Heikensten: Economic conditions for wage formation

I should like to begin by thanking those who have arranged this seminar for providing a good opportunity of discussing the matter of wage formation.

Wage formation is a constantly topical and major issue when discussing economic developments in Sweden. Unduly high nominal wage increases over a number of decades have contributed to our economic problems. And now that the prospect of a broad economic upswing is coinciding with the renewal of wage agreements for large segments of the labour market, wage formation is particularly relevant. In the latest Inflation Report the Riksbank considered that perhaps the greatest threat to a continuation of low inflation is the possibility that wage formation does not function properly.

A further twist to the discussion about wage formation seems to have arisen from the Riksbank’s view in the Inflation Report that in the next two years nominal wages will rise 4–4.5 per cent while the rate of inflation moves up comparatively moderately. Some press commentators have taken this to mean that the Riksbank has given a go-ahead for wage increases of that magnitude.

Against this background I shall be saying something about how the Riksbank perceives the present situation and the outlook for the coming years. I shall be kicking off from the assessment in the Inflation Report. Then I shall comment on why our Report did not envisage a stronger price rise despite wage increases that are relatively high. Finally I shall talk about the frameworks for wage formation I consider should apply in an inflation targeting regime.

In the Inflation Report, which the Riksbank published early in October, it is foreseen that the GDP growth rates in this and the next two years will be 3.6, 3.8 and 3.0 per cent. These rates are higher than we have been accustomed to in recent decades and certainly stronger than the Swedish economy can cope with in the longer run with unchanged, low inflation. Growth of this order is expected gradually to utilise the economic resources in Sweden that are unutilised at present. We estimate in the Report that the output gap will close in the coming two years and that some capacity shortages will arise at the end of the period.

Gradually growing tensions and shortages in the economy lead to upward tendencies in prices and wages. As I just mentioned, in the coming two years the Riksbank counts on an annual wage rise of 4–4.5 per cent. This is accompanied by the prospect of annual productivity growth of just over 1.5 per cent, which is somewhat weaker than in recent years, mainly because the increased employment is foreseen mainly in sectors where productivity gains are comparatively low.

All in all, our assessment in the Inflation Report is that inflation will accelerate in the coming years. CPI inflation is judged to move up from 0.9 per cent at present to 2.0 per cent in September 2001. When inflation is measured excluding interest expenditure and changes in indirect taxes and subsidies (factors that monetary policy normally disregards), the rate is 1.8 per cent at present and is expected to rise to 2.1 per cent at the end of 2001. Adding the Riksbank’s appraisal of the risk spectrum in the Report and attempting to quantify the various outcomes gives an underlying rate of inflation two years ahead of just over 2.2 per cent.

In the long run it is, of course, the case that the real wage trend is bound to mirror the ability of labour to produce goods and services. Real wage increases presuppose improvements in productivity. In time, wage increases in excess of productivity growth contribute to higher inflation. A simple rule for the room for nominal wages at the national level is accordingly that wages can rise at the targeted rate of inflation (2 per cent) plus the rate of change in productivity. There are, however, certain conditions for this to which I shall be returning shortly.

In the shorter run, however, productivity growth is not the only factor behind the relationship between nominal wage increases and prices. The relationship may be altered if that firms choose or perhaps are obliged to offset increased wage costs with decreased profit margins instead of passing them through in higher prices. Another factor is that other price components may be moving in the contrary direction and holding inflation down.

In our latest Inflation Report it is mainly two factors of the latter type that are judged to retard inflation in the coming years:

1. Import price increases are judged to be low, which has to do with a number of things. In the Inflation Report we count on some fall in the price of oil. This is accompanied, moreover, by a low price trend for manufactured products, notwithstanding an international recovery, because initially the global economy is considered to have unutilised resources. Last and probably most important, the krona is assumed to appreciate.

Our import price assumption’s downward effect on inflation is clear when one looks at what happens without this effect. The Riksbank also presents a measure of inflation, UNDINHX, that excludes import prices. Measured in this way and all else equal, the rate of inflation in September 2001 is judged to be 2.7 per cent. For safety’s sake perhaps I should add that this measure should be expected to normally be above broader indexes of inflation, such as the CPI, because it has a larger share of services, the prices of which tend to rise faster than for goods.

2. The outlook for inflation in the coming years is also brightened by a number of deregulations, for example the markets for electricity and telecommunications, as well as in connection with the EU’s agricultural policy. The average effect in these years amounts to some tenths of a percentage point.

For a time, then, the price trend can be either more or less favourable than the underlying paths of wages and productivity warrant. In the long run, however, it is crucial for inflation that wages are in line with the productivity trend. The stronger the growth of productivity, the more room there is for wage increases and vice versa.

Improvements in labour productivity can be generated by investment and/or technical innovations. This creates room for higher real wages. But as registered labour productivity can be affected by other factors of a more temporary nature, the productivity statistics do not necessarily show the economy’s productive and wage-paying capability.

In the 1980s hourly productivity growth in Sweden averaged just over 1 per cent a year. There are now some indications that the trend has improved; since the beginning of this decade the annual rate has averaged almost 2 per cent. It is conceivable that a number of structural reforms and the establishment of a low-inflation regime have contributed to this.

Still, there are strong reasons for being cautious about using the relatively high productivity figures in recent years as a basis for future wage increases. After all, the sharp increase in productivity growth occurred in connection with the recession in the early 1990s. A major reason was probably that the decreased demand mainly eliminated less efficient firms. A conceivable reason why productivity growth subsequently remained higher is that the successive economic recovery lead to an increased utilisation of real capital in existing production facilities. If it was these factors that led to increased overall productivity, it does not follow that there is room for real wage increases, at least as long as a quick reduction of unemployment is the aim.

Another approach to identifying a reasonable rate of productivity growth involves looking at what has happened in other countries. This is reasonable if one considers that the conditions for productivity do not differ appreciably between countries in Europe, for example. In the 1970s and ’80s, productivity growth in Sweden was below the EU average but this has changed during the 1990s and Sweden is now above the average. Note that this refers to productivity per employee.

An analysis of this kind obviously does not provide an unequivocal picture of productivity’s future trend. Still, a productivity growth rate between 1 and 2 per cent does seem reasonable. That in turn would make it possible to combine a 2 per cent inflation target with a national nominal wage rise between 3 and 4 per cent. Note that the wage rise refers here to the combination of negotiated increases and wage drift. It is another matter, to which I shall be returning shortly, that the room for wages naturally varies between sectors.

This rate of wage increases is clearly below the Riksbank’s forecast in the Inflation Report. What I have said shows that the Report’s figures for wage increases are just that, forecasts, not assessments of the rate that would be best for the economy in the longer run.

I noted that the expected rate of inflation, 2 per cent, plus the estimated rate of productivity growth is a natural starting point for assessing the room for wages at the national level. There are, however, clear restrictions in this respect. The economy must be in external and internal balance, which in practice implies full employment and no problems with the balance of payments. Moreover, the overall profit level must be conducive to good long-term growth. If this is not the case and the economy, as at present, is not in a situation with full employment, the average level of wage increases must be lower than the simple rule indicates so that increased employment can be generated. EU countries that have succeeded in restraining nominal wage increases have also been successful in creating new jobs. Good examples of this are Denmark and the Netherlands.

Another point I want to emphasise is that I have argued solely from an overall, national perspective. Such a perspective has distinct shortcomings. Improvements in productivity are achieved by individuals in different firms and production units. So that is where the room for wage increases is created. It can then be risky to focus too much on the average rate of wage increases in the economy as a whole. The wage increases that actually occur in different sectors—the structure of wages—are of central importance for the effects on employment and inflation.

Let me present an example. Suppose that a given average overall rate of wage increases, in line with the inflation target, comes mainly from increased wages for labour that is relatively unskilled and inexperienced. The consequences for total unemployment will be very different and more negative than if the same average rate is a combination of wage increases that mirror differences in productivity between occupational groups. Alternatively, suppose that the average overall increase comes mainly from sectors where competition is comparatively low. This can be expected to show up in higher prices to a greater extent than if the wage increases were distributed more uniformly.

In conclusion, it is encouraging that in recent years the labour market organisations seem to have been increasingly in favour of the inflation target policy. In many ways I find this natural. Low, stable inflation creates good conditions for growth and employment. It is most certainly just the members of your trade unions that would be hit in particular if inflation were to be left to accelerate once more. Furthermore, a policy focused on a clear inflation target should be able to provide a stable long-term foundation for wage formation and thereby simplify matters not least for the parties to wage negotiations.

At the same time, it is important that current assessments and policy are discussed continuously. Along with most other observers, we now count on an acceleration of inflation in the coming years. Assessments of the rate at which inflation will move up may vary, of course. But when the underlying rate of inflation is 1.8 per cent, it is obvious that a repo rate increase ought not to wait particularly long, given that nothing unforeseen alters the economic assessment. Timely action creates the best conditions for a stable development with a longer upward phase.