RaboResearch - Economic Research

The state of play of the Eurozone labor market: Recent developments and expectations


  • Government support and a drop in the labor force curtailed the increase in unemployment in the Eurozone. But the fall in hours worked indicate that the decline in labor was in fact large.
  • The largest blow has been dealt to the young and lower skilled employees. Combined with the rather rapid recovery this should help limit labor market scarring and a loss of skills.
  • In the early stages of the recovery, growth in hours worked will outpace employment growth.
  • As the recovery progresses, employment will also again increase on the back of rising activity, yet it will be held back for some time by battered companies as fiscal support is wound down.
  • Combined with the return of discouraged workers then, unemployment will rise temporarily.
  • Specific industries may struggle to find adequately skilled staff, but we don’t foresee general labor shortages. This, in turn, should prevent broad-based pay rises.

Unemployment figures defy GDP malaise

Unemployment in the Eurozone has increased since the outbreak of the pandemic, but much less than was to be expected considering the extent of the drop in GDP (figure 1). The unemployment rate currently stands at 8.0% (April), up from 7.3% pre-pandemic, but down from its peak of 8.7% in the summer of 2020. In 2013 it reached 12.1% at the peak, while the fall in GDP during the financial and debt crises was less intense. The gap between realized unemployment and the expected rate based on the historical relationship with GDP has shrunk since 2020Q2, but was still substantial in the first quarter of 2021 (figure 2). In this note we establish the main reasons behind this relatively muted rise in unemployment, its implications, and our expectations going forward.

A combination of factors has kept a lid on unemployment. We discuss the two most important in our view: i) fiscal support for both businesses and workers, and ii) a record fall in the labor force.

Figure 1: Unemployment has risen much less than was to be expected based on the GDP drop
Figure 1: Unemployment has risen much less than was to be expected based on the GDP dropSource: Macrobond, RaboResearch
Figure 2: Large gap between unemployment estimates based on Okun’s law and realizations
Figure 2: Large gap between unemployment estimates based on Okun’s law and realizationsSource: Macrobond, RaboResearch model estimate

Public support curtails job losses

Governments across the Eurozone have implemented substantial support packages, including direct transfers to hard-hit households and firms, tax moratoriums and government-backed loans. Liquidity support to firms has prevented a large increase in defaults and ‘voluntary’ liquidations, effectively also protecting jobs at these firms. On top of that, governments have directly sustained employment via bans on layoffs and short-time work schemes (figure 3). The exact implementation of the latter differs per country, but broadly speaking firms can temporarily reduce the number of hours work and pay for their employees. The government then (partly) takes over salary payments for those reduced hours of work.[1] Importantly, people fully or partly registered in the short-time work schemes still count as employed, even though they might not actually work at all.[2]

Figure 3: Furlough schemes have been very effective in protecting jobs, but it isn’t over yet
Figure 3: Furlough schemes have been very effective in protecting jobs, but it isn’t over yetNote: *Last data available for Germany is from March-21.
Source: Macrobond, National government websites, RaboResearch calculations.
Figure 4: Hours worked have contracted and grown fairly in line with quarterly GDP swings
Figure 4: Hours worked have contracted and grown fairly in line with quarterly GDP swingsSource: Macrobond, RaboResearch calculations

In some countries these schemes were non-existent before. In others they were, but either the scope has been broadened and access made easier or the costs for firms using the scheme have been reduced. Hence in all major countries recourse to these schemes was much larger than during prior crises, partly explaining the more muted drop in employment and modest rise in unemployment than in past crises. The number of people in these schemes[3] has decreased significantly from its peak in April (figure 3), but continues to concern more than 5% of employees (over 7 billion people) and contract hours in the four largest Member States. Meanwhile, figure 4 shows that whilst the fall in employment has been very limited when compared to the drop in GDP (in Q4, -2% y/y and -5% y/y respectively), effective hours worked as reported in labor force surveys have contracted fairly in line with GDP. Hence the deterioration in employment has been more pronounced than one would assume when only looking at unemployment figures[4].

An unprecedented fall in the labor force

Another reason behind the low unemployment rates is more statistical in nature: the number of people active in the labor market has dropped (i.e. the labor force, figure 5). People who lost their job did not all become unemployed but some of them turned inactive.

Figure 5: A share of fired workers has turned inactive rather than unemployed
Figure 5: A share of fired workers has turned inactive rather than unemployed Source: Macrobond, Eurostat - ILO harmonised unemployment figures, RaboResearch calculations
Figure 6: The number of discouraged workers has grown with 1.3 million
Figure 6: The number of discouraged workers has grown with 1.3 million Source: Macrobond, Eurostat, Labor Force Survey, RaboResearch calculations

The labor force is made up of both employed and unemployed people (figure 7). Yet a person is only considered to be unemployed if he or she doesn’t have a job, but is willing to work and is actively looking for a job. In the first half of 2020, the number of people without a job not seeking work, also called discouraged workers, grew substantially (figure 6), causing the labor force to shrink in 2020H1. Had the labor force remained unchanged compared to 2019Q4, i.e. had all those who had lost their job officially turned unemployed rather than inactive, the unemployment rate would have increased to 10% in 2020Q2, rather than the observed 7.6%.

To be fair, it is common that the labor force shrinks in times of crisis, because people looking for a job get discouraged as there are  fewer jobs available during a crisis. Figures 5 and 6 show that this also happened after the global financial crisis and subsequent Eurozone debt crisis. Back then it took three years for the labor force to return to its 2009Q1 peak, whilst the number of discouraged workers never declined to its 2009Q1 level.

Yet the magnitude of the drop in the labor force was much larger in 2020Q2 than during previous crises. This probably has to do with the peculiarities of the current crisis: hospitality venues, retail stores, and factories were shut for many weeks during the first lockdown. Effectively making it pointless to look for a job in these places, while at the same time people were confined to their homes, making it difficult or even impossible to actively look or apply for a job elsewhere.

Figure 7: The decomposition of the working age population
Figure 7: The decomposition of the working age populationNote: The unemployment rate is the number of unemployed persons as a % of the labor force.
Source: RaboResearch.

Although the drop in the active population has put a lid on the increase in official unemployment, the longer-term impact of larger inactivity is arguably more negative than that of higher unemployment: the greater the distance from the labor the market, the more difficult it generally is to return to employment. And the smaller the pool of people actively looking for a job, the more difficult for firms to find adequate personnel.         

Fortunately, the participation rate bounced back sharply in Q3. And despite new contractions during the second and third wave end-2020 and early this year, it is now performing more in line with previous crisis episodes. But even though the gap is no longer at record levels, inactivity is still larger than before the corona pandemic, especially among the younger population. To illustrate, if the labor force were to return to its pre-crisis peak without any changes in employment, the unemployment rate would rise from the current 8.0% (April 2021) to 9.3%.

The intergenerational divide

Government support has sustained employment among all age groups, but has nonetheless been unable to prevent the young and the lower educated from being affected disproportionally (figures 8-11). This is because the hardest-hit sectors employ relatively many young and low-skilled personnel. Moreover, temporary contracts are more prevalent in hospitality and tourism. Although countries have made short-time work schemes available for temporary –and yet-to-be-hired seasonal– employees, employers have largely abstained from hiring staff they didn’t need just to put them in a short-time work scheme. On top of that, the broad-based malaise, uncertainty and containment measures have led to fewer apprenticeships and subsequent hiring of graduates. Consequently, employment rates have dropped more extensively among the young, while both unemployment and inactivity have increased more.

Figure 8: Unemployment mostly hit the young…
Figure 8: Unemployment mostly hit the young…Source: Macrobond, RaboResearch calculations
Figure 9: …and the lower educated
Figure 9: …and the lower educatedSource: Macrobond, RaboResearch calculations
Figure 10: Employment and activity dropped much more among the young…
Figure 10: Employment and activity dropped much more among the young…Source: Macrobond, RaboResearch calculations
Figure 11: …as the number of discouraged workers rose disproportionally
Figure 11: …as the number of discouraged workers rose disproportionallySource: Macrobond, RaboResearch calculations

The greater labor market detachment among the young implies a greater detachment of those who have the longest careers ahead of them. This could amplify the negative long-term impact of larger labor market inactivity. That said, a loss of skills might be less of an issue this time, since the hardest-hit sectors, which will also take longest to recover, mostly concern jobs for which the required skill set tends to be relatively limited.

The outlook

Going forward, as the recovery progresses we expect the developments of the past year to reverse: available work should increase, government support should be gradually unwound, in tandem with the recovery, and discouraged workers should return to the labor market to search for a job. The latter, in turn, should contribute to a rise in the labor force. Paradoxically, this implies that while we expect the labor market to recover, we expect unemployment to rise, albeit moderately, towards the end of the year, before it slowly resumes its decline in the course of  2022.

Available work will increase

Figure 12: Employment expectations are rising and very strong in industry compared to pre-crisis
Figure 12: Employment expectations are rising and very strong in industry compared to pre-crisisNote: The balance is derived from positive – negative expectations expressed by businesses in surveys.
Source: Macrobond, RaboResearch.

Labor market performance will improve in line with the booming activity in the manufacturing sector –as far as supply chain disruptions permit– and the reopening of the economy over the coming months (figure 12). Especially in the services sector there is a strong correlation between output and the need for personnel. Just think of restaurants. Food and drinks do not serve themselves and a waiter can only serve a limited amount of tables within a certain amount of time. If restaurants are allowed to open and/ or operate more tables, they will quickly need more waiters to maintain the level of service.


From furloughs to (un)employment

Where possible, firms will first put employees still residing in short-time work schemes back to work. In the statistics, then, this would show as a significant rise in hours worked, but a smaller increase in employment. Moreover,  while we assume that governments will gradually unwind support in tandem with the recovery, this may not prevent part of the currently furloughed workers losing their job.[5] This follows from our view that a share of jobs has only been maintained due to a ban on layoffs[6] and that a share of saved firms will go bankrupt once government support is withdrawn and insolvency procedures are reactivated. [7] [8] In statistics this would imply employment destruction. At the very least, employment growth among battered firms will likely be limited in the short- to medium term. At the same time, demand for personnel in fast-growing companies and sectors will be larger than the pool of people they have furloughed. This is underscored by the rising number of vacancies and will induce growth in both hours worked and employment – once the vacancies are fulfilled.

Inactivity should drop

Meanwhile, as the health risks are getting under control and the probability of success in finding a job improves, we think that many people that currently have an inactive status will return to the labor market[9]. We have already seen the start of that process in several member states at the start of this quarter. This is a positive development as it should help firms finding personnel needed to fill the vacancies. The aggregate pool of unemployed and discouraged workers is over 2 million people larger than pre-crisis, while people currently furloughed but turning unemployed at some point, will even add to that pool of available labor. In statistics, the return of currently inactive persons would prevent the unemployment rate from falling or even lead to higher unemployment rates as long as the re-entered job searchers have not effectively found a new job (figure 13) – again underscoring the weak ability to grasp labor market health from unemployment figures.

Figure 13: From inactive to (un)employed?
Figure 13: From inactive to (un)employed?Note: In scenario 1, the unemployment rate could both fall and increase depending on whether employment or unemployment growth would be stronger. In scenario 2 and 3, the unemployment rate would rise and fall, respectively.
Source: RaboResearch.

Labor market slack versus mismatches

Given the large amount of slack still present in the labor market combined with the fact that the vacancy rate is still lower than pre-crisis, we think that a broad-based shortage of labor is not imminent, and hence will not impede the recovery or induce large-scale wage growth. Yet the uneven recovery among sectors, with some sectors already performing far better than pre-crisis, might make it difficult for specific firms and sectors to find adequately skilled personnel.

Figures 14 and 15 underscore our view that on the macro level there is no shortage of labor, but that there are large sectoral differences. Labor, or better said the shortage thereof, is much less of an impediment to businesses in the service sector than in industry, when compared to pre-crisis. And even within the manufacturing sector there is large dispersion among different product groups. In certain sectors, such as textiles and automotive, production and capacity utilization are still substantially below pre-crisis levels, albeit due to different reasons. In other sectors, such as computer, electronics and optical products, production is already far beyond pre-crisis levels, arguably making it harder for firms to find adequately skilled personnel.

Going forward, apart from some hurdles in certain service sectors in the re-opening phase, it are likely mostly pre-crisis difficulties to find adequately skilled technical and ICT staff that will persist and even intensify given the EU’s plans to digitalize the economy.Also in the area of the energy transition, the other main pillar of the EU’s Recovery and Resilience Facility, there could well be a lack of adequately skilled workers to execute all recovery plans in the coming years.

Figure 14: No broad-based shortage of workers in the service sector so far
Figure 14: No broad-based shortage of workers  in the service sector so far Source: Macrobond, RaboResearch
Figure 15: Risk of personnel shortages rises in industry, but still far less than during 2018 peak
Figure 15: Risk of personnel shortages rises in industry, but still far less than during 2018 peakSource: Macrobond, RaboResearch

Key messages

Unemployment has increased since the outbreak of the pandemic, but to a much lesser extent than we initially expected. While part of the reason is that government support has significantly cushioned the blow of the crisis and softened employment deterioration, it is also true that unemployment figures are not the best statistic to fully grasp the size of the deterioration in employment. Overall unemployment figures mask both the wide differences among age groups, the significant drop in actual hours worked, and the significant rise in inactivity.

Going forward, we expect the labor market to recover in stages, with different forces at play. We expect the unemployment rate to rise from the current 8.0% to 8.8% towards the end of this year, before it slowly falls again through 2022. On average, we have penciled in unemployment rates of 8.5% for both this year and next.

In the very early stages of the recovery, new work will first mostly be filled in by furloughed employees – with no net effect on employment. Accordingly, in the near term, health of the labor market should rather be measured in terms of hours worked. As the recovery progresses, more employment will be created at prospering companies. Yet at the same time, a gradual winding down of fiscal support measures will endanger jobs (creation) at heavily battered businesses, while previously discouraged workers will return to the labor market. This may temporarily raise the unemployment rate – even though the latter driver is actually a positive development.

Meanwhile, specific industries may experience a shortage of labor owing to skill mismatches, but we believe that the pool of currently inactive people is sufficient to avoid more general labor shortages that could impede the recovery or drive wage growth higher. If anything, wage growth is likely to remain subdued for quite some time, given the amount of labor market slack. Plus, collective wage contracts that were signed during the pandemic have mostly resulted in moderate pay rises in the coming years.

Looking at the medium-term prospects, we expect to see smaller scarring effects from the pandemic on the labor market than during the previous crises. Unemployment and inactivity have mainly hit lower skilled workers, and the labor market will recover faster than after the global financial crisis and Eurozone debt crisis. This means that a structural loss of job skills should also be more limited.


[1] In some countries, either employers have to chip in, i.e. complement public support, (the Netherlands) or employees do not receive full salaries (Italy, Spain, Germany, France).

[2] At the start of this year, Italy has adjusted the accounting of short-time work schemes. If people reside in one of the country’s furlough schemes for more than three consecutive months, they are no longer counted as employed.

[3] Data is not easily comparable across countries. Some countries report the number of people receiving support via STWs indifferent from the number of hours; others report the number of hours for which STW support is filed.

[4] Yet, clearly, the short-time work schemes substantially mitigated the blow to households’ income from the deterioration in employment.

[5] Currently, the short-time work schemes in Germany and France are set to run until the end of the year. The schemes and ban on layoffs in Italy and Spain are set to expire by the end of June and September, respectively, but both have been extended before to prevent a cliff-edge drop in employment.

[6] It is estimated that between 0.5 and 2mn people could lose their job if the ban on layoffs in Italy is lifted per 1 July.

[7] Alternatively, governments can decide to keep non-viable firms alive with support for years to come to support employment, but this would prevent both capital and employees to move to firms where they can be more productive, being negative for overall economic activity in the medium to long run.

[8] Insolvency procedures have been put on hold by government decrees in multiple countries. The phasing out of such provisions in for example Germany since the fall have already led to a rise in insolvencies.

[9] Unlike what is happening in the US, it does not pay to sit at home outside short-time work schemes. Benefits are much less likely to reach you than certain enhanced benefits in the current US COVID support program.


naar boven