RBA Annual Conference – 2004 Discussion

1. John Edwards

In reflecting on the theory and practice of central bank transparency and communication Malcolm Edey and Andrew Stone have come up with a subtle, interesting assessment which moves convincingly from issues of presentation to issues of substance. I want to focus my comments in two areas. The first is whether the current Reserve Bank of Australia (RBA) arrangements in respect of transparency and communication are the best that can possibly be devised, given the institutional framework. The implication of Malcolm and Andrew's treatment is that they are, and I am not sure this is so. The second is in the wider area which Malcolm and Andrew open up in the fourth part of their paper. This is the big question of how the Bank's inflation target should now be interpreted by the market. This is one of the central issues raised by the whole conference, and I think Malcolm and Andrew have much to contribute. In particular their description of the RBA's inflation target as a ‘medium-term constraint’ will, I think, be remembered, referenced and pondered for some time to come.

There is no doubt that the RBA is more transparent and communicates very much more effectively today than it did a decade ago. As the authors rightly remark, the introduction and gradual development of the quarterly Statement on Monetary Policy, the twice-yearly appearances by the Governor and other RBA officials before a parliamentary committee, the statements which accompany interest rate changes and – just as importantly – the greater candour and directness of speeches by RBA officials, have created a mass of material which now offers a good guide to the RBA's assessment of Australia's economic circumstances, the issues which it thinks will and won't be important, and the kind of data which might bear on monetary policy decisions. It does not of course reliably tell us what the RBA will do, but it does tell us how the RBA thinks economic circumstances will evolve, and how it might respond to some specified developments. In other words, it tells us as much about the future as the RBA itself knows. There are as yet no detailed forecasts. Except in the twice-yearly Parliamentary Committee appearances by the Governor, the RBA still seems wary of offering a numerical GDP forecast – partly perhaps because it might differ from that offered by Treasury. But it does now offer an explicit inflation forecast.

There are, as Malcolm and Andrew say, several purposes or rationales for communicating more material about the views of the RBA. One is that the RBA now has a conditional independence under which it exercises large powers. Accountability requires it to accept an obligation to be as transparent as possible and open to scrutiny, and is the necessary condition of the great powers it exercises. It is also highly likely that monetary policy will, over the long term, be more effective if other participants in the economy know the central bank's goals, time frame, style, and methods of work. It is not possible, for example, to anchor inflation expectations if there is widespread uncertainty about the real goals of the central bank. This is true of the inflation target, but it is also true of other issues pursued by the RBA – for example, its warnings about house price inflation. It is equally important to know what are not targets of the Bank – for example, the current account deficit, or temporary inflation ‘blowouts’ caused by oil price increases. I think it would be impossible to conduct policy now if the Bank was secretive or circumspect or misleading in respect of its goals, which is why the debate about the cleverness of conducting business through monetary policy ‘surprises’ is one for the journals and modellers alone. A third rationale for transparency is specific to financial markets. It is probably true that interest rates and currencies will be less volatile if the policy approach of the central bank is always well understood, and it is probably true that lower volatility is good for economic efficiency. Financial market deregulation has gone hand in hand with both the increasing authority of central banks, and with their increased transparency. It is difficult to imagine a secretive or deceptive central bank coexisting with very large and highly deregulated financial markets, though the European Central Bank is arguably testing the proposition.

These rationales are listed here to make the point that transparency is no longer discretionary, or voluntary. It is not a gift of the central bank. It is true that it is not a statutory requirement of the RBA, but it is an operational requirement. If this is so, then the more transparency the better, so long as the degree of transparency is consistent with the ability to make decisions effectively.

Is the current degree of transparency enough? I think not. It is not at all clear to me why the RBA cannot publish a summary of discussions at board meetings, and the material which is offered to the board to guide its decisions. At the very least, the RBA should be able to publish a résumé of the discussion, as the US Federal Reserve System now does in respect of the deliberations of the Federal Reserve Open Market Committee.

The RBA has an unusual institutional arrangement in that monetary policy decisions have to be approved by a board. The purpose of the Board has always been to make it harder for governments to exercise control over the money supply. The Board is not intended or required to be a group of monetary policy experts. It is not representative of the community nor is it meant to be. Nor is it representative of business. It does not, for example, include a representative from a media or telecommunications business. By law it cannot include a commercial banking representative. The Board members are selected by the government of the day, with no clear criteria. The function of the board these days is still to help protect the independence of the RBA from day-to-day influence by the government. Because of its composition, in practice it also works as a reality check on RBA recommendations, and a sounding board for the Governor. It does not value technical competence in monetary policy but experience in business leadership. My impression is that these days the ideal board member is someone who has long and successful experience of applying common sense to judgements about the future course of events, and who is not awed by the esoterica of monetary policy. The characteristics of the Board probably contribute quite a lot to the qualitative rather than quantitative style of the RBA, and its pronounced preference for commonsense, judgement, experience, balance and economic insight over econometric work and technical brilliance.

As it happens, there are quite a few excellent people on the Board who can make a contribution to an assessment of the state of the local and global economy. But if the Board is important, why are its deliberations shrouded in secrecy? It is not the board of a business which is in competition with others. It is the Board of what is effectively a monetary policy authority, and it is charged with making the decisions about monetary policy in Australia. Malcolm and Andrew suggest that if the Board's deliberations were published, members would be obliged to argue for the industries which they represent. I think it is very unlikely that board member and global shopping centre developer Frank Lowy, for example, would argue for the property industry if his role in the Board's deliberations was published, and not if it wasn't. Surely the same applies to all board members of substance. One might say that publication of this material would work the other way – board members whose contributions were exposed to view would be wary of peddling an industry line.

Publication of the minutes or a résumé of the minutes would serve several purposes. It would guide markets and the wider public to the nature of the monetary policy debate. This is often not as apparent as it could be. The RBA Statements and so forth, after all, offer only one viewpoint. Even where there are disagreements within its leadership, they are obscured or resolved in the process of drafting the material. But we know there will be different interpretations of the data, different views about the state of the local and global economies, different ideas about what is important and what is not at any particular time. It would be useful, therefore, to know the flavour of the discussion on the Board. It would mean fewer surprises for markets because markets would be more aware of the ranking of the probabilities of alternative policies, and the balance of opinion on the Board. It would enrich the economic debate.

But it would not only be useful to the markets and the public. It would also be useful to the Bank. It would impose a stronger discipline on board discussions. It would make it very difficult to introduce parochial concerns or irrelevant anecdote or political prejudice. It would elevate the discussion. It would better protect the independence of the Governor from the government by precluding political considerations or Treasury views from having too much influence.

What is really involved in releasing board minutes or a résumé of discussions is breaking the monopoly the RBA executive now exercises over the central bank's take on the economic debate. In the US, we not only have résumés released, we also have speeches by the FOMC members. Quite often those speeches will enlarge the debate, without contradicting the leadership of the Fed. In Australia, résumés are not released, and board members other than the Governor and Deputy Governor are discouraged from publicly offering economic views.

The minutes or a résumé of discussion would be very much more useful if published together with the material presented to guide the Board's decisions. This could include such forecasts as are used by the Board. Some of this material is now published in the Statements and some is in the publicly available chart pack, but the full range of material presented to the Board would also be interesting. The RBA should, by now, be quite unconcerned that on occasion its forecasts for GDP, for example, may sometimes be different to those of Treasury.

I have no doubt it would improve understanding of, and interest in, monetary policy, and that sooner or later it will happen. There is no push to further open up the RBA today because Australia is now in the 14th year of uninterrupted low-inflation expansion, and the RBA can claim some credit for that. Sooner or later there will be bad times, however, and that is when the pressure for changes will mount.

I think minutes or résumés of discussion could be useful and important. By contrast, I think detailed forecasts of GDP components and so forth are not important. What really matters in assessing the economy and the likely path of interest rates is the story rather than the projected numbers. Numbers without a story are quite useless. We know from our own experience that Treasury or RBA forecasts, or our own numerical forecasts, rarely pick up the big changes in investment, equity prices, capital flows, currencies, interest rates or the political framework which make a real difference. That is why most serious discussion is framed as risks to a forecast and why the medium-term story, with its risks and possibilities, is the heart of forecasting. Of course it is difficult to separate the story from the numbers. If one thinks it is highly likely, for example, that housing construction will fade in the coming year and that exports will increase, it is important to put some numbers around it. But the numbers are illustrating a view – they are not the view.

I turn now to the much bigger issue raised by Malcolm and Andrew, which is also the issue underlying this whole conference. How should we interpret the central bank's inflation target? We are reflecting on well over a decade of low inflation in most OECD economies. The excellent inflation outcome compared with earlier decades no doubt owes a good deal to the anchoring of inflation expectations by central banks, but it was also assisted by supply-side developments which have nothing to do with central banks. This is part of the ‘divine coincidence’ which Ken Kuttner referred to in his paper. In Australia, low inflation was brought about by the reforms of the eighties and the early nineties, and the unintended deep recession of 1991, which for Australia was the equivalent of the Volcker deflation of the late seventies and early eighties. There are also global influences. These include a halving of the real price of oil over the last decade, constant downward price pressure in manufacturing as the labour force of China engages in the world economy, the lower barriers resulting from the Uruguay Round and unilateral reductions in protection, cheap computing and telecommunication and all the technologies associated with them, and the upswing in productivity growth in the US, Australia and some other economies, which is probably related to cheap computing and telecommunications, as well as the more flexible labour market and other reforms in Australia. It is also the case that for the developed economies, most of the unfavourable shocks over the last decade have been on the demand side – the Asian financial crisis, the LTCM/Russian crisis, the crash in IT and telecommunications equities, September 11, and the Iraq war. In the developed economies, these crises posed the risk of lower output growth and lower inflation simultaneously. In each and every one of these the obvious and clear response was for central banks in Australia, the US, and Europe to lower interest rates.

In those sorts of circumstances it would at least have been possible to think of the inflation target the way in which Goodhart apparently did think of it, as quoted by Malcolm and Andrew. In the early years of the Bank of England Monetary Policy Committee he saw it as a literal and sole target to be achieved within a definite and known time frame. Judgement was important, but only in estimating the path of the policy interest rate likely to see the target achieved at the date required.

We should assume that in coming decades central banks will sometimes be faced with the more typical unfavourable supply-side shocks, which threaten both lower output growth and higher inflation. In Australia, for example, one might see lower labour productivity growth. Globally, one might see sharply higher oil prices or gradually rising manufacturing costs as the value of labour in China rises. It is also true that inflation targeting has been complicated by asset-price bubbles. We have already seen that continued low inflation and low interest rates can give encouragement to inflation in equity or house prices which can become speculative and which can in various ways cause economic contractions when the bubble bursts.

In these circumstances, insistence on achieving an inflation target and only an inflation target, and achieving it only within a specific and defined time frame, would see the whole theory and practice of anti-inflation targeting and perhaps independent central banking collapse, after what would be a stormy period of debate. Malcolm and Andrew's paper is really important in laying out the ways central banks generally, and the RBA in particular, are redefining what the inflation target might be interpreted to mean. As they argue, ‘central banks need to develop ways of bringing into consideration factors that may be relevant to the policy decision but which would not fit into a conventional shorter-term inflation forecast’ (p 90). This might mean, for example, that inflation should return to target after some permitted deviation, and over a longer time frame. This is how one imagines the RBA would respond to a sharp and prolonged increase in the oil price. It may mean an understanding that there will be quite sharp changes in interest rates to deal with the growth of an asset bubble or its aftermath, even though forecast inflation through the period is within target. It may mean, in the case, for example, of a decline in labour productivity growth in Australia with perhaps some coincidental increase in wages growth, that the central bank departs entirely from the notion of an independent focus on inflation alone and proposes or accepts working within a framework which includes fiscal and incomes policy as well as monetary policy. All these outcomes are surely possible, and it will be no response at all for a central bank to say that inflation is its only responsibility. We have not even considered here the scope for even bigger events, such as some problem in the Australian banking system, which would require an immediate and massive response led by the RBA, but which would have to involve the Commonwealth Government and quite possibly other central banks and foreign governments as well.

Given these considerations, Malcolm and Andrew make two points, which I think are quite important statements. They are not entirely new but they are well expressed, and offered with authority. The first is that ‘The general principle, which is recognised in the policy mandate, is that the inflation target is a medium-term constraint, not a deterministic formula that requires information outside the short-term inflation forecast to be ignored’ (p 91). The second is that ‘The general observation suggested by these experiences is that it is unrealistic to expect an inflation forecast on its own to represent all of the information that policy needs to take into account, even though it is obviously an important component of it’ (p 94). These statements seem to me to helpfully clarify how we should think about the RBA interpretation of its mandate. To think about the inflation target as a constraint rather than as a goal, and a constraint that can, if necessary, be met within a flexible time frame, leaves plenty of room to pay attention to asset prices when necessary, growth and employment when necessary, to large discrepancies between headline and underlying inflation, and to other complicating issues which will no doubt come up from time to time in the decade to come. But in reinterpreting the inflation target in a broader and richer way we also need to recognise that the requirement for communication and transparency will increase. If asset prices are going to be important at various times we need to know how important, and in what circumstances. If the time frame for returning to the midpoint of the inflation target is to be extended, we need to know how far, and for what reasons, and how progress towards the other goals of policy are to be assessed. There is otherwise a danger of drifting back to a ‘checklist’ approach, in which the weight and priority and urgency of various goals is undefined. Since most of the new issues will not lend themselves to quantitative forecast, the nature of the monetary policy debate will be increasingly important – and that strengthens the case for publication of board discussions, and perhaps changes in the nature of the Board itself.

2. General Discussion

One participant argued that any review of the appropriate extent and form of transparency and communication should start with a recognition of the success of inflation targeting in Australia and elsewhere, especially when considered alongside the difficulties experienced with earlier monetary policy regimes (such as monetary targeting), and the initial uncertainty about the likely success of inflation targeting. While there had been an element of good luck, with the absence of significantly adverse supply shocks, some of the shocks experienced in recent years – such as the Asian financial crisis and the worldwide recession of 2001 – did not seem all that minor when they occurred. Looking ahead, another participant pointed to the challenge of embedding considerations of asset prices within the general inflation-targeting framework. This was important to avoid giving the impression that the central bank was abandoning the framework. This challenge was similar in some respects to earlier efforts of the RBA to communicate how movements in the exchange rate should be considered in the context of a flexible inflation target, rather than being mistakenly perceived as somehow being outside the framework.

As with Rick Mishkin's paper, there was some discussion about the implications of releasing the minutes of the central bank's policy committee. One participant noted that this decision could not be divorced from the exact role of the policy committee, or the composition of that committee. This was clearly related to the broader issue raised by a number of participants regarding the potential trade-off between greater disclosure of a central bank's internal discussion and the quality of that discussion. On the same theme, one participant thought that an important factor in any decision to release minutes was whether it was desirable to encourage personal accountability or collective responsibility for decisions of the policy committee.

There was considerable discussion of Figure 4 in Edey and Stone's paper, which showed the (relatively small) deviations of inflation forecasts from the target for the central banks of Australia, New Zealand, Sweden and the United Kingdom. As a number of participants noted, these results were surprising given that only the Reserve Bank of New Zealand presented its inflation forecasts on the basis of an endogenous projection for interest rates. One participant suggested that this general result might be explained by the actual publication of forecasts, which may tend to constrain forecast values to be close to the target, while another participant suggested that it implied that central banks' forecasts were being made from a position where the interest rate was already close to its neutral level. In either case, the proximity of inflation forecasts two years out to the target had implications for any lengthening of the inflation-targeting horizon. While it is important, conceptually, to think further than two years out, if the central bank believes it will hit the target in two years it is difficult to comprehend how it should take longer-term events into consideration when deciding the appropriate stance of monetary policy today.

There was also some discussion about the implications for inflation-targeting regimes of the establishment of inflation expectations at low levels. In Australia's case, one participant suggested that this had given the RBA the flexibility to deal with the shocks it had faced in recent years, such as the Asian financial crisis. Another participant argued that inflation targeting was akin to a policy approach of constrained discretion. It was suggested that constraining mechanisms are important for building credibility when the central bank was trying to signal a clear regime change and to lock in low inflation expectations. Once that was firmly in place, the wider objectives of monetary policy could become a larger part of the emphasis. In other words, there was scope for more discretion.