Speech Macroprudential Policy: A Suite of Tools or a State of Mind?
Head of Financial Stability Department
Paul Woolley Centre for the Study of Capital Market Dysfunctionality Annual Conference
I would like to thank the Paul Woolley Centre for inviting me back to speak to this conference. I am particularly grateful to have the opportunity to do so in a session chaired by Ian Macfarlane. Ian was instrumental in setting the tone for the way we at the Bank think about financial stability policy today. His example helped us take the historical insights of authors like Charles Kindleberger and transform them into a practical approach to analysis.
The financial crisis has spurred a huge, multifaceted regulatory response. I know Jim Murphy is going to talk about some of the reforms that have actually happened or are in train. I would like to talk about just one element of that response: the calls for a so-called ‘macroprudential policy framework’. Development of such a framework is at a rather early stage, but it's attracting a lot of attention and energy. For example, following their meeting in Seoul in late 2010, G-20 Leaders called on international agencies to do more work on the subject, which they have done. But what is a macroprudential policy framework? Is it separate from financial stability policy? Does Australia really need a whole new framework?
As background for the IMF's recent Financial Sector Assessment Program review of Australia, APRA and the Bank prepared a paper on how we think about financial stability policy and its macroprudential dimension. In essence, we see ‘macroprudential’ policy as being subsumed within the broader policy framework for promoting financial stability. That seems to be something of a minority view internationally. So we published the paper last month, after the IMF visits had concluded, to put our take on this subject into the public domain.
Today I would like to give a little more detail about our thinking on the issue of macroprudential frameworks, as set out in that joint paper. After giving a brief history of the term ‘macroprudential’, I will organise my remarks around three assertions, which are consciously intended to be a little more sceptical about the idea that a whole new framework is needed than some international commentary has been. In doing so, I hope to provide a bit more clarity about the international policy debate.
First, macroprudential policy is only a subset of the policies intended to enhance financial stability. For a start, it is not really designed for managing crises when they do occur. It does no good to muddy the waters by claiming things as macroprudential for the sake of it.
Second, most supposedly macroprudential policy tools are in fact the usual prudential tools long used by ostensibly ‘micro’ prudential supervisors. What is ‘new’ is the motivation behind their use. Even that is not particularly new in many countries, including Australia.
Third, the build-up to the recent crisis resulted more from a microprudential failure than a macroprudential one. The easing in US mortgage lending standards, the growing reliance on short-term wholesale funding, the low risk weights applied to complex and highly leveraged structured securities were all things that an avowedly microprudential supervisor could have – and arguably should have – noticed and responded to.
Certainly, a more holistic, or system-wide, perspective could help supervisors see if risks are building up. This could happen because many individual institutions are doing the same risky things. Or it could happen because particular risks have become concentrated in a few institutions. But that macro perspective does not necessarily require a whole new institutional or policy framework for regulation. It possibly might not even require new policy tools. I would argue that macroprudential policy is indeed more of a state of mind than a suite of tools. I think it would be a mistake if a focus on macroprudential frameworks were to occur at the expense of much-needed reforms to normal prudential policy arrangements, both at the international level and in some of the jurisdictions overseas where the crisis began.
It would also be a pity if a false dichotomy between macroprudential and microprudential supervision were to be drawn. The two perspectives do not need to be pitted against each other. They can both be parts of one, holistic, view of financial regulation and supervision, which is how we see it in Australia.
What is Macroprudential Policy?
The term ‘macroprudential’ first appeared in internal documents of the precursor to the Basel Committee in the late 1970s. The BIS and others started using it publicly by the mid 1980s. It isn't a new term; neither are the concerns behind it. The underlying idea was that prudential supervisors should take a system-wide view in the way they supervise. They should recognise that the actions of individual firms can collectively generate systemic risk, even if those actions are individually rational. They should recognise that risk can build over time, and that the distribution of risk can matter. And according to this view, they should avoid focusing narrowly on the safety of individual institutions without regard to the feedback from those institutions' behaviour to the wider system.
So the original idea behind ‘macroprudential’ policy was a sensible one. It was about how prudential supervisors should do their job and the perspectives they should have. The policy tools are the tools of prudential regulation and supervision. I would go as far to say that what others think of as macroprudential supervision, the Australian authorities consider simply to be competent supervision.
Not all Financial Stability Policies are Macroprudential
Not every policy that helps financial stability can be reasonably labelled macroprudential. It would be needlessly confusing to shovel everything into that category. Why not just call it financial stability policy, after all? For example, central banks' provision of liquidity to the market is as old as central banking itself. It is clearly helpful for financial stability, both before and during a crisis. But is it macroprudential? I have heard some claim so, but I don't see it that way.
We should also remember that not every policy that helps financial stability was necessarily enacted with that purpose in mind. Sometimes, policies are worthwhile for completely different reasons, but happen to be helpful for financial stability as well. A good example of this is consumer protection regulation around the provision of credit. Consumers generally have less information and expertise in financial matters than the firms that lend to them. Regulators try to redress this imbalance in regulation. For example, they might specify what lenders must tell consumers about the loans they offer, or impose more general requirements that the loan product must be suitable for the borrower. This is seen as worthwhile because it prevents possibly vulnerable people from being ripped off or simply from unwittingly making a decision that is not in their interest.
A well-designed consumer protection regime can be very useful for financial stability, even though that is not its primary intention. Australia's National Consumer Credit Code requires that the lender must be able to show that the consumer can be reasonably expected to repay the loan from their own resources, without having to sell the collateral. (So did the state-level Uniform Consumer Credit Code that preceded it.) If not, the loan can be modified or even set aside by the courts. Lenders then have incentives to extend credit responsibly. They should especially want to avoid the asset-based lending and ‘no-doc’ lending that caused so many problems in the United States in the lead-up to the crisis. It's no surprise that the US regulation was not uniform, and did not cover all borrowers.
Macroprudential Tools are Prudential Tools
As I mentioned before, the original idea behind macroprudential policy is that it is, in essence, prudential policy. Many of the so-called macroprudential tools currently being discussed in international forums are the usual tools available to a prudential supervisor – liquidity rules, risk weights, capital standards, large exposure limits and so on. They might be calibrated differently, and the reasons motivating their use might differ. It is in this sense that macroprudential policy seems new. But in most cases they are the same old policy tools, just as the original proponents at the BIS intended.
Part of the confusion in recent international discussions seems to lie in how some observers think prudential supervisors operate. It has become common for observers overseas to refer to bank supervisors like APRA as ‘microprudential’ regulators. A distorted view of how supervisors do their work has arisen, one I consider quite unhelpful. The misconception is that supervisors charged with ensuring the safety and soundness of banks are thought incapable of looking beyond the individual bank they supervise. It's such a caricature of prudential supervisors that the best way I can think of to explain why it's a misconception is to show you a cartoon (Figure 1).
The view of some observers seems to be that a ‘microprudential’ supervisor is like this self-satisfied looking fellow, complete with a capital adequacy report on his desk. He is saying, ‘The bank I supervise is well capitalised, and that's all I have to care about!’
It might well be that there are supervisors who think in this narrow way. But I labelled the cartoon as the ‘microprudential nutter’ because it reminds me so much of a concept popularised by Sir Mervyn King (King 1997). He made the point that even if a monetary policymaker cared only about inflation – an inflation ‘nutter’, as he put it – monetary policy would still respond to output, because output tells you something about future inflation. Likewise, any sensible prudential supervisor knows that the number one cause of bank failure is a systemic crisis, and that the safety of the bank they supervise depends on the environment around it. Even a microprudential ‘nutter’ would look at the big picture. So I think a better cartoon version of a good prudential supervisor looks something like Figure 2.
It is in this sense that I would argue that macroprudential policy really is about the state of mind, not the suite of tools. The tools are already there. It is in how they are meant to be employed that the differences between micro and macro arise.
In fact, I would dispute that there is much different in macroprudential policy from supervision as it is currently practised in many countries. Like the old joke about people discovering that all these years they had been writing prose, I suspect that a lot of bank supervisors are discovering that they are really macroprudential, and have been all along.
For this reason, I would caution against efforts to invent too many ‘new’ and untried tools, particularly ones designed to be manipulated over the cycle, much like monetary policy. One such tool we've been hearing about from other jurisdictions lately is to set a cap on loan-to-valuation ratios (LVRs) for mortgages, and then vary that cap from time to time. Obviously it is prudent to have some sort of limit to the leverage in collateralised loans such as mortgages. The alternative would be the situation in the United Kingdom or the Netherlands, where one could borrow as much as 125 per cent of valuation and, more importantly, where many people actually did so, not just a small fringe. I don't think that is good practice. People need to provide some deposit when they buy a home. It protects them if something goes wrong for them, like a job loss or illness, especially if it happens at the same time that housing prices are falling.
The question is whether policymakers should go further and set maximum ratios well below 100 per cent, and whether that maximum should be varied periodically. This might be sensible for the jurisdictions that have adopted such a policy. For some of them, their exchange rate regimes constrain them from raising interest rates instead. But that is not relevant in Australia. I also have a number of conceptual reservations about an LVR cap. These lead me to question whether this kind of product regulation is a suitable prudential tool in the Australian environment.
The first issue is that this kind of policy is only relevant for home mortgages. Business lending can be more easily structured around such a restriction, especially the property development loans, which are particularly risky. And for unsecured lending, it's not even relevant. In any case, high LVR loans are still available to at least some borrowers in jurisdictions that set this cap. In Hong Kong, the excess over the cap is covered by the government-sponsored mortgage insurer. This is not so different from Australia's prudential framework, which requires higher risk weights on high LVR loans, especially those that are not insured.
The second issue is that the cap would have to be set very low to be binding on existing home buyers who are trading up. First home buyers would be squeezed out, but most buyers would be little affected. Third, the cap would not prevent boom-bust cycles in housing prices. The evidence from overseas is that it instead limits the increase in arrears rates that occurs when the bust comes. That is because the borrowers who get into difficulty are more likely to have some equity, even if prices fall. They can therefore sell rather than default. That's not a bad thing, but in the end our responsibility for financial stability is about protecting the real economy. Protecting the banks from the real economy is not the ultimate goal.
I also worry that focusing too much on the LVR would be a mistake. Lenders that do that are engaged in asset-based lending, which is exactly the kind of practice we would want to avoid. There are many other dimensions of lending standards beyond LVRs. In my view, the borrower's ability to repay is more important than the collateral. Lenders also need to think about how certain they are about the valuation in that loan to valuation ratio. One of the things that went wrong in the United States during the boom was that the valuations behind the lending decisions were often flawed, or even fraudulent.
The Current Crisis Partly Stemmed from a Microprudential Failure
As with many other reforms that have been proposed since the crisis, new macroprudential policy frameworks are being advocated with the view that these might help prevent similar crises in future. But would they? Well, maybe. I am not convinced that they would be the most effective approach. If we think about what went wrong in the United States in the lead-up to the crisis, it was in essence that a lot of bad assets were created – in this case mortgages and the securities based on them – and that these bad assets were distributed around the global financial system, primarily to entities that were funding themselves at very short terms. How was this possible? Obviously there were lapses in risk management and governance at the banks and other firms involved. In part, though, it was also a failure of prudential supervision that enabled those lapses. Many of the lenders weren't supervised at all. Of those that were, many nonetheless were not stopped from making loans of dubious quality. That was a microprudential problem.
There was a macro-level element to the boom: those poor-quality loans were packaged up and distributed around the world. Each national supervisor thought that it was a good idea to transfer the risk out of their system. It could then be spread across the world, supposedly more thinly. That showed a lack of macro perspective. But it was also a microprudential failure, because banks were allowed to buy these securities without holding enough capital to cover their risk. Risk capture – getting the risk weights broadly right – is very much a micro-level aspect of prudential regulation. Macroprudential frameworks would not have solved this problem.
Nor was a macroprudential institutional framework strictly necessary to avoid the liquidity problems brought about by short-term funding of these security holdings. Funding and liquidity are traditional, micro- or institution-level concerns of prudential supervisors. It certainly needed more attention globally in the lead-up to the crisis. But this didn't require a new macroprudential or systemic risk regulator. It just needed more attention from the supervisors and the Basel Committee.
Where Does that Leave Financial Stability Policymakers?
If all of the tools are in the hands of the prudential supervisor, what role does the non-supervisory central bank have in financial stability policy? There are many things we can do, which we described in the joint RBA-APRA paper. Crisis management is one aspect. Liquidity provision is another. And there is also the one use of the word macroprudential that I actually feel comfortable about: macroprudential analysis. This involves identifying the risks in the financial system that ostensibly ‘micro’ prudential regulators can respond to. This does not mean looking only at macro-level data, by the way; it is usually the riskiest banks or borrowers that cause the trouble, not the average ones.
I can perhaps best explain how this works with another cartoon (Figure 3). I'd like you to imagine that the supervisor from the previous cartoon – I call him ‘Mike Rowe’ – receives a visit from the head of the central bank's financial stability department, ‘Prue Denshall’, sometimes nicknamed ‘Macro Prue’. Prue has news for Mike: all the banks are doing crazy lending – systemic risk is rising. And Mike responds: ‘Gosh, thanks, “Macro” Prue! I'll raise their Pillar 2 capital add-ons right away!’
Of course, it's just a cartoon. In reality, if lending standards were imprudent, Mike would already know about it. And Pillar 2 capital add-ons aren't the only tools at his disposal. That said, I'd like to draw out a few points from this picture. The first is the focus on lending standards. Risk-taking, here in the form of easier lending standards, is what we are watching for. It's about behaviour, not a single number. The second is that a qualitative assessment might well be good enough. Prue isn't telling Mike that her favoured measure of systemic risk has risen to such and such a level, and that this has breached some trigger point. Existing models of systemic risk simply aren't good enough to be relied on that way, and they probably never will be. But informed judgement is probably good enough in most cases.
The third point to note is that Mike already has the tools at his disposal to respond. Prudential supervisors have always had the scope to increase banks' minimum capital ratios above the Basel minimum. That is what Pillar 2 of the Basel framework is for. Not every supervisor used that power, but it was there.
The fourth and final point about this picture is that Mike believes Prue, and is willing to act on it. This is pretty much how it works in Australia. It comes down to the relationship between the agencies and the culture of the supervisor to create that willingness. There doesn't need to be a so-called systemic risk regulator issuing public warnings or directions, telling the supervisor what to do. Some jurisdictions might want to set things up that way, but there are other ways that might be better here. In Australia, we think that a culture of cooperation, dialogue and mutual respect is more important than formalised arrangements. The Council of Financial Regulators has proved itself to be a low-cost, flexible way of coordinating between agencies, alongside bilateral relationships. We think we have the essential elements needed to promote financial stability with a holistic frame of mind. In the end, what is needed is the wisdom to see the problems and the willingness to act in response. No elaborate set of institutional arrangements and rules can manufacture those two things.
Thank you for your time.
Thanks to Sarah-Jane Edis for assistance with graphics formats. I created the cartoons in this speech by modifying original royalty-free artwork from RetroClipArt/Shutterstock.com, which retains the copyright (see <http://www.shutterstock.com/licensing.mhtml>). [*]
See G-20 (2010), ‘Seoul Summit Document’, paragraph 41, available at <http://www.g20.utoronto.ca/2010/g20seoul-doc.pdf>, and the resulting joint FSB-IMF-BIS progress report to the G-20, ‘Macroprudential Policy Tools and Frameworks’, available at <http://www.imf.org/external/np/g20/pdf/102711.pdf>. 
See RBA and APRA (2012), ‘Macroprudential Analysis and Policy in the Australian Financial Stability Framework’, September, available at <http://www.rba.gov.au/fin-stability/resources/2012-09-map-aus-fsf/>. 
See Clement (2010) for a comprehensive review of the history and usage of this word. 
In addition to the RBA-APRA paper released last month, a good description of how APRA takes a system-wide view in its supervisory practice can be found in Laker (2010). 
See Wong et al (2011). 
For more discussion on the different dimensions of lending standards, see Ellis L (2012), ‘Prudent Mortgage Lending Standards Help Ensure Financial Stability’, Address to the Australian Mortgage Conference 2012, Sydney, 23 February, available at <http://www.rba.gov.au/speeches/2012/sp-so-230212.html>. 
Clement P (2010), ‘The Term “Macroprudential”: Origins and Evolution’, BIS Quarterly Review, March, pp 59–67.
King M (1997), ‘Changes in UK Monetary Policy: Rules and Discretion in Practice’, Journal of Monetary Economics, 39, pp 81–97.
Laker J (2010), ‘Supervisory Lessons from the Global Financial Crisis’, AB+F Leaders Lecture, 8 December. Available at <http://www.apra.gov.au/Speeches/Documents/09-ABF-8-December-2010.pdf>.
RBA and APRA (2012), ‘Macroprudential Analysis and Policy in the Australian Financial Stability Framework’. Available at <http://www.rba.gov.au/fin-stability/resources/2012-09-map-aus-fsf/>.
Wong E, T Fong, K-F Li and H Choi (2011), ‘Loan-to-Value Ratio as a Macroprudential Tool – Hong Kong's Experience and Cross-country Evidence’, Hong Kong Monetary Authority Working Paper 01/2011.