Rosenberg: Assessment of monetary policy
Let me begin by thanking you for the invitation to come here and speak to you! This type of conference, with participants from various inflation-targeting central banks, and where we can discuss common issues is very important in helping us to improve our knowledge.
Introduction
Many countries, including the UK, Norway and Sweden, today conduct monetary policy with an explicit inflation target. Moreover, in many countries the central bank has been given an independent position and responsibility for maintaining price stability. Experiences indicate clearly that it is difficult to maintain low, stable inflation unless the central bank can act with a large measure of independence. Maintaining confidence in monetary policy and in the inflation target requires that the central bank can be free from any suspicion that it might give way to pressure from the political system or from various pressure groups to conduct monetary policy aimed at other targets that are not compatible with price stability.
Maintaining confidence also requires an understanding in society as a whole and broad political support for monetary policy. The central bank’s independent position therefore entails a requirement for openness and clarity with regard to the bank's method of working with monetary policy. Independence also means that it is important to be able to assess monetary policy. I therefore really appreciate the opportunity to participate in today's discussion on assessing monetary policy.
I have today chosen to highlight a very fundamental question with regard to assessing monetary policy; is it really necessary to interpret inflation that deviates from target as a failure for monetary policy? It is clear that short-term deviations do not necessarily imply failure. But how should we regard a deviation from target that persists for a longer period of time, perhaps several years? Here I take as a point of departure the changes in the supply side of the economy that have occurred in recent years in a number of countries, including Sweden and Norway, where inflation has for a period of time been below target. I will, however, begin with a more general description of my views on the assessment of monetary policy and how we in Sweden work to create the conditions for a good and fair assessment of monetary policy.
Assessing monetary policy
When discussing and assessing central banks’ target fulfilment, it is important to base the assessment on a realistic picture of what monetary policy actually can and should achieve. To make the most correct assessment possible, it is therefore necessary for central banks to have a clear analytical framework to produce the material on which decisions are made. Similarly, it must be quite clear which principles have been used as a basis for monetary policy decisions and on what information the conclusions were based.
An assessment of monetary policy must be based on how the inflation target is formulated, and one important issue is of course whether the central bank has interpreted the target in a way that is compatible with the original intention. Most central banks take into account not only inflation, but also in one way or another the stability of the real economy. For an assessment of this policy to be correct, there must be a description of how it is done.
An assessment must also be based on what monetary policy can achieve. Inflation is constantly subjected to shocks and as monetary policy affects the economy with a time lag, it is not possible to keep inflation exactly on target all of the time. Monetary policy decisions must therefore be based on forecasts. The success of monetary policy will be partly dependent on the accuracy of the forecasts. An analysis of the forecasts and the reasons behind various forecasting errors must therefore form part of an assessment of monetary policy.
However, there is considerable uncertainty linked to forecasts for the economy a couple of years ahead and there are many reasons why forecasts prove inaccurate. As forecasts are based on the information available at that point in time, it is natural that unexpected events can occur which cause the economy to develop differently than the central bank had expected. There is insufficient knowledge of how the economy functions and of how monetary policy affects the different parts of the economy. Errors and omissions in statistics can also be one reason why forecasts are based on incorrect assumptions. It is therefore also interesting to assess the central bank’s forecasts against those of other analysts. Would it have been possible to make better forecasts than those the central bank produced? In order to answer this question, it is necessary to keep a record of both the central bank’s forecasts and other analysts’ forecasts.
Another important factor in an assessment is whether the monetary policy decisions are well-balanced, given the forecasts and judgements that have been made. The way the strategy has been formulated is important in this context. However, an assessment does not need to stop here. It may of course be more comprehensive and cover not only target fulfilment, but also the monetary policy framework as a whole.
Against this background, let me just say a few words about the monetary policy strategy we have chosen in Sweden and the changes we have recently made to our analysis framework. The strategy has of course developed gradually during the period we have applied inflation targeting and is primarily the result of practical experiences of inflation targeting in Sweden and in other countries. New findings in academic research in this field have also had significance.
The Riksbank’s monetary policy strategy
In Sweden, the Riksdag (the Swedish parliament) has delegated the responsibility for monetary policy to the Riksbank, through the Sveriges Riksbank Act, which came into force in 1999. According to this Act, the objective of monetary policy is to maintain price stability. The Riksbank has specified this as an inflation target, which states that the annual rate of change in CPI should be 2 per cent, with a permitted deviation of ± 1 percentage point. The Riksbank shall also promote a safe and efficient payment system. The preliminary work to the Sveriges Riksbank Act also makes clear that the Riksbank as a public authority under the Riksdag, without neglecting the price stability target, should support the general economic policy objectives of sustainable growth and high employment.
The Sveriges Riksbank Act also established increased independence for the Riksbank. Since 1999, decisions on the repo rate have been made by the Executive Board of the Riksbank and the members of the Board are expressly forbidden from seeking or receiving instructions when carrying out their monetary policy tasks.
Price stability is thus the objective of Swedish monetary policy, but like Norges Bank and most other inflation-targeting central banks we conduct monetary policy that also to some extent takes into account activity in the real economy. The fact that the interest rate is often adjusted in relatively small and predictable steps can partly be regarded as an expression of this. If the aim had instead been to bring back inflation to the target as quickly as possible, without any other considerations, it is probable that larger and more frequent adjustments would have been warranted. A more gradual change in the interest rate can also be justified by the uncertainty of the forecasts and statistics.
Monetary policy is normally aimed at attaining the inflation target within two years. This horizon is justified by the fact that monetary policy's impact comes after a time lag, but also by the desire to contribute to subduing fluctuations in activity in the real economy. By aiming to allow some time before the inflation target is attained, the Riksbank can contribute to stabilising the development of the real economy, but at the same time we must safeguard confidence in the inflation target. Taking into consideration developments in the real economy must not mean that the price stability target is set aside. The two-year horizon provides scope for flexibility but at the same time sets a limit for how large the deviations from target can normally be.
The fact that we have chosen to specify an exact target for inflation is an expression of our desire to be clear and open. This is also why we have explicitly stated that inflation will normally be returned to target within two years. Similarly, the deviation interval around the inflation target aims to explain that monetary policy cannot, and does not, attempt to exactly attain the target all of the time. The deviation interval thus also emphasises that there must be scope for temporary deviations from the target, which may be justified with regard to avoid unnecessary fluctuations in growth and employment.
We thus endeavour to attain the inflation target within a two-year time perspective, under normal circumstances. In certain cases, however, we may need to have even greater flexibility than the two-year perspective allows. If the economy were to suffer very large supply shocks, for instance, in the form of structural changes that have a major effect on inflation, or if extreme situations such as war or natural disasters occur, the process of bringing inflation back to the target may need to take even longer to avoid excessive strain on the real economy. In this type of situation it would be even more important that there was confidence in the inflation target. However, flexibility must not be allowed to undermine the nominal anchor. If confidence were to fail, we would be forced to act more strongly and would not have the same possibility to consider the real economy.
Thus, the objective of monetary policy is to maintain price stability. However, it is possible to give consideration to the development of the real economy. This does not mean having a target for growth or employment. Instead, this type of consideration is made possible by the way in which we have chosen to formulate the principles for attaining the inflation target. However, it assumes that there is sufficient confidence in the inflation-targeting policy.
Changes in our analytical framework
As monetary policy has an effect after a time lag, the Riksbank must look ahead and base its interest rate decisions on forecasts of future developments, for instance, regarding economic activity and inflation.
Like other central banks, we work constantly on improving our analysis work in order to produce the best possible forecasts and analyses, and these improvements in methods also contribute over time to making monetary policy easier to understand and assess. In recent years, for instance, we have focussed on improving the models we use in our forecast work.
Last autumn we changed the interest rate assumption that forms the basis for our forecasts and at the same time we chose to lengthen the time horizon for the forecasts. Previously, up to and including the second Inflation Report published in 2005, the Riksbank’s forecasts were based on the assumption that the repo rate would remain constant over two years. As the repo rate is usually adjusted during the forecast period, and thereby affects the outcome of inflation and other economic quantities, it is difficult to obtain a good perception of the accuracy of the forecast with this assumption. It therefore makes it more difficult to assess monetary policy.
As of the third Inflation Report published in 2005, the forecasts are instead based on the assumption that the repo rate will follow market expectations as reflected in implied forward rates. This makes is easier to compare the Riksbank’s forecasts with those made by other forecasters. They do not usually base their forecasts on an unchanged interest rate. The new assumption makes it easier to assess the forecasts, as the conditions are normally more realistic. The fact that we have chosen to make this change can be partly regarded as a result of academic research, and also partly because this assumption appears to have worked well for other central banks that have chosen to base their forecasts on assumptions other than a constant key rate, such as Norges Bank and the Bank of England.
Basis for assessment
So much for our monetary policy strategy, which is what I believe we should be assessed against. Let me now say a few words about the basis we supply for assessment.
The analysis work on which the monetary policy decisions are based is published in the form of Inflation Reports. These contain analyses of the determinants of inflation and forecasts for inflation developments over the coming years. They also contain analyses of why deviations from target have arisen or are expected to arise, given a particular monetary policy stance. The Inflation Reports thus contribute to spreading knowledge about the Riksbank’s assessments and thereby make it easier for external parties to follow and understand monetary policy. Moreover, we want them to encourage discussion of monetary policy issues.
So far we have published four Inflation Reports a year, but we have now decided to reduce the number to three. This is because the two Inflation Reports in the autumn have come too close together to have sufficient new material to justify publishing two separate reports. The reason for this is that the first autumn report is published in connection with the Riksbank Governor’s hearing in the Riksdag Committee on Finance, which is usually held in October, after the Riksdag has opened for the autumn session.
Let me emphasise that this does not reduce our openness. Openness and transparency involve ensuring that the information provided is clear and relevant, so that those monitoring us can understand what we do and why we do it. Publishing Inflation Reports that contain little new information does not seem very important in this context. Now we can instead obtain a more even spread of the reports across the year and thereby use our forecasting resources more efficiently.
In addition to the Inflation Report, a press release is published after each monetary policy meeting, with an account of the motives behind the decision made. Press conferences are organised when an Inflation Report is published or when an adjustment has been made to the repo rate. The Executive Board’s monetary policy discussions are reported in separate minutes that are published approximately two weeks after the meeting where the decision on the interest rate was made. The minutes also show whether a member has expressed a reservation against the forecasts or the decision, and if so, on what grounds they base this reservation. In this way, all those who wish to closely follow our reasoning can do so and it creates a basis for assessing our evaluations and decisions.
I can also mention that the Riksbank Governor appears before the Riksdag Committee on Finance twice a year to answer questions on monetary policy. On the first occasion every year we publish an Inflation Report that contains a special appendix with material for assessing monetary policy over the past 2-3 years. This contains, for instance, comparisons of our forecasts with other forecasts and with outcomes. In addition, we Executive Board members hold around forty published speeches every year, where we can describe our joint or our individual view of economic developments. We also publish information in our quarterly journal and other publications issued by the Riksbank.
The Riksbank thus publishes a substantial amount of information that enables the general public to examine our analyses and to understand how monetary policy decisions are made. At the same time, it creates the conditions for a serious debate on monetary policy. This is a major change compared with central bank history.
Monetary policy is assessed regularly in Sweden. In 1999, the Riksdag Committee on Finance made the assessment that the Riksbank’s increased independence meant that there was a greater need to examine and assess monetary policy. Since then, the Committee has annually examined the monetary policy conducted over the past three years in connection with a more in-depth review of our activities. The main basis for these examinations is the appendix to the Inflation Report that I mentioned earlier. One might think that it would be even better if more, and independent, agents took an active part in serious assessment.
From time to time, assessments or assessment reports are commissioned. For instance, the National Institute of Economic Research(1) was given the task of producing background and facts prior to the Committee on Finance’s assessment in 2002. Last year the Committee on Finance decided to commission an independent, in-depth assessment of Swedish monetary policy during the period 1995-2005, that is, the ten-year period during which the inflation target has acted as official anchor for Swedish monetary policy. This assessment is being made by two international experts in this field, and it will, for instance, illustrate whether the inflation target is correctly designed to ensure price stability, to what degree the monetary policy conducted has led to attaining the inflation target during this period and the background and forms for the monetary policy decisions. This is something that I and my colleagues in the Executive Board welcome and I look forward to the analyses and debates this assessment will lead to.
Of course, one can regard the general debate as an examination of our activities and as an indicator of how well we explain our monetary policy. However, it is rare that particularly thorough comparisons of, for instance, the forecasts published by the Riksbank and others are made in the debate. The debate thus cannot replace systematic assessments.
The development towards increasingly open central banks appears to be the same in other countries that work in similar ways, particularly in Norway where the Centre for Monetary Economics has organised an annual, independent evaluation of monetary policy for the past few years. As I said at the beginning, this openness and the possibilities for evaluation are necessary conditions for an independent central bank to enjoy confidence and legitimacy.
To summarise, I consider it extremely important that independent central banks should be assessed. To ensure that this is carried out in a purposeful manner, the central banks’ forecasts and other information on which decisions are based should be well-documented and reported openly. It also requires a reasonably coherent explanation as to why the forecasts look the way they do. This means that we need to constantly develop our methods for making and assessing forecasts. However, it also requires that an institution actively commits itself to making thorough analyses of the information compiled so that the assessments and debate can be serious and constructive.
What happens if low inflation persists?
Not only the Riksbank, but also several other central banks have in recent years faced a situation where the economy has grown strongly, while inflation has been below target, despite very expansionary monetary policy. Let me therefore go on to discuss an important issue that has arisen due to developments in recent years, namely how to deal with a situation where structural changes contribute to inflation remaining low over a long period of time.
In Sweden, the upswing in world trade has contributed to strong growth in Swedish exports. Over the past year, domestic demand has also developed strongly, partly due to stimulation from expansionary monetary policy. At the same time, inflation has been kept down by high productivity growth and weak growth in import prices. This is partly related to developments in economic activity. However, there also appear to be factors of a more structural nature behind the low inflationary pressure. This is probably to some extent connected to the fairly comprehensive investment in new technology. However, it is clearly also largely concerned with the effects of increased global competition. Our analyses thus indicate that the low inflation rate has in many cases been the result of a number of changes in the supply side of the economy.
Our forecast is that growth in the Swedish economy will remain high over the coming years. We also envisage inflation rising as capacity utilisation increases. Our most recent inflation forecast indicates that inflation will approach the target a couple of years ahead, on condition that the repo rate is gradually increased.
However, forecasts are of course uncertain. In addition to the usual uncertainty regarding developments in economic activity, it is very difficult to determine the durability of this type of structural change. Is, for instance, the upswing in productivity we have seen in recent years only temporary or is it a more permanent improvement? There is a discussion that inflation could remain low over a long period of time, as a result of structural changes in a globalised economy(2).
The question is what the consequences would be if we were to respond to such a development by conducting a very expansionary monetary policy over a long period of time to bring inflation back on target. What would happen to economic activity, the exchange rate, domestic inflation, asset prices and household and company balance sheets? Probably, such a situation would lead to risks in both the real economy and the financial economy, resulting perhaps in overheating in the economy, high domestic inflation and rapidly rising asset prices. However, if we on the other hand were to allow the inflation adjustment to take a long time, how would this affect confidence in the inflation target? How long can inflation deviate from the target without this affecting confidence and without the inflation target being perceived as meaningless?
These are difficult considerations that I assume many of you have also reflected on. In Sweden our forecast, which we reported in the February Inflation Report, is that increasing capacity utilisation over the coming years will lead to inflation gradually rising. However, continued low inflation despite high growth is a development that requires consideration. As I see it, it is possible to take such a scenario into account, with low inflation over a long period of time, within the framework of flexible application of the inflation target. However, confidence in the inflation target is of central importance here, which makes considerable demands on us, to be clear that the policy we conduct is in line with our strategy to attain the inflation target in the best possible way. In a situation where, for example, there is reason to expect that inflation will be below target for a long period of time and we do not intend to bring it back on target until a couple of years from now, we should explain our intentions clearly. And this should then be taken into account when we are assessed. Once again, it is largely a question of openness and clarity.
Conclusion
Let me say in conclusion that the assessment of monetary policy is an important part of an independent central bank’s endeavour to be open and clear and of the endeavour to attain and maintain confidence in the inflation target. In Sweden, we publish a substantial amount of information that enables the general public to examine our analyses and to understand how monetary policy decisions are made.
The developments in the economy in recent years have given rise to many important issues, not only for us decision-makers, but also for those who assess us. We are several central banks who have faced structural, unforeseen factors contributing to keeping down inflation, while monetary policy is expansionary, the economy is showing robust growth, and household borrowing and asset prices are rising rapidly.
An inflation rate lower than the target level need not be interpreted in this situation as due to excessively tight monetary policy and as a failure for monetary policy. It is interesting to note that this has not been the conclusion in the most recent assessment of Norges Bank.
Developments in recent years have raised issues which are extremely important in the discussion and which require improved analysis. If we were to aim at keeping inflation on target at any cost, without taking into account developments in the real economy, what would the consequences be of trying to stimulate demand further so that inflation comes closer to the target? On the other hand, what effect would it have on confidence in the inflation target if we allowed a long time for the adjustment? How long can inflation deviate from the target without this affecting confidence and without the inflation target being perceived as meaningless?
These are difficult questions that I know we at the Riksbank are not the only ones to be reflecting on. I would very much like to hear your thoughts and reasoning on this. Safeguarding confidence in the inflation target is of central importance for monetary policy. I therefore believe that the best means of managing these concerns is to be open and clear with regard to the analyses produced and the decisions made on the basis of these analyses. I believe that this will in the long run create the best conditions for the policy to be accepted and to win confidence.
Thank you!
Notes:
1) The National Institute of Economic Research (NIER) is a public authority under the Ministry of Finance. The NIER produces analyses and forecasts on the Swedish and international economies and conducts research in these fields. It has an independent position and is responsible for the assessments it makes.
2) See, for instance, Borio, C. and A. Filardo, (2006), “Globalisation and inflation: New cross-country evidence on the global determinants of domestic inflation”, BIS, Work in progress.
