First a word of thanks for the opportunity to attend this meeting. The autumn conferences are a fine tradition to which it is a pleasure to contribute.
On previous occasions I have used the opportunity to look back over the past year and discuss what the Riksbank has been doing, including how monetary policy is conducted. Then I have tried to look ahead and pick out some issues of importance for monetary policy in the longer run.
My approach this year will be broadly the same. First a review of what has happened since last autumn, with comments on how the Riksbank's assessment of inflation has evolved over the year. Then, as previously, a look at the question of how we communicate policy and how it has been perceived in financial markets.
On this occasion we are in the throes of an international financial crisis, with great uncertainty as to what its real economic effects may be. Contagious effects of the crisis have gradually become larger and larger. Looking back, one can clearly see how the economic forecasts of most observers have swung during the year, from a phase of concern about a future increase in inflationary pressure to a phase of strong anxiety that the global economy will be hit by a recession.
Against this background I shall be discussing the current monetary policy appraisal of the Swedish economy, together with some factors of importance for the final assessment. First, however, I shall briefly outline monetary policy's line of thought.
Forecasts and probabilities
Monetary policy's overriding objective is to safeguard the value of money. The annual rate of inflation, measured as the change in the consumer price index, is to be kept at 2 per cent. As it takes time for the Riksbank's monetary measures to work in full, our actions have to be based primarily on an assessment of inflation's future path. The policy horizon—the period in which most of the impact of monetary policy materialises—is around twelve to twenty-four months.
The Riksbank's assessment of future inflation is presented continuously in our Inflation Reports. Since the first Report was presented in October 1993, the content has been developed and the Riksbank has, for example, become increasingly explicit in its assessments. In the first two years the Report contained a more general account of how the inflation indicators and inflationary pressure in the economy were developing. From November 1995 onwards the Riksbank has been attaching numbers to its inflation assessments.
In the June 1996 Report the inflation assessment was underpinned for the first time with an overall picture of the cyclical position. A further step was taken in September 1997, when the inflation assessment was supplemented with a path for inflation in the forecast period. In the next Report our main path for inflation was flanked by uncertainty intervals. The probabilities of inflation following various alternative paths are weighed into the overall picture on which monetary policy decisions are based.
How, then, are the uncertainty intervals calculated? In practice, our forecasters construct assessments of a range of variables, taking into account both their current uncertainty and how this compares with their normal uncertainty. This information is then incorporated in a fairly simple macro model that is used to estimate future inflation. The probability distributions we obtain in this way are, of course, highly subjective but they do give a consistent picture of how the Riksbank evaluates different risks.
While we are on the subject of the inflation assessments, it is particularly important to bear in mind that they are predicated on the repo rate being unchanged throughout the forecast period. A forecast path for inflation that deviates from the target—or the large deviations that the uncertainty intervals represent as conceivable paths—constitutes a diagnosis that may indicate treatment. When called for, repo rate adjustments are used to bring inflation into line with the target. In reality, therefore, the uncertainty about future inflation is not as great as the uncertainty intervals might suggest. Another consequence is that our inflation forecasts become out of date as soon as the Riksbank alters the repo rate.
Assessments in the past year
Much has happened in the past year as regards economic developments and the conditions for monetary policy. The Riksbank, along with other official observers and market forecasters, has revised its assessments in the light of the increasingly widespread effects of the international crisis.
– Last autumn most observers counted on a strong economic trend, both in Sweden and in the rest of the world. The crisis that had erupted in Southeast Asia was expected to lead at most to a welcome easing of international price pressure. In the Swedish economy, some indicators last autumn pointed to rising inflationary pressure. In particular, a gradual increase in capacity utilisation was judged to lead in time to inflation above the Riksbank's target. Inflation expectations, moreover, were tending to move up, having previously stabilised after a falling trend for a number of years. In the December Report it was judged that inflation would exceed the target in 1999. The Riksbank raised the repo rate 0.25 percentage points.
At the same time, the Riksbank underscored that, due to uncertainties in the assessment, the outcome might differ from the main scenario. On the one hand, an exacerbation of the crisis in Asia could result in weaker inflationary pressure; on the other, there were risks that high wage increases and weak productivity growth would fuel the rate of inflation. These two alternative scenarios were judged to be equally probable.
– During the spring the picture of inflation changed by degrees. In the March Report, inflation prospects had altered somewhat in connection with the mounting crisis in Asia and low import prices. Subdued inflation expectations and restrained wage demands pointed in the same direction. In view of a weak development of international activity, it was foreseen that full capacity utilisation in the Swedish economy would not be reached until the end of the forecast period, when inflation was assumed to be marginally above the target. This called for a readiness to undertake some cautious tightening of the monetary stance later in the year.
At the same time, the spectrum of risks had shifted, with less risk of inflation exceeding its expected path but a continued risk of development being weaker than in the main scenario. In view of the uncertainty, the Riksbank chose to await future developments.
– At the beginning of the summer we observed that inflation was following a lower path than we had assumed in the main scenario in March. Weaker import prices were contributing to this but so were other factors. Transitory effects—lower interest rates and changes in indirect taxes and subsidies—were particularly prominent. Moreover, on account of lower wage increases and stronger productivity growth than expected earlier, it seemed that inflation in the longer run would be somewhat lower than the Riksbank had predicted. With the outlook for productivity, it was likely that some surplus capacity would continue to exist during the whole period.
All in all, inflation prospects one to two years ahead were accordingly more subdued than before and predicted inflation was below the target throughout the forecast period. In addition, there was judged to be a large risk of inflation being below the path in the main scenario. Under these circumstances the repo rate was lowered 0.25 percentage points.
Notwithstanding these revisions, the economic outlook for the near future has not changed much in the past year, with expectations of good growth during 1998, to an increasing extent on account of rising domestic demand. These predictions have been fulfilled. During the year, however, the uncertainty about growth prospects further ahead has risen, above all in the late summer and autumn.
Transitory effects and lower import prices
Another matter to note is that although the cyclical picture has proved fairly accurate, the path of inflation during 1998 has been markedly lower, mainly on account of large changes in transitory effects and import prices.
At the beginning of the year it was foreseen that the transitory effects would largely cancel out, with no impact at all on the course of inflation during 1998. With the spring decisions to cut tobacco taxes and property tax, however, it was assumed that the transitory effects would act as a strong curb on inflation. In the June Report it was foreseen that the transitory effects would have a downward impact on inflation during 1998 of as much as 1.2 percentage points; in the September Report this figure was adjusted upwards to 1.3 percentage points. For 1999 and 2000 the transitory effects on inflation were assumed to be marginal.
Import prices have also been more subdued than expected. In the December assessment last year the Riksbank assumed that during the forecast period the annual contribution to inflation from import prices would amount to about half a percentage point. Some downward revision was made in March, followed by a further adjustment in June, when it was envisaged that the impact of import prices on the CPI would not be sizeable. The lower outcome and the adjusted forecasts mainly reflect a price fall for raw materials, though the international price trend for manufactured products has also been weaker than expected, presumably in part as a consequence of the Asian crisis.
Rents are another CPI item that has been more subdued than foreseen in the spring. Since the beginning of the year, the rate of rent increases has dropped to levels that are historically low. This is partly attributable to lower mortgage interest costs and lower expectations for wages and inflation but it also reflects the cut in property tax and decreased heating costs.
Wages seem to have developed more or less as the Riksbank foresaw. Today we count on an annual wage rise of 3.5–4 per cent in the period 1998–2000; that is only a little below our estimate last December before the wage negotiations had been completed. Opinions differ as to the effect which the Riksbank's actions may have had on the latest wage round. In the light of contacts with labour market representatives, my view is that monetary policy's influence on the wage settlements was positive. But this was by no means the only factor. Contributions to the low outcome seem to have been made by this wage round having been conducted both centrally, with efforts to hold wage formation together, if not throughout the labour market then at least in large segments, and decentralised to individual firms. Whether this model can be repeated is another matter.
Transitory effects, import prices and rents account for 2 percentage points of the substantial downward revision of inflation in 1998 by a total of 2.2 percentage points between the Inflation Reports in December 1997 and this September. The same factors are responsible for about half of the downward adjustments for 1999 and 2000 between the forecasts in March and June.
Monetary policy increasingly understood, but problems last spring
For some years now the Riksbank has been working hard to communicate the line of thought behind its decisions both to the general public and to market players. This is intended to heighten the general understanding of how monetary policy functions and what it can do and also prevent, as far as possible, market disturbances on account of uncertainty. The more clearly the market understands our way of thinking, the smoother should short interest rates adjust to new information about inflation prospects.
A comparison of market expectations (as expressed in implied forward interest rates) with the actual repo rate can serve as a indicator of the market's error in foreseeing the repo rate. This predictive error has, on the whole, decreased. With a reservation for other effects on this indicator from various interest rate premia that may fluctuate over time, this suggests a growing understanding of our policy.
With a longer time horizon, the predictive error naturally tends to be larger. With a one-year horizon the error presumably has less to do with imperfect communication or a lack of understanding of how our policy functions; it is no doubt more a matter of shifts in the whole picture of economic activity and inflation. Things are somewhat different in a three-month perspective; predictive errors then presumably also arise from misinterpretations of the Riksbank's signals.
During the past year the repo rate has been raised once and lowered once. This has required a great deal of the Riksbank in the way of communicating policy to the outside world and explaining the reasons for its measures.
A study of the market's prices, predictions and statements shows that the repo rate increase in December 1997 was broadly expected. It was, in fact, just at last year's Conference here that an important piece of signalling took place. As regards the repo rate cut in June, on the other hand, the pattern of market expectations changed relatively late. The implied forward interest rate curves three months earlier show that the June cut had not yet been foreseen by market players. It was mainly in connection with an interview I gave some weeks before the decision that expectations shifted to a cut.
The Riksbank's management of information about the repo rate cut came in for market criticism. We were said to have been indistinct and to have changed our inflation assessment too abruptly. Some players used big words, claiming that the lack of clarity had damaged confidence in our policy and would make it difficult to conduct an effective policy in future. I think that was overdoing it. But the Riksbank should still have been able to make the market more prepared for what the future could have in store.
It should also be underscored, however, that the tendencies during the spring were difficult to assess. We certainly changed our opinion but not so abruptly and for a long time it was highly uncertain what the next step would turn out to be. In such a situation, with a high degree of uncertainty, communication is difficult.
One reason for not reacting to the faulty pattern of market expectations was that the forecasts of market players were in line with our own. Thus it was not that we differed in the assessment of inflation prospects but that, even though the path of inflation was below the target, market prices did not represent a view that the Riksbank would lower the repo rate. To put it bluntly, market players seemed to share the Trade Union Confederation's view that the Riksbank acted with a “bias”.
A clear lesson from events last spring is, in any event, the value for market observers of constructing independent forecasts of inflation as a foundation for their repo rate predictions. For the Riksbank, what happened has prompted us to work even more on continuously monitoring market opinions so as to be able to react in good time if the assessments of market players seem to be deviating from our own.
Main scenario and balanced risks
Now for a look ahead in the context of our current assessment of inflation.
In the latest Inflation Report, published at the end of September, the Riksbank emphasised that inflation prospects contained a large element of uncertainty. The cyclical picture had changed somewhat since June in that growth was judged to be strongest this year and then slacken somewhat. But we still counted on reasonably good activity, with annual growth in the coming years averaging almost 3 per cent. Inflation, measured as the 12-month change in the CPI, was assumed to be below 2 per cent in 1998 and 1998 but to reach approximately 2 per cent at the end of the period.
The composition of demand had also undergone a shift. The international crisis was expected to render net exports weaker than assumed earlier, while domestic demand was likely to be somewhat stronger than foreseen in June. The assessments were based on an assumed appreciation of the Swedish krona from the current TCW index level of 124 to 120 in the coming year and just over 117 in the year after that.
Inflationary impulses from the rest of the world were expected to be low in the forecast period.
The rising domestic demand was not expected to result in an alarming increase in inflationary pressure. Investment growth and labour supply were judged to be sufficient to avoid general problems with capacity during the period. The development of wages and productivity was expected to be favourable for a moderate, though rising, rate of inflation.
Inflation expectations among market agents, firms and households were still low. Inflation was expected to be below the target in the short run and then be in line with the target further ahead.
With the fundamentally positive picture of the Swedish economy, including good growth in the forecast period and a successive return to the targeted rate of inflation, the Riksbank left the repo rate unchanged. As uncertainty had increased since the earlier assessments, however, the uncertainty interval surrounding the main path was considerably wider. But the risk spectrum was symmetrical, that is, the risks of inflation being higher than in the main scenario were judged to be as great as the risks of lower inflation. International economic activity may result in inflation being even lower than in our main scenario, at the same time as a lastingly weak exchange rate that has no counterpart in poorer economic fundamentals could lead to upward pressure on prices.
Increased financial unrest
Financial unrest has grown since the September Report was published. It is difficult to foresee to what extent the crisis that began in Thailand almost a year and a half ago will affect the real industrialised economies in the West.
The direct real economic effects via foreign trade will probably be relatively small.
For economic activity and inflation in Sweden, the effects that could arise from decreased general confidence are more important. Share prices in Sweden have gone on falling since the September Report was published, wiping out not just the gains during 1998 but also a part of last year's increase. Since the high in July, the Stockholm Exchange has dropped nearly 30 per cent, giving a fall of about 8 per cent from the level at the beginning of the year. Since the Inflation Report was published, share prices have fallen nearly 6 per cent.
Continued unrest and falling asset values would affect the behaviour of firms and consumers. If households and firms postpone investment and consumption decisions, economic activity will be lower. Some signs of this can be discerned; the latest figures suggest a minor blow to households' expectations of Sweden's economy, though this has not yet affected their personal economic expectations. There has been some fall in the growth of business credits.
Another channel for effects on the real economy is the banking system. A more cautious supply of bank credits could dampen economic activity. In this respect the international turbulence gives cause for concern. The tendency in the late summer for investors to move to assets that were regarded as generally secure—a flight to quality—has developed into a general aversion to holding risk, with investors fleeing instead to assets that are more liquid. The economic fundamentals in particular countries are currently playing little part in investment decisions. This has hit some Third World countries very hard.
The financial system in Sweden is monitored very closely by the Riksbank. The available statistics suggest that, compared with many international banks, Swedish banks are in a better position, with lower exposures to countries hit by the crisis. The picture also seems to be relatively good as regards exposures to hedge funds. This low exposure is probably a consequence of the bank crisis in Sweden in the early 1990s. So the risk of a domestically generated credit crunch should be small.
Weak currency
Besides increased uncertainty and greater risks of weaker economic activity, developments since the September Report was written have involved downward pressure on the Swedish krona. Since September 28th the krona has weakened almost 2 per cent in terms of the TCW-index, while since the depreciation began last May the exchange rate has fallen about 9 per cent.
Just as on previous occasions, financial unrest has hit the Swedish krona as investors play safe and move to larger and more stable currencies. This time, however, the unrest seems to have less to do with uncertainty about the future for Sweden's economic fundamentals. Instead it appears to involve general risk aversion among investors. Support for this interpretation is to be found in market letters (Swedish as well as foreign). That should point to an appreciation in the next phase, when the general unrest has subsided and investors concentrate once again on economic fundamentals. Judging from Prospera's surveys for the Riksbank, for instance, it seems that market players are also allowing for this in their assessments.
The recent market weakening of the Swedish krona is a reminder of both the advantages and the drawbacks of a flexible exchange rate. The weaker exchange rate can serve as a mechanism for adjustment when international activity is slowing down. This could be valuable if the crisis is profound. At the same time, the recent events illustrate that our vulnerability with a flexible exchange rate is considerable. The krona has fluctuated even though economic policy is basically focused on stability. Other countries with a flexible exchange rate and an inflation target have had similar experiences, for example Australia, Canada and New Zealand. It not only complicates assessments of the future path but also probably entails costs in the form of a poorer real economic trend.
Conclusion
Monetary policy in Sweden has been conducted with an inflation target and a flexible exchange rate since January 1993. The results have been good in many respects. Since the upturn in the summer of 1993, growth has been good compared with, for example, the years of high activity in the late 1980s. In contrast with that period, moreover, inflation has been low. Another point worth noting is the good real wage trend. With the low price rise, the annual real wage increases have averaged 2,5 per cent since 1994, which is more than in the 1980s, for example, when nominal wages rose rapidly but what households actually gained in real terms was very modest.
Monetary policy also seems to be better understood. Presumably this is mainly because, in spite of everything, economic development in recent years has been relatively good. The change in attitudes is evident in surveys of how people perceive the Riksbank and monetary policy. A similar picture is obtained from the treatment of the Bank and monetary policy in the Swedish press. Our basic line of thought has been broadly accepted, though from time to time there has naturally (and sometimes perhaps even rightly) been criticism of what we do. Another point worth noting is that a large majority of the new Riksdag (Sweden's parliament) backs an agreement that includes safeguarding the value of money as a statutory objective, as well as a more independent status for the Riksbank. The earlier discussion suggests that an understanding of the Riksbank's actions has also grown among market players, though that does not rule out further improvements.
In recent weeks, instrumental rates have been cut in other countries. This is, of course, a reflection of the increased concern about a deeper crisis that leads to lower inflation. There is reason, however, to be rather careful about citing interest rate cuts by other central banks as grounds for similar action by the Riksbank, though this is frequently done in the debate. When making comparisons with the United States, for instance, it should be borne in mind that the initial level of interest rates there is 1,15 percentage points higher. Most observers consider, moreover, that the United States is in a different cyclical phase. When making comparisons with Continental Europe, it is particularly important to consider exchange rates. The present weaker exchange rate represents an appreciable economic stimulus. If the TCW index were to remain at the present level in the coming twenty-four months, it would represent an expansionary effect on the economy that is equivalent to an interest rate cut of something like one or two percentage points.
The weeks that have passed since the Riksbank presented the Inflation Report have been dramatic in financial markets. Our main scenario for the years ahead is still relatively good activity in the Swedish economy, with inflationary pressure that gradually rises from a low level. But uncertainty about the future path has been further accentuated in connection with the financial turbulence. There is an increased risk of a weak international development. In addition, the continued stock-market unrest has made domestic demand more difficult to assess than before. Meanwhile, the weakening of the krona is the opposite of the tendency on which our main scenario in the Report was based.
At the Riksbank we are keeping a close watch on developments. That is a self-evident component of our work for promoting financial stability in the economy. It is also a necessary part of the work on monetary policy. At present there are no grounds for altering the direction of interest rate policy.
