Job Polarisation In Europe: Are Mid-Skilled Jobs Disappearing?

If we look at what kinds of jobs have expanded most in recent years, we can split European countries into two: those that experienced job polarization (with ‘good’ and ‘bad’ jobs growing, and the middling type shrinking relatively) and those that experienced upgrading (with a relative expansion of ‘good’ jobs). These patterns can be traced back to the 1970s.

Some argue that technical change and international trade are hollowing out the occupational structures of European economies, destroying mid-skilled industrial and services occupations that traditionally formed the core of the middle class. Others argue that, on the contrary, the information revolution is driving an unprecedented expansion of demand for highly skilled labour that educational systems are struggling to cope with. And yet others argue that recent labour market deregulation in a context of economic stagnation produces an anaemic labour market that can only create low-paid jobs. Which of these claims is closer to the truth?

Our own work for Eurofound´s European Jobs Monitor (for a list of publications see here) shows that all of those claims can be partially valid. In short, in recent years there has been quite a significant diversity in the patterns of occupational change in Europe (see figure 1), which we can summarize in the following two points.

• The most frequent developments have been job polarization and upgrading. Between 1995 and 2007, job polarization was observed in the Netherlands, Germany, France and the UK. Job upgrading mostly took place in the Scandinavian countries. In Southern European countries there was a significant expansion of mid-paid occupations, often simultaneous with a shrinking of the lowest-paid.

• The recent economic crisis reduced somewhat the degree of diversity in the patterns of occupational change across Europe, though not entirely. It tended to generalise the polarisation pattern, particularly within some countries on the periphery that suffered an intense destruction of employment in mid-level occupations.

Figure 1: Yearly average change in employment by job-wage quintiles. The bars represent change in employment in each country for five initially equal categories of jobs, from lowest paid (on the left) to highest (on the right).

It is important to note that most of the diversity across countries and periods concerns middle- and low-paid occupations. In other words, high-paid occupations tended to grow in absolute and relative terms across most countries and periods, which seems to fit reasonably well the argument that recent technological change tends to increase demand for high-skilled occupations. But the observed diversity in developments for low- and mid-paid occupations (which grew in some countries and periods, but not all) does not fit very well with a technological argument. Job polarization seemed to be, instead, associated with a process of deregulation of employment contracts that has taken place in some countries of the core of Europe in the last 15 years (such as Germany, the Netherlands or France).

A problem with the evidence discussed so far is that it covers the whole of Europe but only for a relatively short period of time, coinciding with a single business cycle (of boom between 1995 and 2007, and bust afterwards). Eurostat´s harmonised European datasets do not allow one to go back further than that. For that reason, in the latest annual report of the European Jobs Monitor, we have teamed up with a group of researchers around Europe to extend the analysis all the way back to the 1970s drawing upon national data from the UK, Germany, Sweden and Spain, and using exactly the same approach and methodology as in our earlier work.

Figure 2: Absolute change in employment across job-wage quintiles in 4 European countries, 1970-2014 (thousands). The arrow lines indicate the periods represented by each chart.

Perhaps the most striking finding of this expanded scope of our analysis (see figure 2) is the consistency of the patterns of occupational change over time in the different countries:

1. The German labour market has been “hollowing out” for the whole period, particularly after reunification: mid-paid occupations have been consistently shrinking in relative and absolute terms in all periods after the 1990s. It is important to note that job polarisation in Germany predates the more radical period of labour market reforms of the 2000s (the Hartz reforms), and can be traced back to the erosion of the traditional German employment model in the context of high unemployment from the early nineties.
2. The dominant picture in the UK is also one of job polarisation, although much less stark than in Germany and with significant exceptions in the late eighties and 2008-2013. A rapid and advanced process of deindustrialisation seems to explain to a large extent the phenomenon of job polarisation in the UK, although (perhaps surprisingly) after the 1990s a centripetal redistribution of employment in the public sector had a similar effect in the occupational structure.
3. In Sweden, the long-term trend is clearly one of job upgrading. The success of the Swedish export-oriented social welfare model in recent decades is reflected in a consistent expansion of high-skilled/high-paid occupations, while powerful unions have effectively blocked the development of a low-paid sector. Even in the deep recession of the early nineties, the very significant destruction of employment in Sweden was concentrated in the lowest paid segments of the labour market, with continuing (though negative in this case) job upgrading.
4. Spain is the most cyclical case of all, with sharply different patterns in upturns and downturns. In the first half of the period, the dominant picture is one of negative upgrading, with a massive destruction of employment in mid- and especially low-paid occupations as the economy modernised. In the long economic expansion between 1995 and 2007, Spain experienced an almost flat (though mildly upgrading) expansion of employment across the occupational structure. But in the crises of the early nineties and post-2008, it experienced very drastic processes of job polarisation, shredding mid-paid jobs in construction and manufacturing in large numbers.

These long-term results reinforce the idea of diversity and institutional embeddedness of the patterns of occupational change in Europe. Each country requires its own-context specific explanation: no single factor can account for what happened in all countries.

The implication for policy is that there is no inescapable trend in occupational developments. The pervasive forces of technical change or international trade do not necessarily polarise or upgrade occupational structures: different policies and contexts can significantly alter their effect. And as in many other areas, the Scandinavian countries provide the most attractive example.

This blog post is based on research carried out by the author jointly with John Hurley and other researchers for Eurofound. The full report can be downloaded here.