First a word of thanks for the invitation to visit Umeå and talk about monetary policy.
The Swedish economy is developing favourably. Growth has been impressive in recent years, averaging around 3 per cent in annual terms, and at present the rate may be even higher. This has contributed to a rapid reduction of unemployment and strongly rising employment.
A look at the statistics shows that the present expansion of employment in Sweden is the fastest since the labour force surveys were first published in the early 1960s. And although growth has been strong, inflation has remained moderate.
The explanations for the good circle in which our economy is now moving are not simple. But I do believe that one contributory factor is the realignment of economic policy that has been achieved in the past ten to fifteen years. This has involved deregulating the credit and foreign exchange markets, modifying components of the transfers system and paving the way for stronger competition. All this has helped to improve the workings of the economy. Another factor, perhaps the most important, has been economic policy’s commitment to stability, with price stability as a statutory objective and a successful consolidation of the public finances.
The last time I was in Umeå and talked about monetary policy almost three years ago, I presented a list of eight questions. They covered what I perceived at the time to be the most topical monetary policy issues: What is good about low inflation? What is the reasoning behind the Riksbank’s objective? Why is central bank independence a good thing? And so on. I am glad that not all of those questions are still as topical today. Progress has been made in the debate. But there are reasons for returning to some of them and new questions have arisen; the latter are more specific and have to do with how we operate and in what framework.
Let me first underscore that monetary policy regimes are not hewn in stone. It is important that both monetary policy and the intellectual framework used by the Riksbank are scrutinised thoroughly and continuously. The conditions for monetary policy change as we gain experience and acquire new knowledge. A constructive dialogue, inside the Riksbank as well as elsewhere, reduces the risks of making unnecessary mistakes in monetary policy.
1. Why 2 per cent inflation?
The Riksbank Act states that monetary policy’s objective is to maintain price stability. Partly in order to make the policy intelligible and open for assessment, the Riksbank has decided that inflation — measured as the annual change in the consumer price index (CPI) — is to be held at 2 per cent, with a tolerance interval of ±1 percentage point.
Motivating low inflation is not difficult. A great deal has been written on the subject. But perhaps it will suffice to recall the Swedish experience in the 1970s and ’80s. The high and widely fluctuating rate of inflation had clear negative effects on our economy. Productivity gains and investment growth in that period were low. Sharply rising prices led to an arbitrary redistribution of income and wealth. In order to rectify the corporate sector’s competitive position, the value of our currency had to be written down on a number of occasions.
Bit by bit, however, more and more people, regardless of which political party they supported, realised that this economic policy could not be sustained. So in the early 1990s, Sweden finally joined the growing number of countries that had chosen to re-focus their policy on low inflation.
But while it is not difficult to advocate low, stable inflation, it is less easy to argue that just 2 per cent is the appropriate rate. When the inflation target was adopted in 1992, it was noted that this was approximately the level of ambition that important countries had chosen. Moreover, inflation was currently around that level when the decision was made.
Alan Greenspan has argued that inflation should be so low that people generally do not give it a thought when making economic decisions. That points to 2 per cent as a reasonable level.
There are some objections, however. Some argue that an inflation target ought to be higher. One reason has to do with the downward rigidity of nominal wages. We know from experience that wages rise, sometimes so much that they generate inflation. Downward wage adjustments are less frequent. Seen from this angle, inflation might be regarded as a smooth way of achieving necessary adjustments to real business costs.
Another risk, according to some observers, is that an unduly low inflation target leaves too small a margin for monetary policy. The economy is liable to land in a situation where the central bank is no longer in a position to influence the development of prices by pushing the real interest rate down. It simply is not possible to reduce the interest rate below zero per cent.
As I see it, the Riksbank’s decision to target a specific annual rate of inflation and implement its policy symmetrically has eliminated a large part of the risk of our being caught in a deflationary spiral and the dilemma of zero interest rates. Everyone knows we strive to bring inflation back to the targeted level of 2 per cent. Inflation expectations in recent years have in fact been stable around the 2 per cent target.
A specific target also means that the labour market parties know the rules of the game, which I believe has promoted reforms in wage formation. To a large extent, wage demands are now drawn up on the premise of inflation being 2 per cent.
All in all, I see no grounds today for reconsidering our 2 per cent target. The objections in the literature to a target around 2 per cent seem to be less relevant if anything than they were some years ago. The labour market is on the way to adjusting to low inflation. And the risks of deflation should diminish when monetary policy targets inflation symmetrically.
2. What is the appropriate measure of inflation?
The Riksbank has chosen the consumer price index (CPI) as monetary policy’s target variable. But all indicators of inflation, including the CPI, have drawbacks. These have to be weighed against the advantages. The CPI has the advantage of being an established indicator of inflation among economic agents and familiar to people in general. It is also a comparatively broad price index. Being familiar and intelligible made the CPI particularly suitable for the process of establishing the new inflation targeting policy and stabilising inflation expectations.
Partly in view of the krona’s depreciation and the risk of inflationary impulses from import prices, the Riksbank decided in 1993 that the target would not apply until 1995. This avoided a situation in which meeting the new target might have required a sharp monetary tightening that could have led to an unnecessarily large drop in output and higher unemployment. That shows that right from the start the Riksbank realised there could be reasons for temporary deviations from the target.
A drawback with the CPI is that besides containing information about underlying or trend inflation, it catches transitory price movements that are sometimes difficult or almost impossible to influence. During the years with the inflation target, various transitory factors have affected the CPI. One example is the interest rate fluctuations, which have shown up in the own homes item in the CPI. During 1996–97 in particular, the CPI was relatively far below not just the targeted rate but also the lower tolerance limit. To a fairly large extent, this was because the Riksbank lowered the interest rate markedly with a view to ensuring that the target would be met a little later on.
In the period 1993–98 the Riksbank tried to explain how effects of this kind would be handled. But not everyone considered the policy was sufficiently clear. The Riksdag (Sweden’s parliament), for example, assessed the Riksbank’s performance solely in terms of the CPI outcome even though there had been long periods when monetary policy had focused on other indicators that exclude transitory effects.
In the best of all possible worlds one would prefer to use an indicator of inflation that excludes all those effects one wants to disregard. Unfortunately, there is no such indicator that suits every situation. For example, a tax increase may, under certain circumstances, need to be excluded as a transitory effect. In other cases, however, tax increases should be taken into account when formulating policy because they are a part of the inflation process.
Against this background, in 1999 the Riksbank declared that in connection with every monetary policy decision, any grounds that existed for modifying the policy’s target variable would be clarified. This forces the Riksbank systematically to consider whether there are any transitory factors that ought to be disregarded when deciding the repo rate. It also facilitates a correct evaluation of monetary policy.
One risk with this arrangement is perhaps that the discussion may sometimes become rather technical and focused on details. Should or should not the index we use include a given tax change that alters inflation by one tenth of a percentage point? Continuous minor revisions of our target variable are also liable to make us less intelligible, which may ultimately render monetary policy more difficult to scrutinise.
So we evidently face a difficult choice. Since the clarification, we have chosen so far to start from UND1X, which unlike the CPI excludes house mortgage interest expenditure and price effects of indirect taxes and subsidies. In addition, we present a more detailed picture of how we view certain price movements that could be regarded as transitory, for instance the recent oil price rise.
There are those who argue that the Riksbank ought to disregard downward price effects from structural factors. Examples are effects on the price level from various deregulations and the increasingly strong international competition. Reasons for doing so are difficult to find. What matters is not whether a price movement is ‘structural’ but whether it affects the path of inflation more fundamentally, that is, as a part of the inflation process. If it does, it ought to be taken into account when policy is formulated, regardless of whether its effect is downwards or upwards. Matters would be different if the effects in question were temporary, so that they only affect the price level for a year or so. In that case there might be grounds for disregarding them. But deregulations and increased international competition are presumably as much a part of the inflation process today as were the opposite effects that increased employer contributions and other taxes exerted in the 1970s and ’80s.
3. Why have a rule of action?
The Riksbank formulates monetary policy primarily on the basis of a forecast of inflation one to two years ahead. If inflation, measured as the annual change in the CPI, is expected to be above 2 per cent during this time horizon, the interest rate is normally increased and vice versa. This rule of action is intended to:
- make policy more intelligible
- act as a support for internal decision-making
- enhance external credibility
The rule of action has been an important support for the Riksbank in recent years. Besides helping to make monetary policy more intelligible, it has contributed to a tighter analysis and a more structured internal discussion. But a simple rule such as this clearly cannot be applied mechanically.
A noteworthy aspect is the time horizon to which the rule refers. The Riksbank is required to maximise goal fulfilment, that is, to keep inflation in line with the target, for ever. As I shall be discussing, moreover, it is reasonable that the Riksbank considers developments in the real economy and tries, when making decisions, to avoid unnecessary disturbances and fluctuations in growth and employment.
One reason why the Riksbank did not opt for a shorter target horizon is the time lag that normally occurs before interest rate adjustments affect the economy. I say normally because, for example, the exchange rate sometimes speeds up the effect from the interest rate. Moreover, a large interest rate adjustment is presumably associated with a quicker effect, for instance if household wealth is affected via share prices. Developments beyond the current two-year period are excluded by the target horizon because, for example, it is often difficult to predict what will happen so far ahead. But certain trends may be fairly clear at times and monetary policy should then take them into account. So although policy is guided by the prospects one to two years ahead, there are normally reasons for also paying some consideration to inflation’s forecast path in the coming year as well as beyond the target horizon.
To date the Inflation Report has contained detailed assessments of inflation two years ahead. In the next Report, in December, we intend to include a section with a more general description of inflation prospects two to three years ahead. This is intended to promote a better discussion of the choices that lie ahead.
Another important matter is how strictly the rule should be applied. There are many sides to this. One has to do with whether monetary policy ought to allow for other matters besides inflation, for instance the real economic situation or the risk of problems that might threaten financial stability. I shall be returning to that shortly. Another is the timing of an interest rate adjustment.
It is, of course, primarily the inflation forecast that guides monetary policy. But there may be room at the margin for other aspects. Normally it is a matter of whether a measure can be taken a little sooner or later than otherwise. This may be desirable to ensure that the measure is well received in the financial markets, for example. On other occasions we have prepared the general public for future decisions in order to avoid unnecessary conflicts and discussions in society. When we raised the interest rate in 1997, for instance, the decision was timed with a view to sending a clear message in advance of the wage negotiations at that time.
4. Should we take output and employment into account?
Price stability is the Riksbank’s policy objective. The antecedents to the Riksbank Act state that, without prejudicing this objective, the Riksbank should also endeavour to promote other economic policy goals that the Riksdag has established, for example a favourable development of growth and employment.
One reason for focussing on low inflation is that most things suggest this is a prerequisite for a favourable long-term development of growth and employment. That means it is also the best contribution monetary policy can make to combating unemployment.
A short-run increase in employment can be obtained by stimulating demand, even if this entails a risk of higher inflation. But this short-term exchange can be risky, as we know not least from the experiences in the 1970s and ’80s I referred to earlier. And just the fear of an unduly expansionary policy that may lead to rising inflation is liable to push interest rates up, leading to weaker growth and employment. For monetary policy it is therefore crucial that confidence in the low-inflation regime is strong.
The construction of the inflation target automatically involves some allowance for economic activity. Weaker growth that leads to a capacity surplus normally implies diminishing inflationary pressure from demand. This in turn leaves room for a more expansionary monetary stance: the interest rate is lowered, thereby stimulating demand. Conversely, unduly high demand involves a risk of excessively high capacity utilisation, with upward pressure on prices and wages. To prevent this, the monetary stance has to be tightened by raising the interest rate. In this way, the Riksbank’s endeavour to fulfil the inflation target normally contributes to economic stabilisation.
There are, however, situations in which this relationship does not apply. One case is when inflation is being driven from the supply side, for example because an isolated oil price rise is pushing prices up. There may then be grounds for accepting a temporary increase in the price level instead of acting swiftly to push inflation down. Unnecessary fluctuations in the real economy can then be avoided. As I mentioned earlier, the Riksbank has therefore clarified that if higher or lower inflation is judged to be a transitory phenomenon, attempts to counter it immediately and fully are sometimes not desirable.
From these clarifications it will be clear that in the conduct of monetary policy we do in fact also take the real economic consequences directly into account. It should be underscored, however, that making allowances of this kind does presuppose a basic confidence in monetary policy — that different economic agents count on our fulfilling the inflation target, if not immediately then at least in the near future.
There has been a discussion recently, among economists as well as in the public debate, as to whether the interest rate should be raised already even though a threat to the inflation target is not apparent in the coming two years. In that way we could avoid a build-up of imbalances that in the longer run might necessitate a drastic repo rate increase and thereby expose the real economy to major strains.
I have no objections to this view in principle. We can indeed take such matters into account in our intellectual framework. It is important, however, to be clear about which imbalances one is referring to and what real economic effects they might have. It should also be borne in mind that the rule of action with a two-year horizon should normally give us enough time in which to act.
Being forced to confront one’s observations and assessments with a rule of action has in my opinion helped to focus the Riksbank’s thinking in these matters. When the time comes to make a decision, it is of course a question of balancing the different risks.
5. Should the risks of financial instability be taken into account?
The Riksbank is responsible by law for Sweden’s payment system, which means we have to oversee financial stability and take it into account in our policy. But as I touched on earlier, financial imbalances also affect the general economy indirectly.
An extensive academic literature in this field has grown up in recent years. The financial sector’s role in the national economy is becoming increasingly important. And as financial markets are now global, the horizon has to be widened. This was evident from the Asian crisis.
It is reasonable to assume that stable prices reduce the risk of financial imbalances. With lower inflation, economic agents are less likely to misinterpret prices and make imprudent investments. The wave of property speculation in the 1980s was occasioned in part by the misleading signals which the high-inflation economy transmitted.
A favourable economic trend that is accompanied by a strong expansion of credit which people use in the hope of getting a quick return, is liable to generate financial bubbles. Over the years there has been clear evidence of this in connection with excessively high expectations in the stock market or the property market, for example. In the end, it becomes difficult to justify prices in terms of economic fundamentals. Elements of this have been apparent in recent years, for instance in the United States.
One issue is whether central banks ought to consider asset prices when setting the interest rate. The notion here is that asset prices may influence the path of inflation. An increase in household wealth on account of rising share prices, for example, is likely to affect consumption and thereby demand in general and the price trend. Effects from expectations are also involved. Rising share prices may mirror expectations of permanently higher corporate profits and lead to expectations of higher lifetime incomes. If confidence in the future becomes much stronger in this way, the effects on demand may be considerable. Share price increases may perhaps affect the corporate investment propensity in a similar way. The Riksbank takes this into account as far as possible in the inflation assessment. But much remains to be done, theoretically as well as empirically, when it comes to understanding the nature of these relationships. A lively discussion is in progress, at the Riksbank as well as among those who scrutinise us, about what a given development of asset prices should imply for monetary policy.
There is even less agreement as to whether central banks should conduct policy so that the development of asset prices is affected directly. This immediately raises the difficulty of determining when an asset market can be said to be over-valued. And if one does decide to intervene, how can one be sure of deflating the bubble in an orderly manner? Just mentioning these two problems often leads those concerned with central banking to conclude that intervention is not advisable.
I should like to broaden the argument. In addition to the objective of price stability, as the central bank of Sweden the Riksbank is responsible by law for the payment system and financial stability. Without prejudicing these aims, we should also strive to support economic policy’s other goals, such as high growth and employment. This means, for instance, that we should take pains to avoid unnecessary fluctuations in output and employment.
So if we see what we believe to be mounting financial imbalances and judge that in time they will jeopardise a stable economic development, I consider that in principle, at least, there is a case for acting with the repo rate even if we do not perceive a direct threat to the inflation target. The difficulty in practice lies in deciding when an intervention is called for. To some extent, this is made easier by the risks being normally inherent in the growth of credit in the economy rather than in the actual development of asset prices.
In conclusion I can mention that the Riksbank has been working on these matters a great deal in recent years. Twice a year, for example, we publish a report on financial stability in the Swedish economy. In these reports, any risks we identify are highlighted with a view to creating conditions so that they can be corrected. On the basis of our assessments, we are also in a position to influence the course of events by proposing rule changes and legislative amendments.
6. Why is clarity important?
The Riksbank’s work with an inflation target has been developed over the years. The decision to adopt an inflation target was made in 1993. Inflation indicators began to be published quite soon after that. Comprehensive inflation assessments appeared some years later, followed by inflation forecasts. The rule of action with a target horizon was then added, as well as the clarification of how monetary policy would be formulated in the event of different shocks.
Future developments will no doubt exact additional requirements. Almost every time a new policy decision is at hand, any new information — on oil prices, taxes and price tendencies, for instance — has to be related to the intellectual framework we have gradually built up. As I mentioned earlier, in the next Inflation Report we intend to include a section on inflation in the somewhat longer run. We are also thinking of publishing probability assessments not just of forecast inflation but also of forecast growth.
The approach we have chosen has a number of advantages. In that the inflation assessments are published, we are forced systematically to appraise all the relevant information about future price tendencies. This is done in internal processes, with our staff playing a decisive role initially, after which the Executive Board is free to discuss the material that has been prepared. There are other strong reasons why the Riksbank should be open. As I mentioned earlier, it means we have had to tighten the analysis. It also makes it easier to follow what we do and enforce various forms of accountability.
In the public debate it is sometimes argued that we are too bound by our established principles. I hope it is clear from what I have already said that that is not the case. We are free to extend the target horizon, take any risks of problems in the financial sector into account or take timely action to avoid a marked economic slowdown further ahead. But before doing so, we require good, clear grounds for our actions that really will stand up to scrutiny.
I am sometimes exhorted to be more pragmatic and less tied down by our intellectual framework. The U.S. Federal Reserve is often cited as a good example. That is a little surprising, mainly because in the United States the academic discussion advocates the opposite — that a clearer target for price stability would perhaps be preferable. The background here is that the policy is currently associated so strongly with a single person, Alan Greenspan, and there is concern about what will happen after him.
To me, proposals of this kind about a more pragmatic policy seem rather naive. Policy principles that are less clear-cut, combined with a less transparent basis for decisions, by no means guarantee a successful policy in the longer run. Just the reverse, I would say in the light of many years of decision-making in economic policy. One of the most important things to do in work of this type is try to build up distinct processes whereby a foundation for decisions based on sound economic analysis is produced systematically and then subjected to as broad a scrutiny as possible. One will then at least have provided conditions for consistent and wise decisions. That in turn is fundamental for all credibility in monetary policy, too.
I believe, moreover, that decisions should be made in ways that contribute to insight and afford good opportunities of evaluating them and exacting accountability. These are requirements I consider people in Sweden should be entitled to make of an institution which has been given such a strong and independent status as the Riksbank.
The current monetary policy situation
Let me conclude by briefly presenting my view of current economic developments and their implications for monetary policy.
For a couple of years now the Riksbank has had reason to count on rising economic activity. Future demand growth has been judged to exceed what the economy can cope with in the long term without generating higher inflation. To date, however, price trends — not least in the domestic economy — have been lower than we had expected, even though our forecasts were below those of most other observers.
One recent event is the notable fall in share prices. To date, however, this seems to be more of a desirable normalisation than a fall that endangers economic activity.
There are also more and more indications that the recent strong growth of production and employment is tending to slacken. To date, however, that has been broadly in line with our earlier assessment.
Finally, there is the rather mixed information about the ongoing wage negotiations. This year’s wage outcome is, if anything, somewhat lower than expected. At the same time, it looks as though several of the demands that have been presented are a little bit high in relation to the inflation target, particularly considering the risk of future wage drift in an increasingly strong labour market.
These are matters we will be deliberating in the Riksbank in the run up to the next Inflation Report and monetary policy meeting in December. For my part, there seems to be no reason at present for any appreciable change in the assessment I made in October, namely that a repo rate increase will probably be needed in the future in order to safeguard the continuation of a favourable and stable development, with low inflation and rising output.
