Economics and politics - comment and analysis

Why Sweden copies German wage moderation

Why Sweden copies German wage moderation

By Will Denayer and Heiner Flassbeck 

Wages need to rise in sync with productivity growth plus inflation. If wages do not follow productivity growth, innovation stifles and productivity declines. Sweden also suffers, although it is not a member of the euro zone.

The IMF recently criticised Sweden because its wage increases are too low in relation to how the domestic economy is developing (see here and here). Sweden should look less at Germany, not bother with a largely non-existent wage drift and its social partners should not accept low wage increases. Currently, wage increases of 3% should be possible – the productivity trend is about 1% and inflation expectations are 2%. Instead, wages have increased by 2 – 2.5% in recent years. Sweden also suffers from declining productivity trend growth, which is a problem in all developed countries (see here). In Sweden, the trend declined from over 2% before the crisis to about 1.25% in recent years. The Swedish wage development increasingly follows the German one. The obvious argument is that Sweden’s wage growth must be in line with the rest of the world, otherwise it will price itself out of the market. But the IMF does not agree. Wages should increase with productivity growth, not with what competitors are doing. Swedish employers should pay more. Instead of looking at the Germans – Sweden’s main trade partner – Sweden should look primarily at its own productivity growth and its inflation expectations (see here).

There is no debate about this. Some, such as Anders Rune, chief economist of Teknikföretagen, just say the IMF is wrong (see here). “Why should wages actually increase?” a Swedish Business Life Wage Expert asked (see here). ”High unemployment among foreigners contributes to subdued wages for all wage earners,” Nordea Bank writes. Foreigners’ significantly weaker job prospects, with around 15% unemployment, are a ’forgotten factor’ that retains wages. And on p. 7: ”(E)ight out of ten new jobs go to foreigners, who are likely to take into account their own unemployment risk, rather than the general unemployment when formulating their salary claims.” In short, Nordea sees three main explanations: high unemployment among foreign-born (immigration), low expectations (immigration) and low productivity (partly immigration) (see here).

In fact, in Sweden, the unemployment rate has shown a steady decline, from about 8% in 2012–14 to 6.7% by mid-2017. This is broadly in line with its pre-crisis average and, indeed, far from optimal for a country that has a history of having full employment as a top priority. At the moment, enterprises report labour shortages that approach all-time highs. Notwithstanding a tightening labour market, business sector nominal wage growth has been subdued at 2.3% on average in 2016–17, compared with 2.7% in the 4 years to 2015 and higher than 3% before 2010. In March 2017, the social partners agreed to a central wage agreement that provides for a cumulative wage increase of 6.5% over the next 3 years. There is very low wage drift (the amount by which aggregate wage growth exceeds the centrally-agreed rate owing to agreements at firm level) (see here and here).

In Europe, under German leadership, a process of wage-following behaviour has emerged since the introduction of the euro, leading to persistent wage moderation in many EMU countries. Wage development in Sweden also reflects this evolution. Since the Industrial Agreement of 1997, collective wage formation has been led by bargaining in the industrial sectors, which set the benchmark that is to be adopted everywhere else. Some wage negotiations involve local/firm-level negotiations that take local economic conditions into consideration. Basically, however, industry sets wages – meaning that the sector that is most exposed to international competition defines wage growth for the economy as a whole on the basis of wage developments of its actual or potential competitors. Germany is of course the evidently fateful focal point for Swedish industry and the unions. As said, the conditions relevant to the actors in the industrial production sector have a major influence on wage developments in all other sectors, including the public sector and services. Collective agreements cover over 90% of all employees. Swedish wages have moved very closely to German wage growth in recent years. The proces reflects the employers’ growing emphasis on safeguarding competitiveness and export market share. The broad consensus with the trade unions is motivated by protecting jobs (see here).


Figure 1: Comovement with German wage moderation  (source: IMF).

The IMF construed 3 model series. Models 1–3 include only domestic variables. Significant variables include the unemployment gap, changes in involuntary part-time employment and labour productivity growth. Models 4–6 add indicators of labour market slack in the euro area and Models 7–9 add German wage growth. On the basis of the most sophistacted models (7-9), a scenario analysis suggests that further decline of unemployment, rise in German wage growth and recovery of trend productivity could significantly raise domestic wage growth (see here).

The IMF shows that an increase in productivity growth from the current 1.25 to 1.75% would raise nominal wage compensation growth by 0.5% in the short run. The implication of 1% decline in the unemployment rate and the part-time employment share would lead to a lead to an effect on wage growth of 0.2 and 0.1% respectively. The spillover effect of an increase of German wage growth from 2 to 3% on Swedish wages, on the other hand, would be close to 0.6%. That is much more than the impact of growing employment in the euro area, which is ca. 0.1%. A combination of domestic and foreign factors could increase nominal wage growth by about 1.5% (see here).

A panel regression shows that, compared with other EU countries, wage growth in Sweden follows Germany more closely. For what concerns the impact of productivity growth, expected inflation and euro area labour market slow down on wage growth, there is little difference between Sweden and the EU average. The problem is that the spillover effect of German wage moderation is much larger in Sweden. It is, in fact, three times as big: a percentage point increase in the German wage rate leads to a 0.6 percentage point increase in Swedish wage growth, while the average on EU wage growth is only a 0.2 percentage point. The IMF adds a decomposition exercise to this analysis. It indicates, as can be fully expected, that the recent wage moderation is associated with both lower productivity growth and low inflation (see here).

It is more than noteworthy – because it is not coincidental – that the comovement of German and Swedish wage started in 2008. The financial crisis of 2008-2009 created 150.000 unemployed in Sweden. From 2012 t0 2015, Swedish wage growth completely followed the wage growth in Germany. Since mid 2015, Swedish compensation grew a little bit more than in Germany, only to fall at the end of 2016. Meanwhile, recent German labour compensation growth is at around 2%. That is higher than the pre-crisis average, but it remains grossly insufficient (see Flassbeck’s and Paetz’s latest article here). Hence, Swedish wage growth continues to lagg (see here).

The IMF finishes by stating that there are other structural factors that are not included in its economic analysis that may contribut to recent wage moderation. One is the growth puzzle: we have weak demand, low inflation and low levels of investment.


Figure 2: US and European productivity growth since 1875 (Source: Financial Times).

Notwithstanding all technological advances, US productivity growth is currently at its lowest level since the 1800s – apart from the crisis in the 1880s, it has never been so low (see here). In Europe, productivity growth fell back in the 1980s to the level of 1875, apart from the 1930s and 1940s, when it was much lower. Both US and European productivity growth show a declining trend, with some minor upticks. The IMF also leaves little room for optimism. The researchers comment on the decline of profitability which can be witnessed in the decline in the net return on capital in Sweden. This decline may reflect structural changes that favour less capital-intensive services. It may also be due to global increasing competition in manufacturing that leads to more temporary and part-time work and to more outsourcing (see here).


Figure 3: Gini coefficients of income inequality in several countries, showing Sweden’s big increase (Source: OECD).

What is the impact of insufficient wage growth on productivity? The problem is lacking innovation and investment in a situation of a tight labour market (according to the official statistics) with insufficient wage growth. Productivity is falling because wages do not rise sufficiently. This remains a hypothesis for Sweden, but one which should absolutely be put to the test. In an analysis of 20 countries over 44 years, Kleinknecht estimates that 1% lower wage increase reduces, in the medium term, the growth of value added per labour hour by 0.3 – 0.5% (see here).

Finally, figure 4 shows the financial balances of the economic sectors in Sweden. Household savings remain very high. The balance of companies is barely negative, pointing to an overall lack of investment. In a ’normal’ economy, this balance should be negative: companies take out loans to finance innovation and investment. A market economoy cannot work without innovation. Capital constantly replaces labour and new jobs are being created. The balance of the state is positive. Given the situation Sweden finds itself in, the state should step in and use public money to invest. Interest rates remain historically low anyway. The investments would pay themselves back. The balance with other countries is negative, but this does not seem essential at the moment (the real figures are not big and the current account to GDP ratio is at its lowest point since 2008) (see here).


Figure 4: Financial balances of the economic sectors in Sweden (Source: AMECO).

Unfortunately, politically, this discussion is off-topic (see  here). Today, the European labour parties and the unions do not need concertation. There is complete coordination. They all follow the same road, like zombies, aping the German model. The results are insufficiently rising wages, at best, everywhere, faltering productivity growth and rising poverty and inequality. What would happen if they would work together? It is not that the Germans are doing it better. They are not even doing it right. The German model does not work at home, it cannot work anywhere. It is not proper economics. Now that it is everyone for himself, we all lose. That is the reality of a struggle that is not being fought.