Heikensten: Introduction on monetary policy
Let me begin as usual by thanking you for the opportunity to come here and give an account of the Riksbank’s monetary policy. It is ten years now since my predecessor Urban Bäckström was here for the first time to attend a hearing under such transparent and public conditions. At that time the inflation target had only recently been introduced. Given that this is the tenth anniversary of the hearings in their current form, I think it is appropriate to begin by looking back on the past years. One conclusion is that much has gone well; the inflation rate has been low and stable, and growth in both real wages and the Swedish economy in general has been high. Employment growth has not been quite as robust, however. Against this background, I will be focusing particular attention today on the issue of monetary policy’s ability to influence employment. By way of conclusion, I intend to touch upon the Riksbank’s outlook for economic developments and inflation in the years ahead and to comment on the decision we made yesterday regarding the repo rate.
Ten years of low inflation
New policy
The crisis at the beginning of the 1990s was a dramatic, tragic conclusion to almost twenty years of stabilisation policy problems in the Swedish economy. The 1970s and 1980s witnessed a succession of cost crises. The fundamental problem was that economic policy had been unable to prevent excessively fast rises in wages and prices from ending up time and again on a collision course with the commitment to maintain a fixed exchange rate. As a result, the krona was devalued five times altogether over the period 1976-82. When the fixed exchange rate was finally abandoned in November 1992, following extremely high interest rates and costly attempts to defend the krona, it was obvious that something fairly drastic had to be done to enable the Swedish economy to develop more favourably than during the previous twenty years.
The solution was what could be called a shift in stabilisation policy regime, whereby the framework and rules for both monetary and fiscal policy were fundamentally changed. The shift in regime involved the introduction of a distinct division of roles in Swedish economic policy. The exchange rate remained variable and monetary policy was given the main task of trying to keep inflation at a low, stable level. Fiscal policy, which had often resulted in large deficits under the previous regime of a fixed exchange rate, would now instead be targeted at achieving long-term stability and sustainability in the public finances.
Favourable results
Looking back at the ten or so years that have passed since the change of regime, I think that most would agree with me that the shift in policy has turned out well and that the new stabilisation policy regime has lived up to our hopes. Sweden has gone from having been a high-inflation economy with recurring cost crises, high interest rates and unstable economic performance, to being an economy with low, relatively stable price rises and wage increases that put real money in people’s pockets, and where the risk of overheating and subsequent recession has diminished markedly. In an international perspective, too, the Swedish economy has performed well in the past decade, in stark contrast to the 1970s and 1980s.
The robust performance is evident if you compare developments in a number of central variables during the 1970s and 1980s with their developments under the new stabilisation policy regime. (Chart 1.)
In the 1970s and 1980s the rate of price increases stood at just over 8 per cent a year, while the corresponding rate since 1995 has been just below 2 per cent. The lower inflation rate has not entailed weaker growth in output. On the contrary, GDP growth has averaged around a half percentage point higher a year under the new regime. Productivity growth has also been more robust. The higher economic growth and productivity has contributed to healthy growth in real wages. This is one considerable difference compared with the developments in the 1970s and 1980s. While real wages have risen by about 2.5 per cent a year since 1995, the annual increase in real wages was no more than around 1 per cent in 1970-89. So, even though nominal wages rose more than 9 per cent a year during the 1970s and 1980s, purchasing power improved by only about 1 per cent; the high inflation and devaluations eroded the rest.
In fairness, I should add that in addition to the shift in stabilisation policy regime there were a number of significant changes in other areas too that may have contributed to the robust performance. For example, the rapid advances in information technology are likely to have contributed to the improved productivity. Another important factor underlying the stronger recorded productivity growth in the whole economy is the relatively smaller role of the public sector today compared with ten years ago. Yet another feature of the performance of the last ten years is deregulations and increased international competition, the latter boosted by membership of the EU and the ongoing integration of the world economy.
So, Sweden’s economic performance after the crisis in the early 1990s has been healthy in many respects. This does not mean that there are no problem areas. One of these is employment. There was indeed a relatively fast upturn in employment after the crisis years. It has also proved possible to combine low unemployment, compared with other European countries, with low inflation, contrary to what many experts expected. But open unemployment has only come down towards 4 per cent in a few occasional years, and stands today at 5.5 per cent. The number of employed now is lower than before the crisis, and some 1.2 million people of working age (16-64 years) are not participating in the labour force.
The regime shift and employment
Against this backdrop it's not surprising that a lively debate arises from time to time about unemployment and what can be done about it. What is perhaps more surprising is that some people still assert that the shift in stabilisation policy regime at the beginning of the 1990s is what has caused unemployment to be higher today than it was during the 1970s and 1980s. Sometimes the argument goes that Sweden has given up its previous objective of full employment in order instead, for some more or less obscure reason, to introduce a target for inflation. If we only were prepared to return to the high-inflation policies of the 1970s and 1980s, then everything would be better.
There is much to say about this. Firstly, it is difficult in a more general sense to imagine that the markedly improved stability in economic policy that the shift in regime brought about in the long run could be something negative for employment. Instead, there are many reasons to believe that more stable economic performance over time also leads to conditions that foster high growth and employment.
Moreover, a simple comparison of unemployment data between the periods before and after the shift in regime proves rather misleading. This is because of the attempts in the 1970s and 1980s to meet the target of full employment through a policy that was unsustainable in the long run – it was exactly because of this that the change in regime was implemented. As I mentioned earlier, this policy involved Sweden devaluing the krona as soon as we were hit by a cost crisis that resulted in weaker developments in exports and employment in the private sector. The devaluation of the krona restored competitiveness by lowering real wages for a number of years until excessive increases in prices and wages compared with abroad once again eroded the scope that had been created. The other way in which unemployment could be kept down was through a sharp increase in public sector employment. From having comprised 15 per cent of the labour force in 1960, the proportion of the employed accounted for by the public sector rose to just over 30 per cent by the end of the 1980s. So even though jobs were being lost in the private sector, employment held up through increased employment in the public sector.
However, neither the decreases in real wages due to the devaluations or the increased public sector employment could solve the problems in the long run. As a result of the devaluation policy, it became increasingly difficult to keep inflation in check. Wage earners tried to compensate themselves for the reductions in real wages, and firms had little reason to resist; everyone believed that the krona would be devalued if price and wage inflation became too high and competitiveness was threatened. The expansion of the public sector contributed in turn to a rise in the tax burden that was unsustainable in the long run. One way of putting it is that the devaluations and the increase in public sector employment allowed us to put off dealing with our employment problems.
Monetary policy and employment in the long term
The crucial insight when discussing unemployment and jobs – an insight which has firmly rooted support in decades of economic research and practical experience - is that loose monetary policy cannot be used to bring about a lasting rise in employment. The best that monetary policy can do is to reduce the fluctuations in the cyclically determined component of unemployment. Consequently, it’s not possible to create lasting jobs by systematically keeping interest rates low, weakening the exchange rate and allowing inflation to rise.
In other words there is no opposition between a target for inflation at the levels pursued by most central banks today and a policy of high employment. In order to reduce unemployment permanently, attempts must be made to influence the structural component of unemployment, i.e. the component that is not due to fluctuations in the demand for goods and services in the economy. And this requires completely different measures than those of monetary policy. In particular, it is a question of implementing measures that influence the functioning of the labour market.
One indication that the employment problems we are seeing today are largely structural is the big differences in unemployment that exist between different regions and different occupational categories. (Chart 2.) Unemployment figures are twice as high in some regions as in others, and the percentage of people receiving benefit from their unemployment fund is twice as high in some occupational groups as in others. The fact that these differences remain over long periods, even when there is a shortage of skilled labour, suggests that the labour market doesn’t function as well as might be desired. The problems in the Swedish labour market are also underlined by the fact that open unemployment barely came down to 4 per cent when economic activity was at its strongest a couple of years ago.
Target deviations and employment
How the Riksbank works
So, even though monetary policy cannot be expected to have any tangible impact on employment in the long term, short-term effects can arise. This is because a policy aimed at attaining the inflation target may entail changes in interest rates, which temporarily affect demand so that employment rises or falls. The limit for how long and how much monetary policy can raise employment is determined in practice by the economy’s long-term growth potential, which in turn depends partly on how well the labour market functions. When we conduct monetary policy with an inflation target, this means under normal circumstances that we can help stabilise actual output around its long-term potential. If the Riksbank is to be able to take any account of employment in the short run, however, it is particularly vital that economic agents have confidence in monetary policy and the inflation target; this is a clear lesson from the situation in the 1970s and 1980s.
As you know, the principal instrument we use to try to meet the inflation target is the repo rate. We set this rate with the aid of forecasts of inflation a couple of years ahead. Naturally, a lot can change in the space of two years so that our forecasts do not prove accurate. We can make up for this to some extent by being able to change the rate in the light of new information. But we can’t compensate entirely for such events, of course. Sometimes, developments affect inflation at very short notice: a sharp hike in electricity prices, an equity price bubble that bursts, a war that heightens unease in the world and curbs demand, etc. In such cases it’s not possible to prevent inflation from deviating from the target, and even if it was possible to lessen the deviations, it may require such sharp movements in interest rates and such substantial effects on the economy that it would not seem desirable or reasonable. When inflation deviates from the target, we normally try to bring it back gradually to the target so as to avoid unnecessary fluctuations in the real economy and in interest rates.
When Sweden became one of the first countries to define its objective for monetary policy in terms of a definite figure – even though we were very aware of the fact that it would be difficult to meet such a target exactly – we did so for several reasons. One was to find an anchor for policy - a target around which inflation expectations could be stabilised. But it was also done with a view to facilitating evaluation of our activities and to make it easier to enforce accountability. Further steps were taken in the same direction when we, once again among the first in the world, decided to publish the material underpinning our decisions, mainly our inflation forecasts. This provided a more comprehensive basis on which to analyse the background to our policy. Are deviations from the target a result of incorrect assessments? If so, in what respects? Do we differ from other forecasters or not, etc? The conditions for understanding and evaluating our actions were probably improved further when the Riksbank in 1999 began to publish the minutes from its monetary policy meetings. Now it is also possible to examine the arguments put forward. Were they the right arguments or not in the light of what has happened? On the whole, I believe that we have laid a solid foundation during these years for an in-depth assessment of our decisions.
Inflation and inflation expectations
So how well has the Riksbank succeeded in achieving its target in the past ten years? (Chart 3.) Looking at the average inflation rate since the target was introduced in 1993, CPI inflation has been 1.8 per cent, while the measure of underlying inflation that we in practice have used over a large part of the period to explain our policy, UND1X, has been 2.1 per cent. If we instead make the comparison from 1995, the year when policy began to have its full impact, the corresponding figures are 1.4 and 1.9 per cent, respectively. Even when average inflation has stood fairly close to the target, sharp deviations can of course occur periodically. This has also been the case; during shorter or longer continuous periods inflation has been above or below target. At the same time it is also evident that the size of the deviations depends somewhat on the inflation measure in question. UND1X has fluctuated less as it is not affected as much by changes in interest rates and taxes.
In this context it is also interesting to look at inflation expectations; have they been affected much by the fact that we sometimes have missed the target or that we on average have been slightly below it? (Chart 4.) Even though inflation expectations can be measured in different ways, the overall picture is clear. From around 1996-97 expectations have been fairly well in line with the target a couple of years ahead. Sometimes, of course, they have moved according to the actual inflation rate, but they have not been much above or below the target for any appreciable length of time. This is a much better performance than most expected when the monetary policy regime was changed in the early 1990s, and many doubted that we would attain the low, stable inflation that we actually have achieved.
Allow me to make a comment in this context. Sometimes, it is claimed that the Riksbank acts asymmetrically, being more focused on avoiding exceeding the target than falling short. However, this view is not supported by measurements of inflation expectations: economic agents do not believe that we have a “secret agenda”. Instead, they believe that we try to do exactly what we are doing: trying our best to achieve inflation of 2 per cent and a stable economic performance in Sweden.
Reasons for target deviations
So what has caused the target deviations in the past ten years? Four factors seem to have been especially significant:
- At the beginning of the period with an inflation target, particularly in 1996, the CPI was affected considerably by the dramatic decline in interest rates that ensued when the credibility of the new regime was established.
- On a couple of occasions in the past years, rapid rises in energy prices in particular, but also in food prices, have pushed up inflation. The rises were essentially related to supply factors, such as water shortages in hydroelectric power companies’ reservoirs. The effects were judged to be temporary, which also proved to be the case.
- During certain periods, not least recently but also in the latter part of the 1990s, imported inflation has been unexpectedly low.
- At the end of the 1990s and the beginning of this decade, productivity growth has been unexpectedly robust.
It is interesting to note that during several periods when inflation was below target, monetary policy was clearly loose at the same time as GDP growth was high. But in spite of this, inflation has been too low. What has happened above all is that unexpected supply changes have dampened inflation. This has been a global phenomenon to a large extent; in other words, a situation that we have shared with other countries.
Developments in recent years
With hindsight, we can certainly see that the Riksbank has not always succeeded in predicting inflation correctly and that monetary policy therefore could have been conducted somewhat differently on a number of occasions in the past ten years. This applies, for example, to 1995 and perhaps also to 1999-2000, even though the latter period’s probably too low interest rates did not have any consequences since the upswing in the economy and inflation was broken when the equity price bubble burst. This has also been the case in the past few years. Since the second half of 2003 inflation has been below target, and it is likely that it will continue to be so for at least another half year or so. This suggests that interest rates could have been lower in primarily 2002 and 2003. One conclusion from this is that we must deepen our understanding of factors such as productivity and of the impact on international prices of tough global competition. In these respects we are essentially in the same boat as other forecasters. As far as we can see, no Swedish forecasters have presented forecasts that have differed markedly from our own. In discussions with colleagues abroad, moreover, I have noted that all have been surprised by the low international inflation in the past year. When this kind of shock occurs, whether it is a question of lower or higher inflation, it is reasonable to allow the effects to dissipate gradually and not to act too strongly with the interest rate.
One feature of the debate on monetary policy in the past year has been more or less fanciful assessments of how much higher employment could have been if only the Riksbank had acted more prudently. The figures have varied from 20,000 up to 75,000 more jobs. It is certainly not easy to make these kinds of counterfactual estimates; I admit that much. They require some rather bold assumptions. They presuppose not only that the Riksbank would have been considerably more successful than all other forecasters in assessing where the economy is headed. They also presuppose that we would have succeeded in convincing external parties of this in advance, and would have done so at least partly during a period when inflation in Sweden was clearly above the 2 per cent target.
In order to get an idea of what the effects on employment of a different interest rate policy could have been, we have conducted a couple of experiments ourselves. They show that it would have been difficult to achieve employment effects over 10,000 people even if the repo rate had been a whole percentage point lower in the past year. In order to put this figure in perspective, it may be worth remembering that around 200,000 people in Sweden today are openly unemployed and almost 400,000 people of working age are on sick leave. For safety’s sake, I should perhaps also stress that it is a question moreover of a shift in the timing of an employment upswing, and not a question of jobs having been lost for good.
Allow me to round off this discussion. In my opinion, the shift to low inflation has worked well on the whole, considerably better than most people thought. This does not mean of course that our policy at various times does not deserve criticism, and recently the Riksbank has undeniably fallen short of the 2 per cent target. This spurs us to deepen our understanding of the inflation process even more. Nevertheless, some of the criticism that has been levelled at the Riksbank has been out of proportion and ill-founded. The unemployment problem is so serious that it deserves a better discussion.
The current assessment
By way of conclusion, I shall now comment on the Inflation Report that we published today and the decision we made yesterday regarding the repo rate.
The main scenario
In brief, the main scenario described in the Inflation Report can be summarised as follows. International economic activity is expected to continue to strengthen this year. Over the coming two years, growth is expected to be robust, although it will probably be dampened somewhat. Thus far, the underlying strength of the upturn has if anything surprised on the upside, largely thanks to developments in a number of emerging economies in Asia, Latin America and eastern Europe. There is therefore reason to be slightly more optimistic than we were in the May Inflation Report with regard to short-term prospects for growth in the world economy, despite the signs of weaker developments in the United States. Coupled with the unexpectedly large price increases for oil and other commodities, this means that international price pressures are expected to be slightly higher than previously forecast, particularly this year and next year.
Economic growth in Sweden is also forecast to be robust throughout the forecast period and has been revised up in relation to the May forecast. The unexpectedly high international demand has contributed to surprisingly rapid growth in exports over the summer. The increasingly evident recovery in the IT and telecoms sector has also contributed to the favourable growth in exports. The Budget Bill contained signals of more expansionary fiscal policy in the period ahead, which can be expected to further boost demand in the economy through effects on both private and public consumption. New National Accounts data also indicate that the pickup in investment long expected by the Riksbank and other forecasters now appears to have begun. As in the May Report, gradually higher resource utilisation and rising international prices are expected to lead to a gradual increase in inflationary pressures. So, although inflation is expected to be slightly higher than previously anticipated, our assessment is that it will be in line with the Riksbank’s target towards the end of the forecast period.
Some questions
One could say that the Inflation Report revolves around some questions, which I intend to touch upon briefly:
- One question is to what extent the high oil price will significantly hamper international economic activity. Although we, like most other forecasters, still assume that the oil price will fall again over the coming years, our assessment is that the price will be higher during the forecast period than we estimated in May. However, there are indications that the effects of the high oil price on growth will be relatively limited. For instance, the rise in the oil price appears to be largely a result of the fact that the world economy is expanding, although concern over terrorist acts and other factors also appears to have played a role. In addition, western world economies are today much less dependent on oil than they were in the 1970s, while monetary policy is now conducted in a way that makes it difficult for inflationary impulses from oil price rises to gain a foothold. Our current assessment is therefore that the high oil price will dampen, but not derail, the international economic upswing and that the effects on global inflation will be fairly modest.
- Another question is how inflation could be so low thus far this year, despite growth being relatively high. One important explanation for the relatively low domestic inflation is, as I mentioned earlier, the high productivity growth. Combined with low wage increases, the favourable productivity growth has meant that cost pressures in the Swedish economy have continued to be moderate. This has been the case despite electricity prices failing to fall as sharply as we assumed in May. Imported inflation, which I also touched upon earlier, has been weak as well. The high oil price has been countered by price falls for other imported goods, including clothes and footwear. This question is in turn connected with another: how is it that we expect inflation to remain relatively low during the forecast period when we have revised up our growth forecast? One explanation is that we have allowed our assessment to be affected by the new productivity figures, the lower wage outcomes and the continued weak situation in the labour market. In our current assessment, inflation would probably have been around a tenth of a percentage point higher if we had not taken into account the new information received.
The balance of risks
To sum up, the forecast in the main scenario is that Swedish GDP growth will be 3.6 per cent this year and 3.2 per cent a year in 2005 and 2006. Private consumption and investment are expected to take over from exports as the main driving forces behind the pickup in demand in 2005 and 2006. UND1X inflation is expected to be 1.4 per cent one year ahead and 2.0 per cent two years ahead, while the corresponding figures for CPI inflation are forecast to be 1.6 per cent and 2.5 per cent, respectively.
This time, both the international and domestic inflation risks are judged to be balanced. On the international front, the main upside risk is that the oil price will develop more unfavourably and have greater contagion effects on other prices than is currently foreseen. The downside risks mainly consist of a number of financial imbalances that could contribute both to lower inflation and weaker economic activity. These include high house prices and an increase in household indebtedness in many countries, as well as the large deficits in the US federal budget and current account. On the domestic front, the risks are mainly related to productivity growth; more specifically to what extent the robust productivity reflects a more permanent structural phenomenon and to what extent it is due to cyclical factors. There is considerable uncertainty in this regard and there is a risk both of overestimating and underestimating future productivity growth.
The balanced risk outlook means that the probability of inflation being higher than the forecast in the main scenario is judged to be equal to the probability of it being lower. In other words, the risk-adjusted inflation forecast is the same as the forecast in the main scenario.
Interest rate decision
Given the outlook for economic developments and inflation that I have outlined here, we decided yesterday to leave the repo rate unchanged.
As the economic upturn progresses and resource utilisation increases further, it is reasonable to assume that monetary policy will in future need to become gradually less expansionary. There is nothing dramatic about this; rather, this is normally necessary in an economic upswing to ensure that the inflation target is attained and that economic developments in the period ahead are stable. It is also what most forecasters appear to be expecting. However, it is worth emphasising that it is not possible at present to determine exactly when it may be appropriate to make a first adjustment of the interest rate level. This will depend as usual on how the Riksbank assesses the inflation outlook and on the new information received regarding economic developments.
Thank you.
