Heikensten: Some topical monetary policy issues
Thank you for inviting me to talk here about monetary policy and the economic situation.
The economic seminars arranged by the Centre for Business and Policy Studies are an established tradition that is well worth maintaining. This is my sixth appearance here as Deputy Governor of the Riksbank and for me, these have been good occasions both for looking back and reflecting on all that has happened in the field of monetary policy during the past year and for considering what the future may have in store. My talk today is arranged along similar lines.
From the viewpoint of economic policy and in keeping with the latter part of the 1990s, last year's economic developments gave relatively little cause for concern. GDP growth had provided a series of pleasant surprises at the same time as inflation had been lower than expected. This had been accompanied by a rapid improvement in the public finances.
Signs of poorer tendencies were, however, beginning to appear. Stock markets were already starting to fall in the early part of 2000 and in the late autumn a number of indicators pointed to an economic slowdown, first in the United States and then, early this year, in other parts of the world. To begin with this was perceived as just a temporary dip - a blip in the stocks cycle, a minor adjustment in the IT sector or something else. But stock markets seemed to be clearly over-valued and it would have been odd, particularly after such a long upward phase, if this had not led to a certain amount of over-investment. There were also clear signs of insufficient saving, particularly in the United States. Consequently there were grounds for believing that once a slowdown had started, it could prove to be both relatively deep and long. Against this background, in the March Inflation Report we talked not so much about a V-shaped path as a U-shaped one or a "banana". Since then the outlook has become darker but more of that later.
For monetary policy, turns in the business cycle are not inherently a major problem, apart from the difficulty of foreseeing them. Falling resource utilisation, for example, should normally led to lower inflation in the future. That in turn can pave the way for interest rate cuts. In the past year, however, things have not been quite as simple as that. Two partly interrelated factors have contributed to this. One is inflation, which has been higher than expected and also above our target. The other is the exchange rate, which has been weaker than anticipated and also weaker than the real economy warrants. During a period in the late spring, moreover, the krona fell rather uncontrollably in the absence of similar movements in other currencies. All this has made our work harder than in the preceding years.
Price tendencies
I should like to comment first on the high inflation figures.
Fluctuations in inflation are commonplace
Recently people have been asking whether the higher inflation outcomes constitute a failure. Some years ago, when inflation month after month, not to say year after year, turned out to be lower than we had predicted and below our 2 per cent target, people asked exactly the same thing. I did not see it on the whole as a failure then and neither do I do so now. The lower inflation at that time was mainly due to downward effects of a transitory nature even though the economy was expanding rapidly and capacity utilisation was rising. Similarly, the higher rates at present primarily stem from transitory factors. Earlier we noted that some upward shift in inflation could be achieved in the short run but only by means of drastic interest rate cuts. The corresponding remedy today would be drastic hikes. Such an approach would make interest rate policy erratic and result in unnecessary fluctuations in economic activity.
What we can do - and do - is to aim consistently at our 2 per cent target. There is then a good chance of succeeding in the longer run, particularly if economic agents take us at our word and rely on future inflation being in line with earlier assessments. Judging from the various indicators of inflation expectations, on the whole we seem to have been successful (Diagram 1). Although inflation has actually fluctuated markedly, for a number of years money market players, for example, have counted on future inflation being approximately 2 per cent. Neither have they been greatly mistaken. Since we began to target inflation, the rate measured by the CPI has averaged 1.3 per cent, while UND1X inflation has averaged 1.8 per cent.
If we have anything to reproach ourselves for, perhaps it is that, despite our efforts to the contrary, we seem to have induced many segments of society to believe that we can fulfil the target, if not always then almost so. That is evidently not the case. Inflation fluctuates, sometimes a lot. The same seems to apply in most other relatively small countries. Compare inflation in, for instance, the individual euro countries with the average rate in the European Union. In Sweden's case the fluctuations are possibly accentuated by the price-setting being comparatively centralised in some important respects, for instance the labour market and rents.
Still, there are good reasons for taking a close look at the latest inflation outcomes. The prise rise has come mainly from food and energy but also from telecom prices (Diagram 2). In the case of food and electricity, the price increases, particularly during the spring, largely stemmed from factors on the supply side: foot-and-mouth disease, mad cow disease, a relatively dry winter, production problems at oil refineries and so on. It is, in fact, these CPI items that normally fluctuate most. It is not just a coincidence that in Canada, for example, the inflation index that primarily guides monetary policy does not include either food or energy prices. The rather large price fluctuations that are a normal feature of these goods and services can be taken to indicate that the recent increases do not constitute a ketchup effect, so that the conditions for a return to a rate in line with the target should be relatively good.
Not just transitory effects
At the same time I definitely do not want to make light of the increase in inflation. Even excluding the goods and services I just mentioned, inflation has moved up more than expected and it has done so even though resource utilisation is lower, at least at present, that we counted on earlier. That is the main reason why, in the latest Inflation Report, we have slightly revised our appraisal of how resource utilisation is likely to affect inflation.
A conceivable interpretation of the revision is that today we consider that resource utilisation last year was higher than we reckoned on earlier. That is not self-evident, however; this year's wage outcomes have been lower, which one would not expect if resources have been more strained. Another possibility is that, even if resource utilisation was not in fact higher, price increases have occurred due to insufficient competition.
In the past, the Riksbank has presented a picture of goods and services divided into those that are priced in a market as opposed to administratively. This showed that in 1996-97, for instance, goods and services priced administratively kept inflation up even though resource utilisation was falling (Diagram 3). Perhaps we are now seeing a similar phenomenon. When turnover in a competitive market falls, prices normally decline, too, but in some respects the opposite now seems to apply. Local authorities, for instance, are trying to make up for a loss of tax revenue by raising prices for water, public transport and the like. From the media it looks as though there is a similar tendency in civil aviation authorities and some air transport, for example.
The Riksbank has no direct cause to regard administratively determined price increases differently from market increases: both contribute to inflation. But when the former accompany declining resource utilisation, it obviously complicates matters for monetary policy. In addition, these observations naturally raise other questions that have to be dealt with, not by the Riksbank but by parliament and the local authorities and have to do with competitive conditions in the Swedish economy.
The exchange rate
Recently the exchange rate has also been the subject of a lively discussion. That is only to be expected when the krona has fluctuated markedly, besides being clearly under-valued on the whole in recent years.
Differences in inflation and growth could previously explain most things
In a longer perspective there are explanations for a large part of the krona's depreciation. In the past three decades the nominal exchange rate has weakened about 90 per cent. More than half of this deterioration can be attributed to the higher inflation in Sweden compared with the rest of the world in the 1970s and '80s. Mostly of the remaining part of the real deprecation can be explained in terms of relative GDP growth and worsening terms of trade in the same period. But neither of these two factors is relevant to the period since 1992 with a flexible exchange rate: inflation, growth and terms of trade have been at least on a par with the rest of the world.
Various flows may have played a part in the short run. Looking back, it is evident that export growth and rapidly rising Swedish share prices may have contributed to the high exchange rate with the euro twelve to eighteen months ago. During the past year the opposite has applied. Another factor has been large investments in foreign shares when earlier barriers to risk diversion were removed from the pension system. In September there was also general financial unrest. However, variables such as these have proved to be of little value for forecasting.
So personally I do not think it has always been hard to understand the direction, upwards or downwards, of exchange rate movements in recent years. It is the tendency for the krona's average value to be unduly low that has been more difficult to interpret. To some extent, this has presumably been a consequence of the American dollar being generally over-valued against European currencies. I also believe that some part may have been played by various structural factors, for example the construction of the tax system, the internationalisation of the corporate sector with the attendant relocation of certain functions, the status of the Swedish krona as the importance of the euro grows in Europe and so on.
Some of these notions are supported both by arguments from principle and by somewhat anecdotal information. There is reason to believe, for instance, that the taxation of both capital gains and wealth has had effects. Moreover, at least a part of the residual item in the statistics could have to do with a tendency for firms not to repatriate export receipts to the same extent as before, which in turn may be a consequence of a larger proportion of their central functions now being located abroad. But we do not know how quantitatively important these factors are and I believe that perhaps one should be wary of exaggerating them. Still, these are questions that deserve further scrutiny and we intend to pursue them. To date we have published a report on EMU's conceivable importance for the Swedish krona; the conclusion was: not much, provided - and it is a major proviso - Sweden provides equal conditions in other respects for wealth formation and business enterprise.
The krona's part in the assessment of inflation has varied
One of the topics in the discussion during the past six months has been how the Riksbank reacts to the exchange rate in the formation of monetary policy. We have been accused of being indistinct or dithery during the summer and autumn. First let me say that there is a pattern to this that I find familiar. Conducting and interpreting monetary policy is more difficult when the situation is appreciably affected by other factors than resource utilisation, for instance the exchange rate or inflation expectations. The assessments are then more complex and the Riksbank faces greater educational challenges. The decisions also tend to be more controversial because there is a clearer conflict between different goals, so that more criticism is only to be expected.
Our forecasts are always based on a variety of factors. A traditional model for determining inflation in the context of a Phillips curve, for example, points to resource utilisation, inflation expectations and the exchange rate as central components. However, these variable's relative importance for inflation changes over time.
To illustrate our reasoning over the past nine months, here is a presentation of resource utilisation, our view of international prices and the exchange rate (Diagram 4).
Last December, when it still looked as though resources would continue to become increasingly strained, we were concerned about the exchange rate and its effects on prices. We counted on GDP growth rates last year and this of 4 and 3.5 per cent, respectively. Since then our appraisal of resource utilisation has changed a good deal. We now know that last year's growth stopped at 3.6 per cent and for this year we are counting on 1.3 per cent. This means that as a percentage of GDP, resource utilisation at the end of 2001 will be between 2 and 3 points lower, which is a marked downward revision. Another aspect is that the slowdown is connected with a weak tendency in the rest of the world. Consequently our assessment of international export prices has also been revised downwards and this, too, counters the inflationary impact of the weaker exchange rate. So although we continue to point to the weak krona as an inflation risk, this risk now seems to be smaller, mainly in view of the downward revisions of capacity utilisation.
In June we had some additional factors to consider. The krona suddenly weakened dramatically in the absence of a general increase in financial unrest. This suggested that the causes were specific for Sweden. As inflation had moved up appreciably more than expected in the previous months, this was particularly troublesome. Although a large part of the increased inflation was presumably transitory, the Riksbank perceived a risk of higher inflation expectations. Rather than take that risk, we decided to raise the interest rate and intervene in the currency market.
The forecast
In conclusion I want to make some brief remarks about the future, using our latest Inflation Report as the chief starting point. The main scenario there is based on the prospect of what might be called a normal cyclical development, driven above all by an international recovery, good disposable incomes and a relatively weak exchange rate (Diagram 5).
Unchanged appraisal of the krona's fundamental value
Our assessment of the future exchange rate is somewhat less optimistic than before. This has to do not so much with new information as with the weak tendency in the past year. In the case of a variable that is so notoriously difficult to predict, the assumption about the future is always partly educational. A major aim on this occasion was to avoid an assessment of inflation that could be perceived as unrealistic. But the krona fluctuates sooner than the time it takes to revise our forecast, as is clear from what has happened.
In the discussion that followed the publication of the Inflation Report it has been suggested that the Riksbank has lost faith in the krona. I think that's overdoing it. And regardless of the forecast we presented, I see no strong reasons for altering the appraisal of what is a reasonable fundamental exchange rate in the long run, any more than there were grounds for revising the basic principles for share valuation just because stock markets had shot up to record levels a couple of years ago. That does not mean that some time in the future there could be a case for a different opinion about the exchange rate's reasonable level in relation to the real economy. But that would stem in turn from an altered view of Sweden's relative production potential, terms of trade or some other fundamental variable. In this context it should be borne in mind that an unduly weak exchange rate generates effects that should favour an appreciation further ahead: the balance of trade continues to be strong, corporate profits become higher than elsewhere, making Swedish shares more attractive, and so on.
Calculations of the krona's fundamental value will no doubt become topical at the latest if Sweden is to adopt the euro and we require a picture of the exchange rate that will be serviceable in the longer run. Experience has shown that once such rates have been communicated, they guide the market to a high degree up to entry.
Lower resource utilisation but large risks in both directions
In the Report we count on inflation falling back again next spring. This year's appreciably lower resource utilisation has a lot to do with that but so does the prospect of last spring's rapid price increases coming to an end so that the earlier effects drop out. In the latter part of the forecast period inflationary pressure in the main scenario becomes somewhat stronger as the economy recovers fairly rapidly.
Support for this picture is to be found if one goes behind the price statistics and looks at wages and unit labour costs (Diagram 6). Even with the rather substantial downward revision of productivity growth, these costs rise at a moderate rate throughout the forecast period. That should ensure that even if inflation does not fall back quite as much as calculated, its level will be low. The overall assessment is that in September next year CPI and UND1X inflation will 1.5 and 1.7 per cent, respectively, while the rates in September 2003 will be 2.1 and 2.0 per cent.
The main downside risks are connected, for example, with the possibility that the adjustment in the United States after the long upward phase could take longer than expected. After September 11, moreover, terrorist attacks and military action can clearly cause developments to take very different paths, with markedly divergent outcomes for the global economy. Another aspect is the highly expansionary direction that economic policy has now taken in many countries, not least Sweden. If confidence in the future is restored, the table is really laid for a quick upturn. But all in all, there is considered to be a strong probability of inflation being lower as a result of the slowdown abroad being deeper and more protracted. Such a development would be likely to lengthen the slowdown in Sweden as well.
There are also risks of inflation being higher than in the main scenario. One has to do with the fact that inflation has risen more than expected. Our main scenario makes some allowance for the higher outcomes but the revision is rather small. It was just with such small, cautious revisions that we coped with the surprisingly good inflation outcomes earlier. Without more knowledge it is difficult to see how one can act very differently. In addition to the marginal revision in the main scenario, the risk spectrum now contains a further upside component. Another upside risk for inflation lies in the exchange rate, even though a persistently weak currency is presumably the most likely alternative in conjunction with slacker economic activity and thereby lower resource utilisation.
There are always grounds for highlighting the spectrum of risks as an important element in our assessment. I want to end by emphasising this particularly at present. In many respects the situation is unusually difficult to assess. It is easy to see that developments could quickly take a different turn from the main scenario. That in turn could led to interest rates becoming either higher or lower than at present.
Diagram
