RBA Annual Conference – 2018 Discussion

1. Sally Auld

Introduction

Good afternoon. Let me begin by thanking the conference organisers for the opportunity to participate in today's session. I have to admit that when I read in the newspaper earlier this week that the Bank of England's (BoE's) Chief Economist had organised for Billy Bragg to come and speak to BoE officials, I was a little worried that Ben's paper might have taken a radical turn! But, as he notes at the beginning, his paper is sympathetic with the prevailing orthodoxy on the topic, and so our session is defined more by evolution than revolution.

The paper argues the case for the continued separation of monetary and macroeconomic policies. On one level, the case for separation of these powers relies on an assumption that spillover effects are small.

Another argument for separation rests on the quantification of objectives. It's easy to assess the performance of an inflation-targeting central bank relative to its policy objective, as both the target and the target variable can be easily and credibly observed. Macroprudential objectives, in contrast, are harder to define and harder to measure. The theory around incentive design suggests that it might be optimal to separate jobs relating to easily measured objectives from those that relate to more opaque objectives.

In discussing Ben's paper, it's only fair that I state from the outset that I am not an expert in the academic literature on this topic. So I thought I would frame my comments with reference to the local experience, given that the topic is particularly pertinent in both Australia and New Zealand. I want to add a couple of additional arguments to the case for separation, and then will talk a little about what might happen when tension arises between monetary and macroprudential policy objectives.

Some more thoughts on separation of macroeconomic and macroprudential

The first argument for separation that I want to address starts with something called Maslow's hammer, which states that if you only have a hammer, then every problem looks like a nail. For the topic at hand, this is just saying that interest rates are not necessarily the right tool to deal with financial stability issues. This concept of the right tool for the job was espoused by Ben Bernanke in a speech back in 2002 (Bernanke 2002). Bernanke doesn't really argue one way or the other for separating responsibility for economic and financial stability objectives; rather, he simply argues that interest rates are not the right tool for dealing with financial stability problems.

I think his argument can be extended in favour of the case for separation. Effectively, it comes down to taking a view on what it is appropriate – or even sensible – for a central bank to do.

It goes without saying that central banks pay close attention to financial markets when thinking about the economy's trajectory and the optimal path of interest rates.

But this is a very different proposition to thinking about whether macroprudential tools should be deployed for dealing with financial stability issues. For a start, this would require the central bank to take a view on the valuation of an asset class, whether it be stocks, houses, or something else. Moreover, it also probably relies on an assumption that the central bank is better, on average, than the market at valuing assets, meaning it can identify a bubble in an asset class before others. In my experience, central bankers generally tend to shy away from such pronouncements, and for good reason.

Central bank officials have sometimes bemoaned the overanalysis of their comments on the economy and outlook for monetary policy. If the market was also trying to second guess the central bank's view on a particular asset class, then this would probably be highly detrimental to achieving an optimal allocation of capital across asset markets.

If the experience in New Zealand in recent years is any guide, market participants also become unnecessarily obsessed with the trade-off between policies, trying to gauge what any particular tweak to macroprudential policy is worth in terms of interest rates. This seems almost reminiscent of the days of monetary conditions indices.

To be fair, some might argue that the benefits of coordination outweigh potential communication difficulties, especially given that acute financial stability issues are relatively rare. But as Guy Debelle noted in his paper at this conference, the communication demands on central banks have increased over the past decade or so, and so, in light of this, I think there's a compelling practical case for keeping it simple.

Another benefit of separation derives from the fact that we cannot rely on a stable relationship between macroprudential and monetary policy. Over time, the relationship between these policies and the associated spillover effects will essentially depend on: the extent to which the business and credit cycles are synchronised; structural factors that determine the extent to which lending occurs outside of the regulated sector; and global influences on domestic interest rate and currency settings.

We don't have to look too far to find examples where inflation has been lower than desired but the credit cycle robust, thanks to a globally induced regime of low rates and a strong currency. One might argue that policymakers in both Australia and New Zealand have found themselves facing this mix of outcomes in recent years. In both jurisdictions, the credit and inflation cycles have de-coupled, allowing the possibility that macroprudential policy and interest rate policy might be in conflict. In such circumstances, monetary policy settings would be aimed at generating above-trend growth, in order to return inflation to target. In contrast, macroprudential settings would be trying to restrain or slow credit growth. Macroprudential policies could therefore potentially thwart the objective of monetary policy, if slower credit growth became a headwind to gross domestic product (GDP) growth.

If so, it might make sense from a communications perspective if the policies were owned by separate entities. It is relatively simple to explain policy actions in terms of the deviation of inflation or output from desired levels. But how does a central bank explain the trade-off between pursuing macroeconomic and macroprudential objectives? In the event that there is tension between the two policies, this might invite questioning of who sets the parameters around the trade-off, and could, in some circumstances, unsettle otherwise well-anchored inflation expectations.

And, at a time when the two objectives were complementary to one another, the distributional impact of using one policy in preference to the other might be quite different. In such circumstances it might be quite uncomfortable for a central bank to defend its actions.

The local experience

Here in Australia, financial stability is now explicitly mentioned in the Statement on the Conduct of Monetary Policy. This is not problematic when monetary policy is set such that it is sympathetic with achieving both macroprudential and macroeconomic objectives. But policymakers should be aware that this is not always the case. As I have alluded to, the current environment is a case in point.

In the short run, good (inflation) target design can help minimise trade-offs. For example, it's no secret that the inherent flexibility in our inflation-targeting regime has provided policymakers here in Australia with the timely opportunity to focus on financial stability concerns in the last 18 months, particularly as they relate to household balance sheets and the types of mortgage lending.

The Reserve Bank of Australia (RBA) has been quite explicit about this, and has explained that a period in which inflation is a little below target and growth a little below trend is an appropriate cost to wear in order to ameliorate risks around household balance sheets. Were the inflation target less flexible then, arguably, the RBA could have been forced to pursue policy outcomes that exacerbated financial stability issues.

However, it is also important to remember that this flexibility is not infinite – that is, inflation outcomes eventually have to be consistent with the target. After all, one of the stated objectives of the inflation-targeting regime is the anchoring of inflation expectations.

Locally, I don't think it's too hard to see how this scenario could become somewhat uncomfortable for policymakers. In Australia at present, macroprudential policy is working effectively to tighten credit to mortgage borrowers – particularly to investors – and has delivered both a slowing in credit growth and house price growth. To be sure, these are desired outcomes.

But to the extent they could also compromise household consumption outcomes over the next little while, and thus see the economy sustain a period of below-trend GDP growth for longer, they might also delay the return to trend growth and target-consistent inflation outcomes. This outcome is acceptable for a while, but not for a long time. As noted above, the inflation target is not infinitely flexible.

A priori, it's not clear which path policymakers should take in such circumstances. Does the inflation target, by virtue of its importance to the anchoring of inflation expectations, eventually take primacy over macroprudential considerations? Or alternatively, as Ben suggests in his paper, is the body tasked with macroprudential objectives asked to internalise its spillovers (as was the case in the United Kingdom in 2013)?

Clearly the circumstances matter, as does the policy at hand. Some macroprudential policies are potentially complex in their application, and hence their effect on the overall economy is subtle and takes longer to occur. In contrast, policies aimed at restricting growth in a particular type of lending, or to a specific borrower, can be more effective over a shorter time period.

Perhaps the right view is the long view. In the long run, failure to deal with financial stability issues will ultimately mean failure on the inflation target and, potentially, as we have seen in the global economy in the past decade, legitimate fears around the ability of central banks to anchor inflation expectations. In this context, if the cost of dealing with financial stability issues is low (but not too low) inflation, then that's probably an acceptable price to pay. Again, it might just come down to communication – if a credible central bank can communicate a sense that it knows flexibility in the inflation target isn't infinite, then it can buy itself a lot of time and minimise the cost of any tension between macroprudential and macroeconomic objectives.

In summary, I think the case for the separation of macroeconomic and macroprudential objectives makes a lot of sense. But even with separation, it's not clear what happens when the policy objectives are in conflict. Target flexibility, credibility and strong communication all help. In Australia, we are lucky enough to have all three, and hopefully this will be just enough to buy us the time to deliver success on both our macroeconomic and macroprudential objectives.

Reference

Bernanke BS (2002), ‘Asset-Price “Bubbles” and Monetary Policy’, Address given to the New York Chapter of the National Association for Business Economics, New York, 15 October.

2. General Discussion

One of the major topics of the discussion was the accountability of the macroprudential authority. Following Ben Broadbent's presentation advocating separation of monetary and financial stability powers, a participant asked whether the responsibility for the two policies should be in two different institutions. The discussion covered several points in favour of having both committees remain at the same institution: the ease of coordinating policy, overlapping areas of interests, the ability to have joint briefings and economies of scale in monitoring the same developments. Participants preferred this arrangement to a diffuse system of multiple agencies looking at various parts of the system, as is the case in the United States. The dedication of a group of people to the issues of financial stability was said to be more important than whether that group of people is within the central bank or outside of it. The separation of powers enhances accountability by giving monetary and macroprudential issues different forums.

One participant questioned why there wasn't greater demand for some quantitative measures of financial stability. Governments generally provide the target of monetary policy, but no central bank has been given a target or even definition of financial stability. Another participant agreed, preferring a situation where the central bank does not have discretion over the definition of a target or the instrument used to achieve the target. The participant did suggest that practice should create, de facto, a commonly used target and instrument.

This led to some discussion on transparency and communication. A participant noted that these had previously been agreed as two key pillars of an effective inflation-targeting regime, and that they should be similarly important, or even more important, for financial stability policy. Macroprudential policy requires central banks to sustain policy arguments over a longer time and the policy effects may be higher on households – for example, macroprudential policies in the United Kingdom have prevented some people from getting mortgages. Good, clear communication was said to be necessary when explaining why financial stability policy is acting to address a risk that some event may occur in the future, perhaps many years ahead. Participants also commented that the public and the media have more experience and understanding of monetary policy and that may be why macroprudential issues receive less coverage.

The other major topic covered by the discussion was the trade-off between price stability and financial stability. This discussion opened with a question regarding whether the zero lower bound makes this trade-off more acute. The effective lower bound may cause interest rates to be ‘stuck’ at a very low level for an extended period of time, which could make monetary policy very ‘predictable’ and also lead to financial imbalances. There was agreement that the zero lower bound could complicate the task of policy coordination in the case, for example, where inflation was weak, interest rates were low and yet there was very strong credit growth. In such a case, a tightening response by the macroprudential authority may be uncomfortable for the monetary authority. Participants also noted that there is a small range for interest rates, close to the zero lower bound, where the goals of financial stability and monetary stability could be said to be in conflict. In this range, institutions that have a lending book with thin spreads over the bank rate would face depressed earnings and possible insolvency if the policy rate was lowered beyond the effective lower bound.

Participants discussed the implications of this trade-off for the organisation and actions of the monetary and macroprudential authorities. One participant, citing the example from Broadbent's paper, stated that the conflicts raised are not sufficient reason to merge the two policy committees into one because they can coordinate instead. In 2013, the Bank of England issued formal monetary policy guidance that interest rates would not rise until the unemployment rate fell below a certain level. However, they qualified the guidance by noting that it would not hold if the macroprudential authority judged that low interest rates posed risks to financial stability that could not be mitigated. Another participant shared the view that the overall policy package implemented in the euro area had counterbalancing effects with respect to bank profitability, and that the implementation of non-standard measures was said to improve the overall macroeconomic outlook for the economy. While negative policy rates were said to reduce net interest income for banks, this was offset by lower loan provisions and the asset purchases program led to a capital gain for banks. These circumstances highlighted that coordination was possible while retaining the benefits of separate bodies discussed earlier.

The discussion also covered the role of money as a unit of account and how the trade-off between price and financial stability might evolve with possible changes to the means of payments. A participant suggested that these changes to payments may result in a currency with a constant real value and that the central bank would not consider adjusting the real value of money to address financial stability. While there was some sympathy for the view that a constant real value of the currency could help people make good decisions, it was noted that this is not so different from a world with a successful inflation-targeting regime. Participants reflected on the period of the classical gold standard, which showed that a relatively stable real value of money is no guarantee of financial stability. One participant stated that financial stability can be better dealt with by other tools and there was little support to use the ‘hard-won’ credibility of monetary policy to adjust domestic credit supply.

Lastly, the point was made that the role of fiscal policy and the fiscal position of the government should play a larger role in the debate regarding financial stability policy. This reflects the recent experience when financial stability policy actions relied on the balance sheet of the public sector in the form of quantitative easing and government bailouts of banks. Participants agreed that the three-way interaction between fiscal policy, monetary policy and macroprudential policy could be rather complicated but that research tends to show that commitment and the separation of policy assignments can assist operationally.