Öberg: The Riksbank and the Swedish economy
I would first like to thank you for the invitation to this lunch meeting. Today I intend to speak about monetary policy and wage formation. This is a very topical subject, particularly given that most of the collective wage agreements in the labour market are due to be renegotiated next year and that wage formation has such significance for developments in inflation and the stability of the Swedish economy. Wage formation has functioned much better over the past 10-15 years than it did during the 1970s and 1980s and it is essential that this good development can continue.
I will divide my speech into five parts: First a section on how monetary policy is conducted, followed by a section containing a long-term perspective on economic developments, and then a section on how wage formation has changed, a section on the conditions for wage bargaining and finally I will conclude with a summary. Allow me to begin by describing how we work at the Riksbank.
How monetary policy is conducted
In Sweden we now conduct monetary and exchange rate policy based on inflation targeting and a floating exchange rate. This is different from the decades up to autumn 1992 when we had a fixed exchange rate as target. The way we at the Riksbank conduct our monetary policy is described in the brochure “Monetary policy in Sweden”, which was published in May this year and is available on our website. I shall here reproduce the main features of the brochure and provide some comments of my own.
The target for monetary policy is to maintain a low and stable level of inflation. The assignment we have been given by Parliament is to “maintain price stability”. The Riksbank has specified an inflation target whereby the annual change in the consumer price index (CPI) is to be 2 per cent with a tolerance interval of plus/minus 1 percentage point.
The target for monetary policy is thus formulated in terms of CPI. However, when we analyse the development of inflation, we look at several different measures. Underlying inflation measured as UND1X is one such measure. This excludes household mortgage interest expenditure and the direct effects of taxes and subsidies. The former is particularly important, as an increase in the policy rate pushes up the CPI. It then looks as though inflation increases when we raise the policy rate, despite the fact that this measure will in the long-term dampen inflation. There is no such problem with UND1X. However, we also look at other measures. For the purpose of more clearly explaining the analysis, it may sometimes be necessary to disregard energy prices, or to divide inflation up into price developments on imported and domestically-produced goods.
The effect of monetary policy on inflation has a considerable time lag. It is therefore necessary for us at the Riksbank to make forecasts of economic developments. Previously, we made these forecasts on the assumption that the policy rate would remain constant during the forecast period. The simple policy rule was then described as being that if inflation was forecast as being lower than the target of 2 per cent, the policy rate would normally be cut. We now make forecasts based on the policy rate developing in accordance with market expectations. If the forecast shows that inflation will be lower than 2 per cent, this does not necessarily mean we must cut the policy rate, but that we should perhaps raise it at a slightly slower pace than the market is expecting.
However, forecasts are uncertain, and become increasingly uncertain the longer the forecast horizon extends. I have spent a lot of time working on forecasts in my career and I perhaps therefore tend to emphasise the uncertainty of forecasts more than many others. In recent years the strong productivity growth and low import prices, both surprising outcomes, have contributed to inflation being below target. We analyse economic developments in a broad perspective and try to identify the factors that could lead to a higher or lower inflation rate than implied in the forecast. This “risk outlook” also affects the monetary policy stance.
We thus have one objective for our monetary policy: CPI should increase by 2 per cent a year in the long term. However, we are not “inflation nutters”. We do not concentrate on inflation to the exclusion of all else; we also give consideration to real economic and financial developments. We conduct what is known as flexible inflation targeting. It is made fairly clear in the preparatory works to the Sveriges Riksbank Act that the idea is that we shall also give consideration to developments in the real economy as long as the inflation target is not threatened.
A long-term perspective
I will now move on to talk about the development of the Swedish economy in a long-term perspective. Let me begin by observing that I personally consider monetary policy to have been very successful since the crisis at the beginning of the 1990s. There is a substantial difference between the economic developments over the past 14 years and during the decades prior to this period. Sweden had an inflation rate much higher than that in Germany every year of the 1980s (see Figure 1).
The 1980s began with two devaluations of 10 per cent and 16 per cent. For a devaluation to be successful, inflation must be kept down afterwards so the country does not lose competitiveness again. However, this was not successful in the 1980s. After the devaluations, Sweden’s competitiveness declined by around 30 per cent up to the crisis breaking out in 1992 (see Figure 2).
The 1990s also began with a depreciation of the Swedish krona amounting to 20-25 per cent when we abandoned the krona peg to the European ecu in autumn 1992. Unlike the 1980s, inflation was afterwards pressed down to the same level as Germany (see Figure 1). We have even continued strengthening our competitiveness through strong productivity growth and moderate wage increases (see Figure 2).
Why did we not succeed in bringing down inflation to the same level as other countries during the 1970s and 1980s? The most important explanations are the history of repeated devaluations, insufficiently tight fiscal and monetary policy, an overheated labour market and deregulation of the credit market with ensuing credit expansion. An important difference compared with the present situation with a clear inflation target is that monetary policy under the regime of that time, with a fixed exchange rate, could not help subdue the overheating as it was detailed to maintain the fixed krona exchange rate.
During the 1990s, on the other hand, we succeeded in bringing down inflation to the same level as other countries. However, this required drastic measures. The 1990s began with the most severe crisis in the Swedish economy since the 1930s. Open unemployment rose to a high point of 9 per cent in 1993. This had a dampening effect on prices and wages. In addition, fiscal policy and monetary policy were very tight. The budget consolidation in the mid-1990s was very extensive and interest rates were kept at a very high level during a transition period. Moreover, the conditions for the Riksbank to take action were improved as the bank was gradually given increased independence and in 1999 was given a statutory objective of safeguarding price stability.
Developments since the crisis have been good on the whole, even regarding other aspects than inflation and competitiveness:
- Growth has been high. On average, GDP has increased by 2.9 per cent a year during 1994-2005 as compared with 1.8 per cent a year during 1970-1993 (see Figure 3).
- Employment has increased and unemployment has declined, although not as much as might be desired (see Figure 4). However, the Riksbank’s assessment in the most recent Inflation Report was that the labour market situation would continue to improve over the coming years.
- Public finances have improved to an extent where they now show stable surpluses of around 2 per cent of GDP (see Figure 5).
- The current account balance has improved and now shows very large surpluses (see Figure 5). Exports as a percentage of GDP have increased from 27 per cent in 1992 to 49 per cent in 2005.
Wage formation has improved
At the end of the 1980s it had become increasingly evident that Swedish wage formation did not function in a manner compatible with economic balance. It was common that central wage agreements signed in the 1980s contained price development guarantees, wage development guarantees, earnings development guarantees, control stations, etc., which led to prices and wages chasing one another in an upwards spiral. The work on changing wage formation can be divided into three important steps.
The first step entailed the Government appointing the Rehnberg Commission in March 1990. On the basis of the Commission’s proposals, stabilisation agreements were signed covering the greater part of the labour market, which meant that the rate of wage increase fell from around 10 per cent in 1990 to 3-4 per cent in 1992-1993. The fact that a lower rate of wage increase was not yet fully established in the mid-1990s is, however, shown by the fact that wage increases rose again to around 6 per cent in 1996, after the 1995 wage bargaining rounds.
The second step entailed the social partners in the industrial sector signing a new cooperation agreement in March 1997, known as the Industrial Agreement, whereby the parties took joint responsibility for wage formation in their part of the labour market and had agreed on a regulatory framework for negotiating wage agreements. The agreement was followed by similar, but not as far-reaching, agreements in other sectors.
The third step meant that the new National Mediation Office began its operations in 2000. This established a third, independent party in the labour market, which could actively promote an efficiently-functioning wage formation process. The National Mediation Office also received increased powers of authority to take action in connection with wage bargaining rounds. The National Institute of Economic Research was given the task of annually producing reports on the economic conditions for wage formation.
The results of these changes are clear and positive. Since the crisis at the beginning of the 1990s, wage formation has functioned much better than before. The agreements no longer contain flexibility clauses and wage drift is much more limited. The number of labour market conflicts has been lower than before. Real wage increases, which reflect what we can buy for our wages, have been better than before, while the nominal increases have been markedly lower than before (see Figure 6).
The social partners have contributed in a constructive manner to this positive development. The positive institutional changes in several areas -–e.g. new regulations for fiscal policy, monetary policy and wage formation - are strongly contributing factors to the favourable developments since the crisis in the early 1990s.
Conditions for wage bargaining rounds
The inflation process has changed in nature in recent years, not only in Sweden but also in many other countries. As the inflation-targeting policy has gained credibility, inflation is increasingly determined by inflation expectations. Inflation has become lower and also appears to be less dependent on variations in capacity utilisation (the short-term Phillips curve has shifted downwards and become flatter). Capacity utilisation mainly affects inflation through wage formation.
However, wage costs still play an important role in the inflation process. For example, the unexpectedly strong productivity growth, combined with moderate wage increases, has contributed to inflation being below target in recent years. Wage formation will remain a focal point in the monetary policy analyses, particularly next year when the majority of the collective wage agreements are due to be renegotiated.
The level of capacity utilisation and the impact it has on wage formation are also important and topical issues. The OECD calculates what is known as the GDP gap. Their interpretation is that capacity utilisation in the Swedish economy is already higher this year than is compatible with a low and stable inflation rate. However, calculating the GDP gap and similar measures is a very uncertain process. Nor are the measures stable; they change over time.
The National Institute of Economic Research (NIER) calculates the labour market gap and equilibrium unemployment. The calculation in the most recent wage formation report shows the labour market gap as being close to zero in 2007 and then becoming slightly positive in 2008 and 2009. Equilibrium unemployment is calculated at 4.5 per cent; it is predicted that this level will be reached next year and that open unemployment will thereafter be slightly below equilibrium level. However, according to the NIER, the social partners have the possibility in the central wage negotiations next year to choose different alternative actions that could either reduce or increase equilibrium unemployment.
Capacity utilisation can also be followed more directly in the NIER’s business tendency surveys and through analyses of labour supply. The business tendency surveys indicate that capacity utilisation is not yet so high seen over the economy as a whole. However, it has risen to high levels in the construction sector and in certain services sectors.
Analyses of the labour supply also indicate that there are substantial labour reserves. Open unemployment is still relatively high. In addition, there will be many latent job-seekers, including full-time students looking for work, and an increasing number of people will continue to work after they reach the age of retirement. Labour force immigration will also increase.
To summarise, I can see that the economic conditions for next year’s wage bargaining rounds look good. Growth is high and employment is rising, but the economic upturn is, in my view, not so strong as to prevent capacity utilisation remaining at a relatively normal level next year. Profits in the business sector are good, but not as exaggerated as in the mid-1990s when we last faced problems with overly high nominal wages. Inflation is moderate and the conditions for attaining the inflation target over the coming years are good, on condition that the Riksbank gradually increases the repo rate. This was also the message in our most recently-published Inflation Report.
Conclusion
Let me conclude with a summary of what I have said so far. The experiences of the 1970s and 1980s were very negative. Repeated crises and poor support from inflation expectations did not give good results in terms of production growth and real wages. The experiences of the past 10-15 years have been good, however. Fiscal and monetary policy have laid the foundations for better economic growth and wage formation has improved. The social partners have actively contributed to this good development. The results are positive. Growth has been high and real wages have risen, despite the moderate increases in nominal wages.
It is of course extremely important that this positive development can continue over the coming years. The economic conditions for next year’s wage bargaining rounds are good and there thus is every reason to believe that wage formation will continue to function well. However, we will of course closely follow developments.
The contribution that the Riksbank can make is to conduct a monetary policy aimed at ensuring that inflation and inflation expectations develop in accordance with our target. This in turn means that the social partners can continue to sign agreements based on the assumption that the inflation target will be met a couple of years ahead. Judging from the surveys we make of inflation expectations, monetary policy in this respect has a very high level of credibility. It is the task of myself and my colleagues on the Executive Board to ensure that this remains the case.
