Heikensten: Monetary policy
Thank you for the invitation to talk about monetary policy. If my memory serves me right, this is the fifth autumn conference I have addressed as deputy central bank governor. For me, the annual opportunity of discussing monetary policy with some of our country’s leading economists and economic observers has become a very enjoyable tradition.
I shall begin, as usual, with a brief account of the past year, taking 1999’s October assessment as my starting point. Even though our appraisal of inflation at that time was relatively optimistic, the outcome has been better than we foresaw. Developments in the past year raise questions about the economy’s inflation propensity. That topic was already up for discussion at last year’s conference, though the reason then was that several leading observers considered the risks of inflation were being underestimated by the Riksbank. A third matter I shall be talking about is our monetary policy guidelines. In the past year we have been criticised for basing repo rate decisions on appraisals that are too narrow. Finally, I shall be looking ahead and commenting on our view of inflation in the coming years.
Inflation in the past twelve months
In the third Inflation Report last year the Riksbank’s growth forecast for Sweden in the coming years was revised upwards. This mainly mirrored a more favourable outlook for international activity; it also seemed likely that domestic demand would be stronger than had been expected earlier. Our overall growth forecast agreed with most other observers’. Some predictions, however, did include an even stronger increase in household consumption, the reason being that rising asset prices in general and the stock market boom in particular were judged to lead to a markedly increased consumption propensity.
Our higher growth forecast was accompanied by little change in the assessment of inflation we had presented earlier in the year. Thus, we judged that the relationship between growth and inflation would be more favourable than we had previously calculated. In this respect we differed from many other observers and that occasioned a good deal of discussion.
The background to our forecast was that inflation’s path had been surprisingly favourable and that, instead of being just a result of occasional factors, this had to do with more fundamental features in the economy that would continue to influence inflation up to our two-year horizon.
Here is a picture of our inflation assessment last October and the outcome (Chart 1). I have chosen to present it in terms of UND1X because that is the indicator of inflation on which monetary policy has been focused in the past year. The chart also shows the price of imports and an indicator of underlying domestic inflation, UNDINHX. As you will see, we believed inflation would move up gradually in the wake of rising activity. This would be countered by low, even falling, import prices in connection with a stronger exchange rate, lower oil prices and generally strong global competition.
Even our inflation forecast was evidently on the high side, at least to date. It is perhaps even more noteworthy that underlying domestic inflation has turned out to be appreciably lower. This is something that observers of the Swedish economy cannot ignore, particularly as the past decade has provided so many surprises of this sort, though not as dramatic. The main point to consider is, of course, how permanent the changes are likely to be. To what extent should we abandon earlier relationships and commit ourselves to a continuation of the relationships that have applied in recent years?
Why has inflation been low?
The first thing to note is that this development has not been confined to Sweden. Inflation in many other countries has also been unexpectedly low. There, too, the reasons for this are not entirely clear. Some pieces of the puzzle have been identified but not enough to make up a clear picture. As decision-makers, however, we cannot simply shrug our shoulders and say we don’t know. Decisions have to be made and it would not be reasonable to go on acting as though nothing had changed until sufficiently long time series are available for econometric analyses.
Recently I had the pleasure of chairing a central bank seminar that BIS arranged on just these issues. My impression was that it has proved difficult to produce definite evidence of any structural shift in the relationships that govern inflation, such as various versions of the Phillips curve. In many countries, including Sweden, appraisals of resource utilisation, for example, are now more optimistic. As you are no doubt aware, in the past five years our output gap estimates have in fact been revised a number of times. One explanation I find fairly plausible is that after the profound crisis in the early 1990s we simply misjudged the size of the gap. There is also some support for the possibility that productivity has followed a more favourable trend than expected, thereby reducing the risk of bottlenecks. Total factor productivity, for instance, has been on a rising trend since 1993. Another possibility is that, for various reasons, the labour market is functioning more efficiently. One indication of this is that rapidly falling unemployment in recent years has not been accompanied by acute labour shortages. Perhaps I should mention that we have also detected certain changes in inflation relationships. In a box in one of last year’s inflation reports, for example, we found that the impact on inflation from a given output gap has decreased; one explanation for this could have to do with inflation expectations being more stable.
At the BIS seminar, moreover, two papers provided indications of changes in relationships behind the prices of imported goods. That agrees with the picture we have formed in recent years. The pass-through to inflation from exchange rate movements, for example, is lower than before. This may be connected with stronger competition at various levels. It also seems clear that permanent exchange rate adjustments have price effects that differ from and are greater than those elicited by movements that are a natural part of today’s flexible regime.
In any event, in the last few years there have also unquestionably been favourable effects from the deregulation of electricity and telecom markets. Even though similar changes of equal importance for the rate of price increases in the coming years are not easy to identify today, both deregulations and generally growing competition will presumably be important features in the picture of inflation.
Another point worth making is that in a world where monetary policy with a more independent mandate is focused on low inflation in a growing number of countries, low inflation as such is presumably an important factor. That is really what the textbooks teach us to expect, with their standard message that inflation is a monetary phenomenon, determined precisely by monetary policy. Inflation expectations that are low and stable clearly help to combat inflation.
But I believe the low-inflation policy has wider implications. It is presumably one factor behind the recent improvement in trend productivity in Sweden. Firms today are exposed to considerably stronger pressure to change than they were in the 1980s. The conditions for labour market negotiations have also changed in ways that may well have positive effects on unit labour costs and thereby on inflation. Another consequence is that measures for strengthening competition have suddenly become politically relevant in ways that were not the case at least when I was working on such questions in the 1980s. Deregulation and the enhancement of competition can now have positive effects that are directly evident in the form of low inflation and thereby lower interest rates and higher economic growth. Moreover, with the division of responsibilities that now applies in stabilisation policy, fiscal policy is freer to focus on long-term issues and structural matters.
The Riksbank’s way of working
I should now like to comment briefly on how the Riksbank works and on the implementation of our so-called rule of action. These matters have featured in the general debate in the past year, presumably just because inflation and our inflation forecasts have continued to provide relatively weak arguments for raising the repo rate even though activity is strong.
In simple terms, the Riksbank can be said to formulate monetary policy on the basis of inflation’s forecast rate, in the first place one to two years ahead. If inflation is expected to be above the 2 per cent target during this time horizon, the interest rate is normally raised and vice versa.
This rule was formulated with a view to making policy more intelligible, to ourselves as well as to others. We also believed a rule of this kind would serve as a structure for decision-making and help to enhance confidence in monetary policy. In my opinion, the rule has served these purposes. It has contributed, not least, to a more structured discussion of monetary policy issues and thereby to a deeper understanding of the choices we have to make.
Have we observed our rule? Judging from this chart, I would say we have (Chart 2). The chart shows the repo rate in practice together with the path indicated by the Riksbank’s inflation assessments. You will see that we have followed our forecasts fairly closely and thus adhered to our rule of action. Deviations have occurred, mainly in the early years with the inflation target; I believe they were largely connected with concern about the exchange rate.
While I accordingly consider that a rule of this kind is of considerable value, I must underscore that we never intended to implement it mechanically. What matters is that we secure our objective in the longer run, really for ever.
We chose to focus on the period from twelve to twenty-four months ahead for a number of reasons. One reason for not having a shorter target horizon is the time lag that applies before an interest rate adjustment normally affects the economy and inflation. But effects do arise in the shorter run, for instance via the exchange rate. We pay less heed to developments more than two years ahead partly because forecasting the longer term is normally difficult. Still, although monetary policy is guided mainly by prospects one to two years ahead, there are normally grounds for paying some attention to forecasts both for the coming year and for the period beyond the two-year horizon.
Neither is it reasonable in practice to focus solely on inflation. For one thing, the Riksbank is also accountable for the proper functioning of the payment system in general, for financial stability. For another, without prejudicing these two objectives, the antecedents to the Riksbank Act require us to promote other economic policy goals, such as good economic growth and increased employment. This amounts in practice to trying to make decisions that avoid major shocks and fluctuations in growth and employment.
Highly respectable reasons can thus be found for deviating from the rule of action. If we decide to do so, however, we must have good arguments and present them particularly clearly. General talk of financial imbalances, for example, will not do. As I see it, we must be clear about which balances we are referring to and why we consider they threaten the economic future. Another point to bear in mind is that the rule of action with a two-year horizon should normally give us sufficient time in which to adjust the monetary stance.
Before turning to my last topic, I can mention that from time to time we are exhorted to be more pragmatic and less tied down by our intellectual framework. It is quite often suggested that we should follow the example of the Federal Reserve in recent years under Alan Greenspan. I find that basically naive. Policy principles that are less clear-cut, combined with less transparency about motives and so on, would by no means guarantee a more successful policy in the longer run. Just the reverse, I would say, in the light of quite a few years in the vicinity of decision-making in economic policy.
Economic outlook in the coming years
A suitable starting point for a discussion of the economic future is the October Inflation Report. Our assessment there assumed that international economic activity will tend to slacken but still remain relatively strong. We counted on GDP growth in the OECD area being 3.8 per cent this year, followed in the next two years by 3.0 and 2.7 per cent, respectively. Inflation would remain relatively low, partly in view of strong competition in the international markets.
Growth in Sweden was judged to be 4 per cent this year, 3.7 per cent next year and 3.0 per cent in 2002. Domestic demand growth would be relatively high for consumption as well as investment. A damper was expected, however, from an appreciation of the krona. A gradual increase in inflation was predicted, so that two years ahead the rate would be approximately in line with our target.
As usual, we identified various risks in our forecast. One was the possibility of the oil price staying at a high level. We also considered that the exchange rate might be weaker than economic fundamentals in Sweden would warrant in the longer run. Another risk was that of higher wage increases. It was conceivable that the wage rise would be even higher than the annual rate of just over 4 per cent we had assumed for the coming two years, due, for example, to far-reaching demands for compensation and parity. The main downside risk — lower inflation than in the main scenario — lay in favourable surprises of the kind we have seen earlier in the relationship between growth and inflation.
So how does this picture from October stand up today? Fairly well, in my opinion.
Price movements to date have been broadly in line with the forecast. And as we expected, there have been signs that international activity is slackening. This is most evident in the United States but even some slowing of growth in Europe seems likely in the coming years.
In Sweden, too, there are indications that activity has become calmer. Growth still seems to be high but households as well as firms have become less optimistic about the future. Orderbooks are not growing as drastically as before. In general, this is in line with what we foresaw in October. But of course we are keeping a close watch on developments. A more dramatic course cannot be ruled out, particularly in the United States. The oil price remains high and the krona has weakened. So in these respects, developments are more in line with the risk spectrum than with our main scenario.
Share prices, as we all know, have been volatile in the past six months but the fall to date is not a major cause for concern. As I mentioned earlier, during 1999 and early this year a number of observers saw rising share prices as a strong reason why the Swedish economy might become overheated. At the Riksbank we were more inclined to wait and see. There is no reason now to adopt a different view and exaggerate the risks of the share price fall. Consumption is being sustained in the first place by a strong development of income that stems from increased employment, a good increase in real wages and a fiscal stance that is less restrictive than before.
There is some cause for concern in the present round of wage negotiations. In a sense, that is inevitably the case when many settlements are being concluded simultaneously and the wage share is approximately two-thirds of the economy. That means that no other single factor is anything like as important as the wage trend for the path of the Swedish economy in the coming years. The picture of the negotiations that are now beginning is mixed. The outcome to date has been a positive surprise. It also looks as though the negotiating parties are formulating their bids on the basis of the inflation target. Moreover, the available picture of future wage expectations looks relatively good (Chart 3). Here we see the wage outcome, wage expectations and unfilled job vacancies. It is evidently becoming increasingly difficult to find the right people for the jobs that are advertised. This presumably makes the risk of wage drift greater than in previous years. Moreover, the wage demands that have been presented are higher than before the previous round of negotiations.
When it comes to judging future inflationary pressure, the degree of resource utilisation is the crucial concept, at least as long as inflation expectations are parked around the 2 per cent target. The next chart is instructive here (Chart 4). As you know, we estimate total resource utilisation econometrically in terms of the output gap. If we take the October estimates as a starting point and project them with the accompanying GDP forecast, we find that for the first time in a number of years we look like having full or more than full resource utilisation in the years ahead. Perhaps the picture exaggerates the risk of future inflationary tensions. The cyclical position, for example, may be somewhat calmer than we counted on in October. But even with a growth rate of just over 3 per cent, the Swedish economy would be expanding more rapidly than we believe it is capable of in the longer run.
Let me conclude by underscoring what I mentioned initially, namely that forecasting inflation proved difficult in the 1990s, in Sweden as well as elsewhere. There have been substantial changes in our economies, partly because policy now focuses on stability and competition has become stronger in various ways. Inflation in Sweden has been a positive surprise in recent years even though activity has been rising strongly, so further positive surprises should naturally not be ruled out. Still, there are now many indications that resource utilisation is relatively high and will go on rising. That implies that, little by little, inflationary pressure will grow.
We will be presenting a comprehensive assessment of inflation prospects in the December Inflation Report. In that context we will also be deciding the repo rate. Recent developments have not altered the picture I formed in October. Although it is still too early to specify when, there will presumably be grounds for increasing the repo rate in the future.
