Grant Spencer: Getting the best out of macro-prudential policy
Speech by Mr Grant Spencer, Deputy Governor of the Reserve Bank of New Zealand, to INFINZ, Auckland, 13 March 2018.
The views expressed in this speech are those of the speaker and not the view of the BIS.
Introduction
It is five years since we first introduced macro-prudential policy in the form of loan-to-value ratio (LVR) restrictions in October 2013. It is time now to take a step back and review the experience with this new policy framework.
Indeed we are committed to undertake such a review under the Memorandum of Understanding (MoU) on Macro-prudential Policy that was signed with the Minister of Finance in 2013. The review will be undertaken jointly with Treasury as part of the Government's broader review of the Reserve Bank of New Zealand Act.
You will hear a wide range of views on the LVRs from a number of perspectives - some favourable and some less so. From the Reserve Bank's perspective, the LVRs have assisted us in achieving our economic and financial stability objectives. The LVRs have reduced housing-related risk in the banking system and also helped to ease housing market pressures. The policy has been particularly useful in the current environment of globally low interest rates, where domestic monetary policy has been unable to counter the strong housing cycle of recent years.
But do we have the most appropriate macro-prudential framework for our circumstances in New Zealand? What role could the other instruments in our toolkit play to enhance and/or complement the LVRs? Should we be considering other macro-prudential instruments? Is the governance structure sensible? There are diverse views on these questions.
With the review of the macro-prudential framework ahead of us, my intention today is to consider the key issues that might be addressed and to offer some thoughts, based on our experience, for options to improve the policy framework. I have been closely involved in the Reserve Bank's development and implementation of macro-prudential policy and I am keen to see macro-prudential policy continue to develop as a credible and sustainable policy over the longer term.
I will first discuss the New Zealand and international experience with macro-prudential policy, and some of the lessons we have learned. I will then make some observations and suggestions aimed at putting macro-prudential policy on a more systematic and sustainable footing. These are my own views and are not a formal Reserve Bank position.
The New Zealand Experience
As set out in the MoU, the objective of our macro-prudential policy is to increase the resilience of the financial system to credit, asset price and liquidity shocks, with a secondary outcome of moderating the credit cycle. The four macro-prudential instruments - the countercyclical capital buffer (CCB), sectoral capital overlays, the core funding ratio, and limits on high LVR mortgage lending - are designed to provide additional buffers to the financial system on top of the baseline prudential settings in periods of cyclical pressure. Given the potential impact and public profile of the framework, the MoU requires the Reserve Bank to consult with the Government prior to the use of the macro-prudential instruments. Accountability has been supported by our regular reporting on macro-prudential developments in our biannual Financial Stability Reports.
The first set of LVR limits in October 2013 was motivated by our concern about increasing risks in the housing market and the associated trends in high-LVR lending by banks. Our view was that increasing household indebtedness and a build-up of housing credit risk on banks' balance sheets, combined with an increasing degree of stretch in house prices, was making the financial system increasingly vulnerable to a sharp correction in the housing market. The initial response of the market to this first round of limits, in terms of housing credit growth and house price inflation, was encouraging. But further pressures and growing risks, particularly in Auckland, led the Bank to refine the LVR policy in 2015 and again in late 2016. More recently, the moderation in the housing market and associated financial stability risks have allowed us to begin a staged easing of the policy, although we remain cautious in the face of ongoing housing shortages.
Based on this experience, our assessment is that the LVRs have had transitory effects on the rate of house price inflation but have had sustained beneficial effects on the resilience of banks' and households' balance sheets. The share of outstanding mortgages with an LVR greater than 80 percent has steadily trended down as a result of the LVR policy, from 21 percent in September 2013 to 7 percent in December 2017. Recent Reserve Bank estimates suggest that banks' credit losses from a severe housing downturn could be reduced by around 20 percent.1 Also, with larger equity buffers, fewer households would be under pressure to sell their house or cut back on consumption in such a downturn.
Overall, I believe the LVR policy is regarded as a qualified success. While it has its critics, it has also won considerable support amongst policy makers, the public and the banks. The banks have recognised the benefit of an external industry constraint that prevents an escalation of mortgage risk as banks compete for market share through greater risk taking. I take five broad lessons from the experience with macro-prudential policy to date.
First, it is a policy that can reliably improve banking system resilience, on a sustained basis, but that has limited capacity to influence asset prices. The latter should not surprise us given the wide range of factors that influence asset prices more generally.
Second, and arising from the first, macro-prudential policy should always have a prudential purpose, and be modest in its ambition. It should not seek to control housing or financial cycles. It should seek to build resilience to these cycles while at the same time lean against the extremities of those cycles.
Third, to improve transparency and understanding, macro-prudential policy should have a well-understood and systematic framework for policy adjustments that is fully consistent with the underlying prudential framework.
Fourth, there is a need to build and sustain support for the macro-prudential framework - amongst politicians, the public and the banks themselves. To succeed long term, macro-prudential policy must gain broad acceptance, along the lines of monetary policy. Part of the challenge here is that many observers assume the main objective of macro-prudential policy is to stabilise asset price and credit growth rather than build financial system resilience.
Fifth, a good infrastructure is required to support macro-prudential policy. The Reserve Bank has invested heavily in better measurement and analysis of housing and rural credit risk. This has had positive spin-offs for prudential policy more broadly.
International Experience
The countries we look to for lessons about our own policy are those with similar banking systems and with a propensity for housing cycles - such as Australia, Canada, Ireland, and the UK. The latter three have been active in macro-prudential policy for several years while Australia has been a more recent starter. The experience of these countries tends to reinforce the lessons I have mentioned.
In Ireland, the Central Bank sees borrower-based macro-prudential policies as an integral part of mitigating systemic risk from housing. Over time the Irish policy has been tailored to address both the collateral (LVR) and serviceability (loan to income ratio (LTI)) dimensions of borrower risk, with separate limits on each dimension for different categories of borrowers. The flexibility of the policy means the Central Bank can proactively respond to risks it sees in different segments of the mortgage market. The CBI commits to annually publish a review of the effectiveness and calibration of its macro-prudential policy.2
In the Bank of England's case, they place a limit on high LTI lending. This is seen as an 'insurance' policy, with the calibration currently set at a non-binding level. The aim of the measure is to prevent a deterioration in lending standards, not necessarily to actively lean against mortgage lending at present.3 In a similar vein, the UK Financial Policy Committee's approach to setting the CCB is to maintain it at a positive, neutral level, and to stand ready to either raise or lower it in line with the Committee's assessment of aggregate risks to the system.4 Further the CCB is seen as an important signalling device - underlining the FPCs view on the evolving shape of systemic risk in the UK system.
We should also consider the evidence and analysis of the IMF and other international bodies which have reviewed macro-prudential experiences across a number of countries.5 While few macro-prudential regimes have been tested through a full financial cycle, evidence suggests that macro-prudential policies have limited success in leaning against asset price and credit growth cycles, particularly during strong upswings. The key benefit of capital and borrower-based macro-prudential tools is from the buffers they create which help to build system resilience against potential shocks.
Regarding the shape of operating and governance frameworks for macro-prudential, there is no clear convergence on operating models and the sharing of responsibilities between different parts of government. However, there is a growing consensus on the desirability of:
- a clear assignment of the macro-prudential mandate and powers;
- well-defined policy objectives that are modest as to what macro-prudential policy can and should achieve;
- the need for high quality analysis and monitoring of systemic risks to support decision-making; and
- transparency and accountability mechanisms that support policy legitimacy.
Taken together, these elements of a macro-prudential framework will tend to support both an ability and a willingness to act on the part of the macro-prudential authority.
It is also important to clearly articulate the processes underlying the policy framework, including the links between the tools, the intermediate targets of a policy intervention, and the ultimate stability goal. In doing so, macro-prudential authorities can improve public understanding and build an enduring constituency for financial stability.6
The IMF's recent Financial Sector Assessment Programme (FSAP) benchmarked the New Zealand macro-prudential framework against their view of emerging best practice.7 The report noted approvingly the Bank's clear mandate for financial stability, independence of the Bank's decision-making process, and a demonstrated willingness to act. Key recommendations of the report were to further clarify the respective roles of the Bank and the Minister when considering variations to the framework, such as inclusion of a debt to income (DTI) instrument in the toolkit, and to continue to enhance accountability mechanisms.
What does all this mean for the New Zealand framework going forward? I will first discuss the policy instruments and then turn to the governance framework.
The Policy Instruments
I doubt there is a single set of instruments that will always be relevant to meet evolving macro-financial risks. However, based on our own and international experience, I expect that both LVRs and some form of debt servicing instrument will become part of the macro-prudential "furniture" over time. They represent the two key risk factors in mortgage lending, and I expect housing cycles will continue to be an ongoing (but variable) source of systemic risk in the NZ financial system.
The Counter Cyclical Capital Buffer (CCB) is likely to remain as part of the Basel 3 framework and is commonly used internationally, including as a signalling device. While it has not yet been activated in NZ, it could be given closer consideration when the Reserve Bank comes to look at minimum capital ratios as part of its broader review of bank capital requirements. The Core Funding Ratio (CFR) is an existing (micro) prudential policy instrument. It currently requires banks to have at least 75% of their funding in "Core" instruments such as retail deposits. While not yet used as a macro-prudential tool, it could easily and usefully be activated in response to an increase in system-wide liquidity risk. The Sectoral Capital Ratio (SCR) has not been used to date but would seem a potentially useful additional buffer against increasing risk in a particular sector such as housing or dairy.
Whatever instruments are relevant for the toolkit through time, I believe the upcoming Review should consider a structure for macro-prudential instruments which allows them to be "off' or "on" through time, but where they remain established within banks' reporting and compliance systems. These instruments relate to risk metrics that we expect banks to keep track of, whether or not they are subject to ceilings or floors. Macro-prudential policy would adjust back to 'neutral' or non-binding settings when no heightened risk is present and no policy constraint is intended. This sort of structure already exists for the CFR and the CCB, since changes to these instruments would overlay existing core funding and capital requirements. Neutral settings for the LVRs and other instruments would be a matter for future analysis and consultation.
With 'neutral' settings, banks would not need to remove/add the operational frameworks as policies are turned on and off. Rather they would just need to adjust the settings of existing policies. The public would become increasingly familiar with the policies and the potential for changes to settings. Policy changes would become easier to implement and to communicate. In principle, macro-prudential policy would follow a more transparent and systematic process that looks more like our monetary policy process. Such an approach would be more amenable to the analysis and assessment necessary for effective policy accountability.
While we stated at the outset in 2013 that LVRs would be temporary, I believe there is a case to consider maintaining a policy infrastructure of this sort, with policies being adjusted through time between binding and non-binding settings. This is similar to the 'insurance' approach adopted by the Bank of England where a loan to income ceiling has been set at a non-binding level. One aspect that would need to be considered carefully before moving in this direction would be any increased risk of avoidance activity.
Regarding the contents of the policy tool-kit, the mix of instruments should slowly evolve through time. In the present economic and financial environment I support retention of the four instruments listed in the MOU (LVRs, CCB, CFR, SCR) and the addition of a carefully designed debt to income (DTI) or debt servicing instrument. You will recall that the Bank consulted on including a DTI instrument in the toolkit last year. That process was overtaken by the general election and a decision on inclusion or otherwise was deferred to the upcoming Macro-Prudential Review. The DTI is a natural complement to the LVR, focused on reducing the risk of borrower default. Many macro-prudential authorities overseas view some form of debt servicing ratio as a key anchor and safeguard for macro-financial stability. The Review should give serious consideration to adding such an instrument to the toolkit.
Governance
The current macro-prudential MoU gives decision making powers to the Governor, but commits the Reserve Bank to prior consultation with the Minister and Treasury. The development of any new instruments is undertaken in consultation with the Treasury and also includes public consultation. In the case of policy adjustments, consultation with the banks is required ahead of implementation. So the present governance model for macro-prudential policy has the Reserve Bank in the driving seat, but requiring extensive consultation with Government and the public, in particular when considering new instruments.
In the upcoming review of macro-prudential policy, and of the Reserve Bank Act, it will make sense to integrate the intent of the current MoU into the Reserve Bank Act. At present, macro-prudential policy operates under the same high-level financial stability objective, and relies on the same prudential powers, as conventional prudential policy. This has worked effectively to date and I believe it is important to retain this common ground between the two policies, in particular the same high level prudential purpose. However, the Review will be taking a fresh look at the detail of the governance architecture, including operational objectives, policy guidance, prudential powers, decision making and accountability mechanisms. In working through these areas it will be important to allow for the differences between micro and macro-prudential policy.
What are these differences, and how might they be reflected in a revised governance structure? I see two key differentiating characteristics of macro-prudential policy: first, macro-prudential only addresses significant systemic shocks, not idiosyncratic day-to-day shocks; and second, macro-prudential has a stabilisation role (albeit a secondary one) as it seeks to moderate cyclical extremes in credit growth and asset prices.
Thinking about where the micro/macro-prudential differences might be reflected in a revised Act, the first is in objectives. Current objectives for the Reserve Bank's broad financial stability role (sections 1A and 68 of the Act) are focussed on "promoting the maintenance of a sound and efficient financial system". The Bank is also required to have regard to any statement of Government policy on the Bank's functions issued by the Minister (section 68B). The objectives and policy guidance for macro-prudential are consistent with this but more specific. The current MoU objective is:
"to increase the resilience of the domestic financial system and counter instability in the domestic financial system arising from credit, asset price or liquidity shocks."
Guiding principles for macro-prudential in the MOU include the principle of supporting monetary policy where possible and being fully consistent with micro prudential policy. The objective and guidance from the MoU should be lifted into the Act and potentially enhanced, for example with the principle that macro-prudential policies should be applied uniformly across all banks.
Further elements of the governance regime might include: a list of areas that macro-prudential policies could be applied to8, e.g. capital, loan to value ratios, debt servicing capacity; and potentially a PTA-like document that would express operational objectives agreed from time to time with the Government of the day. I expect the technical details of the macro-prudential instruments would continue to be set out in the Banking Supervision handbook.
The specific characteristics of macro-prudential that I have mentioned ̶ the common ground with micro-prudential policy and its stabilisation role ̶ point to the macro-prudential mandate remaining with the Reserve Bank. There are significant synergies across the micro and macro-prudential functions, and the stabilisation aspect requires a macro-analysis capability and degree of coordination with monetary policy. The IMF in its 2017 FSAP report underlines this point and supports the Reserve Bank having responsibility for macro-prudential - but with full accountability. They point out that shared responsibility models, where macro-prudential decisions are made jointly by councils (e.g. made up of the ministry of finance, central bank and regulatory agencies), often suffer from inaction bias.
An important further aspect of governance is decision-making. Given the planned introduction of a new decision making committee (MPC) for monetary policy, the Review should consider establishing a financial policy committee (FPC) for decisions relating to both micro and macro prudential policy. The Reserve Bank has supported a two-committee (MPC/FPC) model in place of the current single Governing Committee, for example in the Bank's 2017 "Briefing for Incoming Minister".9
I would expect an FPC to have overlapping membership with the MPC, at least including the Governor. However, it would likely include internal financial policy experts who do not sit on the MPC. Externals could be included on the FPC, although the complexity of the subject matter, the high volume of prudential decisions, and the need to avoid conflicts of interest would severely limit the number of viable candidates. Unlike in the UK where there are separate micro and macro-prudential committees, it would be sensible in the New Zealand context to have a single FPC that covers all prudential/regulatory policy issues. This would help to ensure consistency between the micro and macro-prudential policy frameworks. The FPC would be responsible for all decisions involving changes to the prudential/regulatory framework and also give guidance to the Governor on the shape of the Reserve Bank's supervisory framework. Potentially the FPC could be a sounding board for important supervisory actions with respect to individual financial institutions. However, such operational decisions can be highly complex and, in stress situations, very time-critical. They should therefore continue to be the final responsibility of the Governor.
The FPC would need to be held accountable for its micro and macro-prudential decisions. This would continue to be achieved through scrutiny by the Reserve Bank board on behalf of the Minister; by the Parliament through FEC; and by the public and financial markets through policy consultations and transparent reporting. The six-monthly Financial Stability Report would continue to be an important accountability document, supported by speeches and, potentially, published FPC minutes.
As I mentioned, the information infrastructure will play a crucial role. There have been significant improvements in the information collections since the advent of macro-prudential policy in 2013. The increasing richness of data and analysis will continue to play a crucial role in underpinning macro-prudential decision making and the associated accountability structure.
Conclusion
In conclusion, there is a broadly held view that macro-prudential policy has been a useful addition to the financial policy toolkit since its introduction in 2013. While having limited scope to sustainably influence credit and asset price cycles, macro-prudential rules have significantly improved the resilience of banks' balance sheets to potential housing market shocks.
With the upcoming Review of the Reserve Bank's financial policy framework, it is timely to consider ways in which macro-prudential policy may be improved and more broadly accepted. We need to get the best out of macro-prudential policy and position it as a lasting policy framework.
We have close to five years of experience to draw on. Key lessons, reinforced by the international experience, are that macro-prudential policy must be modest in its ambitions and always have a prudential rationale. Macro-prudential policy cannot be used simply to try and manage the housing cycle. It must be applied consistently and backed by sound analysis, with transparent reporting and accountability.
In anticipation of the Review I have presented some thoughts on how the macro-prudential framework may be improved, through a more systematic approach to policy adjustments, and a clearer articulation of the governance structure within the Reserve Bank Act. Such changes, including potentially a new Financial Policy Committee (FPC), should assist in gaining public understanding and help to put macro-prudential policy on a sound footing for the future.
Macro-prudential is a very new policy framework that has filled an important policy gap. It needs to be given the best chance of future success.
1 See Box A in the November 2017 Financial Stability Report. Since banks' risk weights are designed to be proportionate to the riskiness of their lending, the decline in the high-LVR share has reduced the regulatory capital buffers held against these loans. To counter this, the Reserve Bank has separately increased capital requirements for high-LVR and investor mortgages since 2013.
2 See Central Bank of Ireland (2017), Review of residential mortgage lending requirements.
3 See, for example, Bank of England (2017), Financial Stability Report, November 2017.
4 See Bank of England (2016), The Financial Policy Committee's approach to setting the countercyclical capital buffer.
5 See IMF-FSB-BIS (2016) Elements of Effective Macroprudential Policies: Lessons from International Experience.
6 See, for example, CGFS (2016), Objective-setting and communication of macroprudential policies, and Gai (2017), The Design, Implementation and Governance of Macroprudential Policy.
7 IMF (2017), New Zealand Financial Sector Assessment Program: Technical note - Macroprudential institutional framework and policies.
8 Along the lines of section 78 in the current Act.