Rosenberg: Monetary policy in Sweden
Thank you for the invitation to come here and speak. The topic of the conference is economic prospects and I will of course provide my views on these. However, I will speak in fairly general terms – a more detailed analysis will be presented in the Riksbank’s Inflation Report, which is published at the end of October.
Before I move onto economic developments, however, I would like to take this opportunity to say a few words about how monetary policy is conducted in Sweden in general, or the monetary policy strategy, as it is usually called. More specifically, I intend to take up some questions that have arisen in the discussions in the media and elsewhere as there still seems to be some confusion.
This applies in particular to the objective of monetary policy. Do we attach less importance to the inflation target now, or have we even introduced other objectives in addition to the inflation target? Let me emphasise right now that no such changes have been made. Our focus is on the inflation target and the central aim is to anchor inflation expectations in the economy around this target. However, when we formulate our monetary policy we also give consideration to the way the rest of the economy is developing. It is our reasoning on this that I intend to begin with.
The inflation target and consideration of the real economy
The Riksbank has the statutory objective of maintaining a low and stable inflation rate, or of “maintaining price stability” as it is expressed in the Sveriges Riksbank Act. We have specified this task as ensuring that CPI inflation is around 2 per cent a year, with a tolerated deviation interval of +/- 1 one percentage point. As long as confidence in the inflation target is not threatened – or, as it is usually expressed, as long as the nominal anchor for setting prices and wages is not at risk – we also have the opportunity to take into consideration how the real economy, such as production and employment, develops.
Although the latter has not been written into the Act, it was stated in the preliminary work to the new Sveriges Riksbank Act that came into force in 1999 that it was natural that the Riksbank should support the economic policy objectives of sustainable growth and high employment without neglecting the long-term price stability objective. This was considered a natural consequence of the Riksbank being an authority under the Riksdag (the Swedish parliament). It was therefore considered unnecessary to specify any further objective than price stability in the law text. There were probably also some fears that statutory objectives for the real economy could give the wrong impression – that monetary policy could also govern long-term developments in growth and employment, for example. It is now widely accepted that these are determined by factors that are not directly affected by monetary policy, such as the rate of technological advances and the efficient functioning of various markets.
In general terms, one can say that the policy conducted by the Riksbank is aimed partly at limiting the inevitable deviations from the inflation target and partly at ensuring that the real economy does not fluctuate excessively. The objective of monetary policy is price stability, but the Riksbank also considers real economy considerations to be important. This is characteristic of what is known as flexible inflation targeting, which as far as I know, is conducted by all central banks that have inflation targets.
Monetary policy affects the economy with a certain time lag. Monetary policy decisions must therefore be based on forecasts of future inflation. Normally, monetary policy is aimed at achieving the inflation target within two years to avoid the deviations from target from being too large and too protracted. This two-year horizon can be regarded as a restriction that the Riksbank has placed upon itself as a means of upholding confidence in the inflation target. At the same time, it provides an opportunity to show consideration for the real economy. If, for instance, there is a shock that leads to inflation being pushed upwards despite weak demand, a very rapid return to the inflation target would entail large negative effects for the real economy – an even greater slowdown in the already weak demand – than if the recovery was allowed to occur at a slower pace. If inflation is instead held back despite good growth in demand, attempts to quickly return it to target through very expansionary monetary policy could lead to the economy overheating. Over time a policy that quickly subdues inflationary impulses would perhaps result in more stable inflation, but there would be substantial fluctuations in the real economy.
We have chosen a two-year horizon for monetary policy because this is considered to give sufficient scope in most cases to ensure acceptable developments in the real economy. However, the rate at which it is desirable to bring inflation back on target depends on what shocks the economy has suffered and how substantial they are. Sometimes deviations from the target may be so substantial that there is reason to allow inflation to return to target beyond the normal two-year horizon and in this case we explain the reasons when we announce our decision. However, the ambition remains the same; that is, a desire to ensure that inflation is low and stable a few years ahead, while giving some consideration to the stability of the real economy.
What significance do house prices and household indebtedness have?
How do developments in house prices and household indebtedness come into the picture? House prices must be given consideration in the forecasts as they affect, through the wealth effects, households’ decisions regarding consumption and saving. The size of household debts is important to the way in which interest rate changes affect consumption.
However, one problem is that it can be difficult in the normal forecasting and analysis work to capture the risks that ensue from large fluctuations in house prices and other asset prices. One reason for this is that the models usually used to make forecasts do not have such sophisticated descriptions of the financial sector and its interplay with the real economy. The problems are particularly great when house prices and credit growth have over a long period developed in a manner that is not sustainable in the long term. It is simply not possible, with the analysis and forecasting tools currently available, to forecast with sufficient accuracy whether, for instance, there will be a sudden correction in house prices, or whether there will be a calmer, more protracted adjustment.
At the same time, the Riksbank must take into account the risks when conducting its monetary policy. If the rate of increase in house prices and household debts remains at a high level over a long period of time, the risk increases that imbalances will build up and will force an adjustment later on, which could in the worst case have significant consequences for the real economy and inflation. The way in which the Riksbank can manage this type of risk when making monetary policy decisions is, in principle, to make changes in the interest rate sooner or later than would otherwise be assessed as the most appropriate timing, given the “normal” forecasts for inflation and the real economy. It is hoped that such a policy will contribute to a smoother adjustment process for house prices and help to avoid a more abrupt correction.
One example of when this type of reasoning has been applied is at the monetary policy meeting in February. We then chose to raise the interest rate by 0.25 percentage points, which was also in line with market expectations, despite the fact that the forecasted inflation path showed the rate as being at the lower edge of the target margin two years ahead, and low for the majority of the forecast period. The inflation forecast as such could therefore have warranted delaying a few months before raising the rate.
It is important to realise that this type of decision with preventive interest rate changes does not mean the inflation target has been “downgraded” or that we now have a special target for house prices. The reason is that we believe it is the best way to safeguard that we have been set to safeguard – stable prices while avoiding large fluctuations in the real economy.
I do have some understanding for those who consider this type of consideration to mean that monetary policy becomes “discretionary”. This is of course true to some extent. However, at the same time I find it difficult to see that we should act in a different way when we are now, as far as I know for the first time since introducing the inflation target, experiencing a period when house prices and household indebtedness have developed in a way that raises concerns that monetary policy has been too expansionary. The important thing is that the Riksbank in these situations reports the motives for the decisions made as clearly as possible. However, the considerations are very difficult to make and it is not always possible to provide simple, precise explanations.
One might say, somewhat provocatively perhaps, that it appears as though many have interpreted our objective of maintaining price stability, or more generally, safeguarding confidence in the inflation target as meaning that the Riksbank should conduct a policy where the inflation forecast always shows inflation as being on target two years ahead. However, this is an overly narrow interpretation and I am convinced that if we were to follow this principle in all situations the final result, seen over a long period of time, would be worse than with a policy that also gives consideration to other prevailing circumstances. In this context it can also be observed that, despite the debate that monetary policy has become less clear and that there is a risk that confidence in the inflation target will weaken, none of this is reflected in inflation expectations. These remain firmly anchored around 2 per cent.
Much of my reasoning here on the Riksbank’s monetary policy strategy is described in a booklet entitled “Monetary policy in Sweden”, which we published a few months ago(1). The most important aim of this booklet is to explain how the Riksbank, when setting its interest rate, has scope to take into consideration developments in both inflation and in the real economy.
One step in this direction was taken already back in 1999 when the clarification of monetary policy was published by the Executive Board for the purpose of explaining that the Riksbank conducted flexible inflation targeting (2). In the period since 1999 we have learned even more about how flexible inflation targeting is conducted. There has also been considerable progress in the academic research on this subject. New insights led to a decision to issue a new document on monetary policy strategy. This is essentially a question of developments and changes in the method of describing and communicating inflation targeting policy rather than in the basic strategy itself.
Inflation in focus, not the interest rate
Let me conclude this section of my speech with a few words on a concept that has been used in the discussions on monetary policy recently, namely that the Riksbank has begun a “normalisation” of the repo rate. The way the term ”normalise” is used in this context usually means that the policy rate is being increased after having been very low for a period, that is, it is simply a way of pointing out the natural direction for the policy rate in future.
At the same time, there is an implication in the term “normalise” that there is a level for the policy rate where it normally should be. Some analysts appear to have interpreted this to mean that the Riksbank’s increases in the repo rate point reflect that there is an end in itself to achieving such a "normal" level now that the economic upturn is well established and most curves are pointing upwards, quite regardless of how inflation may develop. However, this is not the way we reason.
The fact that the interest rate is below the "normal" level is not in itself a reason to increase the rate, even if real economic developments are favourable. It is not an argument that can stand for itself. The decisive factor for future interest rates is the result of our forecasts and assessments of inflation and economic developments as a whole. It may be the case, for instance, that different factors are expected to hold back inflation over a relatively long period of time and that there may thus be reason, even in an economic boom, to allow interest rates to “normalise” – if we are to use this term – at a slower rate than would otherwise be justified. In recent years weak growth in import prices and rapid productivity growth have been such factors. On the other hand, it may also be the case, if inflation is high and held up by special factors for a fairly long period of time, that interest rates must be maintained at a high level even when economic activity is not so strong.
This means that it is also very difficult in practice to determine what is a "normal” level for the policy rate. It is possible to use different methods to calculate an approximate interval for what can be regarded as a normal repo rate. However, this type of interval will of necessity be fairly broad – in a box in the June Inflation Report it was estimated at 3.5-5 per cent. Experiences also show that the repo rate can fall outside of this rather broad range for fairly long periods of time.
Given this, it can be observed that estimates of the normal level for the repo rate have a fairly limited value when assessing interest rate developments in the near future and that interest rate decisions cannot be justified merely because the repo rate is deviating from a level regarded as normal. Ultimately, it is the inflation forecast in relation to the target that decides monetary policy. How quickly the target should be attained is also decided by developments in the rest of the economy.
The economic situation
Let me go on to say something about the monetary policy situation and the prospects for inflation and economic activity. It is the inflation assessment on which we based the decision to raise the repo rate by 0.25 percentage points at our most recent monetary policy meeting on 29 August that I intend to discuss. An account of the discussion held at the meeting is contained in the separate minutes published last week.
We observed in the Inflation Report published in June that economic activity in Sweden had strengthened, largely because of strong international developments, high productivity and expansionary monetary policy. Our assessment was that these factors would also continue to stimulate the economy. GDP growth this year was therefore expected to be high, which was supported by both the national accounts figures for the first quarter and the economic indicators for developments during the second quarter. There were also signals from the labour market that economic activity was continuing to improve, with an increasing demand for labour.
We could also note that inflation had begun to increase, after having been strikingly low. The high inflation rate was partly explained by rising energy prices, but prices of certain other goods had also been raised more quickly than before. Our assessment was that inflation would not change very much over the coming year and would then gradually increase as economic activity improved and there was a rise in the rate of wage increase. Our forecast was that inflation would be on target a couple of years ahead, on the assumption that monetary policy gradually became less expansionary in line with market expectations in June.
International statistics received over the summer indicate that economic prospects look more favourable now than we had assumed in June. The signals regarding the US economy are mixed, however, and some indicate that growth will be weaker than expected next year. However, the picture for the euro area and other parts of the world is that developments have been slightly stronger than assumed in our earlier forecasts.
With regard to the Swedish economy, the national accounts for the second quarter have confirmed our assessment in June that growth would remain strong. The statistics show that GDP growth during the second quarter was actually higher than our forecast. However, the development in the number of hours worked and the number of persons employed was weaker than expected, which reflected the fact that productivity growth had been much higher than anticipated. The new indicators received point to continued stable economic activity and to an improvement in the labour market in future. All in all, this indicates that growth will be higher than we had assumed in June, even taking into account the fact that expectations of repo rate increases have been adjusted upwards over the summer.
The upward revision in growth prospects has not led to any major revision in our inflation assessment. Although the high GDP growth in itself indicates that inflation may rise more rapidly in future, the strong productivity growth is at the same time expected to contribute to cost pressures being lower this year than was assumed in June. Cost developments in the somewhat longer term are very difficult to assess. If productivity continues to develop strongly, companies may postpone recruiting new staff, which would mean a delay in the labour market upturn and continued low cost pressures. However, there are a number of indications that the labour market is now strengthening and a more rapid increase in employment will probably contribute to rising cost pressures in future.
During the summer inflation has been slightly higher than expected, which was mainly due to rising energy prices. The inflationary impulse this entails is to some extent counteracted by the low unit labour costs, but inflation this year will probably be slightly higher than we had assumed in our June forecast. Our assessment is that the effects of the energy prices will subside at the beginning of 2007. However, as with productivity growth, there is a risk that we are underestimating the effects – this could mean that energy prices may continue to rise and that there could be contagion effects pushing up inflation further ahead. During 2007 rising unit labour costs will provide an increasing contribution to inflation and a couple of years ahead underlying inflation is expected to be in line with the target, given a gradual increase in the repo rate.
Prospects for economic activity and inflation thus indicate that less expansionary monetary policy is called for in future. This will ensure that inflation is close to the target and that developments in the real economy are balanced. Given this, we chose, in line with the repo rate path expected by the market, to raise the repo rate by 0.25 percentage points at the monetary policy meeting in August. The assessment of the prospects for economic activity and inflation over the coming years that we made then showed it would be reasonable to also assume that the gradual repo rate increases will need to continue. I see no reason to change this assessment today.
Thank you.
Notes:
1) See Monetary policy in Sweden”, Sveriges Riksbank, 2006 (can be downloaded as a file from the Riksbank's website, www.riksbank.com, or ordered in booklet form).
2) Memorandum “Riksbankens inflationsmål – förtydligande och utvärdering”, 4 February 1999, registration no. 1999-00351 DIR, or Heikensten, L., “The Riksbank’s inflation target – clarifications and evaluation” Sveriges Riksbank Quarterly Review No. 1, 1999.
