paid by banks wholly owned by New Zealand entities—but these effects will be relatively minor in terms of total quantum, though potentially significant for the individual banks affected. The Reserve Bank does not appear to have taken these direct economic costs of its proposals, or the potentially uneven impact of tax effects, into account.

12 www.thinkSapere.com 4. References Cline, W., 2015. Testing the Modigliani-Miller theorem of capital structure irrelevance for banks. Peterson Institute for International Economics WP 15-8. Cline, W., 2016. Benefits and Costs of Higher Capital Requirements for Banks. Peterson Institute for International Economics WP 16-6. Cummings, J. and L. Nguyen, 2019. Impact of the Basel III capital reforms on bank funding costs: Australian evidence. Cummings, J. and S. Wright, 2016. Effects of higher capital requirements on the funding costs of Australian banks. Australian Economic Review 49, 44-53. Miles, D., Yang, J., and Marcheggiano, G. (2012). Optimal bank capital. Economic Journal, March, pp. 1- 37. Modigliani, F., and Miller, M. (1958). The cost of Capital, Corporation Finance and the Theory of Investment. American Economic Review 48 (3): 261-297. Reserve Bank of New Zealand (2019). Capital Review Background Paper: An outline of the analysis supporting the risk appetite framework. Scott, G., G. Boyle, M. Lubberink, and K. Murray, 2019. How much capital is enough—a review of Reserve Bank Tier 1 capital proposals. Sapere Research Group Shleifer, A. and R. Vishny, 1997. The limits of arbitrage. Journal of Finance 52 (March): 35-55. Titman, S., 2002. The Modigliani and Miller theorem and the integration of financial markets. Financial Management 31 (Spring): 101-115.

© 2019 Sapere Research Group About us Sapere Research Group is one of the largest expert consulting firms in Australasia, and a leader in the provision of independent economic, forensic accounting and public policy services. We provide independent expert testimony, strategic advisory services, data analytics and other advice to Australasia’s private sector corporate clients, major law firms, government agencies, and regulatory bodies. ‘Sapere’ comes from Latin (to be wise) and the phrase ‘sapere aude’ (dare to be wise). The phrase is associated with German philosopher Immanuel Kant, who promoted the use of reason as a tool of thought; an approach that underpins all Sapere’s practice groups. For more information, please contact: Kieran Murray Phone: +64 915 7592 Mobile: + 64 21 245 1061 Email: kmurray@thinkSapere.com Wellington Auckland Sydney Melbourne Canberra Level 9 1 Willeston Street PO Box 587 Wellington 6140 Level 8 203 Queen Street PO Box 2475 Shortland Street Auckland 1140 Level 18 135 King Street Sydney NSW 2000 Level 2 161 Collins Street GPO Box 3179 Melbourne 3001 PO Box 252 Canberra City ACT 2601 P +64 4 915 7590 F +64 4 915 7596 P +64 9 909 5810 F +64 9 909 5828 P +61 2 9234 0200 F +61 2 9234 0201 P +61 3 9005 1454 F +61 2 9234 0201 (Syd) P +61 2 6100 6363 F +61 2 9234 0201 (Syd) www.thinkSapere.com independence, integrity and objectivity

MEMORANDUM FOR Geoff Bascand FROM Financial Policy and Financial Systems Analysis MEETING DATE 31 July 2019 SUBJECT Briefing for meeting with the Minister of Finance – Impact of capital proposals on agri sector FOR YOUR Information Background

1. This note provides background material for the meeting with the Minister of Finance (MoF) on the potential impact of the capital proposals on the dairy sector. We propose sending a brief slide pack to the MoF’s office, which covers: a. Irrespective of the Capital Review, banks have formed the view that current dairy debt levels are unsustainable and banks have in recent years tightened their lending standards and increased their margins. b. The potential impact of higher bank capital requirements on the agri sector is highly uncertain, as this is determined by many factors other than capital (e.g. competition, farms’ access to other sources of financing). In some cases, The purpose of this memorandum is to brief you on the potential impact of our proposed higher capital requirements on the agricultural sector. We recommend that you:  Note that the agri sector, particularly dairy, remains highly indebted. However, the dairy sector’s financial position has improved over the last 3 years, as evidenced by the decline in the proportion of high-LVR debt and increase in the proportion of dairy debt that is being paid down (rather than on interest only terms).  Note that credit availability for the dairy sector has been tightening over the last 3 years. Tighter credit conditions have been driven by both price (i.e. banks charging higher margins) and non-price factors (i.e. banks’ tighter policy and risk appetite settings for dairy lending).  Note that based on our estimates of an average lending spread increase of 20 to 40 bps and based on the current composition of the big four banks’ risk weighted￾assets, we estimate that on average, the farm lending spread could increase by 30 to 75 bps. High LVR loans are likely to experience higher repricing.  Note that we suggest providing the Minister with a high-level slide pack, which contains key messages on the potential impact of the capital proposals on the agri sector.

Figures were adjusted to incorporate the overlays applied to ANZ's operational risk, housing and farm lending portfolios.

ANZ