Extending Reserve Bank Liquidity Policy BS13 to overseas bank branches

Consultation Paper: Extending Reserve Bank Liquidity Policy BS13 to overseas bank branches The Reserve Bank invites submissions on this Consultation Paper by 30 March 2012. Submissions and enquiries about the consultation should be addressed to: Jeremy Richardson Senior Adviser, Financial System Policy Prudential Supervision Department Reserve Bank of New Zealand PO Box 2498 Wellington 6140 Email: jeremy.richardson@rbnz.govt.nz Please note that a summary of submissions may be published. If you think any part of your submission should properly be withheld on the grounds of commercial sensitivity or for any other reason, you should indicate this clearly. January 2012

2 1: Background

1. In October 2009, the Reserve Bank published a near-final version of its liquidity policy (BS131
2. We announced at the time that we would finalise the details of the liquidity requirements for the other registered banks by the end of 2009. The reasons for the staged approach to implementing BS13 were: ) and imposed conditions of registration on all but one of the locally-incorporated banks, to implement the minimum one week and one month mismatch ratios, the minimum core funding ratio, and requirements on liquidity risk management, with effect from 1 April 2010. Implementation went ahead on that timetable, with a few further minor changes to BS13 before April 2010. a. the key priority was to get the requirements in place for the big four banks, given their systemic importance; b. we also felt it was desirable to include the stand-alone local banks which do not belong to overseas banking groups subject to home country liquidity requirements; c. BS13 allows the scope for a more tailored approach to branches (see paragraphs 16- 19 of BS13), which requires further consideration; and d. there is one overseas banking group that is not of systemic size in New Zealand but is “dual-registered” (that is, the parent bank has both a subsidiary and a branch in New Zealand), and devising the best combined approach to that subsidiary and branch also needed further consideration.
3. The planned work to adapt the liquidity minimum ratios to the branches has been repeatedly postponed, because of other higher priorities in the Reserve Bank’s policy development work for registered banks. While we believe that most of the banks that are not currently subject to BS13 have generally adequate liquidity in New Zealand, we think that we should not delay any longer in ensuring that all of them meet common minimum standards.
4. This paper therefore sets out the Reserve Bank’s views on how the rationale for applying our liquidity policy to locally-incorporated banks adapts in the case of branches. It makes corresponding proposals on the extent to which, and with what variations, the policy requirements should be applied to branches.

1 Liquidity Policy (BS13) http://www.rbnz.govt.nz/finstab/banking/regulation/3675928.pdf and Liquidity Policy Annex: Liquid Assets (BS13A) http://www.rbnz.govt.nz/finstab/banking/regulation/3675953.pdf

3 5. Section 2 considers whether and how the BS13 mismatch ratios should be applied to branches. Sections 3 and 4 do the same for, respectively, the core funding ratio and the BS13 requirements on banks’ own liquidity risk management. 6. Section 5 covers regular reporting to the Reserve Bank on liquidity risk, using the standard template. Section 6 touches briefly on public disclosure of a bank’s liquidity risk (although does not propose any changes at this stage). Section 7 refers to the costs and benefits of the proposals, without including any detailed estimates: we are seeking information from the branches to allow us to produce those. Section 8 sets out the future steps in the process of extending BS13 to the branches, and proposes the timing of those steps. 7. To assist the consultation process, the Reserve Bank has posed a series of consultation questions at the relevant points in this consultation paper. For convenience, these questions are repeated in an Appendix. However, submissions should not necessarily be limited to addressing these questions, and are not required to address all (or any) of them. The Reserve Bank welcomes all relevant comments on the proposals outlined below. 2: Applying the mismatch ratios to branches General rationale for applying mismatch ratios to branches 8. The aim of the one week and one month mismatch ratios is to ensure that a bank has sufficient readily liquefiable assets to be able to cope with a short-term disruption in its ability to roll over funding. 9. A key difference between branches and locally-incorporated banks is that a branch is part of a single legal entity that also runs the risk of being unable to meet immediate cash claims in its country of incorporation, or indeed in any other jurisdiction where it operates a branch. The bank’s survival therefore depends on its ability to manage this risk across the entity as a whole, so in theory it only needs to own sufficient liquid assets on its whole balance sheet to meet potential outflows across that balance sheet. However, while centralised liquidity management has efficiency benefits for global banks in normal market conditions, as the prudential supervisor we want to design a policy that ensures the liquidity is immediately available where it is needed, in stressed market conditions and when the affected bank is suffering a temporary loss of confidence. 10. We believe that in the type of short-term crisis that the mismatch ratios are designed to address, both the currency of the liquid assets, and the jurisdiction where control of them

4 resides, matter. A bank with a central pool of liquidity, faced with a sudden call on NZD cash, might for instance have to liquidate home country government debt held in the home country, convert the proceeds into NZD, and make them rapidly available in New Zealand. The types of friction that may prevent the bank achieving this can include: a. time zone differences; b. temporary disruptions in FX swap or spot markets; c. short-term uncertainties in the home country about where liquidity is most urgently needed in the bank, with consequent delays imposed either by head office or the home country supervisor. 11. And the above describes the case where head office is willing in principle, but temporarily unable, to fund cash calls on the branch. In a more extreme crisis, if the parent bank as a whole is facing imminent default on its payments, it may decide its best option is to prioritise continuing to meet its liabilities in its home country, and pull in funding from wherever it can overseas in a last-ditch attempt to achieve that. Head office might cease to roll over maturing funding to the branch. At that point the Reserve Bank might be looking at putting the branch into statutory management, in which case any liquid assets still held in the branch at the point where we can ring-fence it would be beneficial. 12. Some additional considerations arise with dual-registered branches. There are possible scenarios in which only the locally-incorporated subsidiary suffers a loss of confidence, not the branch (and vice versa). For instance, a rumour about account security problems at the subsidiary’s internet banking operation could spark a run of retail depositors, without undermining confidence in the parent bank and its New Zealand branch (at least initially). A stock of liquid assets held by the branch can thus be a valuable local source of additional liquidity support for the subsidiary (and the lack of such a stock may result in an additional call on the liquid assets of the subsidiary in a crisis). We believe that this provides an additional argument for imposing mismatch ratios on the dual-registered branches by comparison with the stand-alone branches. 13. Overall therefore, we think there is a strong starting presumption that the overseas bank branches in NZ need to hold an adequate stock of (mainly NZD) liquid assets in New Zealand against some metric of the potential short-term claims. Q1. Do you agree with the arguments why branches as well as locally-incorporated banks should be required to hold a minimum stock of liquid assets in New Zealand?

5 Interaction with home country/ Basel III requirements 14. The Basel III liquidity framework2 15. We plan to issue a consultation paper in the first half of 2012 which will air the issues raised by Basel III liquidity for all registered banks. As a first general point, we note here that adoption of the Basel III framework is an option rather than an obligation for the Reserve Bank, since (i) we are not the home supervisor of any internationally active banking group and (ii) New Zealand is not a member of the Basel Committee. will require each home country supervisor to apply the Basel Liquidity Coverage Ratio (LCR) on a consolidated basis to each overseas bank group represented in New Zealand. This affects all overseas bank branches, as well as subsidiaries of overseas banks. The LCR is due to come into force on 1.1.2015. LCR versus BS13 mismatch ratios 16. The main high-level question that Basel III raises for branches is whether the objective of requiring them to hold a minimum stock of liquid assets should be achieved by imposing the Basel III LCR on them, rather than the BS13 mismatch ratios. 17. The locally-incorporated banks have considerable sunk systems costs in implementing the BS13 ratios. Also, subject to further, more reliable quantitative impact comparisons, it appears that the BS13 mismatch calculations generate a total liquidity stock requirement that is broadly comparable with that of the LCR. For both these reasons, we have already indicated that we will keep our current liquidity policy (BS13) in place for the foreseeable future. 18. The branches are part of legal entities that will in due course be required to calculate the LCR, and will therefore need to categorise the branch’s business on the basis of the LCR as implemented by their home supervisor. So other things being equal, a more efficient longer-term solution might be for the Reserve Bank to impose the minimum LCR on the branch, calculated on the basis of the branch balance sheet and using the home supervisor’s methodology. 19. However, the relevant home supervisors are currently some way off having the details of their LCR implementation finalised. APRA, who are the earliest to come out with proposals, plan to have their framework finalised by mid-2012. For supervisors in EU states (for example), the requirements are still in the form of a draft Directive. So it seems unlikely that the branches will be able to put the systems in place to contribute to their home country LCR calculation until much closer to the switch-on date of January 2015. Waiting to rely on the home country’s LCR measurement details could thus mean a significant further delay in imposing a liquid asset requirement on the branches. Given

2 Basel III: International framework for liquidity risk measurement, standards and monitoring (Basel Committee on Banking Supervision, December 2010) http://www.bis.org/publ/bcbs188.pdf

6 that this is already two years later than originally planned, we are reluctant to build in further delay. 20. An alternative would be to transpose the LCR into our own separate policy document sooner, to run alongside BS13. However, we think this would take considerably more Reserve Bank staff time than the BS13 option, and so might not actually be achievable much sooner than the home country LCR option. 21. There would also be some potential problems in locally-incorporated banks remaining subject to the BS13 mismatch ratios while the branches were subject to the LCR. For instance – a. since New Zealand does not have deposit insurance, the lowest run-off rate for retail deposits in the LCR would be 10%, compared to 5% in BS13; b. the LCR is only a one month ratio, so we would have to decide whether, and if so how, to adapt it to also impose a one week equivalent; c. the dual-registered banks would have a different measurement basis in their branch from that in their subsidiary, which might require adaptation of either or both frameworks to remove perverse interactions between the two; d. in particular, high quality liquid assets (HQLA) in the LCR are more narrowly defined than liquid assets eligible for BS13 (although there are options to separately address that, discussed below); and e. we have not completely ruled out switching locally-incorporated banks from BS13 to LCR eventually, and if we do, we think it would be safer to approach that switch at the time in a more measured way that suits all the affected banks, rather than risking locking in specific treatments for branches now that would be harder to unwind later. 22. On the other hand, if we opt for implementation via BS13 now, there is the possibility that branches will bear that cost now and still face the cost of LCR implementation at some later date. The biggest systems work in implementing BS13 appears to be in the determination of which size band a given provider of non-market funding falls into, and this approach does not feature in the LCR. But a number of the branches have little or no non-market funding. Other aspects of the calculation, such as allocating funding to maturity buckets, apply under both the LCR and BS13, and should in any case already be captured in banks’ liquidity risk management systems. In fact, we understand that several of the branches already regularly calculate some form of the BS13 mismatch ratios, and at least one already estimates its LCR as well.

7 23. Section 7 below seeks information on the costs to the affected banks of both our preferred BS13 approach, and of the LCR alternative: more exact estimates of the relative costs will be an important input into our final decisions. Q2. Do you agree that on balance the best option for delivering a liquid asset stock requirement in the near future is to implement the BS13 mismatch ratios? LCR definition of liquid assets 24. Regardless of whether the branches are subject to BS13 mismatch ratios or to the LCR, an issue which will affect most of them is which NZD securities will meet the LCR’s stringent criteria for HQLA (high quality liquid assets). The precise impact of this on the branches will depend on the details of home country implementation of the LCR. 25. Our preliminary view is that, of the liquid assets eligible in the BS13 mismatch ratios, the only ones which will also meet the standard criteria for HQLA will be cash, Reserve Bank settlement balances, and New Zealand Government debt. There are therefore substantial amounts of liquid assets currently held by locally-incorporated banks to meet their mismatch ratios which will not pass the HQLA tests. 26. For most if not all of the branches, it is likely that the scale of the NZD mismatch run by the branch will not be material for the banking group as a whole, and that the home supervisor will not therefore impose a hard requirement on the group to hold NZD￾denominated HQLA. We also do not think it likely that home supervisors will deem that NZD liquid assets outside the specific Basel criteria can qualify as HQLA for the group LCR calculation. (This is in contrast to the position of the big four Australian subsidiaries, whose NZD mismatch position is material for their groups as a whole, and whose home supervisor (APRA) is proposing to allow all BS13-eligible liquid assets to count for the LCR3 27. Other things being equal, the likely outcome for the branches is that there will be liquid assets that are BS13-eligible but not LCR-eligible, which will therefore contribute to a branch’s compliance with the BS13 mismatch ratios, but not to its banking group’s compliance with the LCR. .) 28. The Basel III liquidity framework makes three options available to jurisdictions that can demonstrate they have an insufficient supply of HQLA, to enable them in effect to add additional HQLA4

3 See section 4.6 of Implementing Basel III liquidity reforms in Australia, APRA Discussion Paper, November 2011. . Option 1 is a contractual committed liquidity facility from the relevant central bank, with a fee. At this stage, the Reserve Bank does not want to raise expectations that it will put such a facility in place: if we do decide to take this option http://www.apra.gov.au/adi/Documents/ADI_DP_IBLR_November_2011.pdf 4 See paragraphs 45-49 of the Basel III liquidity framework.

8 further, we will do so as part of the forthcoming wider consultation on the issues raised by Basel III liquidity. However, we would be interested in feedback from the branches on what this would mean for them. 29. If the Reserve Bank were to put such a facility in place, a bank could in effect convert specified non-HQLA, BS13 liquid assets into HQLA by paying an annual fee on pre￾committed amounts of those assets. This might be advantageous for the branches, whether we require them to use the LCR or the BS13 measurement framework. Q3: How do you expect that the proposed application of the BS13 mismatch ratios to your New Zealand branch will interact with your home country supervisor’s application of the LCR to your banking group? Q4: If you have identified problems in answer to Q3, would a contractual committed liquidity facility from the Reserve Bank help address those problems? Details of proposed implementation via BS13 30. To apply the BS13 mismatch ratios to the branches, we would impose the standard condition of registration already in place for locally-incorporated banks, that requires the bank to meet the minimum one week and one month mismatch ratios calculated in accordance with BS13. We raise here two issues for branches with this standard approach. If we decide, following consultation, that either of these issues requires a variation for the branches from the standard approach, we will consider further whether to update BS13 itself, or to reflect the variations only in branches’ conditions of registration. Head office funding compared to other related party funding 31. In the BS13 mismatch ratios, funding from related parties is all treated as market funding, with a 100% run-off assumption at its earliest contractual maturity. This makes the severe crisis assumption that other parts of the group are only able to look out for themselves rather than trying to help the New Zealand subsidiary. 32. But in practice this existing treatment (within BS13 as applied to locally-incorporated banks) only currently applies to funding provided by separate legal entities in the same banking group. We have therefore considered whether any different treatment should apply to funding provided within the same legal entity, particularly to the branch from head office. 33. Our view is that we should assume the most extreme case, where the parent as a whole is facing imminent default on its payments, and may decide its best option is to prioritise meeting its liabilities in its home country. We included this scenario in the rationale for requiring a stock of liquid assets (discussed in paragraph [NN] above).

9 34. This treatment will also have the advantage of incentivising banks to lengthen the term of head office funding provided to the branch. While there is always some doubt how far head office would in fact be bound by the terms of intra-entity lending in extremis, we believe there is some benefit in this. 35. In sum, it appears safer to treat funding from head office to the branch as market funding, in line with the existing policy. Q5: Do you agree that funding from head office to branch should have the same run￾off assumption as funding from parent bank to subsidiary? Committed lines received 36. In the BS13 mismatch ratios, 75% of the undrawn amount of a committed funding line granted to the bank can be included as an inflow, up to a limit of 3 percentage points of the ratio from any one provider, and up to 9 percentage points from all providers in total. The effect of this is to allow a funding line from the rest of a bank’s group to count towards the mismatch ratio up to the 3% limit. 37. A possible question is whether to treat intra-entity commitments (from head office to branch) more generously than BS13 currently allows. The argument for this would be that a commitment is more reliable intra-entity than intra-group. But as already discussed in relation to existing intra-entity funding, there is the risk that under extreme liquidity stress the parent will not be well-placed to fund the branch. On the other hand we do not want to entirely disincentivise banks from providing such lines to their New Zealand branches. 38. We therefore propose to stick with the standard treatment, giving some allowance for a head office commitment to the branch, but only up to the 3% limit. Committed lines provided by third parties specifically for the benefit of the New Zealand branch of the legal entity would also be allowed, within the overall 9% limit. Note that this means that a dual-registered bank would be able to benefit from both a committed line from head office to branch, and from parent to subsidiary, each up to their respective 3% limits. Q6: Do you agree that there should be no variation from the standard treatment for committed lines from head office to branch? 39. There is a separate technical point to address. BS13 is not symmetrical in its cash inflow and outflow assumptions for commitments received and granted. This would give the dual-registered banks the chance to improve both entities’ mismatch ratios without any actual reduction in liquidity risk: if a branch was granted a $100mn committed line by its associated subsidiary, that would require the subsidiary to hold $15mn more in liquid assets but would allow the branch to hold $75mn less.

10 40. To remove this asymmetry, we propose to make the following adjustment to the treatment of commitments provided by one registered bank (Bank A) to another registered bank in the same banking group (Bank B): the amount of such commitments that qualifies as a committed line for Bank B’s mismatch ratios would have a 75% run￾off factor in Bank A’s mismatch ratios, rather than the standard 15%. Q7: Do you agree that this is a necessary and workable adjustment to the standard mismatch ratio calculations? Exceptions 41. The mismatch ratios provide a broadly risk-sensitive measure of liquidity risk, so that one branch that is subject to a lower level of liquidity risk than another will generally need to hold a lower level of liquid assets. So we do not expect commonly to need to make exceptions to the minimum mismatch ratio requirements. 42. However, the New Zealand branch of ANZ is currently subject to a condition of registration that has the effect that it can only obtain its funding from within the same legal entity. We do not therefore believe that the branch poses any threat to New Zealand financial stability, and so while this condition remains in force, we do not intend to subject it to the mismatch ratios. This is unlikely to be a common situation, but we would consider exempting any other branch in future which was subject to similarly tight constraints on its business. International comparisons 43. The 1983 Basel Concordat5 44. Overseas banking supervisors that impose specific quantitative liquidity ratios have commonly applied those ratios to foreign bank branches operating in their jurisdictions. This has been the case for example for APRA and the UK FSA. Under its Basel III implementation plans, APRA proposes to apply the LCR to foreign bank branches. Under its old mismatch regime for banks, the FSA exempted some branches from its standard requirements, provided it was satisfied on a number of points, including the sets out the principles for the division of responsibility between home and host country supervisors in the supervision of international banking groups. It remains a live document despite the many subsequent Basel Committee developments such as the various Capital Accords, and now the proposed liquidity framework. It is noteworthy that the Concordat sees the liquidity of foreign bank branches being a shared responsibility between home and host supervisors. This is in contrast to solvency requirements, which only make sense on a whole legal entity basis, and are firmly the responsibility of the home country supervisor.

5 Principles for the Supervision of Banks’ Foreign Establishments (“Concordat”), Basel Committee on Banking Supervision, May 1983 http://www.bis.org/publ/bcbsc312.pdf

11 home supervisor’s liquidity policy, and the bank’s liquidity management approach, as applied to the branch. The FSA has retained the possibility of exemptions in its new regime, but conditional on obtaining far more extensive regular information from the bank and the home supervisor about the whole bank’s liquidity. Conclusions on the mismatch ratio 45. We conclude therefore that: a. there is a strong case to require overseas bank branches to hold a stock of liquid assets, based on both first principles arguments and international precedents; b. the only exceptions would be cases where the business of the branch is tightly constrained by other conditions of registration; c. to achieve the required liquidity stock in a reasonable timeframe, the preferred option is to apply the BS13 one week and one month mismatch ratios to the branches; and d. the interaction between the BS13 mismatch ratios and the Basel III LCR when applied to the overseas banking groups of the branches (and of some locally￾incorporated subsidiaries) may have some implications for the stock of eligible liquid assets defined in BS13. 4: Core funding ratio General micro-prudential rationale compared to mismatch ratios 46. By contrast with widening the scope of the mismatch ratios to include the branches as well as the locally-incorporated banks, our assessment is that there is no good case for doing the same with the core funding ratio (CFR). The key difference between the mismatch ratios and the CFR is the timescale. The aim of the CFR is to reduce banks’ reliance on short-term wholesale funding, essentially to buy time if term funding markets are closed for a while. It addresses a slower-burning problem, not an immediate crisis. 47. As noted above, head office and branch are part of the same legal entity, and the failure of the branch to meet its obligations could spell disaster for the bank as a whole. We therefore expect head office to do its utmost to continue ensuring that its branches are funded. Nevertheless, we have noted above the kinds of short-term disruption that may mean that head office is unable, even though willing, to provide the branch with immediate funds. And we have noted more extreme cases in which head office may have no option but to prioritise meeting immediate claims in its home country, and so may in effect becoming unwilling to fund the branch.

12 48. In the case where head office is willing but temporarily unable to fund the branch, a stock of liquid assets tides the branch over the short-term problem. If the branch continues to lose third-party funding over the medium term, then provided that the bank as a whole legal entity does not face a crisis, head office can continue to replace that funding. 49. Imposing the CFR on a branch would among other things result in head office providing funding to the branch at longer maturities. But we do not think that much reliance can be placed on the maturity of intra-entity lending, since it is not governed by a contract between two separate parties. If head office ever gets to the point where it needs to pull money out of the branch, the maturity at which that funding has been provided is probably of little relevance. So as long as head office remains willing to fund its New Zealand branch, the maturity of the funding doesn’t matter; and if it ever reaches the point where head office is unwilling to do so, the whole legal entity is under severe threat and the maturity of the funding is irrelevant anyway. 50. If we compare the reasons why we have judged it worthwhile to impose the CFR on locally-incorporated subsidiaries of overseas banks, we can highlight the contrast with the position of branches: a. it is more likely that an overseas parent under pressure would cease to roll over maturing funding or cease to provide new funding to its New Zealand subsidiary, since it can do so without triggering its own default; b. if a parent was able and willing to replace third-party funding that was draining out of its subsidiary, in many cases it would soon run up against home supervisor limits on related party lending (such as APRA’s APS 222 limit). Such limits are not usually applied intra-entity; and c. there is more benefit in incentivising longer maturities on funding provided from parent to subsidiary. Those maturities are part of contractual arrangements between two separate legal entities, and the subsidiary’s consent will be needed to any early repayments. The subsidiary may refuse, if to agree would threaten its survival: and the likelihood of this is increased by existing RBNZ policies aimed at increasing the chance that the subsidiary can survive a crisis independently of the parent bank (particularly those on corporate governance and outsourcing). Level playing field / dual registration considerations 51. This section considers whether there are any separate arguments for applying the CFR specifically to the dual-registered branches.

13 Transfer of core funding need 52. A dual-registered branch can be used to help strengthen the subsidiary’s CFR position. For instance, the subsidiary can transfer part of its long-term lending portfolio to its associated branch, with funding provided by the parent. The result is that the subsidiary’s existing core funding goes further, as it has fewer loans and advances to fund, while the parent bank has more. 53. We do not think this justifies applying the CFR to the branch. First, the parent bank has assumed some of the subsidiary’s liquidity risk, which thus becomes a matter to be addressed by the home country supervisor’s liquidity requirements on the parent. Second, the RBNZ’s local incorporation policy limits the total value of loans that the branch can acquire, so such a transfer only provides a one-time benefit. And third, imposing the CFR on the branch would simply result in head office increasing the maturity of its funding of the branch, which, as argued above, would not achieve any real benefit. Intra-entity maturity transformation 54. A second potential concern is that a dual-registered branch could lend at long maturities to its related subsidiary while funding itself short-term (from third parties as well as from head office). The branch thus “manufactures” core funding for the subsidiary, which could undermine the effectiveness of the CFR as already applied to the subsidiary. A subsidiary that does not have the benefit of a dual-registered branch would have to ensure that at least 70% of this amount of funding from its parent bank or third parties was at long maturities6 55. We could in principle prevent this happening by applying the CFR to the branch. To directly address the arbitrage opportunity, the CFR calculation would need to be adapted, to require that any lending from the branch to the subsidiary that qualifies as core funding for the subsidiary must be added to the branch’s core funding requirement. The CFR would then ensure that a sufficient amount of the funding from head office to branch would be at long maturities. However, as already noted, we do not see material value in a rule affecting the maturity of intra-entity transfers. The main value of the CFR is in lengthening the maturity of funding provided to the locally-incorporated bank as a legal entity, from other legal entities. The CFR applied to the subsidiary bites on the maturity of funding provided to it by the parent bank as a legal entity, regardless of whether that funding is booked in its head office or in its New Zealand branch. .

6 This assumes the minimum CFR is set at 70%, as at present. The RBNZ plans to increase the minimum CFR to 75% for all locally-incorporated banks with effect from 1.1.2013.

14 56. In conclusion, there are some advantages for an overseas bank that has a New Zealand subsidiary in setting up a dual-registered branch, but the objectives of the CFR can only be effectively achieved by applying the CFR to the subsidiary, not to the branch. Macro-prudential considerations 57. The CFR is aimed not only at reducing the risk of failure of an individual bank, but also reducing the vulnerability of the financial system as whole. New Zealand runs a persistent current account deficit, and a significant proportion of the deficit is funded via banks’ offshore funding. By lengthening average maturities of that funding, the minimum CFR requirement reduces the amount that has to be re-financed over a given period. So we also consider here whether applying the CFR to branches might contribute to this macro-prudential objective. 58. If pressure arose on the funding of the current account, an overseas branch would tend to lose any third-party offshore funding as it matures, which its head office would then have to replace. The core macro-prudential vulnerability is then that part of the current account deficit is being funded by short-term lending from overseas head offices to their New Zealand branches. That undoubtedly poses some risk, but the CFR would not usefully address the risk by requiring that the intra-entity funding to be pushed out to longer terms. The Reserve Bank’s local incorporation policy limits the scale of the risk by limiting the size of overseas branches. International comparisons 59. We are not aware of any current equivalents of the CFR overseas, so there are no precedents for applying it to branches. The planned Basel III Net Stable Funding Ratio (NSFR) is broadly equivalent to the CFR in its objectives and likely impact. The NSFR is not specifically required to apply to branches. However, in their consultation paper APRA propose that they will apply the NSFR to foreign bank branches that are currently subject to APRA’s “scenario analysis approach”: these tend to be the bigger and more sophisticated branches. We do not yet know what other supervisors’ plans are for applying the NSFR to branches. Summary on CFR 60. Although there is a workable way to apply the CFR to the branches, we are not persuaded that doing so would achieve any material benefit, whether in limiting individual bank liquidity risk, addressing macro-prudential concerns, or dealing with unequal treatments between banks. Q8: Do you agree with this assessment of applying the CFR to overseas branches?

15 5: Qualitative requirements 61. BS13 includes extensive guidelines on how registered banks should manage their liquidity risk. These guidelines hang off the following standard condition of registration, which currently applies to all banks that are subject to the minimum liquidity ratios: That the registered bank has an internal framework for liquidity-risk management that is adequate in the registered bank’s view for managing the bank’s liquidity risk at a prudent level, and that, in particular: (a) is clearly documented and communicated to all those in the organisation with responsibility for managing liquidity and liquidity risk; (b) identifies responsibility for approval, oversight and implementation of the framework and policies for liquidity-risk management; (c) identifies the principal methods that the bank will use for measuring, monitoring and controlling liquidity risk; and (d) considers the material sources of stress that the bank might face, and prepares the bank to manage stress through a contingency funding plan. 62. At the time that BS13 was finalised, we envisaged that we would impose these guidelines on the overseas bank branches at the same time as the liquidity ratios. However, there is a technical challenge in applying these requirements to the branches. Not just in the above condition, but also throughout the guidelines there are regular references to the “registered bank”. The problem is that for a branch, the term “registered bank” means the whole bank as a legal entity. While we certainly believe that it would be beneficial for any overseas bank present in New Zealand to comply with the above condition, and follow the guidelines, on a global basis, requiring that is beyond the scope of the Reserve Bank’s objectives, and is firmly the responsibility of the home supervisor. 63. We could in principle adapt the standard condition and the guidelines, to set out our expectations of liquidity risk management in the New Zealand branch alone. What we would expect is that the parent bank and/or banking group has a sound group-wide framework for managing its liquidity risk, and that liquidity risk is separately managed in the branch in a way which reflects the group framework. The branch should have some of its own capability for addressing liquidity risk, but can also rely to some extent on the group framework. But what we can reasonably expect of any given branch will depend on the exact nature of its business, and also on whether the group liquidity risk framework follows a more devolved or a more centralised model.

16 64. So given how diverse the overseas bank branches are in New Zealand, we think that applying meaningful liquidity risk management guidelines to them would require us to negotiate more or less individually tailored guidelines. We do not think that the cost of going through that exercise would justify the benefits. Our view is therefore that achieving adequate liquidity risk management in the branches is best left to Reserve Bank supervisory monitoring and judgement. Q9: Do you agree with our assessment that liquidity risk management guidelines would have to be individually tailored to suit the diversity of the branches? Q10: Do you agree that the burden of applying the guidelines to the branches would not justify the benefits? 5: Regular reporting to Reserve Bank 65. The banks that are already subject to the BS13 liquidity ratios submit monthly liquidity reports to the Reserve Bank using a standard template Excel workbook. A summary of the contents of the report is included in BS13, and an electronic copy of the template is available on request. In brief, it covers compliance with the minimum ratios, a breakdown of liquid assets held, comprehensive data on contractual cash inflows and outflows broken down by maturity, details of new funding raised over the latest month, and the cost of that funding. 66. This information allows the Reserve Bank to understand how the policy is being complied with, to see whether it is having its desired effects, and whether it is having any unexpected or perverse consequences. It also allows our supervisory analysts to monitor liquidity risk in their banks more generally, and helps to provide a picture of funding pressures and costs across the banking system in aggregate. 67. We propose that overseas branches should also start submitting monthly liquidity reports using the same template, with one exception: if it is agreed that branches do not have to meet a minimum core funding ratio, then they will not report the CFR calculations which are on the summary page of the standard template. Q11: Do you have any comments on extending the liquidity reporting requirements to the branches? 6: Public disclosure 68. The Reserve Bank’s disclosure regime is an important component of its approach to banking supervision. In the original consultation leading up to the introduction of BS13, we proposed that each bank subject to the policy should disclose some additional items related to the policy, including the ratio values for the bank. However, those proposals

17 were put on hold for the time being. Existing disclosure relating to liquidity risk is mainly that required by applicable financial reporting standards (NZ IFRS), which applies equally to branches and to locally-incorporated banks. 69. However, a registered bank is required to disclose any breach of any of its conditions of registration, so if a bank breaches a minimum liquidity ratio under BS13 at the end of any business day during a particular quarter, the bank must disclose that breach in the disclosure statement for that quarter. 70. We are still keen to explore further the practicalities of requiring banks to disclose more information on their compliance with the minimum ratios. But we are aware of the commonly voiced concern that disclosing liquidity risk metrics too frequently or too immediately could trigger a sudden loss of confidence in a bank. (APRA in its consultation paper is proposing that a bank would publish the average, high and low of its Basel III ratios over the previous six month period, starting with the LCR in 2015.) We also note that the Basel Committee has still not reached firm conclusions on what the best approach is in this area. 71. We are therefore not making any liquidity disclosure proposals specifically for branches, but we intend to revisit the subject for all banks subject to BS13 as soon as practical. Q12: Do you foresee any problems with the existing disclosure requirement, that is, having to disclose any breaches of the mismatch ratios in the branch’s next disclosure statement? 7: Costs and benefits 72. The Reserve Bank is required to publish a Regulatory Impact Assessment (RIA) when it implements a material policy change. We plan to publish a RIA if and when we put into effect the changes proposed in this consultation (with any refinements made in light of feedback). 73. We published a RIA in October 2009 on the impact of imposing the BS13 requirements on the locally-incorporated banks. The discussion in that RIA about the rationale for the policy and the benefits it delivers (in terms of reducing the risk of potential liquidity stress scenarios) will carry across to the branches, modified by the discussion above on how the components of the policy relate to branches as opposed to locally-incorporated banks. The RIA will also reflect any feedback we receive on the validity of the arguments above. However, as in the earlier RIA, the benefits will be hard to quantify in dollar terms. 74. The costs of the proposals to the branches can more realistically be quantified, and we will include the best available cost estimates in the RIA. To come up with those

18 estimates we need data from the branches on what they expect the costs to be, in the following areas: (a) initial implementation: the one-off costs of putting in place the necessary systems to measure cash inflows and outflows in the way defined by the policy, in order to be able to calculate the mismatch ratios; (b) the continuing annual costs of monitoring compliance with the minimum ratio requirements and submitting the associated monthly reports to the RBNZ; (c) the continuing annual cost (if any) of any change in the quantity and composition of the branch’s portfolio of liquid assets as a result of the policy; and (d) any other type of cost not included in the above. 75. We request that each branch submits as much information to us as possible to allow us to reach a good understanding of the expected burden of extending BS13 to the branches as proposed. 76. In providing this data, it would be very helpful if branches could use the following approach to derive the annual cost of the additional time their staff would need to spend on compliance: estimate the amount of additional hours employees would spend each month on assessing compliance with the ratios and preparing the liquidity returns for the RBNZ; multiply by the estimated hourly total cost to the bank of employing each of those employees (including wages and overhead costs); and then multiply that figure by the number of times per annum the exercise would be carried out (ie in this case, twelve). (This follows the so-called “Standard Cost Model” methodology.) 77. More generally, we would appreciate any other background information that will help us to understand the robustness of the estimates. This could include for instance – • high and low values of assumptions; • assumptions about the rate of return on various liquid assets held, compared to that on any alternative assets that would otherwise be held; • current values of the mismatch ratios for the branch, before any balance sheet adjustments made to meet the expected minimum requirements; • detail on which particular classes of liquid assets the branch expects to hold (and changes from current holdings); and • any assumptions about the impact of home supervisor Basel III implementation. 78. Central case cost estimates should assume our preferred implementation options as indicated above, namely – • the BS13 methodology applies;

19 • branches are subject to the one week and one month mismatch ratios, but not to the CFR; • liquid assets are as defined in BS13, and no committed liquidity facility is provided by the RBNZ; • head-office funding is treated the same as other related party funding; • treatment of committed lines received is unchanged from BS13; and • a committed line granted has a 75% rather than a 15% run-off, to the extent that it qualifies as a “committed line received” for a registered bank in the same group. 79. Branches should also provide information on the comparative costs of the main alternative option discussed above, namely to apply the Basel III LCR to the branch balance sheet: as noted above, having this information will help inform our choice between the two options. If a branch makes a submission suggesting any other variation from the preferred implementation options set out in this paper, it would be very useful to have estimates of how that alternative would affect the costs to the branch. 8: Timing and next steps 80. This consultation closes on 30 March 2012. 81. To implement the proposed policy requirements for the overseas bank branches, the Reserve Bank will add a new condition to their existing conditions of registration. Depending on the feedback received, we may need to consult on some changes to BS13 to implement branch-specific calculation of the mismatch ratios. Our aim will be to formally consult the branches on the proposed new condition and on any revisions to BS13 by the end of June, in time to allow the change to come into effect on 1 October 2012. The branches will have a further chance in that consultation to comment on whether the timetable is feasible, in light of the final details of the policy. 82. The Reserve Bank will impose the monthly liquidity reporting requirement on each bank by issuing it with a formal notice under section 93 of the Reserve Bank of New Zealand Act 1989, which will specify the month from which that reporting must start. We will agree with the branches the lead time they need to put the reporting processes in place, before we formally confirm the start date. 83. Our intention is that full reporting will be in place from the month in which the minimum ratio requirements are imposed, so that the first reports would be as at 31 October 2012, to be submitted by the fifteenth working day of November. We would also expect branches to submit back data for July to September 2012 at that point. If necessary, to avoid further delay in implementing the policy, we could alternatively require summary reporting from the implementation date, to be followed by full reporting at a reasonable interval afterwards (6 months at most). The summary reporting

20 would include the first three worksheets of the reporting workbook, covering the values of the ratios, primary liquidity assets held, and secondary liquidity assets held. Q13: Do you think that the proposed timetable is realistic for your bank?

21 Appendix: Summary of consultation questions Q1. Do you agree with the arguments why branches as well as locally-incorporated banks should be required to hold a minimum stock of liquid assets in New Zealand? Q2. Do you agree that on balance the best option for delivering a liquid asset stock requirement in the near future is to implement the BS13 mismatch ratios? Q3: How do you expect that the proposed application of the BS13 mismatch ratios to your New Zealand branch will interact with your home country supervisor’s application of the LCR to your banking group? Q4: If you have identified problems in answer to Q3, would a contractual committed liquidity facility from the Reserve Bank help address those problems? Q5: Do you agree that funding from head office to branch should have the same run-off assumption as funding from parent bank to subsidiary? Q6: Do you agree that there should be no variation from the standard treatment for committed lines from head office to branch? Q7: Do you agree that this [as set out in paragraph 40] is a necessary and workable adjustment to the standard mismatch ratio calculations? Q8: Do you agree with this assessment [as set out in Section 4] of applying the CFR to overseas branches? Q9: Do you agree with our assessment that liquidity risk management guidelines would have to be individually tailored to suit the diversity of the branches? Q10: Do you agree that the burden of applying the guidelines to the branches would not justify the benefits? Q11: Do you have any comments on extending the liquidity reporting requirements to the branches? Q12: Do you foresee any problems with the existing disclosure requirement, that is, having to disclose any breaches of the mismatch ratios in the branch’s next disclosure statement? Q13: Do you think that the proposed timetable is realistic for your bank?