Consultation Paper: Modifications to framework for restrictions on high-LVR residential mortgage lending

Consultation Paper: Modifications to framework for restrictions on high-LVR residential mortgage lending The Reserve Bank invites submissions on this consultation paper by 14 February 2014. Submissions and enquiries about the consultation should be addressed to: Head, Macro Financial Department Reserve Bank of New Zealand PO Box 2498 Wellington 6140 Email: macroprudential@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. 20 December 2013 Ref #5558957

2 1.0 Background

1. The Loan to Value Ratio (LVR) policy framework (BS19) was introduced in August 2013 to operationalise the use of LVR restrictions. LVR restrictions are one of four macro-prudential tools that the Reserve Bank has agreed (in a Memorandum of Understanding signed with the Minister of Finance in May) that it will use where necessary to help manage financial system risk. LVR restrictions have been in place since 1 October 2013.
2. The framework for implementation of LVR restrictions is outlined in chapter 19 of the Banking Supervision Handbook (BS19). When the Reserve Bank published its response to submissions on the LVR framework in August, the Reserve Bank noted that the final definitions used for LVR restriction purposes would depend on the outcome of Housing Review Stage II, most notably the measurement of mortgage loans and the definition of LVR. The policy outcome of Housing Review Stage II has now been published, although the formal changes to BS2A/BS2B are still to come (in a pending consultation paper).
3. One outcome of Housing Review Stage II is that that the definition of loan value, for both standardised and IRB banks, will not include most non-mortgage debt that the bank holding the mortgage has extended to a borrower, as had been originally proposed. For example, the customer’s credit card debt (at the same bank as their mortgage) will not typically be included in the computation of the LVR for their mortgage.
4. The Reserve Bank wishes to ensure that the intended approach to LVR measurement does not create an avoidance channel where customers can obtain substantial amounts of non-mortgage credit (including funding of their house deposit) outside of the LVR restrictions. The strictest approach to achieving this would be to redefine loan value in BS19 in order to incorporate these additional sources of credit. However, the Reserve Bank’s preferred option is to introduce a deposit identification/verification requirement, together with strengthening of the BS19 anti￾avoidance clauses to set strong expectations on banks regarding excessive provision of non-mortgage credit to mortgage customers. The Reserve Bank also proposes that additional data will be collected data to monitor compliance with this policy.
5. Concurrently, the Reserve Bank has been investigating the impact of LVR restrictions on construction lending. While high-LVR borrowers are a relatively small portion of the new-build market, the Reserve Bank wishes to ensure that the LVR policy does not constrain housing supply, since timely increases in housing supply are needed to meet the strong demand for housing presently seen in many parts of New Zealand. For this reason, the Reserve Bank announced on 11 December that new residential construction loans would be exempt from LVR restrictions. This paper sets out the motivation for the construction exemption, the nature of the exemption, and some operational considerations.
6. The remainder of this consultation paper provides detail on the construction exemption and the anti-avoidance and deposit verification changes we propose to make to BS19. The proposed revisions to BS19 are also available in a separate marked up copy of the chapter. When the formal definitions of BS2A/BS2B terms such as loan value and residential mortgage are released for consultation, the Reserve Bank will also release any consequential amendments to BS19. We expect that these terms within BS19 will continue to largely map to the related BS2A/BS2B definitions. Ref #5558957

3 2.0 Construction lending exemption 2.1 Exemption of residential construction lending from LVR restrictions 7. The base framework for LVR restrictions included several exemptions, which were designed to reduce the efficiency costs of imposing LVR restrictions without unduly undermining their effectiveness. The number of exemptions was deliberately kept small, in order to limit the complexity of the framework and to reduce the risk of policy leakages. The Reserve Bank was also conscious that banks would be able to engage in some high-LVR lending within their speed limit capacity.1 8. When the Reserve Bank consulted on the LVR restriction framework, there was little feedback on the subject of construction lending, and no substantive data provided by submitters. The information available to the Reserve Bank on high-LVR residential construction lending suggested that banks did very little such lending, which could easily be funded out of banks' speed limit capacity. Accordingly, there was not a strong case to exempt such lending. However, as explained in a recent Bulletin article, the Reserve Bank does not see macro-prudential instruments as ‘set and forget’ tools.2 Since the deployment of LVR restrictions, the Bank has conducted on￾going assessments of their effectiveness, which consider both quantitative data (such as housing lending and house price growth) and qualitative data. Changes in market practices and industry concerns are actively monitored by the bank. 9. In response to bank and industry feedback, the Reserve Bank sought additional information on high-LVR construction lending from the banking sector and the building industry. This new information suggests that banks' residential construction lending may be more significant than previously thought. Based on the information received, the Reserve Bank estimates that, while high-LVR construction lending accounts for just 1 percent of banks' total housing lending, it may fund up to 12 percent of residential building activity. These estimates suggest that LVR restrictions could potentially result in 200 fewer homes being built each month than may otherwise be the case (the equivalent of 2400 homes annually from a current annual total of 20,000 homes). These are material, but lower than has been suggested by the building industry. 10. Feedback from the banking sector further suggests that banks might be less likely to prioritise such lending within their 10 percent speed limits, owing to the greater uncertainty and complexity associated with such loans. Banks also suggested the risk of overruns could make them wary of undertaking moderate LVR (e.g. 65 – 80 percent) construction loans. In addition, survey data collected by the building industry is indicative of an initial disproportionate impact on new build enquiries, at least in some segments. The building industry notes the possibility of buy/sell chain effects, where some prospective vendors may be unable to sell existing homes and, hence, commission new ones. 1 Refer “Response to submissions received on the consultation on the framework for restrictions on high-LVR residential mortgage lending” (paragraph 70), August 2013, available at http://www.rbnz.govt.nz/financial_stability/macro-prudential_policy/5302711.html 2 Refer ‘A new approach to macro-prudential policy for New Zealand’, Rogers, L, Reserve Bank Bulletin, 76(3), June, pp. 12-22, available at http://www.rbnz.govt.nz/research_and_publications/articles/details.aspx?id=4364 Ref #5558957

4 11. Overall, the Reserve Bank now believes the supply side impact of not exempting construction lending may be somewhat higher than originally thought. As such, the Bank considers that adjustment of the LVR restrictions framework is appropriate. Extending the exemptions framework to construction lending should help to support new building and to moderate house price pressures. This is consistent with reducing systemic risk in the banking system. 12. Feedback suggests that there can be substantial risk associated with construction lending due to such factors as the potential for cost overruns, variable build quality or construction delays. However, in practice loan performance is typically similar to standard property purchases. This reflects tight credit policy around construction lending, which the Reserve Bank expects banks to continue. Banks are expected to apply their normal prudential lending criteria to construction loans and to set their own maximum LVR limits in accordance with this. The Reserve Bank will be actively monitoring the exemption data provided by the banks for any evidence of a relaxation in lending standards. 13. In formulating the exemption, the Bank has balanced a number of competing considerations, which include limiting the risk of policy leakages, not making the framework overly complex, and providing clarity and certainty around the regulatory requirement. • Clear boundaries are important for monitoring and enforcement of the exemption, and to provide clarity to banks, industry and households on what is exempt, and what is not. In drawing the boundaries, a prime consideration has been the underlying objective of the exemption, which is to support the building of new homes. • The Bank recognises that, as with any of the exemptions to LVR restrictions, the construction lending exemption might be challenging for banks to implement. It is therefore expressed in a permissive rather than a prescriptive way. • Although a compelling case has been made for exempting construction lending, the available evidence does not point to any further broadening of the exemption framework in the foreseeable future. 2.2 Nature of the construction lending exemption 14. The Bank proposes to exempt any residential mortgage lending to finance the construction of a new residential property. The construction loan would be for a property where the borrower has made a financial and legal commitment to buy in the form of a purchase contract with the builder, prior to the commencement of construction of the property. This could be traditional ‘construction lending’ where the loan is disbursed in staged payments, or it could be a loan to finance the purchase of a property, which will be settled (in one payment) once the build is complete. The exemption would also apply to top-ups to the loan arising from construction cost overruns during the build. 15.The purpose of the exemption is to support the supply of new housing. This being the case, the bank does not propose to exempt residential mortgage lending for the purchase of newly built houses (e.g. houses that have had a code of compliance issued but have never been lived in), or where construction has already commenced, as the build of these has already been funded. Developers and investors who finance such construction are not subject to the constraint of needing a buyer to fund the build. Combined with their greater market power, there would be a higher risk of distortions in the housing market, in the form of over-building or raising prices. Ref #5558957

5 Similarly, extending the exemption to new homes could raise the risk of investors “flipping” completed properties to new buyers (who could obtain an exempt loan). 16. The Reserve Bank also does not propose to exempt residential mortgage lending for the following purposes: • renovations of existing properties, including extensions • remediation work, [including the remediation of leaky homes and earthquake￾damaged properties] • borrowing for discretionary expenditure, such as furnishings. 17. The following table provides additional guidelines to the scope of the exemption. This does not provide an exhaustive list, but it is intended to provide a guide to the general boundaries. In scope Not in scope Construction loans – both progress payments and single payment tranche Purchases of newly built houses Construction cost overruns Discretionary expenditure e.g. furnishings Buyer commits before build Buyer commits when construction underway New home, existing home demolished Remediation and renovation Loan to buy section and build Loan to buy section only Builders' land/build package, provided construction of the property in question had not already started at the point the loan was being taken out. Loan to buy section and move house on to it 18. As discussed, the exemption is for lending for the construction of the property and (where applicable) the funding of the land on which the property is built. It is not intended to cover the costs of purchases of items that are not typically considered part of the house, and are not retained in the house if it is sold. For example, the purchase of a fitted kitchen or carpet may be included, but additional lending for furniture should not be included in the exempt loan. Banks retain the capacity to approve non-exempt lending (e.g. for the purchase of furniture) out of their speed limit capacity. 19. The same guidelines on scope apply to overruns – any lending which is necessary to cover additional costs of completing the construction of the property and associated chattels may be exempt. The reason for the overruns (e.g. variations in design or weather) does not affect the applicability of the exemption. 20. The rule about the buyer being committed before the ‘commencement of construction’ is intended to ensure that the individual property is built with a particular buyer in mind and committed from the start of the process. It is not intended to mean that every property in a subdivision is ineligible for the exemption once some property development work commences. In this regard, it is acceptable for the services to a section (e.g. roading, power and phone lines) to have been provided prior to the buyer for a home on that section being found. It is also acceptable for the build to have been consented, and for minor additional site works to have been carried out, prior to the buyer being found. Ref #5558957

6 2.3 Operation of the construction lending exemption 21. The exemption is applicable to all construction lending flows from 1 October 2013. Draft changes to implement the construction lending exemption have been made in the accompanying BS19 chapter of the Banking Supervision Handbook. These changes can be found in the following sections: • s10 – Definitions: this section has been changed to cater for the possibility of top-ups in the case of construction exemptions; • s12 – Exemption categories: this section has been changed to cater for the possibility of top-ups in the case of construction exemptions, and to set out the definition of the construction exemption; • s13 – Guidelines – exemption categories: this section provides guidelines to some of the features of the construction exemption, with a view to clarifying some boundary issues. • s14 – Guidelines – verification of third-party information: this section requires the bank to determine the purpose of the loan, and that the borrower has firmly committed to the purchase. 22. When calculating the speed limit, only exempt high-LVR flows will be counted, so there will be no adverse impact on the speed limit by reporting low-LVR construction lending flows. Banks are encouraged to report flows of construction lending for all LVR buckets where possible. 23. Banks that wish to take advantage of the exemption will need to furnish management reporting data on their construction lending flows, similar to that already provided as part of the LVR new commitments exemption reporting. Interim reporting measures will be in place until April 2013, during which time banks are encouraged to provide indicative data on construction lending, should their systems not be ready to provide finalised data. Banks will have the opportunity to re-submit construction lending data for the first LVR measurement from 1 October-31 March up until 16 April, and the final submission will count towards their speed limit calculation over the period. 24. The exact form of the data returns relating to non-mortgage consumer debt commitments to mortgage customers is still to be decided. One option would be to extend the new commitments template to include statistics on the borrowing limit of the non-mortgage consumer debt lines that are available to customers who are receiving new mortgage commitments, and the extension of new non-mortgage consumer debt facilities to existing mortgage customers. The Reserve Bank would monitor these data for signs of growth in this class of debt for mortgage customers that is not matched by more general growth in the market at large. The LVR bands would allow us to see if the debt is mainly being acquired by customers who are near or above the threshold for being a high LVR customer. 3.0 Risks of new definitions generating avoidance activity 25. The final BS2A/BS2B definitions adopted mean that the non-mortgage debt of a mortgage customer does not generally need to be included when computing a customer’s loan value and loan-to-value ratio. The Reserve Bank has made this change (relative to the initial Housing Review Stage II consultation on this topic) in response to feedback from the banks. Amongst the points raised by the banks were that credit card and personal loans were already treated as unsecured lending (and Ref #5558957

7 risk-weighted accordingly) so their inclusion in the calculation of the LVR could lead to a form of “double counting”, and the risk that the credit card business of mortgage customers would simply be shifted to other banks. Several banks noted that including unsecured lending as part of the loan was unnecessary, as they would generally be unwilling to aid customers who were seeking to raise a deposit by borrowing unsecured funds from the bank before obtaining a mortgage. 26. Banks further suggested that the objective could be achieved without the inclusion of unsecured lending in the LVR calculation by other means. Examples included the obligation on banks to abide by the spirit of the LVR policy, suggestions for deposit identification/verification requirements, and a proposal that the Reserve Bank track the development of other loan commitments by extending the reporting requirements under BS19. A bank noted that it would be relatively easy and appropriate for the banks to provide data returns to the Reserve Bank that demonstrated that the non￾mortgage debt of mortgage customers was not excessive relative to similar debt for other customers. 27. For the purposes of loan-to-value restrictions, the Reserve Bank is interested in ensuring that the non-mortgage debt provided to mortgage customers does not (either directly or indirectly) provide material assistance in generating a deposit for a house purchase. A second issue is that the Reserve Bank wishes to avoid a situation where non-mortgage consumer debt is used excessively by existing mortgage customers in order to avoid ‘topping up’ their residential mortgage. “Excessive” here means that the debt is being provided is in a quantity and/or on terms that would be outside normal credit origination practices or would be effectively relying on the all￾obligation mortgage to be justifiable origination practice. 28. The Reserve Bank’s preferred option for achieving these policy goals leverages banks’ suggestions, by strengthening the anti-avoidance provisions of BS19, and introducing a deposit identification/verification requirement. 3.1 Strengthening of anti-avoidance provisions 29. Two items would be added to the list of potential avoidance measures in the anti￾avoidance section of BS19: a) “providing other lending that is not for business purposes, such as consumer loans or credit cards, to a borrower who has a residential mortgage loan with that same bank where the terms of the loan are substantially different to that which would apply under normal risk management practice or are determined in reliance of the security provided by an all-obligation mortgage”; or b) “providing a residential mortgage loan to a borrower where the bank is aware that the borrower’s deposit for the purchase of the residential property to which the residential mortgage loan relates, has been, or will be, funded by unsecured lending, including credit cards, provided by that bank in the case where, had the unsecured lending been part of the residential mortgage loan, the residential mortgage loan would have counted as a high-LVR loan for the bank”. 30. This preferred policy approach provides banks with substantially more flexibility in handling non-mortgage debt of mortgage customers than the originally contemplated concept of loan value for BS2A/BS2B. Banks are able to continue to offer products to mortgage customers without them counting against the mortgage LVR, so that they Ref #5558957

8 are not disadvantaged (relative to other banks) in competing for that customer’s business. However, in order to avoid compromising the objectives of the LVR policy, the anti-avoidance clause requires banks to limit the additional non-mortgage consumer debt they provide to mortgage customers so that the latter does not become a substantial channel for the avoidance of LVR restrictions. The Reserve Bank proposes to monitor this by obtaining some additional data from the banks. 31. The exact form of the data returns relating to non-mortgage consumer debt commitments to mortgage customers is still to be decided. One option would be to extend the new commitments template to include statistics on the borrowing limit of the non-mortgage consumer debt lines that are available to customers who are receiving new mortgage commitments, and the extension of new non-mortgage consumer debt facilities to existing mortgage customers. The Reserve Bank would monitor these data for signs of growth in this class of debt for mortgage customers that is not matched by more general growth in the market at large. The LVR bands would allow us to see if the debt is mainly being acquired by customers who are near or above the threshold for being a high LVR customer. 32. One exceptional class of non-mortgage consumer debt that would be included in the LVR for BS2A/BS2B and thus BS19 purposes is lending that is not branded as a mortgage but is offered on terms that reflect it being ‘effectively’ secured by the mortgage. These products would have posed difficulties for the BS19 policy if they were not included in loan value. These products are not understood to be common in the market, but could have become more common if they were a useful alternative to high LVR mortgages. Because they are reflected in the mortgage LVR, the proposed anti-avoidance clause is not intended to restrict their provision. 33. In addition to the draft changes to BS19 outlined above, section 7 (Anti-avoidance) has been amended to reflect that arrangements designed to circumvent LVR restrictions through funding of the deposit by unsecured lending would not be considered legitimate under any circumstances. 34. Presuming that the Reserve Bank finalises the revisions to BS19 along the lines suggested above, the additional anti-avoidance clauses will be intended to apply immediately. 3.2 Deposit identification/verification requirement 35. A deposit identification/verification requirement would require banks to check that the loan deposit was not being funded through other credit card or unsecured lending at the (same) bank, and thus artificially holding the LVR below the LVR restriction threshold. Submissions to the Housing Review Stage II consultation noted that deposit identification/verification requirements were already a part of bank credit policy, suggesting that such a requirement would be relatively easy to meet. 36. A regulatory requirement around deposit identification/verification could be implemented in a number of ways: a) by adding a new guideline to BS19, which would set out the expectations around deposit identification/verification of the source of deposit for a residential mortgage loan, b) by adding a requirement in the conditions of registration that banks follow the process requirements set out in BS19, and adding a process requirement regarding deposit identification/verification, c) by adding a condition of registration relating to deposit identification/verification. Ref #5558957

9 37. The Bank proposes the first option. This is a light-handed approach that would be relatively simple to implement, and allows banks discretion in how they meet the requirement. At this point, the Reserve Bank does not see a need to take a stronger approach to enforcing guidelines around banks’ processes, as banks appear to be meeting the spirit of LVR restrictions. 38. Draft changes to implement the first option have been made in the accompanying BS19 chapter of the Banking Supervision Handbook. These changes can be found in the following sections: • s7 – Anti-avoidance: this section has been amended to proscribe the circumvention of LVR restrictions through funding of the deposit by unsecured lending, as described above • s14 – Guidelines – verification of third-party information: this section requires the bank to determine whether the borrower has sourced the deposit from unsecured lending at the same bank. 3.3 Other anti-avoidance options 39. Should the consultation reveal that the proposed anti-avoidance clause and data collection are not workable, an alternative course of action would be to formally redefine loan value within BS19, making it a broader concept than BS2A/BS2B loan value. The most natural possibility would be to define BS19 loan value as BS2A/BS2B loan value, plus (if not already included) other non-mortgage consumer debt owed to the bank by that customer. As stated above, this is not the Reserve Bank’s preferred option at this stage. 4.0 Questions 4.1 Construction lending exemption On the magnitude and characteristics of construction lending (for banks that have not already reported on these questions): a) What share of your new residential mortgage lending is residential construction lending (by value and by number?) What share of your residential construction lending is high-LVR (by value and by number?) b) What is the risk profile of your bank's residential construction lending relative to your other housing lending? What are your current risk management practices around construction lending? c) Do you charge a risk premium for high-LVR residential construction lending, relative to a) low-LVR construction lending b) other types of high-LVR housing lending? d) How does the asset quality of your construction lending portfolio compare to your other housing lending? On the proposed construction lending exemption: a) Do you foresee any unintended consequences of the proposed exemption? b) Are you likely to take advantage of such an exemption? If so, how much effort do you expect it to involve? Ref #5558957

10 c) Do you think the proposed changes are the most practicable way of framing an exemption on high-LVR residential construction lending? d) Are there other boundary cases on which it would be useful to provide guidelines? e) Do you think the proposed drafting of this exemption in BS19 achieves what it is intended to? f) Do you think the proposed exemption is workable from an operational point of view? g) Do you have any suggestions that might help make such an exemption easier to operationalise? h) Do you think the definition of ‘start of build’ is sensible? If not, is there another definition that would be more workable? 4.2 Risks of avoidance activity a) Do banks ever rely on the security provided by the all-obligation mortgage when providing consumer debt, or offer better terms on consumer debt to mortgage customers? Is it appropriate that BS19 notes that this could be construed as avoidance of BS19? b) Would banks’ credit origination practices typically reveal it if a customer has used other unsecured credit from that bank in order to generate a mortgage deposit? Are there any issues with BS19 noting that this could be construed as avoidance activity? c) Will it be possible for banks to provide data on the non-mortgage consumer credit originated for new and existing mortgage customers? 4.3 Deposit identification/verification requirement a) Does your bank have deposit identification and verification processes in place? If so, what are they? Are customers required to attest to the source of deposits? b) What would be involved in implementing a deposit identification/verification requirement? Time/dollars/resource constraints? c) What would be a sensible common deposit identification/verification approach across the banking system? d) Are the proposed changes sufficient to ensure that banks will not knowingly allow customers to use non-mortgage consumer debt to generate a deposit, or is one of the stronger options we have outlined (but not listed as a preferred option) more appropriate? e) Are there other ways that this could be addressed? Any other comments? Ref #5558957