Making monetary policy : a look behind the curtains
Address to the Canterbury Employers' Chamber of Commerce
Mr Chairman, as most New Zealanders know, the Reserve Bank operates monetary policy with the aim of maintaining price stability, currently interpreted as keeping movements in the Consumers Price Index between 0 and 3 per cent per annum after allowing for special factors. Another way of expressing this, paraphrasing the words of Alan Greenspan, the chairman of the US central bank, is to say that our core task is to make decisions on monetary policy in such a way that the public and business community can, for all intents and purposes, confidently ignore the risk of future inflation, and go about their affairs confident that the general level of prices will remain stable.
It took the ravages of the inflation-prone period from the late 1960's through until the late 1980's to remind us of the damage that inflation does to economies and therefore to the welfare of individuals and communities.
And I want to stress that inflation damages people, and not just "economies" in some abstract sense. I am sometimes rebuked by critics who say that the Reserve Bank should pay more attention to the impact of its policies on people, and less to inflation. But this is to totally misunderstand the situation: "economies" are simply groups of people, organising their lives to best advantage, and inflation has a huge and often unpredictable effect on their ability to do that. I have often told the story of my uncle who retired in 1971 after a life-time growing apples in Nelson, who then sold his orchard and invested the proceeds in what he perceived as a low risk asset – 18-year government bonds, then yielding 5.4 per cent – only to see inflation steal 90 per cent of his life's savings over the next 18 years.
So make no mistake: controlling inflation is about helping people, and making it easier for people to plan their lives. And to the surprise of many people, it turns out that it is not so much the rich that are most advantaged by controlling inflation as the poor, since, like most taxes, inflation is a tax that the poor find particularly hard to avoid.
Unfortunately, after several decades of high inflation, people's fear of inflation does not easily subside. Therein lies a significant part of the cost of inflation. That is a cost we were paying through the late 1980's and early 1990's as we fought to bring inflation back to the sorts of levels our trading partners were accustomed to, and indeed to the levels that we ourselves had enjoyed for most of our history prior to the 1960's.
The Reserve Bank of New Zealand Act passed in 1989 set out to build price stability into the very fabric of our society. In particular, it set out to build an institutional framework in which the longer-term benefits of on-going price stability were given emphasis in monetary policy decisions, providing a bulwark against the inevitable temptation to grab the short term gains of an unduly easy monetary policy while discounting the longer term costs of such a policy.
The Reserve Bank Act achieves that by employing some fairly standard principles of management:
- The objective of monetary policy is specified clearly and unambiguously – price stability
- The responsibility for meeting that objective is also clearly allocated – to the Governor of the Reserve Bank
- The authority required to perform that role is clearly vested in the Governor
- The decisions made in pursuit of the objective are required to be made with a high degree of transparency, through the publication of regular Monetary Policy Statements
- And having been given the objective, the responsibility and the authority to carry out the task, the Governor is held accountable for his decisions – primarily by a board appointed by the Minister of Finance to monitor the performance of the Governor on his behalf, but also through appearances before Parliament's Finance and Expenditure Select Committee.
The product of the 1989 Act was the form of monetary policy we now have, generally known in the trade as "inflation targeting". Inflation targeting is one of the most influential innovations to hit the world of monetary policy in the past decade. I think it is fair to say that it originated in New Zealand, and has now spread to be the monetary policy framework of choice in many industrial countries and, increasingly, in emerging market economies as well. It has become one of the most researched, written about and talked about subjects in monetary economics.
The 1989 Act was pioneering in many respects. It and the way the Bank has implemented monetary policy have at times been controversial. With more than a decade of experience with the Act, it is not surprising that the Government has decided to seek a review of the way it has been working. Professor Lars Svensson of Stockholm University has been contracted to undertake that review, and we expect his report to be made public within the next few weeks. As I said when the review was announced, I welcome that review and look forward to the release of the Svensson report.
One aspect which featured strongly in submissions to Professor Svensson relates to the way in which the Reserve Bank Act places all monetary policy decision-making in the hands of a single person – the Governor. This is unusual internationally, although not unique. The Act was quite deliberate in opting for the single decision-maker model. It did so in order to heighten the sense of accountability felt by the Governor for monetary policy decisions. The architects of the Act felt that the main alternative, that of assigning decision-making authority to a board or to a committee, would undermine accountability for decisions and thereby weaken the resolve to place price stability clearly at the top of the objective tree. (Indeed, when I asked the Minister responsible for shepherding the legislation through Parliament in 1989 why the Bill envisaged an agreement defining price stability between the Minister and the Governor, rather than between the Minister and the Bank, he explained simply that "We can't fire the whole Bank. Indeed, realistically we can't even fire the whole Board. But we can certainly fire you!") The non-executive directors on the current Bank board, in their submission to Professor Svensson, have reiterated that concern about accountability in recommending the retention of the single-decision-maker model.
Given that this question of the appropriate decision-making structure is likely to feature prominently in the forthcoming Svensson report, I thought it would be helpful to spend some time today dwelling on the processes by which I, as Governor, currently go about the task of making monetary policy decisions. In doing so, I am particularly anxious to dispel criticisms I hear occasionally that the Bank is insufficiently in touch with the "real world", or that it is excessively "ivory tower" or academic in its approach to policy-making.
I will also try to show that, while I am the sole decision-maker for monetary policy, and remain unambiguously accountable for monetary policy decisions, considerable care has been taken to create processes and structures within the Bank to ensure that I receive comprehensive advice before decisions are taken.
A great deal of the task of formulating monetary policy revolves around gathering, processing and interpreting information. Inevitably, the decisions we take today must involve some form of forecast. And of course forecasts must inevitably be drawn from the information we have available to us today and some sense of how the economy has performed in the past. In particular, we have to make judgements about whether the way people collectively behaved in the past, in the circumstances then prevailing, is a good guide to how they will behave in the future in different circumstances. And if we think that people will behave differently, not only do we have to work out how and why, we also have to make a stab at what it means quantitatively for the economy. None of this is, or can ever be, an exact science. We apply all the science we can muster. But that only takes us so far. After that, we must rely on instinct and intuition, like every other central bank.
Like other central banks, we gather and process an enormous amount of data. The focus of our gathering and processing of data is the production of our quarterly Monetary Policy Statements, to be published this year in March, May, August and November. Let me briefly describe the process that we will employ in producing the next Statement, due for release on 14 March.
The production of this Statement and the accompanying interest rate decision will occupy us intensively for about 8 weeks – and the process is already underway. Much of the background work will be undertaken by our economists, but their analysis is presented to and processed through an internal Monetary Policy Committee. The MPC meets every Tuesday afternoon at 3 o'clock under the chairmanship of a Deputy Governor, Murray Sherwin. Its usual agenda includes a review of new data releases and of developments in financial markets here and abroad. The MPC also considers more substantial pieces of research relevant to monetary policy and sets the priorities for future analysis and research.
But as we approach our quarterly Monetary Policy Statements, the work of the MPC becomes more intensive. At key parts of this 8 week process, the MPC may meet for as many as 5 or 6 times a week, for up to 3 hours at a time.
In the first week or two, our economists travel the country to speak with business and sector representatives in the major cities and provincial centres. We have a contact list of nearly 400 businesses and organisations and typically speak with about 200 of these contacts each year, or 40 to 50 in preparing each Monetary Policy Statement (rather fewer for the March Statement because of the difficulty of finding suitable people willing to talk to us in the last week of January!). These interviews are usually conducted with senior executives and cover questions such as investment and employment intentions, cost pressures, sales expectations and more general impressions of the state of business. The information gathered is collated and used as a cross-check against official statistics as we form our views on economic developments. The information potentially provides us with insights not available from the raw data, and provides clues to how we should interpret the official data series. As such, the effort we put into these business visits is regarded as a core part of our work, though of course we can never hope to get other than a very partial view of what is happening in the economy as a whole from such a small sample.
Incidentally, starting with our next Statement in March, we will be publishing a list of the companies and organisations we talk with in the preparation of each Statement. We will be doing this because the fact of these visits seems to be largely unknown to most people, even in the business sector. Late last year, the council of one large business association expressed concern that we were "no longer" undertaking visits of this kind, while shortly afterwards the chief executive of a very large company told me personally that we should undertake such visits as a way of enriching our understanding of what was happening out in the real world – apparently unaware that my staff had called on two of his senior colleagues just two weeks earlier!
The second and third weeks of our process are typically concentrated on data analysis. We monitor and analyse around 6000 data series in total, and pay particular attention to about 90 of these. Some, of course, are much more important than others, but all help build a picture of how the economy is developing. A major complication in our work is the inevitable time lags in the publication of data. Key numbers such as the CPI are usually published about two or three weeks after the end of the quarter. But other series, such as GDP, are unavailable to us for almost three months. So, as we release our next interest rate decision on 14 March, we will still be two weeks from knowing how rapidly the economy was growing between October and December last. Or in other words, the latest period for which we will have GDP data available when I make the next OCR decision ended in September, almost six months earlier! That means that not only are we having to contend with lags between when we make a decision on interest rates and when that will have its effect on inflation, we are also in the position of making those judgements about the future of the economy with very incomplete information about its current, and even its recently past, state.
We attempt to counter these data lags to some degree by looking for reliable leading indicators of activity. Over the years we have explored all manner of suggestions – everything from sales of typewriter ribbons to GST receipts, road user charges, petrol and diesel sales, orders for workplace overalls, and the production of cardboard packaging material. But even where these indicators provide useful insights, and some of them do sometimes, they are not the economy-wide, comprehensive, measures which ultimately matter for monetary policy. Like our business visits, therefore, they are at best a very imperfect substitute for official statistics.
Having gathered the data – and recall that, because monetary policy works on inflation with a considerable lag, it is the future which must be our primary focus and not the past – the big task is to use that mass of data, and our understanding of how economies work, to project forward how the economy will evolve in the months and years ahead. An important aid in doing this is a forecasting model of the economy.
Most central banks make use of these computer-based models to help them forecast and understand developments in the economy. The core model used by the Bank is probably as sophisticated as any. By feeding in our assumptions and the latest data, we are able to get an indication of how the economy is likely to behave, based on past relationships.
We are also able to run tests to explore the "what ifs" of our various judgements and assumptions. What if key past relationships have changed, if business confidence improves, if the international economy weakens, if the exchange rate rises or falls by 10 per cent, if the public begins to save more, or if we delay or accelerate a shift in interest rates by a quarter or two. By exploring issues of this sort in a formal and structured way, we try to better inform our policy judgements and to lower the risks of wrong or costly decisions.
But while our forecasting model is important, it is not a magic black box that tells us whether to tighten or ease monetary policy. It is a tool, one amongst many, that helps to inform our decisions. It provides a useful starting point for our assessment, but does not provide the answers.
By the fourth week of our process, our Monetary Policy Committee is reviewing developments in the international economy. We don't attempt to do independent forecasts of other economies. Instead, we use a survey of the major forecasters in each of our important trading partners – the Consensus Forecasts. Those forecasts will not always be correct of course, but we expect that, on average, they are likely to be more reliable than anything we could hope to achieve ourselves, even at very considerable cost. The forecast of the international economy becomes a key input into our assessment of likely developments in New Zealand, and so needs to be agreed relatively early in the forecast process.
While we rely on the Consensus Forecasts to establish the "house view" on the international economy, on occasion – the best recent example occurred during the emergence of the Asian crisis in 1997 – we will depart from the Consensus Forecasts and superimpose our own view of the international economy. This is usually relevant when events are unfolding rapidly and it seems likely that the forecasts collected via the Consensus survey are lagging developments in the real world. In late 1997, we recognised this by incorporating the forecasts of only the more pessimistic of the forecasters.
Also in the fourth week, we consider what is known internally as the "hawks and doves" exercise. This is a process whereby we ask four of our analysts who are not directly involved in developing the main forecasts to each write up two views of the current economic situation. Each is required to provide an assessment in which they seek out the lines of analysis that could take us plausibly to a "dovish" policy stance – typically to an easing of interest rates. But the same people are also required to write up the lines of analysis that could plausibly take us to a "hawkish" policy decision – typically in the direction of tightening. This exercise is intended to open out our thinking by alerting us to alternative interpretations of the current economic situation – to lower the risk that we fall quickly into a single view of the emerging situation and lapse quickly into a presumed policy decision. Incidentally, the exercise also illustrates that, because of our incomplete knowledge of what is happening and why, it is usually possible to envisage quite different plausible outcomes on the basis of the information to hand.
Week 5 is our heavy duty week. This is the week in which the MPC goes over the initial forecasts produced by our Economics Department. Tuesday is given over to two lengthy presentations – the morning to discussion of the outlook for the next couple of quarters and the afternoon to exploring the possible evolution of the economy over the next one to three years. These are sessions for listening and asking questions – no conclusions are sought at this stage.
After reflection overnight, Wednesday is when we seek reactions from committee members. We are looking for those parts of the forecast that members feel uncomfortable with, those parts they wish to explore further and alternative views that they wish to have examined. By the end of this session, we are likely to have sent our forecasters away to incorporate modified assumptions, different weightings on some factors, and some alternative scenarios that merit further exploration.
By Thursday, that further work is reported back and debated. By the end of our Thursday meeting, we are aiming to have an agreed set of forecasts that will become the central track to be published in our Monetary Policy Statement, an agreed set of alternative scenarios that may also be published, and generally to have settled on the house view of our economic prospects.
Friday of week 5 is a key day. Our Friday meeting is limited to what is known as the OCR (Official Cash Rate) Advisory Group – a subset of MPC consisting of the two Deputy Governors, the Chief Managers of the Economics Department and the Financial Markets Department, three other senior staff chosen for their ability to contribute to the process, and me. All the members, except me, will have produced and circulated to the other members of the group a brief written note summarising their assessment of the economy and their recommendation for the policy decision to be taken. The task for this Friday meeting is to help me reach a provisional view on what should happen to the Official Cash Rate to be announced with the Monetary Policy Statement. The formal announcement of the OCR is still more than two weeks away of course, so I can change my view up to the day of announcement should there be some surprise development.
In many respects, the projected future track for interest rates also decided at this meeting is as important as the provisional OCR decision. Important too is how the decision compares with current market expectations – a decision to change the OCR that has been fully anticipated by the markets and business community is likely to have little impact on announcement. By contrast, a decision not to change when the prevailing sentiment was that we would or should change is likely to provoke a sharp reaction in markets. For that reason, assessing what markets are expecting is an important part of assessing what impact our decisions might have.
I have mentioned that each member of the OCR Advisory Group provides written advice (excluding me – I generally keep my views to myself until my advisors have had their say). At our Friday meeting, we can explore the views and judgements reached by each member of the committee. Occasionally, views diverge quite significantly; often we will see a consensus emerge from the discussion fairly readily. But we take no votes. Under the structure of the Reserve Bank Act, the Governor is the sole decision-maker. As I said a moment ago, that is intended to make absolutely clear where the responsibility and accountability for monetary policy decisions rest – with the Governor.
I won't dwell on the next 2 ½ weeks, although in many respects they are as important as what has come before. During week 6 we are engaged in drafting the Statement. That is important because explaining why we have taken our decision is in many ways as important as the decision itself. You can see that the timeframe is tight and it is a tribute to my staff that we are able to get from scratch to a published policy document on time and to standard so quickly. The tone of the Statement is important – what "bias" does it convey with respect to the future path of monetary policy, what risks do we choose to identify and dwell on, which alternative scenarios do we choose to identify. This is important in shaping expectations about future monetary policy decisions. As already discussed, having expectations broadly aligned with the policy decision will generally help reduce volatility, disruption and the damaging effects of uncertainty.
The highly structured process I have outlined is designed to ensure that monetary policy decisions are well considered and balanced. Decisions should be informed by the best information we can gather, both in the form of hard data from official sources, and in the form of the "softer" or anecdotal information we gather from companies and organisations up and down the country. We seek the views of outside analysts, sometimes directly and always indirectly, by reading the myriad of commentaries that are produced. And of course we get the views of our own board members, both during the course of preparing each Statement and in the increasingly thorough review which non-executive directors undertake after the event.
The process by which we use all this information is also crucial. For that reason, we have invested a good deal of thought in developing the ways in which we process information and form our views.
But of course none of this guarantees that we get to the right decisions all the time. Monetary policy inevitably requires us to make decisions in the face of considerable uncertainty – uncertainty about the current state of the economy, uncertainty about how the economy works, uncertainty about how old relationships are changing, and uncertainty about the really big sources of surprise, such as the weather or international commodity prices. As The Economist magazine observed recently, "monetary policy is often likened to driving with a blacked-out windscreen, a wonky speedometer and a cracked rear-view mirror". Precisely.
And the recent past provides a good illustration. In late November, as we finalised the text for the Monetary Policy Statement published on 6 December, it still seemed likely that the economies of our main trading partners would remain reasonably robust this year; the New Zealand dollar was near its all-time lows; petrol prices were high; and business confidence had recovered strongly. It seemed entirely appropriate to signal the likelihood that, if those circumstances continued, some increase in the Official Cash Rate would be needed in the first few months of this year. In the next seven or eight weeks, evidence accumulated quickly that the US and possibly the Australian economies were slowing quite rapidly; the New Zealand dollar appreciated significantly; and petrol prices fell, suggesting that the spike in the headline CPI would be of rather shorter duration than previously expected. So when we reviewed the Official Cash Rate earlier this week, I announced that it would remain unchanged in the meantime.
Somewhat to my surprise, I have heard some criticism of the change in our tone between the two statements, the one in early December and the other this week. Surely, it has been suggested, we were "wrong" in early December? I would prefer to say that circumstances changed, and changed quite substantially, between early December and late January. It was clear in our December Monetary Policy Statement that the policy moves which we foreshadowed at that time were conditional on the world economy, on the exchange rate, and on all the rest. As John Maynard Keynes once remarked, "When the facts change, I change my mind – what do you do, sir?"
One of the great advantages of the transparency with which we set out the rationale for our policy view is that financial markets can assess the strength of the arguments, and can modify their views about likely changes in policy as new information comes to hand. It was no accident that, in the week or so prior to my announcement of no change in the Official Cash Rate this week, literally every commentator surveyed had come to the conclusion that we would make no change in the rate. They had read our December Statement, had understood the conditionality of the views expressed in that document, and had seen the emerging data on the world economy, on the exchange rate, and so on. Indeed, it should be, and indeed is, unusual for a Bank monetary policy decision to surprise financial markets, and that is undoubtedly beneficial. Of course, that does not mean that the Bank should be a prisoner of the markets' view: there will be occasions when the Bank will differ from the view of the market.
In summary, I believe that the monetary policy framework we have in New Zealand has served the country well. Clearly, we should be constantly seeking ways to improve it still further, and for that reason we look forward to reading Professor Svensson's report with great interest.
 The Economist, 27 March 1999, p. 19.
 Quoted in The Economist, 24 October 1998, p. 57.