Bäckström: Inflation and the interest rate
Thank you for the invitation to talk at this conference about the outlook for inflation and the interest rate. It is now five years since the inflation target was instituted; so before considering the current situation for monetary policy, I should like to say something about the intervening period.
Five years with an inflation target
In January 1993 the Riksbank's Governing Board decided to adopt an explicit target for inflation and specified that as of 1995 and measured in terms of the consumer price index, the rate of inflation was to be limited to 2 per cent, with a tolerance interval of ±1 percentage point.
The commitment to an inflation target originated from the Riksbank but both the Government and the Riksdag (Sweden's parliament) expressly support monetary policy's focus on price stability as well as the inflation target. Last year five of the political parties in the Riksdag agreed that the Riksbank's objective, to safeguard the value of money, should have the force of law; a Government bill with provisions to this effect was presented to the Riksdag in November. In recent years, monetary policy's commitment to low inflation has thus been established increasingly firmly.
Inflation target and transparency
When the Riksbank originally decided to adopt an inflation target, time was short and did not permit extensive analyses. In my opinion, experience to date of the inflation target policy has been favourable.
One major advantage of having an established and distinct inflation target for monetary policy is that this can be combined with transparency. Market agents can form a clear picture of the grounds for the Riksbank's actions. It is, moreover, comparatively easy to understand how the policy functions, which should help in gaining public support for the target. The inflation target is an instrument for stabilising expectations about the conduct of monetary policy; a lack of transparency and distinctness in monetary policy makes it more difficult for inflation expectations to adjust to the target. This in turn can make it more difficult to achieve high, stable growth.
In recent years the Riksbank has accordingly been aiming for greater transparency. A highly qualified staff has been built up, intensive efforts have been devoted to developing models and analyses, routines have been constructed for monetary policy preparations and decisions, and a new interest rate management system for monetary policy has been launched.
The endeavour to promote transparency is evident in the Riksbank's participation in public hearings by parliamentary committees, in inflation reports, speeches, discussions and seminars of various kinds, and in the publication of specialist reports. This has led, I believe, to a better awareness—among market agents as well as a wider public—of how monetary policy works. One indication of this is that agents' expectations about monetary policy have become increasingly congruent with the Riksbank's monetary policy intentions. In addition, the five-party agreement on the new Riksbank Act is a manifestation of broad political support for the low-inflation policy. It is our ambition, moreover, to continue the process for increased transparency as far as is reasonable and feasible.
Criticism of monetary policy
Monetary policy has, of course, come in for some criticism. A common issue initially was whether striving for 2 per cent inflation was reasonable and feasible. That discussion seems to have dried up. Inflation is currently around 2 per cent and more and more people understand the advantages of this.
Another question has been why the Riksbank does not have the task of promoting other objectives, growth ,for example. Here, too, there is a growing awareness, though it does not include everyone. In the long run monetary policy is only capable of influencing inflation and there is absolutely no long-term trade-off between inflation and growth. If anything, high inflation tends to impair real economic development.
In the short run, however, the situation is somewhat different. In this perspective, inflation and growth are largely two sides of the same coin, at least as long as inflation expectations are stable around the target. Weaker growth implies a diminishing risk of inflationary impulses from demand, which normally warrants a less restrictive monetary stance. Conversely, demand that is excessively strong in relation to potential output implies a risk of rising inflation and a need to raise the repo rate.
The current discussion seems to be about whether the Riksbank acts symmetrically. In other words, is the Riksbank as concerned about under-shooting the inflation target as it is about preventing inflation from exceeding the target? My answer to that question is definitely "Yes". The Riksbank aims to keep the rate of inflation at 2 per cent, neither more nor less.
The issue of monetary policy asymmetry seems to be linked mainly to two matters. One is the question of monetary policy's central concern. Is this inflation's prevailing rate or the rate that assessments point to one to two years ahead? The answer is that the Riksbank is guided by the assessment of future inflation. If inflation is currently running at 2 per cent and the predicted rate (based on an unchanged instrumental rate) one to two years ahead is, say, 1 per cent, then the instrumental rate should be lowered. The reason for this is quite simply that monetary policy acts with a time lag of one to two years. Using the prevailing rate of inflation as a guide would involve adjusting the instrumental rate more frequently and markedly, thereby entailing undesirable economic instability. But how can we be sure that the assessments are always correct? Of course we cannot. But the Riksbank does not simply assess the future, adjust the instrumental rate and then sit back and wait two years for the outcome. Instead the assessments are continuously reconsidered in the light of new information and the monetary stance is adjusted accordingly. In this way we work systematically to fulfil the 2 per cent inflation target and aim to avoid unnecessarily large changes in the repo rate. Even so, the monetary stance may have to be altered more markedly from time to time if a situation arises that is considered to call for this.
The other matter that may underlie the discussion about monetary policy asymmetry is, I believe, the question of transitory effects on inflation. Such effects may come, for example, from changes to taxes and subsidies, rising or falling house mortgage interest costs in connection with monetary policy measures, or external shocks that affect the price level via sudden changes in the price of oil, for instance. Such changes in the price level show up as a temporary increase or decrease in the rate of inflation; with the methods that are used to measure inflation, the impact on the price level will last for twelve months. The Riksbank has declared that such transitory effects should be of secondary importance for monetary policy in so far as they do not affect inflation expectations and thereby generate a more lasting impact on inflation.
Two episodes with transitory effects
The issue of transitory effects first arose around the turn of 1994, when the effects had to do with altered indirect taxes and subsidies in connection with the consolidation of government finances. This entailed a transitory increase in inflation.
In a number of addresses that winter I considered the issue at some length. In a speech to the Economics Association in February 1995, for instance, I noted that there are shocks, such as altered indirect taxes and subsidies, which may affect the price level at short notice and cause inflation to deviate from the target. While such effects are usually absorbed inside the tolerance interval, they may sometimes be so pronounced that inflation fluctuates more widely. In such situations it is up to monetary policy to ensure that the deviations from the target do not have permanent consequences for the rate of inflation and inflation expectations. While drastic countermeasures might produce quick results, they would destabilise the economy, which is not desirable.
In various contexts I made it clear that monetary policy would be conducted so that price effects from the consolidation of government finances would be left to materialise in the price level. At the same time, it was important to ensure that the transitory effects were, in fact, temporary and did not colour inflation expectations.
At that time, however, survey data indicated that the level of inflation expectations was already around 4 per cent. The overall assessment of inflation, based on an unchanged instrumental rate, accordingly pointed to a risk of the inflation target being exceeded as economic activity became stronger. Moreover, the high inflation expectations contributed to excessively high wage settlements in the spring of 1995. It was partly this—not the price effects from budget consolidation—that warranted the upward adjustments of the instrumental rate from August 1994 to July 1995.
Transitory effects again conditioned the rate of inflation during 1996 and 1997 but now in a downward direction via falling house mortgage interest costs. The krona's appreciation, in connection with the rapid enhancement of confidence in budget policy, also tended to lower the CPI.
Had the Riksbank continued to reduce the instrumental rate—after the earlier cuts had had a transitory downward effect on inflation's current rate—it would have tended to destabilise the economy. The Riksbank would, as it were, have been chasing its own tail, with interest rate cuts that became increasingly large and frequent. Inflation expectations and actual inflation would ultimately have risen, obliging the Riksbank to swing to a markedly restrictive stance. As interest rate increases also have transitory effects on inflation, the Riksbank would again start to chase its own tail but now in an upward spiral. This illustrates how absurd it would be to raise or lower the instrumental rate as a response to the direct, but transitory, effects of the Riksbank's own monetary measures. As long as inflation expectations are unaffected, there is no reason to suppose that the effects on inflation will be permanent.
Transitory effects account for inflation target deviations
An explicit inflation target has the advantage of enabling us to evaluate how monetary policy is conducted. Consequently the Riksbank cannot implement an asymmetric approach to the inflation target without this being detected.
Two curves are presented in the diagram below. One represents CPI inflation and the other is one of the various indicators of underlying inflation, in this case the CPI excluding transitory effects from house mortgage interest costs, changes in indirect taxes and subsidies, and prices of imported goods.
Since 1995, when the inflation target first applied, the annual increase in the CPI has n have been somewhat stronger than the upward effects. In the absence of any transitory averaged 1.5 per cent. This outcome is thus relatively close to the targeted rate and well inside the tolerance interval. Underlying inflation, as measured in the diagram, has averaged 2 per cent. In other words, the transitory effects that lowered inflatio effects, inflation would accordingly have fluctuated around 2 per cent.
After only three years with an inflation target policy, it is too early for a final evaluation. But when allowance is made for transitory effects, it is difficult to demonstrate that the Riksbank has acted asymmetrically.
What this does show is that CPI inflation sometimes deviates from the target and may even move outside the tolerance interval. It is then important that the Riksbank can make it credible that inflation will gradually return to 2 per cent. This can be done either by showing that the forces at work have been mainly transitory effects or, if the effects in question are more permanent, by using the instrumental rate to pilot inflation back to the target rate in the time frame with which monetary policy operates.
At times during 1996 and 1997 the 12-month change figure for the CPI was below the tolerance interval's lower limit of 1 per cent. On a number of occasions the Riksbank pointed out that this was mainly an effect of the repo rate cuts during 1996. In that it was not a question of more permanent effects and the inflation assessment pointed to a future increase, there was no need for further repo rate cuts. During the summer and autumn of 1997 inflation did, in fact, move up fairly rapidly, as expected, in that these transitory effects successively dropped out of the 12-month figures. As a result, actual inflation has moved closer to the underlying rate. In describing this tendency, the Riksbank has been far from alarmist.
The latest inflation report
This brings me to the future development of inflation and the current situation for monetary policy.
Our December inflation report provides a natural point of departure. The assessment there was that during 1998 inflation would somewhat exceed the target, with a gradual acceleration to a rate of about 2.5 per cent in 1999. Underlying inflation was likewise expected to move up and reach just over 2.5 per cent at the end of 1999. This assessment was broadly in line with the more or less concurrent forecast from the National Institute of Economic Research, for example, as well as with the figures that forecasters in general presented last November and December.
The basis for the Riksbank's assessment included the prospect of a continued strengthening of economic activity, now with increased support from domestic demand. GDP growth was put at about 3 per cent in 1998 and over 3 per cent in 1999. Inflation expectations were considered to be largely in line with the inflation target.
Compared with the September report, the Riksbank's assessment in December represented a minor upward adjustment (by some tenths of a percentage point) of inflation in 1999, mainly in view of higher wage costs and lower productivity. Of greater practical importance for monetary policy was the forward shift of the time perspective by one quarter, which meant that the inflation forecast now included a period when capacity utilisation and inflationary pressure were calculated to be somewhat higher.
The picture of the real economic future had not changed appreciably during the autumn.
The December report also presented two alternative scenarios that were judged to be equally probable. One envisaged stronger inflationary pressure due to higher wage increases and weaker productivity growth. The other involved stronger effects than expected from the Asian crisis and consequently a smaller increase in inflation.
It has been claimed in the public debate that the Riksbank played down the Asian crisis. That is not so. Certain effects were allowed for in the main scenario for the December report. But above all, the risk of the crisis becoming more profound was highlighted in a risk scenario as an important factor for monetary policy to consider.
A general assessment was that in the Swedish economy's present cyclical phase, it is important that monetary policy does not remain expansionary for too long. Considering the time lag associated with monetary policy's effects, a change to a less expansionary stance must be made in good time before resource utilisation becomes excessively high. The stated aim was that this would provide better conditions for more sustained increases in production and employment.
Against this background, when the inflation report had been presented, the repo rate was increased in December by 25 basis points, from 4.10 to 4.35 per cent.
In periods of financial unrest in international markets there is often a "flight to quality"; savers and investors experience uncertainty and rearrange their portfolios in favour of what are perceived as secure assets. In connection with the Mexican crisis in the spring of 1995, Sweden's currency depreciated sharply and bond rates rose substantially; this amounted to a flight from Sweden. In view of the krona's recent weakening tendency, it would be going too far to talk of a flight to Sweden but the fact remains that the situation differs from earlier occasions with international financial unrest. Even so, a correction of the present under-valuation of the krona would be a good thing.
The conclusion from this is that the increased confidence which Swedish economic policy has earned in recent years must be nurtured continuously, not least to make us resilient to turbulent times. All those with a say in price and wage formation must also realise that the Riksbank's endeavour is to fulfil the inflation target. The repo rate increase in December was therefore important when the overall assessment of future inflation pointed to the target being exceeded.
Inflation outlook somewhat more favourable
Information of some relevance for the assessment of future inflation has become available since the December report. It concerns the crisis in Asia, the early stages of the wage negotiations, and the rent negotiations.
The crisis in Asia has deepened both in terms of its effects on financial markets and as regards its probable consequences for global economic growth and inflation in the coming years. The extent of the effects cannot yet be assessed reliably but there are many indications that they will be greater than most people envisaged at the end of last year. This suggests that in the coming months the statistics will point to a more subdued growth of Swedish exports and industrial production.
Some wage agreements have been concluded, for instance for the paper industry. They still cover just a small segment of the labour market but many observers consider that the existing settlements will serve as a guide elsewhere. It is difficult to arrive at a definite assessment of what these settlements imply for total wage increases but a cautious conclusion could be that the outcome will be somewhat below the Riksbank's estimate in the main scenario, at least for 1998. It should be added, however, that agreements are still pending in many cases and the degree of wage drift in various industries also has to be considered. Still, the probability of the report's alternative scenario with higher inflation should have decreased.
Rent costs account for 11 per cent of the CPI and are the largest main item. For 1998 and 1999, final agreements are in place for about 65 and 25 per cent, respectively, of the stock of rented dwellings. They give average increases of just over 2 per cent for 1998 and almost 3 per cent for 1999. The 1998 figure is somewhat lower than expected. There are signs, moreover, that rents for 1998 may be re-negotiated, which has some downward implications for the assessment of inflation this year.
No reason at present to alter the repo rate
All in all, then, in the period of relevance for monetary policy, global economic development will probably be somewhat weaker than estimated earlier. For this reason the outlook for inflation is somewhat more favourable. To some extent, stronger competition from imports has a similar effect. Somewhat lower inflation in 1998 may also result from a lower increase in rents.
For the construction of monetary policy it is equally important at present to consider the altered spectrum of risks. The probability of an outcome for wages that is appreciably above the main scenario has diminished. At the same time, the effects of the Asian crisis are highly uncertain. Much depends on how the crisis is managed. On the other hand, the fundamentals for growth are in place, particularly in Europe.
The conclusion is that at present there is no reason to alter the repo rate.
