Nyberg: The Riksbank and the inflation target
Thank you for the invitation to come and speak at Nordea about Swedish monetary policy! Today I shall mainly concentrate on the content and interpretation of our inflation target. One special aspect of the inflation target and monetary policy that I shall discuss towards the end of my speech is the household sector’s growing debt burden and house price increases.
A clear framework for monetary policy
Sweden is currently enjoying what is usually described as an economic boom. Growth last year was almost 3.5 per cent, one of the highest figures in 30 years, and households’ real disposable incomes are expected to rise by around 3 per cent this year. Inflation is at its lowest point in living memory, as are interest rates. However, there is a problem. Employment has not risen at all as most forecasters, including the Riksbank, expected. We therefore find ourselves in a discussion of whether interest rates should possibly be even lower. Given the intensive debate on this subject, I consider there is good reason to describe and discuss our monetary policy framework and its significance for the way we work.
It is well known that one of the Riksbank’s fundamental objectives since the beginning of 1993 has been to maintain price stability. This is inscribed in the Sveriges Riksbank Act of 1999 and is thus a parliamentary decision. However, it was left to the Riksbank to determine the practical application of this objective. Our choice was to specify a quantitative target for the inflation rate: The annual increase in the consumer price index, CPI, should be two per cent with a tolerated deviation of +/- one percentage point. The purpose of a tolerated deviation is to make it clear both that inflation will deviate from the two per cent level from time to time and that the Riksbank's aim is to limit such deviations. The target is symmetrical, which means that it is as undesirable to have an inflation rate below the target level as above it.
Many other countries also conduct monetary policy where they have formulated a price stability target in one way or another. The euro countries have chosen to maintain inflation at a level close to, but below, two per cent, in other words, a non-symmetrical target. The United States has taken a different view and chosen a vaguer wording for its target, with no figures expressed, as well as taking greater account of general economic trends. In New Zealand they have a target for inflation expressed in terms of an interval (1-3 per cent) but no exact figure. However, we have seen two clear advantages of having a clearly expressed target to aim for; firstly it provides a good foundation for inflation expectations and secondly it makes it easier to assess the policy conducted.
Since 1997 we have applied a simple policy rule in order to make our monetary policy even easier to understand: If the forecast shows that inflation will be above target level two years ahead, we usually raise the repo rate, and we lower it if the forecast is below target level. But, as I intend to argue here, the policy rule cannot be applied mechanically. It should instead be regarded as a rule that on average provides a good picture of the policy, but where departures from the rule can and should be made under certain circumstances.
The first argument is connected with the time aspect. As there is a time lag before monetary policy has an effect – it takes around 1-2 years before interest rate adjustments achieve full impact on inflation – the Riksbank’s policy must look ahead. Our interest rate decisions are therefore based on forecasts of, for instance, the prospects for inflation and economic activity over the coming two years. It goes without saying that these forecasts, like other assessments of future events, are uncertain. Forecasting errors are often due to genuinely unexpected things, which we have not been able to assess in advance.
At the beginning of 1999 the Riksbank made a decision to clarify the principles for monetary policy, as a complement to the simple policy rule. This clarification states that monetary policy is normally aimed at bringing inflation onto the targeted path one to two years ahead. However, it also identifies two circumstances under which the Riksbank can depart from this aim; namely transient effects on inflation and large deviations from target. The main idea behind this is that we do not focus solely on inflation in every situation; we should also be able to regard developments in growth and employment when these do not conflict with the inflation target.
Transient effects refers to certain determined factors which affect the inflation forecast one to two years ahead, but which will then subside without having any lasting effect on inflation trends. This was the case, for instance, in 2002-2003, when large fluctuations in oil prices had a significant effect on inflation. We made it clear then that monetary policy should disregard these transient effects.
Large deviations from target refers to when inflation deviates so far from the target level that there may be justification for bringing it back to the target level over a longer period of time than the usual two-year horizon. A more rapid return to the target could cause undesirable fluctuations in, for instance, production and employment and thereby lead to large costs for the real economy.
Since the decision on the clarification was taken in 1999, it has become increasingly evident, in my opinion, that there are a number of good reasons for allowing deviations from the inflation target during short or long periods of time. It is important that monetary policy, if necessary, can be allowed some scope and flexibility when the economy is subjected to various forms of shock and disturbance. If this is to happen, it is essential that there is a high degree of confidence in monetary policy. I believe that we have achieved this, largely thanks to a large degree of openness and transparency with regard to the bases for our monetary policy. Inflation expectations two years ahead and further have been well in line with our target since the regime stabilised in 1997-1998. The way we have clearly defined and explained our considerations has enabled us to create room for manoeuvre, which we have also utilised over the years. In other words, we have had fairly substantial freedom to deviate from target without affecting confidence in our policy.
Another factor significant for the interest rate decisions is the length of the forecast horizon. As I mentioned earlier, the Riksbank works on the basis of a time perspective of two years, as this is how long it takes before interest rate adjustments have full impact on inflation. However, since December 2000 we have often found reason to supplement this horizon with a qualitative assessment covering a slightly longer period. The March 2005 Inflation Report took a further step forward, presenting more detailed forecasts beyond the two-year horizon and using the assumption that the repo rate would develop in line with implied forward rates, with the aim of increasing understanding for the Riksbank’s reasoning.
A longer time horizon can contribute to clarifying monetary policy considerations and thereby enable us to better explain the way we work. If, for example, we see that inflation is slightly above two per cent at the end of the forecast horizon, our policy rule says that we should raise the repo rate to hold back demand and thereby inflation. However, if demand nevertheless falls and inflation beyond the forecast horizon is below the two per cent level, the conclusion is not so obvious. Then it may be time to discuss lowering the interest rate. Extending the forecast horizon may in some situations explain why increased flexibility is needed in monetary policy.
These are some reflections on the time aspect and the monetary policy operational framework. However, there are also other important factors indicating that the inflation target and the policy rule must be interpreted with a measure of flexibility. It is easy to have a blind faith in monetary policy’s scope for economic fine tuning, but we should be sceptical of this for at least two reasons.
One is that our forecasts and other reports on which the interest rate decisions are based contain a large measure of uncertainty, because of gaps in statistics, information lags regarding economic trends, incomplete knowledge of how companies and households will react and structural changes, such as increased globalisation. Reactions to monetary policy may also vary considerably over time, partly, but not completely, due to economic activity. Our knowledge is quite simply too brittle to enable us to lead the economy exactly by fine tuning monetary policy at perfectly chosen moments.
The second reason why fine tuning of monetary policy is doomed to failure is that the repo rate is only one of the factors affecting the economy. The most important economic decisions in our society, such as house purchases and corporate investment, are mainly due to developments in long-term interest rates, and monetary policy decision-makers can only have an indirect and very limited effect on these.
This reasoning therefore comes to the conclusion that we should not have overly high expectations of what the Riksbank can achieve to influence real economy developments, as has been the focus of recent debate. There is in actual fact every reason to be humble with regard to what we can actually achieve. But if we can keep inflation within our target area of 1-3 per cent and clearly anchor inflation expectations around two per cent, this will have great significance for growth and employment in the long term.
Monetary policy is not an exact science. Sometimes monetary policy is described as a car, where gentle acceleration or easing on the gas adjusts the speed of the journey down a winding road. This is a very misleading image. It is not possible to see the road ahead clearly and nor is it possible to know exactly what will happen if one puts one’s foot down on the accelerator pedal – only that the effect will come after a fairly long period of time.
Why haven’t we cut the interest rate?
Let me now return to the present situation and the question of whether it is reasonable to lower the interest rate in order to rapidly bring the present low inflation rate back in line with the target. The Executive Board’s reasoning at its most recent monetary policy meetings has been roughly as follows.
Firstly, as I mentioned at the beginning, the Swedish economy has developed at an increasingly strong rate in recent years, and our analyses during the winter and spring have come to an assessment that this will continue. We have thus seen a rise in inflationary pressure at the end of the forecast horizon. If our assessment during this period had been that demand was declining, I believe the situation would have been different.
Secondly, the present low inflation rate can be largely explained by what is known as positive supply shocks, such as increased competition, lower import prices and surprisingly rapid productivity growth. These factors benefit economic developments. To attempt to push up prices for domestic goods and services to counteract a decline in prices for imported goods is not necessarily a wise choice.
Thirdly, monetary policy’s impact comes through influencing demand, primarily corporate investment and household consumption. Investment has begun to rise, albeit at a slightly later stage than we had assumed. However, it is unlikely that the problem of a low investment rate now is connected to interest rates being too high. Given the low interest rates we see now and the good access to credit in the financial system, financing costs can hardly be regarded as a factor holding back investment. With regard to households, it is possible that even lower interest rates could make them increase their debt burden and borrow in order to increase consumption, but to me this is not an obvious conclusion. In addition, there are risks here. This brings me onto the question of how households have reacted to the low inflation and low interest rates, and this issue deserves more detailed discussion.
Households’ growing debt burden and rising house prices
In recent years, households’ debt burden has increased gradually and now amounts on average to 125 per cent of their annual disposable income, which is almost as high as immediately prior to the property crisis in 1992. However, interest payments are now less than half what they were then, approximately 4 per cent of disposable income. The fact that households are loaning at an increasing rate, both in Sweden and many other countries, can be regarded partly as an adjustment to lower inflation, which is also expected to be permanent. Falling interest rates, perhaps combined with greater predictability as to how interest rates and inflation will develop in future, explain a large part of the increased willingness to borrow. One can thus say that households’ growing debt burden is to some extent a natural consequence of the successful inflation-targeting policy pursued over the past decade. As I see it, it is reasonable to assume that we are heading towards a new equilibrium, where households have adapted their borrowing to what they perceive as a lower long-term level for interest rates, although it is not yet possible to say when we will reach this stage. There is still a possibility that households are underestimating the uncertainty of future trends and therefore borrowing more than would correspond to a long-term equilibrium position.
In Sweden, as in many other countries, households have primarily borrowed in order to buy accommodation, which has led to house prices rising at roughly the same rate as the borrowing.
The question now is whether there is reason to take this development into account in monetary policy decisions. A strictly-applied inflation-targeting model would not take this into account, and there is no mention of it in the monetary policy clarification published in 1999. The general problem is whether low inflation can make various asset prices, such as equity prices or house prices soar way beyond what is reasonable and if so, what monetary policy can do to remedy this. Should the repo rate be used to slow down this development to try to prevent prices becoming so high that the fall, when it actually comes, will be too fast and too far? There are many examples of how the air has gone out of asset markets and thereby triggered or aggravated a downswing in economic activity. Or is it more reasonable to wait until the “bubble” has burst and then try to counteract the effects the best one can?
This question is neither a simple one nor a new one. It is discussed by central banks around the world, whether or not they have specific inflation targets. And there is hardly a clear-cut answer.
As I see it, the seriousness we in Sweden should attach to households' increased debt burden is closely connected with how we assess price trends in the housing market. Although there has so far been a large rise in prices, particularly in metropolitan areas, we have not seen the clearly speculative tendencies apparent in other markets, for instance Australia and the United Kingdom. In Sweden we still usually buy housing to live in it ourselves. In some other countries, where the tax legislation and housing loan markets are different from in Sweden, people buy one or more additional properties in the hope that they will rise in value, rather like buying equity or other financial assets. This of course puts upward pressure on prices, but also downward pressure when prices rebound and many people want to sell.
In both Australia and the United Kingdom, prices have levelled out in recent months after short-term interest rates rising in these countries as their central banks increased their key rates. However, in Sweden both household borrowing and house prices appear to be continuing upwards, partly as a result of expectations of continued low interest rates. Moreover, the competition between banks has increased, which has led to more aggressive marketing, lower collateral requirements and longer amortisation periods. It is possible to begin to detect a scent of the reckless 1980s, one cannot deny it. There is therefore every reason to increase our vigilance, and we have recently emphasised the developments in credit and house prices as a further argument, although in my opinion perhaps not a particularly strong one, against more expansionary monetary policy. It is nevertheless important to point out that so far we have seen no signs of unwarrantedly large price rises and overheating in the housing market; in Sweden prices are still essentially governed by the normal supply and demand relationship.
However, taking into account developments in asset prices is a further argument in favour of interpreting the inflation target with some degree of flexibility.
Households and financial stability
Let me now for a moment leave the inflation target and concentrate on some reflections regarding the Riksbank’s other objective, to promote a safe and efficient payments system. Here the concern is whether household borrowing entails any risks for the banking system and thereby for financial stability. When it comes to rapidly rising property prices and growing bank borrowing, it is easy to relate it to the developments at the beginning of the 1990s and the banking crisis they triggered. However, developments were different then in certain essential aspects.
Firstly, the majority of the banks’ losses were not caused by housing loans – this constituted a very small part – but by lending against collateral in commercial properties. Commercial property prices rose more than housing prices prior to the crisis. This has not been the case in recent years; quite the reverse. The commercial property market has been subdued since the economic downturn after 2000, with low prices and high vacancy ratios and only began to improve over the past year. The banks’ risks in this market appear very moderate at present.
Secondly, although households are borrowing substantially, they have a good ability to pay on average. Moreover, Swedish households, unlike US households, have substantial savings, both real and financial savings. As shown in the Riksbank’s Financial Stability Report, which was published earlier today, the risk of households as a collective suffering payment problems if interest rates rise or housing prices fall, is slight. And even if a large number of households were to experience difficulties in meeting interest and mortgage payments on their loans, it would not threaten the banks’ capital adequacy or the stability of the Swedish financial system.
Although households will manage well on average, I believe there is still reason to question whether individual households are not borrowing more than is reasonable in the long term and thereby increasing their sensitivity to unforeseen events. At the Riksbank we have expressed concern, in various contexts, that households have excessive confidence that interest rates will remain low for a long period of time. The average period for fixing interest rates has fallen to 12 months, despite the fact that long-term interest rates have been at lower rates than ever. It is easy to be seduced by a low monthly cost and forget that even minor increases in interest rates will lead to a proportionately large increase in loan costs. If, for instance, the interest rate rises from 3 to 4 per cent, it means that the monthly cost will increase by one third.
There is reason for the Riksbank to closely monitor developments in household borrowing. However, this is a question of concern over developments in individual households and over the effects on house prices rather than a question of stability within the banks and the financial system.
What will be the future direction for monetary policy, given the current situation?
The inflation target, the simple policy rule and the clarification of the principles behind monetary policy together provide a clear intellectual framework, which has guided our actions in recent years. In my opinion, this intellectual framework provides a sufficient degree of flexibility for monetary policy. The policy rule shall not be applied mechanically under all circumstances. This is particularly important given that the economic reality appears to be taking on increasing complexity and becoming more difficult to define.
The Inflation Report published in mid-March contained the assessment that economic growth in Sweden and abroad would remain high over the coming years. Our forecast for Sweden was a growth rate of around 3 per cent a year, which is higher than we consider to be sustainable in the long term. The new information received since then has been mixed. The oil price has been unexpectedly high, although recently it has fallen slightly. Moreover, industrial activity appears to have weakened slightly, both in Sweden and abroad. We therefore came to the conclusion in April that growth in Sweden would be somewhat weaker than was anticipated in the March Inflation Report. Since then, we have seen statistics showing that growth in Germany was slightly stronger than expected during Q1, and that the downturn noted in the United States in March appears to have been temporary. We had anticipated lower growth in Sweden’s net exports, but exports of goods appear to be increasing even less than was forecast. However, good growth in real incomes and low interest rates will contribute to favourable growth in household consumption. The labour market remains relatively weak, and our forecast is based on employment strengthening during 2005.
At the same time, we can see rising price pressures in some industrial nations, particularly the United States, which are partly – but not solely – a consequence of the high oil prices gaining greater impact in the production and consumer segments.
It is important to monitor whether the slackening tendencies seen in economic activity recently are temporary or more persistent. If we were to see a notably poorer development than we envisage now, such as much lower growth than forecast, the arguments in favour of cutting interest rates would in my opinion take on greater substance. We will be considering all of these factors at our next monetary policy meeting on 20 June.
Thank you!
