16 Oct 2018 11:22

Wage formation at a new crossroads

Pay increases in major advanced economies have speeded up in recent years. It thus looks like it was premature to announce the death of the Phillips correlation between the labour market situation and price and wage formation. But in Sweden, we have not seen any similar pay acceleration despite clear signals of recruitment problems in both the private and public sectors, says Håkan Frisén and Olle Holmgren in this week's Reflections.

Partly due to large recent migration, equilibrium metrics are more uncertain than in other countries, but pay increases will probably accelerate a bit in Sweden as well. Although more than a year remains until the next big national wage round takes off in earnest, unions and employee associations are starting to position themselves on important matters of principle, such as the pace-setting role of the export industry and the position of the Riksbank’s 2% inflation target as an anchor for collectively agreed pay hikes.

Although we foresee certain risks, contractual pay hikes are likely to end up a bit higher than in 2017 but compatible with macroeconomic balance, both with regard to the inflation target and competitiveness. However, highly centralised agreements, possibly including special provisions for low-paid groups, may complicate labour market integration and resource re-allocation to expansive sectors.

Despite a relatively strong labour market, Swedish wage and salary increases remain subdued. This development is mainly following an international pattern, but the Swedish situation also has special features. For example, due to very large recent migration, labour market equilibrium metrics are harder to interpret than in other countries.

In addition, Sweden’s strongly centralised wage formation, with both a high level of union membership and closer compliance between different sectors, means that interactions between various players in the Swedish collective bargaining system are especially important. This article discusses how international trends and shortage situations in the Swedish labour market will affect the rate of pay hikes in the coming year. Another factor to keep an eye on is positioning ahead of the upcoming national wage round, which will begin in earnest next autumn.

Slowly accelerating pay in other countries

The sensitivity of price and wage formation to the labour market situation has been a focus of attention in recent years. The question has been to what degree the Phillips curve, which describes the correlation between unemployment and prices/wages, has changed and become less clear. Recent developments have shown that it was probably too early to declare the Phillips correlation dead. A weighted index of pay increases in the four largest advanced economies accelerated from 1.7% in the spring of 2017 to 2.6% today (see chart).

Excluding a short-term peak in the spring of 2008, we are now back at about the same rate of pay increases as during the economic boom before the Lehman Brothers crash of September 2008. This is still a modest rate, considering that unemployment is now generally lower than in 2007 and – in many countries – at levels not achieved for several decades. From an inflation perspective, though, we should also take into account that underlying productivity growth since the Lehman crisis has clearly fallen. If this trend persists, a lower rate of pay increases than before will be required in order to achieve 2% inflation. For the United States and the United Kingdom, for example, our estimates indicate that the correlation between pay and prices is fairly intact, if we adjust for the lower productivity trend.

From a Swedish or Nordic perspective, developments in Europe are of the greatest interest. Wages and salaries in the euro zone have accelerated over the past year, but the rate of increase is still relatively low. In Germany, it has been in the 2.5-3% range for some time. The latest collective agreement in the IG-Metall union field is difficult to interpret, but overall it appears capable of providing pay increases of around 3.5%. Yet it remains unclear whether this will trigger secondary effects that may cause the rate of German pay hikes to take a further upward leap.

Pay increases remain subdued in Sweden

In recent years, Swedish wage and salary increases have largely followed the pattern in other countries. At present, pay is rising roughly in line with the average for major industrialised countries, that is, around 2.5% year-on-year. However, official wage and salary figures from Statistics Sweden do not show as clear an acceleration as in the major industrialised countries, although a lot of anecdotal information indicates that wage drift is speeding up on a relatively broad basis.

Generally speaking, large-scale migration – dominated by heavy immigration from outside Europe – has made the Phillips curve especially difficult to interpret in Sweden. Registered unemployment today is around 6%: well above the rate in most northern European countries, and also high in a historical perspective. In spite of this, the National Institute of Economic Research (NIER), the Swedish government and the Riksbank describe the economy as overheated, with a level of production and employment above the long-term equilibrium. Considering the slow rate of increases in pay and in core inflation, as well as Sweden’s expansionary monetary policy, this seems a bit paradoxical, even though it is supported by the high level of the Riksbank’s resource utilisation indicator, for example. The country’s very high labour force participation rate can also be interpreted as a signal of a tight labour market situation.

In its recently published Wage Formation Report, the NIER drew the conclusion that traditional explanations of low pay increases – such as low productivity increases, low inflation expectations and increased global competition – cannot fully explain the relatively low rate of wage growth.

Another factor may be that the increase in employment largely consists of people born outside Sweden, with generally low educational and productivity levels, which helps push down average pay level. This effect is believed to slow the overall rate of pay increases in the economy by a few tenths of a percentage point. It is also possible that the influx of new arrivals into the Swedish labour market is helping to push down pay levels on a broader front, despite large general problems in the integration process. The NIER also points out that the 2017 national wage round resulted in unexpectedly low pay settlements in relation to the economic situation. Influences from low pay hikes in important competitor countries, plus the unexpectedly low inflation prevailing during the previous collective agreement period (2014-2017) – and thus high real wage growth – helped hold back union demands.

We are sticking to our forecast that pay increases will gradually accelerate over the next couple of years, as a consequence of higher pay increases in other countries. Measured as annual averages, we expect the rate to accelerate from 2.7% this year to 3.0% in 2019 and 3.5% in 2020. The labour market situation will remain relatively tight and unemployment will fall somewhat further in the coming year. The composite resource utilisation indicator has fallen a bit in recent months, but this is entirely attributable to a decline from 75% to 40% in the proportion of companies in the construction sector that are reporting labour shortages. Elsewhere in the private sector, no significant weakening is apparent, and the public sector’s recruitment problems remain at record-high levels. Given our rather optimistic main economic forecast, wage and salary risks are mainly on the upside, but on the other hand downside risks dominate the global macroeconomic outlook.

Preliminary skirmishes ahead of the next wage round

One year remains before the next national wage round takes off in earnest, and nearly 1½ years until most of the current agreements expire, but both the unions and employer associations have already begun to position themselves. Although the playing field will probably not change in a way that radically alters the conditions for future pay increase levels, manoeuvring related to wage formation may affect economic developments on a broad front and have an impact on both Riksbank monetary policy and economic policy in general.

Weaker competitiveness, mainly against Germany, is one of the arguments for continued low pay increases recently cited by the Confederation of Swedish Enterprise. Looking at developments over the past 10-12 years there has been a tendency for unit labour costs, especially in industry, to increase faster than in key competitor countries. This is partly because since 2006-2007, Swedish companies can no longer demonstrate clearly faster productivity increases than other countries.

As we ourselves have emphasised, this trend indicates that the krona’s equilibrium exchange rate against the euro may have weakened to levels of around 9.50 and that the country’s shrinking current account surplus is a signal of this. Although today’s exceptionally weak krona makes Sweden strongly competitive, the unions and employer associations agree that room for pay increases should not be determined by exchange rate fluctuations.

But it is possible to describe some aspects of the competitive situation in more nuanced fashion, especially regarding the choice of starting point for comparisons. Around 2006-2007, Swedish competitiveness was extremely strong, as clearly indicated by a current account surplus of nearly 10% of GDP. Starting at such a level, it is not strange for competitiveness to weaken a bit. Over the past five years, the productivity of Sweden’s business sector has again increased a little faster than in other countries. Especially when the public sector is included (as in the above chart), the trend of Swedish unit labour costs has been well above that of competitor countries. However, this trend is affected by Statistics Sweden’s special calculation methods, which tend to show weaker productivity growth in the public sector than using the methods employed in other countries.

Industrial sector benchmark being questioned again

As for the positioning of the union side ahead of the upcoming wage round, the pace-setting role of industry in wage formation will be one focus of attention. The unions in the domestically dominated “6F” group, with the Building Workers as an important member, have clearly declared that they no longer wish to allow unions and employer organisations in the industrial (exporting) sector to determine the pay increases for the entire labour market. Among other things, they believe that the relatively low pay increases that the export sector tends to agree on are hampering mobility in the Swedish labour market, thus making it harder to transfer resources to expansive domestic sectors. This is partly based on a study that Professor Lars Calmfors recently conducted for the 6F group’s think tank, Katalys. Similar attempts to end the dominant role of the Agreement on Industrial Development and Wage Formation (Industry Agreement) have previously failed, among other things because at the national level, the Swedish Trade Union Confederation (LO) and the Confederation of Swedish Enterprise have firmly backed the current system. But perhaps the conditions are somewhat different this time around.

Other players in the wage formation field have adopted a clear wait-and-see attitude. For example, the NIER has gradually taken major steps backward in terms of its ambition to influence pay increases during the wage round. Its 2018 Wage Formation Report states that the “NIER expresses no opinion on how wages, corporate earnings and employment should develop in either the short or long term. It is up to the social partners to draw conclusions based on their preferences and analysis of the economy.” This is in striking contrast with the Institute’s approach a decade or so ago, when it gave clear recommendations down to the tenth of a per cent – for example in 2003 when the NIER’s report summary included the heading “Annual increase of 3.2 per cent appropriate for the economy”. The same gradual backward steps in macroeconomic analysis also apply to the Economic Council of Swedish Industry. The Council, which works on behalf of both parties to the Industry Agreement, previously published analyses that more or less explicitly suggested suitable ranges for agreed pay increases. Since then it has gradually distanced itself from this approach. The report that it published early this October had the extremely general but typical 2018 title “Industry Under Pressure of Change – Regional Presence, Climate Challenges and Criticisms of Globalisation”.

Disagreements on role of inflation target as anchor

Viewed from a monetary policy perspective, it thus does not appear that the Riksbank will receive much help from other players in emphasising the role of its inflation target as an anchor for wage formation. But the question is how the unions and employer organisations currently view the situation. For years, the two sides were both critical of the Riksbank, arguing that because of its tendency to “lean against the wind”, the key interest rate was being set unnecessarily high.

This, in turn, resulted in a clear bias against a below-target inflation rate, which undermined the role of the inflation target as an anchor for wage formation. Combined with downward-trending inflation expectations, plus the fact that the Financial Supervisory Authority and not the Riksbank was given the main responsibility for financial stability and macroprudential tools, this criticism contributed to the Riksbank’s radical monetary policy shift in the summer of 2014.

In this respect, the union side is now in a more comfortable situation. It can argue that because of the Riksbank’s policy change, the inflation target is again worth taking seriously. To ensure decent real wage increases even in an environment of slightly higher inflation, the unions are likely to make somewhat higher demands than they did in the last wage round. This will probably happen regardless of whether the industrial benchmark survives or not. Although this type of backward-looking perspective on inflation is not logical, it may create problems for the Riksbank, which in recent years has had to struggle mightily against the consequences of the low pay increases in the 2017 agreements.

Today it cannot be ruled out that pay hikes will be boosted a bit too late, viewed from a central bank perspective. This dilemma may be even clearer when it comes to Sweden’s centralised residential rent agreements. Very low interest rates have pushed down the costs of capital, enabling the tenant and owner organisations to limit annual contractual rent hikes to less than 1% in recent years. Assuming that interest rates rise in the future, rent increases are likely to contribute more to inflation.

The employer organisations have perhaps ended up in a more difficult position than the unions. This time around, it is perhaps not enough to criticise the Riksbank for sub-optimal monetary policy implementation. Instead, employer organisations may resort to questioning the inflation target as such. Even earlier, the focus on the consumer price index (CPI) has also been singled out as a problem, since the price trend for companies’ products, not consumer prices, is what determines their wage-paying capacity. But the wide gap in producer price index (PPI) increases from one sector to another is a dilemma. In principle, this should lead to sectoral differences in pay agreements, but after the discouraging experiences of the 1990s, wage formation has again been centralised. Last spring, employer organisations aggressively declared that no divergences from the industrial benchmark for pay hikes would be allowed. It is thus not so easy for the employer side to find a new coherent strategy.

Will pay hikes provide macroeconomic balance without solving structural problems?

To summarise, pay increases in Sweden are now rather well in line with those in other countries, although recently we have not yet seen such clear signs of acceleration as in major industrialised countries. Although the tight labour market will probably cause pay increases to accelerate somewhat, it is unlikely that the picture will change especially much before the next national wage round starts next autumn.

The new positioning attempts now being made indicate that important disagreements in principle are about to flare up, for example regarding Sweden’s competitive situation, the role of the Riksbank’s inflation target and the pace-setting role of export industry in wage formation.Despite certain risks, both upside and downside, contractual annual pay increases will probably end up at about half a percentage point above the 2.2% achieved in 2017. Such a level should be manageable, both from a competitive and inflation target perspective.

This will not, however, prevent the mechanisms of wage formation and the viewpoints expressed by unions and employer organisations from generating other tensions. In many respects, Sweden has applied very liberal regulations, for example concerning labour immigration and the organisational structure of tax-financed activities previously run by the public sector. Such changes also require greater flexibility in the labour market. This may include providing better opportunities for weaker groups to establish themselves in the labour market, but may also require changes in relative wages to make it easier to recruit people for activities that, for various reasons, face an urgent need for expansion. An increasingly centralised system of labour agreements, where sector-specific differences have been eliminated and where unions often demand supplementary provisions for low-paid groups, is not especially consistent with these requirements.