1. A tale of two crises: Recent labour market developments across the OECD

Andrea Salvatori

Russia’s war of aggression against Ukraine is first and foremost a human tragedy with many losses of innocent lives and huge economic and social consequences including the millions who fled their country to escape violence and hunger. The war has also sent shockwaves through the world economy. Europe has seen the largest and fastest growing inflow of refugees since World War II as millions left Ukraine. The economic fallout of the war is threatening the strength of the recovery from the COVID-19 crisis which had been surprisingly robust till the early months of 2022 – in many countries supported by massive recovery plans.1

The impact of the war on energy, food, and commodity markets is adding to the significant inflationary pressures that had already emerged at the end of 2021 because of supply chain disruptions. In the first half of 2022, inflation reached levels not seen in decades in many OECD countries eroding workers’ living standards as nominal wage growth remained generally modest despite the tight labour markets. The impact of inflation is felt more by the same lower-income households which have already borne the brunt of the COVID-19 crisis.

The economic fallout of the war in Ukraine threatens the strength of the economic recovery from the COVID-19 crisis. However, even before the new shock and uncertainty introduced by the war, the labour market recovery from the COVID-19 crisis appeared uneven across countries. The impact of the pandemic continues to shape employment dynamics across industries, which in turn affect the fortunes of the groups of workers that are more likely to work in them. While some of the initial unequal impact of the crisis across workers has been reabsorbed, young people, workers without tertiary education, and racial/ethnic minorities have been lagging behind in the recovery in many countries.

This chapter provides an examination of the latest developments across labour markets in the OECD and is organised as follows. Section 1.1 reviews the latest labour market developments across the OECD. Section 1.2 assesses the progress made in the recovery from the COVID-19 crisis till the first quarter of 2022, when the new crisis generated by Russia’s aggression against Ukraine. Section 1.3 reviews employment developments since the onset of the COVID-19 crisis across industries, laying the ground for Section 1.4 to review the progress made by different socio-economic groups during the recovery. Finally, Section 1.5 describes the labour market experience during the COVID-19 crisis of frontline workers.

The economic recovery from the COVID-19 crisis has been faster than expected thanks to the prompt and massive policy support for firms and household deployed throughout the crisis and the rapid rollout of effective vaccines (OECD, 2021[1]). OECD output returned to pre-crisis levels already in Q3 2021 and continued to grow – albeit at slower pace – into the second quarter of 2022, climbing to 3.4 percentage points above its Q4 2019 level. The economic disruptions from the wave of the pandemic driven by the Omicron variant in late 2021 and the early months of 2022 generally proved mild in most countries, despite some weakness in the United States and Japan, where GDP declined in Q1 2022, and the Euro area, where growth slowed. Preliminary data for Q2 2022 suggests that GDP grew in the Euro Area, Mexico and Japan but contracted slightly in the United States – with positive growth recorded for the OECD as a whole.

The recovery in GDP was uneven across OECD countries (Figure 1.1). In Q1 2022, GDP remained below pre-pandemic levels in eight countries – with output in Iceland, Spain and Mexico more than 1 percentage points below the Q4 2019 reference level. By contrast, GDP was at least 2.5 percentage points above pre-pandemic levels in 22 countries, with particularly large gains in Ireland, Chile, Colombia, Türkiye, Israel, and Poland.

As the economy recovered, total employment in the OECD returned to pre-crisis levels at the end of 2021 and continued its growth – albeit at a slower pace – into the first half of 2022, reaching a level 1.3% higher than before the crisis in July 2022 (Figure 1.2). Employment growth was particularly strong in Australia – where in July 2022 employment was 4.6% higher than at the end of 2019 – and Mexico – where in July 2022 employment was about 4.5% above its pre-crisis level. Employment recovery was less pronounced in Japan – where employment was 1% lower than pre-crisis in July 2022 - and in the United States – where employment reached pre-crisis levels in August 2022. In the Euro Area, employment growth slowed down in the spring of 2022, and total employment reached a level around 2.3% higher than before the crisis in July 2022.

The OECD unemployment rate gradually fell from its peak of 8.8% in April 2020 and stabilised in the first months of 2022. In July 2022, the OECD unemployment rate stood at 4.9%, slightly below the 5.3% value recorded in December 2019 (Figure 1.3). In July 2022, unemployment was below pre-crisis levels in 24 countries, and above that level by more than 0.5 percentage points only in Finland and Estonia. The peak increase in unemployment rate differed substantially across countries: unemployment increased by a larger amount and more quickly in countries that made limited use of job retentions schemes such as the United States, Colombia, Costa Rica and Chile. However, by early 2022 the unemployment rate had returned close to its pre-crisis levels in all countries (Figure 1.4).2 The reliance on unemployment compensation does not necessarily imply that workers in those countries were worse off compared to workers in countries with job retentions schemes. For example, the United States significantly boosted and expanded the cash support and eligibility criteria during the first year and a half of the pandemic.

The Russian invasion of Ukraine has generated a humanitarian crisis affecting millions of people and caused a new set of adverse economic shocks.3 Commodity prices have risen substantially, reflecting the importance of supply from Russia and Ukraine in many markets, adding to inflationary pressures and hitting real incomes, particularly for the most vulnerable households. Supply-side pressures have also intensified as a result of the conflict, as well as the impact of continued shutdowns in major cities and ports in China due to the zero-COVID policy.

More than 6.5 million people have already been forced to flee Ukraine to other countries in Europe, and an even greater number have been displaced within the country.4 The number of people who have already fled Ukraine since the start of the war is several times greater than the annual flow of asylum-seekers into Europe at the height of the Syrian refugee crisis in 2015-16. The refugee flows caused by the war will result in additional public expenditure in the short-term in host countries, although this will be offset over time as refugees enter the labour force. Box 1.1 reviews lessons learned from recent experiences from across the OECD that can help facilitate the process of integrating the refugees into the labour market of the host countries.

The OECD economic projections from June 2022 point to a slowdown in global GDP growth as a results of the economic fallout from Russia’s aggression against Ukraine. Indeed, global GDP growth is now projected to be around 3.0% in 2022 – against the previous projection of 4.5% from December 2019 – and to remain at a similar pace in 2023 (OECD, 2022[4]).

The normalisation of labour markets is projected to continue during 2022-23, despite the new negative shock of the war in Ukraine, which nevertheless makes the outlook more uncertain (OECD, 2022[4]). As the public health situation improves further, based on rising vaccination rates and improved COVID-19 treatments, labour force participation is projected to increase in almost all economies. Across the OECD, as seen in Figure 1.2, total employment returned to its pre-crisis levels already at the end of 2021, but its growth is now expected to slow. In particular, total employment in the OECD is projected to be above its Q4 2019 level by 1.5 percentage points by the end of 2022 and by 2.5 percentage points by the end of 2023. The unemployment rate is expected to stabilise remaining just above 5% both at the end of 2022 and 2023 (Figure 1.5).

There are a number of prominent downside risks that could lead to a further deterioration of the economic situation with potential repercussions on labour markets. These risks are linked in particular to an abrupt interruption of flows of oil and gas from Russia to Europe, stronger disruptions to global supply chains or financial contagion. Inflationary pressures could also prove stronger than expected, with risks that inflation expectations move up further away from central bank objectives and become reflected in faster wage growth amidst tight labour markets. Sharp changes in policy interest rates could also slow growth by more than projected. Risks also remain from the evolution of the COVID-19 pandemic: new and more aggressive or contagious variants may emerge, while the application of zero-COVID-19 policies in large economies like China has the potential to sap global demand and disrupt supply for some time to come.

The labour market indicators for the first quarter of 2022 – which were only marginally affected by the consequences of the Russia’s invasion of Ukraine – show that the labour market recovery from the COVID-19 crisis was generally stronger than expected, but some countries were lagging behind.

At the beginning of 2022, total hours worked remained below pre-crisis levels in many countries. On average across the OECD countries with data available, hours were 0.2% lower in Q1 2022 compared to in Q1 2019 (Figure 1.6).5 The recovery in total hours worked was slowed down or even set back in some countries as new restrictions were adopted in the final quarter of 2021 as the Omicron variant drove a new aggressive wave of the pandemic. In early 2022, total hours worked remained below pre-crisis levels in 19 of the 35 countries with data available. In Finland, Japan, Estonia, the Check Republic, the Slovak Republic, and Iceland the gap was particularly large, exceeding 5%.

Employment and inactivity rates in early 2022 had generally improved relative to pre-crisis levels, but some countries were still lagging behind (Figure 1.7). According to the most recent available data (Q1 2022), the employment rate of the working age population was above pre-pandemic level in 28 of the 38 OECD countries by an average of 1.5 percentage points. In the remaining ten, the employment rate was below its Q4 2019 level by an average of 1.6 percentage points, with the gap exceeding 2 percentage points in Colombia, Costa Rica, Chile, and Latvia.

The initial increase in inactivity that took place in all countries in 2020, as the pandemic discouraged active job search (OECD, 2021[1]), had largely been reabsorbed by early 2022. In the most recent data, inactivity rates were lower than just before the crisis by an average of 1.3 percentage points in 27 OECD countries. In the other 11 countries, inactivity was above pre-crisis levels by an average of 1.2 percentage points with the largest increases in excess of 2 percentage points in Colombia, Costa Rica, and Chile.

At the onset of the crisis, long-term unemployment (i.e. 12 months or more) edged down in several countries (OECD, 2021[5]). This was largely the result of a fall in job search activity in the context of the initial lockdowns that were often accompanied by the suspension of job search obligations, leading to many people being classified as inactive rather than unemployed. Over the course of 2021, however, as job search picked up again, long-term unemployment increased in many countries despite the general improvement in labour market conditions. By Q1 2022, long-term unemployment was still above pre-crisis levels but generally receding in most countries (Figure 1.8).6 In particular, the long-term unemployment rate was above pre-crisis levels in 20 of the 32 countries with data available, but the OECD average had already returned to pre-crisis levels. The increases were above 50% in the United Sates (from 0.5% to 0.7%) and Canada (from 0.5% to 0.8%) – both countries that featured comparatively low levels of long-term unemployment before the start of the crisis.7 Declines in excess of 15% in the long-term unemployment rates were recorded in Greece, South Korea, Latvia, Australia, and Denmark.

The unprecedented rebound of economic activity recorded in many countries in 2021 was coupled with a surge in labour demand, as indicated by the steep increase in labour vacancies in many countries (Figure 1.9). Indeed, in most countries considered, vacancies reached pre-crisis levels already one year after the on-set of the crisis in Q2 2021 and then continued to increase steadily for the remainder of the year. In the first quarter of 2022, the growth of vacancies generally slowed down, but they remained at historically high levels in many countries. By Q1 2022, vacancies were at least 50% higher than before the crisis in Australia, Austria, Sweden, the United Kingdom, and the United States. Vacancies increased relatively less in Germany and Poland, still reaching a level just under 20% higher than before the crisis by Q1 2022. Among the countries not included in Figure 1.9, vacancies reached record highs in in Canada (80% higher in Q4 2021 than in Q4 2019)8 and New Zealand (+31% in March 2022 relative to two years earlier).9 In Italy, the vacancy rate reached record levels in the second half of 2021, stabilising around 1.9 in Q1 2022 (ISTAT, 2022[6]).10 Also in Q1 2022, vacancies were at least 40% higher than before the crisis in Luxembourg and Portugal, and only slightly above pre-crisis levels in Hungary and the Czech Republic.11 Data for Q2 2022 are only available for a few countries at the time of writing, but generally confirm that vacancies remained high throughout the first half of the year. By contrast, two years after the start of the Great Financial Crisis, vacancies remained depressed in all countries reported in Figure 1.9 – highlighting the profound difference in the nature of the two crises.

Two main factors have likely contributed to the widespread surge in vacancies. First, vacancies rebounded after two or three quarters of unprecedented depression in 2020 when turnover in firms had slowed down considerably due to the health situation. As economies reopened and uncertainty surrounding the economic and health situation decreased over the course of 2021, firms and workers likely pursued (and continue to pursue) hiring and job-moving decisions that had been placed on hold. In countries that made limited use of job retention schemes to preserve jobs – like the United States – the rebound was particularly robust due to the need to re-fill temporarily closed positions after the various waves of the pandemic.

A second factor fuelling the surge in vacancies is the strong growth in product and service demand of the second half of 2021 and early months of 2022. The generous support deployed by many countries during the crisis helped keep many firms in operation and preserve the spending power of many consumers, thus creating the conditions for a jump-start of the economy as restrictions became progressively more targeted and vaccination rates quickly increased. The strong economic recovery was then fuelled by massive recovery plans in many countries. In addition, demand was also supported by the savings accumulated by many consumers in the first part of the crisis as they reduced spending on services in particular due to the restrictions in place or out of fear of contagion (McGregor, Suphaphiphat and Toscani, 2022[7]).

As already seen in Figure 1.3, unemployment rates fell throughout 2021, but the speed of the decline did not match that of the surge in vacancies. Indeed, while vacancies were well above pre-crisis levels by early 2022, unemployment was instead close to pre-crisis levels in all countries. While vacancies do typically grow faster than unemployment falls during recoveries, the unprecedented speed of the vacancy surge during the COVID-19 recovery means that labour market tightness increased in most countries to levels typically seen much later in the cycle (European Central Bank, 2019[8]). Also, many of the Beveridge curves reported in Figure 1.9 – which capture the negative relationship between unemployment and vacancies – exhibit a pronounced outward shift in the second half of 2021, signalling a decrease in the matching efficiency of labour markets. Two notable exceptions are France and Germany where the increase in vacancies has been less pronounced and unemployment fell below pre-pandemic levels at the start of 2022.

The increase in tightness in the labour market and the decrease in matching efficiency is clearly reflected in the growth in the number of firms reporting production constraints from labour shortages (Figure 1.10). In Q2 2022, the proportion of firms in manufacturing that lamented labour shortages was, on average, 8.5 percentage points higher (at about 26%) than before the crisis in the 22 OECD countries that are members of the European Union and Türkiye. In services, the proportion of firms reporting labour shortages was 27.5% on average across the same countries – or more than 11 percentage points higher than before the crisis. Among these countries, reports of labour shortages did not increase only in Hungary, the Czech Republic, the Slovak Republic (in manufacturing) and Türkiye (in services). The proportion of businesses reporting labour as the primary constraint was also at a record high in New Zealand in January 2022.12 In Canada, in the first quarter of 2022, 37% of firms expected to face labour shortages in the coming three months.13 An economy-wide indicator of labour shortages in Germany compiled by the Institute of Employment Research (IAB) grew above pre-crisis levels in early 2022, after rebounding from the low levels of 2020 and early 2021.14

EU-level data by finer sectors indicate that recruiting difficulties have been widespread across sectors in recent months, but they are particularly pronounced in relatively low-pay sectors (Figure 1.11). For example, the share of firms reporting production constraints from labour shortages increased by 13 percentage points relative to its pre-crisis level of 20% in accommodation and food services and by 12 percentage points (relative to a pre-crisis level of 23%) in administrative and support services. Firms in accommodation and food services have also been more likely to report labour shortages in the first few months of 2022 in the United Kingdom as well (37% vs an average of 14% in April 2022).15 In Canada, the proportion of firms expecting labour shortages in the first quarter of 2022 was 65% in the accommodation and food services vs an average of 37% across the economy.16

In the United States, after hovering below pre-pandemic levels for over a year, quits reached record highs in the second half of 2021 and then remained high in the first few months of 2022, prompting talk of a “Great Resignation”.17 Increases in quits were recorded in almost all sectors, but – relative to the size of the sectors – they were particularly pronounced in manufacturing, retail trade and finance and insurance.18 The evidence on which workers have been quitting varies somewhat depending on the methodology and timing of the survey. A survey covering 4 000 US companies in the summer of 2021 suggested that quits increased more among prime-age workers (Cook, 2021[9]). A recent survey by the Pew Research Center (Parker and Horowitz, 2022[10]) found that workers under the age of 29 were more likely than all other age groups to have quit their job at some point in 2021, but the study does not provide pre-crisis baseline figures to assess which groups saw the largest increases. According to this survey, men and women were equally likely to have quit their jobs in 2021, but quits were more frequent among racial/ethnic minorities groups.

There is no indication that the increase in quits is driven by people falling out of the labour force. Indeed, the employment-to-population ratio in the United States continued its steady growth in the first quarter of 2022 even as quit rates remain elevated and GDP growth turned negative (see Section 1.1).19 In addition, at the end of 2021 hiring rates were higher than quit rates in all industries, including in low-pay services (Gould, 2022[11]). This suggests high mobility within sectors in a tight labour market, rather than significant outflows from specific industries because of a change in workers’ preferences. A survey by the Pew Research Center finds that the vast majority of those who quit their job in 2021 report having found a new job without significant difficulties and with similar or better conditions than their previous employment (Parker and Horowitz, 2022[10]).

Beyond the United States, the evidence of a significant increase in quits is limited. In the United Kingdom, job-to-job transitions remained below pre-pandemic levels until the summer of 2021 and then reached a record high in Q4, at a level about 30% higher than in Q4 2019 driven by an increase in resignations. In Q1 2022, job-to-job moves declined slightly, while still remaining over 20% higher than in the same quarter of 2019.20 However, there was no indication of an increase in cross-sectoral mobility that would be expected if the pandemic had motivated workers to leave certain sectors in particular.21 In France, after a long depression, quits of permanent workers climbed above pre-crisis levels in the third quarter of 2021 – driven by an increase among workers formerly on job retention schemes – and then remained elevated in the last quarter of the year.22,23 In Germany, however, there was no indication of an increase in quits relative to before the crisis at least until March 2021 (Rottger and Weber, 2021[12]). Similarly, in Australia, the proportion of businesses with open vacancies reporting the need to replace leaving employees was stable over the course of 2021. By February 2022, the figure stood at 79.7% – only 1 percentage point higher than just before the pandemic in February 2020.24

The increasing labour market tightness seen in many countries is likely mostly the result of the sheer speed of the surge in labour demand fuelled by the strong uptake in economic activity as economies reopened. The pervasiveness of reports of labour shortages across countries and sectors suggests that the current situation is not driven by the scarcity of a specific type of labour that could arise, for example, from the asymmetric impact of the crisis across sectors (see Section 1.3). In fact, recent studies have found that the mismatch between types of workers and the types of jobs available grew substantially at the onset of COVID-19 crisis but was short-lived and generally smaller than during the Great Financial Crisis (Shibata and Pizzinelli, 2022[13]; Duval et al., 2022[14]). Instead, these studies suggests that the sluggish response of employment to the surge in vacancies in the second half of 2021 was in part explained by a contraction in labour supply of low skilled and older workers. Indeed, in the United States and the United Kingdom, the vacancy surge occurred even as inactivity rates remained above pre-crisis levels. Another potential factor limiting the availability of labour overall might have been the protracted weakness of net migration in many countries. Preliminary evidence suggests that in Q3 2021 the overall size of the labour force in Europe was still below the levels that would have been expected given pre-crisis trends largely due to the fact that net migration remained depressed (European Central Bank, 2022[15]).

The tightening of the labour market per se might stimulates job-to-job moves – as evidenced by the uptake in quits in some countries – and might encourage jobseekers to search for longer for better opportunities. The generous income support provided by many countries during the crisis might have helped jobseekers to prolong their search for better opportunities – though the evidence from the United States point to mostly small effects (Holzer, Hubbard and Strain, 2021[16]; Coombs et al., 2022[17]; Petrosky-Nadeau and Valletta, 2021[18]). The lingering pandemic might have made frontline low-paid jobs that typically involve direct contact with customers less appealing and might have accentuated the perception of the lower quality of these jobs. Pizzinelli and Shibata (2022[13]) argue that an increase hesitancy to return to these jobs might play a role in the United States and the United Kingdom.

In many sectors – both high and low skill – however the current exceptional circumstances exacerbate pre-existing difficulties in recruiting workers. In their responses to an OECD questionnaire (see Chapter 2), over 70% of the countries reported that labour shortages were an issue in the long-term care and health sectors during the COVID-19 crisis – with most indicating that the crisis has aggravated existing problems. Across Europe, reports of labour shortages had been steadily increasing in the aftermath of the financial crisis (Eurofound, 2021[19]). The Beveridge Curve of several countries had gradually shifted outwards after the Great Financial Crisis, signalling increasing difficulties in matching a large number of vacancies to a large number of unemployed because of skill mismatches or unsatisfactory working conditions (European Central Bank, 2019[8]). As the pandemic broke out in 2020, labour shortages were quickly aggravated in agriculture and in the health and ICT sectors in Europe (Eurofound, 2021[19]; Samek Lodovici et al., 2022[20]).

The coming months will help clarify if underneath the vacancy tide affecting all industries – new tensions are arising (or adding to pre-existing ones) in specific industries linked to qualitative mismatches between labour demand and supply. As discussed below in Section 1.3, industries that have expanded since the onset of the crisis are very different from industries that have seen employment fall, pointing to a potential misalignment in skills between labour demand and the supply that has become available. Geographical mismatches could also be an issue if expanding and retreating sectors are located in different places and as result of changing consumption patterns (for example due to more online spending or to increases in teleworking that shifted some consumption away from urban centres). There is currently very little evidence of mismatches arising in the aftermath of the COVID-19 crisis. Preliminary evidence based on data for Australia, Spain, the United Kingdom, the United States, Canada and Japan suggests that the problem is limited due to the fast rebound of the most-hardly hit sectors (Duval et al., 2022[14]). Finally, in addition to the pressures arising from changes that might have been triggered or accelerated by the pandemic per se, many countries intend to use their recovery plans to accelerate digitalisation and the transition towards a climate-neutral economy – further accelerating structural transformations of the labour market which might also contribute to rising qualitative mismatches.

Despite the strong labour markets, workers’ wages have declined in real terms in recent months. Indeed, while by the end of 2021 or early 2022 nominal wage growth reached high levels relative to pre-pandemic levels in some countries, the nominal increases have generally remained well below the fast-growing inflation generated by increasing commodity prices (Figure 1.12).

In the United States, nominal wage growth edged up already in the second half of 2021. Even so, real wages fell. Indeed, in the last quarter of 2021, nominal wage growth in the private sector reached 5% – about 2 percentage points higher than in the quarters just before the crisis – while inflation jumped to 6.7%. In the first quarter of 2022, annual nominal wage growth remained stable but inflation reached 8%. Nominal wage growth was particularly strong in leisure and hospitality, reaching 9% in Q1 2022 – in part as a result of increases in minimum wages implemented in a number of states and localities (Box 1.2) – while in the quarters before the start of the pandemic it had hovered around 4% (Figure 1.13).25

In Europe, the ECB index for negotiated wages in the Euro Area picked up slightly in the first quarter of 2022 (+2.8%) but remained well below the rate of inflation of 6.1%. In France, nominal gross hourly wages for non-managerial employees grew by 1.9% in Q4 2021 and 2.5% in Q1 2022, outpaced by inflation rates of 2.7% and 3.7% respectively. In Q1 2022, nominal wage growth was above average but still below inflation in two low-pay industries, retail and food and accommodation.26 In Canada, nominal hourly wage growth remained below pre-pandemic levels for most of 2021 and reached 3% in the first quarter of 2022, remaining well below inflation at 5.8%. In the United Kingdom, growth in nominal average weekly earnings was below inflation both in Q4 2021 and Q1 2022 – but measures of pay including bonuses increased more in line with inflation. Data by sectors for the United Kingdom show similar patterns in wage changes between low-pay service sectors and the whole private sector until the end of 2021, but larger wage growth in low-pay sectors in the first months of 2022 (Figure 1.13).27 In Japan, the annual growth rate of total cash earnings was slightly below inflation in Q4 2019, but reached 1.1% in Q1 2022 against an inflation rate of 0.9%.

A tight labour market might help improve working conditions in low-pay sectors. Indeed, as mentioned above, there is some evidence that nominal wage growth has been stronger in some low-pay sectors (see Figure 1.13) and Duval et al. (2022[14]) find that wages in low-pay sectors were more responsive to the increasing labour market tightness over the course of 2021. More generally, tight labour markets are associated with improvement in labour market outcomes for vulnerable groups in particular – both in terms of better working conditions for those employed and higher participation to the labour market (Bergman, Matsa and Weber, 2022[21]; Aaronson, Barnes and Edelberg, 2022[22]). In addition, tight labour markets increase opportunities for labour reallocation across firms with a potential beneficial effect for productivity.

Improving working conditions for the most disadvantage groups need not generate significant widespread inflationary pressures (especially in markets where monopsony power is significant – see Chapter 3). Duval et al. (2022[14]) argue that the overall impact on economy-wide wage pressure of rising tightness among low-pay industries in 2021 was limited due to the overall small share of such industries in total labour costs (in the United Kingdom and the United States). Inflationary pressures could arise from the combination of persistent labour shortages across sectors and the high or rising inflation driven by increases in energy and food prices. Faced with increasing wage demands, firms that have seen their profits increase over the pandemic due to an expected increase in demand might be able to accommodate them without significant price increases. However, firms whose profits have instead been eroded by the pandemic or by the increase in the cost of other inputs might not have much room for increasing wages without driving prices up.

OECD (2022[4]) expects real wages to continue to decline over the course of 2022, as inflation is projected to remain elevated. Indeed, the war in Ukraine has already pushed inflation well above the level expected at the time of collective bargaining to set wage rates for 2022. In addition, nominal wage pressures are likely to ease as international migration picks ups and refugees are absorbed into the labour market of the host countries. For the OECD as a whole the pace of wage increases in nominal terms is projected to decline from around 4.25% in 2022 to 3.5% in 2023 (OECD, 2022[4]). In real terms, wage growth over 2022-23 is projected to be negative in most countries (Figure 1.14).

The impact of rising inflation on real incomes is larger for lower-income households which have already borne the brunt of the COVID-19 crisis. Indeed, the increase in expenditure resulting from recent food and energy price changes represents a larger proportion of total spending for lower-income households, and those households have limited scope to offset this by drawing on savings or reducing discretionary expenditures (OECD, 2022[4]). These households disproportionally include low-pay workers who were more likely to have their income reduced during the COVID-19 crisis either through job loss or a reduction in hours worked (OECD, 2021[5]).

Beyond their role in facilitating collective bargaining, governments have a range of complementary policy tools available to cushion the impact of inflation on low-income households. Available evidence suggests that governments have acted swiftly through temporary energy bonuses and the tax and benefit system, although often with costly, untargeted interventions (see Chapter 2 for a discussion of recent interventions by OECD governments). Statutory minimum wages have also been adjusted in many countries, although they tend to continue to lag behind inflation (Box 1.2).

The markedly asymmetric impact across sectors is a distinctive feature of this crisis that is well documented (OECD, 2021[1]). Industries where telework was not feasible – such as accommodation and food services, arts, and transportation and storage – saw large reduction in hours and employment losses across countries. By contrast, other service industries such as information and communication, as well as financial and insurance activities, saw an increase in activity already over the course of 2020. As the pandemic protracted into 2021, industries with limited teleworking possibilities continued to be affected disproportionally by more targeted restrictions and persistent changes in consumer’s habits even as the overall economic impact of each successive wave became smaller. In the vast majority of countries that made significant use of job retention schemes, the initial impact of the crisis was largely absorbed through reduction in hours, but, as the crisis lingered on, the burden of adjustment moved to the extensive margin, with many on short hours returning to work while jobs destroyed were not fully recovered (OECD, 2021[1]).

The deeply asymmetrical impact across industries and the substantive changes in consumption patterns and in the organisation of work that it prompted raise the concrete possibility that the crisis might lead to some structural and persistent changes in the distribution of employment across firms and sectors. The current phase of rapid developments in the labour markets documented in Section 1.1 makes it difficult to distinguish persistent structural changes from temporary distortions that might subside once the labour market returns to a more ordinary state. Nevertheless, monitoring trends in employment across industries is crucial to highlight possible forthcoming tensions between labour demand and supply. Importantly, the differential impact of the crisis and recovery on different industries remain a significant driver of the impact of the crisis across different groups of workers, as Section 1.4 documents.

To document how different industries and groups of workers have fared in the recovery from the COVID-19 crisis, this section and the next use data from Q1 2022, the most recent data point available for the largest number of OECD countries. Since seasonally adjusted data are not readily available for the outcomes of interest at a disaggregate level, these sections use unadjusted data and take Q1 2019 as the pre-crisis reference point. Checks performed with data on overall employment indicate that results based on seasonally adjusted data for Q4 2019 vs Q1 2022 are consistent with those based on unadjusted data for Q1 2019 vs Q1 2022.

For the countries covered by Eurostat, all the employment series are affected by a statistical break in Q1 2021 (see Eurostat (2022[28])). Whenever available, break-adjusted series provided by Eurostat are used in the analysis. In the other cases, a correction described in Annex 1.B has been applied.

Two years since the onset of the crisis, employment changes by industry across OECD countries are still very clearly shaped by the pandemic (Figure 1.16). Relative to the same quarter of 2019, in Q1 2022, lower-pay industries exhibited employment losses or modest growth, while higher-pay service industries reported larger employment gains. Construction and Manufacturing – two sectors that employ many medium earners – also recorded employment losses. Employment also increased in Health and Education – two medium pay sectors that have been heavily affected by the pandemic.

In order to offer a manageable overview of employment changes by industry across countries given these aggregate results, Figure 1.17 presents results for selected industries aggregated in four broad groups: low-pay service industries (Accommodation and Food Service Activities, Administrative and Support Service Activities, Arts, Entertainment and Recreation, Wholesale and Retail Trade, and Transportation and Storage), Health and Education, Manufacturing and Construction, and high-pay service industries (Professional, Scientific and Technical Activities, Information and Communication, and Financial and Insurance Activities).

Employment gains in high-pay service industries and losses in low-pay services were widespread across countries (Figure 1.17). Indeed, high-pay service industries gained employment in 31 of the 33 countries for which data are available, with particularly large changes in the Netherlands, Hungary and Lithuania. Employment in low-pay service industries was below pre-pandemic levels in 21 countries, with the largest falls seen in the Slovak Republic, Switzerland, and Latvia. The loss of employment in manufacturing and construction was also widespread (22 countries) and particularly large in Switzerland,28 Luxembourg, Slovenia, and the Slovak Republic.

Given the lack of timely and internationally comparable data on workers’ transitions, there is no simple way to assess the extent to which these differences in employment performances across sectors are associated with significant reallocation of workers across industries (possibly through unemployment spells).29 The few studies that have looked into cross-industry transitions for specific countries report mixed results. Rottger and Weber (2021[12]) find an increase in transitions to other industries for workers who had lost employment in accommodation and food services in Germany towards the end of 2020, but not at the time of the first lockdown in the spring of the same year. In April 2021, Aaronson et al. (2021[29]) found no increase in the United States – a country that relied on temporary layoffs rather than job retention schemes – in the probability that unemployed workers move to new industries, nor an indication of an increase in direct flows from heavily impacted industries towards healthier ones. Similarly, in the United Kingdom – where a new job retention scheme was used massively to preserve jobs (OECD, 2021[1]) – Brewer et al. (2021[30]) found that even as job-to-job transitions reached a record high in Q3 2021, the fraction of such transitions occurring across industries was actually the lowest since the early 2000s. They also found no increase in the share of workers who had changed industry within a given year (including through intermediate unemployment spells) which had remained stable at around 5% since 2014. Basso et al. (2021[31]) use data from before the pandemic from Italy to highlight that, because of their skill profile, workers in the hardest-hit sectors have little reallocation potential if demand for in-person services remains depressed. In France, thanks to the massive use of the country’s job retention scheme, the number of workers leaving the accommodation and food services between the months of February 2020 and 2021 increased only marginally relative to the years before (DARES, 2021[32]).

The limited evidence of cross-sectoral transitions highlights the risk of growing mismatches in the labour market if the differential employment performance across industries persists. The growth in long-term unemployment might be a symptom of these developments (Section 1.2.1), but there are also indications of a particularly strong growth in labour demand in recent times in industries that have been lagging behind, at least in some countries (Section 1.2.2). While this strong growth might have been somewhat tamed by the Omicron wave affecting many OECD countries in late 2021 and early 2022, the broad trends suggest that these industries might recover some of the lost ground if the general epidemic and economic situation continues to progress towards increasing normalisation. Indeed, as discussed in Section 1.2.2, labour supply – rather than structural changes in labour demand – is likely to have slowed down the recovery of these industries in recent times. Aaronson et al. (2021[29]) observe that much of the disequilibrium in the United States labour market is driven by the severe impact of the crisis on accommodation and food services, expressing scepticism that the crisis might permanently set back a sector that has steadily grown over the past 70 years.

In addition to the possible reallocation of employment across industries, the pandemic might also have seen reallocation of employment within industries towards firms better equipped to withstand the pandemic shock. Indeed, there is some evidence of employment reallocation among small businesses towards high-productivity and tech-savvy firms despite the deployment of new job retention schemes in Australia, New Zealand and the United Kingdom (Andrews, Charlton and Moore, 2021[33]). This type of reallocation – especially when occurring on a large scale over a short period of time – can also present challenges for workers if the type of labour demanded by expanding firms is different from that normally employed in the same industry. In this context, a concern is that labour demand might have shifted towards more highly skilled workers who might be better equipped to deal with the new changes in the workplaces. Again, timely and internationally comparable evidence on this is hard to come by. A first tentative exploration of the data available on the education level of new hires across countries reveal no clear increase in the share of workers with higher education hired in different industries compared to the years immediately before the COVID-19 pandemic. Nevertheless, changes might take more time to appear clearly in aggregate data, or they might affect workers with different skills within the same educational groups. Monitoring the evolution of the demand for different types of skills is an important task for future research that can help inform policies aimed at supporting workers that stand to lose from these potential transformations.

The highly sectoral nature of the crisis has meant that some groups of workers shouldered the bulk of the burden when the crisis broke out. OECD (2021[1]) documented how low-paid workers, those with lower education and young people paid a high and more persistent price during the crisis over the course of 2020. As the pandemic continued to shape employment dynamics across industries in 2021, different groups of workers have benefitted to different extents from the unexpectedly robust recovery described in Section 1.2.30

Young people were particularly affected by the initial ravages of the crisis (OECD, 2021[1]). Youth unemployment in the OECD surged at the onset of the pandemic, and hours worked by young people fell by more than 26% – close to double the fall seen among prime-aged and older workers (15%).

At the start of 2022, on average across the OECD, young people had recovered much of the lost ground, but were still lagging behind older adults. Indeed, on average across the OECD the youth employment rate was 0.1 percentage points above its pre-crisis levels (as measured by employment rates in Q1 2019), but remained below that level in over half the countries by an average of 2.2 percentage points (Panel A of Figure 1.18). By contrast, the employment rate for workers aged 25 to 54 years was on average 1 percentage points higher than before the crisis and still recovering only in eight countries. Among those aged 55 to 64, the employment rate was 3 percentage points higher than before the crisis and lagging behind only in five countries.

In the countries where the employment rate of young people was still below pre-crisis levels, this was mostly associated with an increase in inactivity rather than unemployment. Declines in the employment rate of young people exceeded 2 percentage points in nine countries, and exceeded 4 percentage points in Portugal, Iceland, and the Slovak Republic. In the 15 countries where youth employment grew above pre-crisis levels, this mostly resulted in a decline in inactivity. Employment rates were above pre-crisis levels by 3.5 percentage points or more in France, New Zealand, Australia, Norway, and Ireland.

The large declines in youth employment are mostly accounted for by losses of employment in low-pay service sectors and to a lesser extent in manufacturing and construction (Panel B of Figure 1.18). While results vary across the 15 countries where the employment of young people increased, on average the broad industry groups that contributed the most to these gains were health and education, low-pay services and high-pay services.

The share of young people not in employment, education or training (NEET) in Q1 2022 was below its Q1 2019 level by 0.2 percentage points on average (Figure 1.19), having re-absorbed the increase seen at the beginning of the crisis to return to historically low levels (OECD, 2021[1]). This average across 29 countries, however, conceals large cross-country differences and results from declines in 18 countries and increases in 11. Increases in excess of 1.5 percentage points were recorded in the Slovak Republic, Lithuania, Estonia, Slovenia and the Czech Republic.

The declines in NEET rates in 19 countries are in contrast with the increases seen at the start of the crisis – driven by the sudden large drop in job search – but are consistent with evidence of increasing engagement in education during periods of labour market difficulties (Carcillo et al., 2015[34]). Indeed, for some countries – like Spain, Portugal, Belgium and the United Kingdom – the differences between changes in overall inactivity (Figure 1.18) and those in NEET rates (Figure 1.19) suggest that the increase in overall inactivity is explained by an increase in the number of young people in education.

The continuing disadvantage of young workers in some countries is particularly concerning in light of the large body of evidence pointing to particularly significant scarring effects for them. Even in many of the countries where employment has picked up, young people are more likely than older workers to have experienced periods of joblessness over the course of 2020. Studies have found large and persistent reductions in earnings for young people who enter the labour market in a typical recession in the United States, Canada, and Australia (Altonji, Kahn and Speer, 2016[35]; Oreopoulos, von Wachter and Heisz, 2012[36]; Andrews et al., 2020[37]). The scarring effects are particularly significant for lower skilled youth (Kroft, Lange and Notowidigdo, 2013[38]; Altonji, Kahn and Speer, 2016[35]) and might extend to health and well-being (Garrouste and Godard, 2016[39]). High-skill workers might take up lower-skill jobs during a downturn – which might lead to skill depreciation and negatively impact their ability to move to higher skill jobs over time. However, evidence from France suggests that the negative effect of entering the labour market during a recession is short-lived, pointing to a potential important role of labour market institutions – and in particular of the minimum wage (Gaini, Leduc and Vicard, 2013[40]). An important focus of future analysis will be to monitor the evolution of the quality of the jobs held by the young workers who have been affected by the COVID-19 shock.

One possible concern is a further increase in the incidence of temporary contracts among young people from already high levels (34% in Q1 2019 across the 30 countries in Figure 1.20), as many firms deal with the protracted uncertainty surrounding the health and economic situation and young people struggle to find alternative options. However, the share of young people on temporary contracts was similar in Q1 2022 and Q1 2019 on average across the 30 countries with available data (Figure 1.20). This represented a rebound as the incidence of temporary contracts declined at the beginning of the crisis when workers on such contracts were more likely to lose their job (OECD, 2021[1]).31 There is no indication that employment growth for young people was linked to an expansion of temporary contracts, as the correlation between the changes in the two indicators was very weak across countries.

The pandemic has been particularly disruptive for young people well beyond its immediate labour market impact. Although international evidence is still developing, there is some indication that the pandemic had negative effects on learning outcomes of children in schools (Thorn and Vincent-Lancrin, 2021[41]) and particularly so for children from more disadvantaged backgrounds, at least in the United States (Dorn et al., 2021[42]). These facts may have longer-term implications for the labour market outcomes of the young people affected. Many work-based learning and apprenticeship opportunities, which can help smooth school-to-work transitions, have been disrupted, while many young people are experiencing financial insecurity, housing instability and mental health issues. Among the youth, those bearing the brunt of the crisis are those who were already facing difficult circumstances prior to the pandemic (OECD, 2021[43]).

As the risks linked to COVID-19 grow with age, the pandemic brought the concern that older adults might choose to leave work earlier in large numbers. Two years since the on-set of the pandemic, this prediction has not come to pass as employment rates for the 55-to-64 and 65-to74 age groups are back to or above pre-pandemic levels in most countries (Annex Figure 1.A.3). In particular, for the age group 55-to-64, the share in employment was up 3 percentage points in Q1 2022 relative to the same quarter of 2019 on average across the OECD. Employment rates were above or only slightly below pre-crisis levels in 29 of the 34 countries with available data. For the 65-to-74 age group, the employment rate was 0.1 percentage points above its pre-pandemic level on average across countries, and above that level in 26 of 34 countries. Exceptions to these trends included Chile, Mexico, the United States and the United Kingdom (for the 55-64 age group only) – where employment rates for these age groups were still below pre-crisis levels in Q1 2022. Data for these countries show that the lower employment rate was mostly associated with an increase in inactivity rather unemployment. Inactivity figures for these older groups are not readily available for the majority of the other countries considered here due to a break in the series affecting all European countries, but the overall result on employment is highly suggestive that these countries have not seen significant increases in inactivity rates for older adults.

The initial impact of the crisis differed dramatically across education groups (OECD, 2021[1]). The initial reduction in hours was more than double for workers with low and medium education compared to those with higher education. The contraction in hours worked among the low educated was also more frequently experienced through losses of employment. In fact, already in the second half of 2020, hours worked for those with high education had returned to pre-crisis levels and employment had even begun to increase, while hours and employment remained heavily depressed for workers with less education (OECD, 2021[1]).

By Q1 2022, on average across 34 OECD countries, the employment rate of people with tertiary education was above its Q1 2019 level by 0.4 percentage points, while that for people with low and medium education were still down by 0.3 percentage points (Figure 1.21) and 0.2 percentage points (Annex Figure 1.A.2) respectively. These changes are quite significant for workers with less than tertiary education, as their employment rates are typically much lower than those of the highly educated. Indeed, on average across the countries considered, the pre-crisis employment rates for those with low and middle education were 37% and 64% respectively, against a much higher 78% for the high educated.

Employment rates among people with low education were down in 21 countries with respect to pre-crisis levels, mostly in association with an increase in the share of inactive people. The largest falls in employment for the low educated were recorded in Chile, the United Kingdom and Slovenia. Net employment losses were driven primarily by reductions in low-pay service sectors and in manufacturing and constructions.

By contrast, in 13 other countries, the proportion of low educated people in employment increased by an average of 1.7 percentage points. This was mostly the result of a reduction in inactivity, while the share of jobseekers generally remained in line with pre-crisis levels in these countries. In the countries with largest increase in employment for the low educated (Norway, Germany, and Denmark), this was mostly the result of growth in manufacturing and construction and in health and education – as in general low educated people did not benefit from the growth of high-pay service sectors.

Results for people with medium education are qualitatively similar to those reported for the low educated (Annex Figure 1.A.2). Indeed, employment rates for the middle educated were below pre-crisis levels by an average of 1.1 percentage points in 21 countries, mostly in the context of rising inactivity rather unemployment. The employment fortunes of the middle educated were largely determined by changes in low-pay service industries, manufacturing and construction.

The initial impact of the pandemic was felt more strongly among women than men across the majority of OECD countries, but already in the second half of 2020 women’s employment had recovered some of the lost ground relative to men in most countries (OECD, 2021[1]).

By Q1 2022, the proportion of women in employment was 1 percentage point higher than two years earlier on average across the 30-four countries considered (Figure 1.22), with most of the gains accruing from a reduction in inactivity. Over the same period, the proportion of men in employment increased by 0.1 percentage points – resulting in a narrowing of the gender employment gaps (Figure 1.23). Overall, between Q1 2019 and Q1 2022, the employment gap between men and women declined in 23 of the 34 countries considered. Reflecting the general strengthening of the relative position of women, unemployment and inactivity gaps (measured as the difference between men and women) improved in 16 and 26 countries respectively – though this resulted in an average increase (of 0.7 percentage points) across all countries considered only for the inactivity rate, while cross-country average unemployment gap edged down by 0.1 percentage points (Figure 1.23).

The average results for women conceal some variation across countries. In ten countries employment for women was down by an average of 1 percentage points. The lower employment levels mostly derived from increases in the proportion of inactive women rather than unemployed. Employment losses were mostly driven by falls in low-pay service industries and health and education. While the share of women in high-pay service industries in these countries was generally stable and sometimes increased, the gains were too modest to offset the losses in other industries.

In the other 24 countries, women’s employment was up by an average of 1.8 percentage points, mostly driven by falls in inactivity. The employment progress for women in these countries was largely driven by gains in high-pay service industries and health and education and – in some countries – manufacturing and construction.

These labour market developments took place in a context that laid bare the negative consequences of longstanding gender gaps and norms around caregiving (OECD, 2021[44]). The OECD Risks that Matter (RTM) 2020 survey reveals that when schools and childcare facilities closed, mothers took on the brunt of the additional unpaid care work – and, correspondingly, they experienced labour market penalties and stress (OECD, 2021[45]). Mothers of children under age 12 were far more likely to report they took on the majority or entirety of the extra care work than fathers (61.5% vs 22.4%). They were also the group most likely to lose employment at the start of the crisis on average across OECD countries. Studies on the United States also point to a slower recovery than average for mothers with young children (Furman, Kearney and Powell, 2021[46]; Shibata and Pizzinelli, 2022[13]) – especially those with lower education (Goldin, 2022[47]). By contrast, however, in the United Kingdom the employment rate of women appears to have grown quickly above pre-crisis levels over the course of 2021 (Shibata and Pizzinelli, 2022[13]).

The distribution of unpaid work remained unequal even when mothers were in paid employment. Consistently with existing literature (Hupkau and Petrongolo, 2020[48]; Del Boca et al., 2020[49]), the results of the RTM survey also show that non-employed mothers took up a disproportionate amount of unpaid household work when fathers were employed, but the relationship was not reciprocated where the father was out of work and the mother employed (OECD, 2021[45]). Goldin (2022[47]) finds that in the United States the proportion of total parental childcare hours born by the mother increased substantially (from levels already well above 50%) compared to pre-crisis levels in households where both the woman and the man are employed and have tertiary education.

Public supports may have helped to lessen gender inequality at home. The gap in the distribution of additional care of children during COVID-19 was smaller in countries with historically higher levels of spending on family supports (OECD, 2021[44]) and in a number of countries that have introduced job retention schemes, or specific care leaves, women have been able to request to move to reduced hours to avoid being pulled from the labour market by home schooling and care responsibilities (OECD, 2021[1]) – see also Chapter 2.

The labour market implications of the increased burden from unpaid care work over the past two years might reveal themselves over a longer period of time. Indeed, increased caregiving responsibilities might lead women to move to part-time work, stay away from assignments with more responsibilities or search for jobs with more flexibility or a shorter commute. These choices often translate into slower wage growth – through limiting the pool of jobs, weaker bargaining power and greater exposure to monopsony (see Chapter 3) and scarcer opportunities for promotion once in situ – see e.g. (OECD, 2018[50]) and Chapter 4. An important focus for future research will be to monitor the evolution of different dimensions job quality for women to highlight potential sources of gender differentials that might reveal themselves over time.

The COVID-19 crisis struck after a decade of progress for migrants in the labour market. In all OECD countries except Türkiye and Colombia, which had seen large inflows of refugees, immigrants became more successful in finding and keeping jobs over five years before the crisis, although they were still lagging behind native-born in most countries (OECD, 2021[51]). The crisis hit migrants particularly hard due to their sectoral concentration causing a fall in employment and an increase in inactivity in Q2 2020 that was larger than for the native-born in most OECD countries – leading to a temporary widening of the employment gap between the two groups in many countries (OECD, 2022[52]).

There is also evidence that migrants were disproportionally affected by job losses within the sectors. For example, at the onset of the crisis, in the EU27, the number of migrants employed in hospitality dropped by nearly 15% between 2019 and 2020, compared with 12.5% for the native-born. In the United States, the fall in employment in domestic services was respectively 28% for migrants compared with 12% for the native-born (OECD, 2021[51]). Auer (2022[53]) finds that, in Germany, migrants were less likely to be placed on job retention schemes than native-born at the start of the crisis. Yet, patterns are not uniform across countries: for example, Hijzen and Salvatori (2022[54]) find no significant differences in the risk of losing employment or being placed on a job retention scheme between foreign-born and native-born in Switzerland.

By Q1 2022, the labour market situation of migrants across OECD countries had considerably improved. Indeed, on average across 28 countries, the share of migrants with a job was above its Q1 2019 level by 1.2 percentage points, while both the share in inactivity and unemployment had declined (by -0.9 percentage points and -0.3 percentage points respectively) (Figure 1.24). For the native-born, the employment rate was 0.3 percentage points above its Q1 2019 level (Annex Figure 1.A.4), implying that the average employment gap between the two groups had narrowed relative to just before the crisis (Figure 1.25). However, this average result masks some variation across countries. In fact, in nine of the 28 countries the employment gap between natives and foreign born increased (by an average of 1.9 percentage points), with particularly large changes in Latvia, Lithuania and Estonia.

In seven of the 28 countries considered, migrants’ employment was still below pre-crisis levels in Q1 2022 by an average of 2.9 percentage points (Panel A in Figure 1.24) – mostly associated with an increase in inactivity. The fall in migrants’ employment was over 2 percentage points in four countries, and exceeded 4 percentage points in Latvia and Lithuania. In most of the countries where migrants’ employment was still down in Q4 2021, employment had not fully recovered for the native-born either, but the deficit was generally larger for migrants, resulting in a widening of the employment gap between the two groups by 1.9 percentage points on average (Figure 1.25).

In the other 21 countries, the proportion of migrants in employment was up by about 2.5 percentage points in Q1 2022 relative to the same quarter of 2019 – mostly as a result of a reduction in inactivity (-2 percentage points). To some extent this is due to a change in the composition of the pool of migrants due to selective departures and arrivals during the pandemic – as migrants without employment were more likely to leave and those who arrived were more likely to already have a job (OECD, 2021[51]).32 However, in most of these countries inactivity decreased among the native-born as well (and therefore in the whole population – see Figure 1.7). Inactivity can decline in a recovery when the improvement of the labour markets activates previously discouraged workers. The increase in the proportion of migrants in employment exceeded 2 percentage points in 11 countries and was above 5 percentage points in Greece, Denmark, and Poland. On average across the 21 countries where the proportion of migrants in employment was higher than before the crisis, the employment gap with the native-born declined by 1.8 percentage points (Figure 1.25).

The share of migrants in low-pay services contracted in most countries. In the countries where their employment increased above pre-crisis levels, this was the result of employment growth in the other sectors – particularly in Health and Education. However, migrants also appear to have benefitted in many countries from the expansion of high-pay services (Panel B in Figure 1.24).

A recovery characterised by a significant amount of reallocation of employment across industries and occupations poses significant challenges for all workers who face the need to reskill to find viable new jobs. These difficulties are compounded for migrants. They are more likely to be affected by the need for reallocation due to their lower tenure and less stable contract situation, among other issues. At the same time, migrants typically have fewer networks to facilitate transitions to different jobs (OECD, 2020[55]; OECD, 2021[51]).

Few OECD countries collect data or information on the labour market performance of racial/ethnic minorities, in some cases to avoid classifying individuals by race or ethnicity for historical reasons. Different to immigrants, racial/ethnic minorities have long been citizens of, or at least been rooted in, their country. There is not a simple definition of racial/ethnic minorities that fits all OECD countries. Racial/ethnic groups are most often characterised by a shared culture or other factors, including language or religion, as well as their physical appearance (for example skin colour) or the country of origin of their ascendants (Balestra and Fleischer, 2018[56]).”

In the United States, the main racial/ethnic minorities were more affected by the initial impact of the crisis, and lagged behind in the recovery until December 2021. At the onset of the crisis in April 2020, the (seasonally adjusted) employment-to-population ratio fell by 13 percentage points for Hispanic/Latino people and 10.5 percentage points for Black people (Figure 1.26). For white people, the corresponding figure was 9.5 percentage points. Women and young people belonging to racial/ethnic communities in the United States were particularly hard hit by the pandemic, experiencing exceptionally high levels of unemployment, and slow employment gains in the recovery (OECD, 2021[57]).

The higher likelihood of employment loss for racial/ethnic minorities was only partially explained by their sectoral and occupational concentration, i.e. minority workers were more likely to lose their employment than white workers in the same industries and occupations over the course 2020 (Cortes, Forsythe and Forsythe, 2021[58]). Indeed, more generally, observable characteristics can explain very little of the highly persistent labour market disparities between Black people and white people in the United States (Cajner et al., 2017[59]).

Both Black and Hispanic/Latino people lagged behind white people for most of the recovery (Figure 1.26). In particular, relative to white people, employment losses for Hispanic/Latino people remained larger until Q3 2021 and those for Black people until Q1 2022 (1.3 percentage points against 1.1 percentage points). In the second quarter of 2022, the recovery of the employment-to-population ratio slowed down or even receded marginally for all groups. In June 2022, the figure was still below pre-crisis levels for all three groups, standing at 58.6% for Black people, 59.9% for white people and 63.7% for Hispanic/Latino people.

In the United Kingdom, racial/ethnic minorities saw a larger increase in unemployment during the crisis and the differential had not returned to pre-crisis levels by early 2022.33 The unemployment rate of minorities peaked at 9.8% in Q4 of 2020, with an increase of 4 percentage points (1.1 for white people) relative to a year before. After the peak, the unemployment rate for white people declined slowly but steadily, while that for minorities plateaued in the second half of the year. By the first quarter of 2022, the unemployment rate for minorities stood at 7.1% and that for white people at 3.1%, with a differential 0.5 percentage points larger than in the same quarter of 2019 (ONS, 2021[60]).

In Latvia and Estonia employment of racial/ethnic minorities fell more as the crisis hit and was still lagging behind in 2021. In particular, in Latvia the annual employment rate of minorities stood at 57.1% – 3.7 percentage points lower in 2021 than in 2019 – against a gap of -1.4 percentage points for the large ethnic group. In Estonia, the employment rate of the racial/ethnic minority stood at 63.1% in Q4 2021 – or 2.5 percentage points lower than in the same quarter of 2019, against a drop of 1.1 percentage points for the larger group.34

In Australia, Indigenous people were more exposed to the initial shock of the pandemic due to the relatively high levels of casual employment and the relatively young age profile of their population.The high incidence of casual employment among them also meant that Indigenous Australians were less likely to qualify for the Australian job retention scheme (JobKeeper) (Mindaroo Foundation, 2021[61]). Further research is needed to assess the longer-term implications of this shock to Indigenous employment.

In some countries, however, racial/ethnic minorities saw significant improvements in their labour market outcomes over the recovery. In Denmark, employment of descendants from other countries – who are often second (or higher) generation migrants – fell more as the crisis hit in 2020, but by early 2021 it had recovered relative to that of persons of Danish origin. In New Zealand, racial/ethnic minorities have benefitted from the recovery more than people of European origin, the largest racial/ethnic group in the country.35 Indeed, employment rates of Maori and Pacific people were respectively 1.4 percentage points and 2.6 percentage points higher in Q4 2021 than in Q4 2019, while the corresponding figure for Europeans was 0.7 percentage points. Nevertheless, the employment rate remains higher for Europeans at 69.5%, against 64.8% for Maori and 63.5% for Pacific people. The employment gains were mostly linked to a reduction in unemployment for the Maori and in inactivity for the Pacific people. The significant decline in inactivity for Pacific people reduced the corresponding gap with the Europeans from 5.5 percentage points to 4.1 percentage points, with the respective rates standing at 32.8% and 28.7 in Q4 2021.

In Canada, employment recovery among Indigenous people was initially slower, but more recently, the employment rate among Indigenous people surpassed its pre-pandemic level, reducing the gap previously seen between Indigenous and non-Indigenous people.36 As of the three months ending in August 2021, the employment rate among Indigenous people was 57.7% against its pre-pandemic level of 56.2% (the three months ending in February 2020). Among non-Indigenous people, it reached 61.2%, a level similar to the pre-pandemic rate. However, employment recovery among older Indigenous adults (55 or older) since the fall of 2021 was much weaker compared with Indigenous youth and core-aged adults. Also, employment recovery has been slightly slower among First Nations, especially First Nations women whose unemployment rate was still much higher (15.5%) in the quarter ending with August 2021 than its pre-pandemic level (4.8%).

While the crisis has had a significant impact on the life of many through loss of income or employment, it has also deeply affected the experience of many who remained employed throughout the crisis. Some were able to quickly adapt the organisation of their work and carry out their tasks from home. For a large multitude, however, teleworking was never an option. Many workers involved in the delivery of essential goods and services had to continue to work in their physical workplace and in proximity of other people through the various waves of the pandemic. Indeed, the pandemic has highlighted the extent to which society depends upon these “frontline workers”. This section offers a characterisation of these workers and of their experience during the pandemic.

Some studies have attempted to identify frontline workers using ad-hoc lists of “essential” workers who were exempts from restrictions in different countries (Basso et al., 2022[62]; Blau, Koebe and Meyerhofer, 2021[63]), the. Typically frontline workers are identified as the subset of essential workers in industries or occupations that before the pandemic had a low-incidence of telework. This approach poses significant challenges in an international comparison because the definition of essential workers varies across – and even within – countries and over time.

This section takes a different two-step approach. First, following Basso et al. (2022[62]), the analysis uses Labour Force Survey data to describe the personal and job characteristics of workers in occupations, that, based on pre-pandemic information, could not be performed remotely and involved considerable interactions with other people. During the crisis, the group of workers employed in these occupations – which is broader than that of frontline workers – was exposed to a higher risk of income losses (through reductions in hours or job losses) and, when they remained employed, to a higher risk of contagion. For this reason, and similarly to Basso et al. (2022[62]), these occupations are labelled here “at risk”.37

The second step of the analysis exploits unique data from the Eurofound survey “Living, working and COVID-19” to identify the frontline workers who actually worked in their physical workplace and in close contact with other people during the pandemic.38 While this survey lacks some of the personal and job information typically available in Labour Force Surveys (including occupation), it offers a range of well-being indicators that provide important insights on the experience of frontline workers during the pandemic.

At the onset of the pandemic, across the OECD, 44% of workers were in “at-risk” occupations – those that, based on pre-pandemic information, could not be performed from home and required physical proximity to other people (Annex Figure 1.A.1).The figure ranges from 40% or less in Lithuania, Germany, the Czech Republic and Luxembourg, to 50% or more in the United States, Spain, Ireland and Greece. Examples of these occupations include health care workers, cashiers, personal care workers, food processing workers, building workers, and assemblers.

Compared to safer jobs that offered the possibility to telework already before the pandemic, in all countries these at-risk occupations employed more low-pay workers (37% vs 15%), more young workers (12% vs 5% on average across the OECD) and a much lower share of workers with tertiary education (on average 34% vs 67%) (Figure 1.27). Foreign-born workers also held a higher share of at-risk jobs than teleworkable ones in almost all countries (16% vs 13% on average), with the exceptions of Luxembourg and Portugal.

On average across countries the share of at-risk jobs held by women was slightly lower than that of teleworkable jobs (51% vs 53%), but the opposite held true in Finland, Sweden, Denmark, Norway, Switzerland, the United States, the Netherlands and the United Kingdom. In the United States and the United Kingdom, the only two countries where the information is available, ethnic minorities were disproportionally represented in at-risk jobs, although to a much larger extent in the United States than in the United Kingdom. In fact, in the United Kingdom, ethnic minorities held 14% of at-risk jobs and 12% of the teleworkable ones – while in the United States the respective figures were 44% and 31%.

In general, Labour Force Survey data do not allow to verify what proportion of workers in at-risk occupations actually continued to work in their physical workplace during the pandemic. An exception is the United States for which CPS data show that only 11% of workers in at-risk occupations who remained employed were able to telework in the second half of 2020. Other surveys show that the types of workers over-represented in at-risk occupations, such as those with lower qualifications and lower earnings, were much less likely to telework in a number of countries (Ker, Montagnier and Spiezia, 2021[64]; OECD, 2021[5]).

To gather further information on the experience of at-risk workers during the pandemic, this section uses Eurofound data to identify the frontline workers who actually worked in their physical workplace and in close contact with other people during the pandemic. The demographic profile of these workers matches that of workers in at-risk occupations in the LFS data based on the characteristics available in both sources, suggesting that they are likely to be employed in the occupations identified in the LFS data. Indeed, both groups feature higher shares of younger workers and workers with lower levels of education, while the gender composition is in line with that of other jobs. In their work on the United States, Blau et al. (2021[63]) use a list of essential industries issued by the Federal Government and offer a very similar characterisation of the group except for the higher representation of men. They also find that migrants and racial/ethnic minorities are over-represented among frontline workers. While the Eurofound data do not provide information on these characteristics, minorities and migrants are over-represented in at-risk occupations in the United States and the United Kingdom, as noted above (Figure 1.27).

Frontline workers were more likely than teleworkers to feel that their job was insecure (12% vs 7%) and to report bad general health (6% vs 4%) (Figure 1.28). They also reported slightly lower levels of mental well-being (53 vs 55) measured using the WHO-5 mental well-being scale (0-100 – with people with a score below 50 considered at risk of depression), based on the frequency of positive feelings over the previous two weeks (Eurofound, 2021[65]).

While it is certainly plausible that the pandemic might have exacerbated existing differentials in job security and well-being, the hypothesis cannot be tested due to the lack of comparable information for the same workers from before the pandemic. Whether or not the hypothesis holds, however, these results are consistent with the hypothesis that workers who are likely to have been on the frontline during the pandemic have lower quality jobs and well-being in general.39

Indeed, this is consistent with the conclusion reached by other studies that have considered other dimensions of job quality, despite differences in the definition of frontline workers. Amossé et al. (2021[66]) find that frontline workers in France have a (historically) higher risk of job loss and enjoy limited opportunities for career progression. Samek Lodovici et al. (2022[20]) find that, across Europe, frontline workers are more likely to be on temporary contracts and are over-represented in sectors – such agriculture, domestic care and road freight transport – where undeclared work is widespread. Low wages and poor job quality (including a high incidence of non-standard employment forms such as shift or temporary work) have been linked internationally to labour shortages in the long-term care sector, an important “frontline” sector typically included in the list of essential ones across countries (OECD, 2020[67]). Chapter 3 shows that the labour markets of at-risk occupations tend to be more concentrated, thereby contributing to worsen job quality. Eurofound (2021[68]) finds that collectively agreed weekly working hours are longer than the EU average of 37.8 hours in sectors that have been considered essential in many European countries during the pandemic, reaching 39.2 hours in transport. Many frontline workers saw their working hours increase during the pandemic. For example, Finland, France, Italy, Luxembourg, Poland and Portugal, implemented provisions to extend working hours, limit rest periods and delay annual leave in the health care, transport and logistics sectors (Eurofound, 2021[68]).

Those workers who worked in their physical workplace and in proximity with other people certainly felt like they were on the frontline of the battle against COVID-19. Indeed, they were much more likely to feel at risk of contracting the COVID-19 virus because of their job than teleworkers (60% vs 29%) (Figure 1.28). Available evidence indicates that this was far from an exaggerated perception. In Italy, the COVID-19 work injuries claim process by the national Work Injury Insurance (INAIL) were strongly concentrated in at-risk occupations (Basso et al., 2022[62]). Similarly, sick leave claims increased at the same time as COVID-19 cases only in industries characterised by a high incidence of at-risk jobs. In the United Kingdom, those working in occupations requiring close proximity to others had higher COVID-19 death rates, with the highest rate found for men in elementary occupations (Windsor-Shellard and Nasir, 2021[69]). In the United States, workers in essential businesses were far more likely to test positive for COVID-19 – an effect that was not driven by health workers only (Song et al., 2021[70]).40

The higher risk of infection experienced by many frontline workers is likely influenced by a wider set of factors associated with their broader socio-economic situation (Windsor-Shellard and Nasir, 2021[69]). Low-income workers are more likely to live in crowded housing and with other people also employed in occupations with a higher risk of infection.41 People experiencing poor working conditions are more likely to attend work while sick (Bryan, Bryce and Roberts, 2020[71]), a phenomenon observed even when paid leave sick is available but likely to be more pronounced in places with limited availability of such benefit.42 When they get sick, people from low-income households report more difficulties accessing health care even in countries with near-universal access (OECD, 2019[72]). These difficulties are often compounded for migrants and undeclared workers (Samek Lodovici et al., 2022[20]).

In Q4 2021, almost two years after the onset of the pandemic, the share of workers employed in at-risk occupations was down by an average of 3.5 percentage points across the 27 countries with data available. The decline occurred in most countries and exceeded 10 percentage points in the Slovak Republic, Ireland, the United Kingdom, and Estonia. In part, the relative decline in the size of these occupations reflect the strong employment performance over this period of high-pay service industries that employ relatively few workers in these occupations (see Section 1.3). However, by drawing a spotlight on the existing working conditions of these occupations and increasing the risks associated with these jobs, the pandemic likely reduced the labour supply to these occupations, exacerbating labour shortages that already affected many of these occupations before the crisis, most notably in health care occupations (see Section 1.2). In fact, in most OECD countries, public employment services report experiencing greater difficulties in filling frontline job vacancies since the start of the COVID-19 pandemic (see Chapter 2).

More than two years since the abrupt start of the COVID-19 crisis, the recovery in economic activity has been stronger than many expected. The strength of that recovery is now threatened by the economic fallout of Russia’ aggression against Ukraine which is projected to slow down economic growth and continue feed inflation over the course of 2022.

European countries in particular face the immediate challenge of integrating the largest number of refugees since World War II into their labour markets. More than 6.5 million people have already been forced to flee Ukraine to other countries in Europe, and an even greater number have been displaced within the country. The refugee flows caused by the war will result in additional public expenditure in the short-term in host countries, although this will be offset over time as refugees enter the labour force. Recent experiences from various OECD countries provide valuable lessons to facilitate the labour market integration of refugees and to ensure that their skills do not remain idle for too long.

The impact of the war on energy, food, and commodity markets is adding to the significant inflationary pressures that had already emerged at the end of 2021 because of supply chain disruptions. The impact of rising inflation on real incomes is larger for lower-income households which have already borne the brunt of the COVID-19 crisis. Indeed, the increase in expenditure resulting from recent food and energy price increases represents a larger proportion of total spending for lower-income households, and those households have limited scope to offset this by drawing on savings or reducing discretionary expenditures (OECD, 2022[4]). These households disproportionally include low-pay workers who were more likely to have their income reduced during the COVID-19 crisis either through job loss or a reduction in hours worked (OECD, 2021[5]). Going forward, it is crucial to monitor closely the differential impact of inflation across household income levels.

Governments have a range of complementary policy tools available to cushion the impact of inflation on low-income households, including facilitating collective bargaining agreements, adjusting statutory minimum wages and the tax and benefit system, or implementing temporary energy bonuses (see Chapter 2 for a discussion of recent interventions by OECD governments).

Even before the new negative shock from the war in Ukraine, the labour market recovery from the COVID-19 crisis remained incomplete and uneven across countries. While some of the initial unequal impact of the crisis across workers has been reabsorbed, young people, and workers without tertiary education have been lagging behind in the recovery in many countries.

There is currently no indication of qualitative mismatches between supply and demand caused by the asymmetric impact of the crisis on different sectors. These mismatches could however emerge more clearly once the current vacancy tide affecting all industries withdraws. This chapter shows that industries that have expanded since the onset of the crisis are very different from industries that have seen employment fall. Furthermore, in addition to the pressures arising from changes that might have been triggered or accelerated by the pandemic per se, many countries intend to use their recovery plans to promote digitalisation and the transition towards a climate-neutral economy. These policies are likely to accelerate further structural transformations of the labour market which might also contribute to rising mismatches.

In this context, monitoring the evolution of skill demands and of labour market outcomes for different workers remain essential to ensure the fine-tuning and targeting of policies aimed at ensuring good matches between workers and jobs to promote an inclusive labour market.


[29] Aaronson, D., R. Lewers and D. Sullivan (2021), “Labor reallocation during the Covid-19 pandemic”, Chicago Fed Letter, https://doi.org/10.21033/cfl-2021-455.

[22] Aaronson, S., M. Barnes and W. Edelberg (2022), A Hot Labor Market Won’t Eliminate Racial and Ethnic Unemployment Gaps, The Hamilton Project Blog, https://www.hamiltonproject.org/blog/a-hot-labor-market-wont-eliminate-racial-and-ethnic-unemployment-gaps (accessed on 16 February 2022).

[35] Altonji, J., L. Kahn and J. Speer (2016), “Cashier or Consultant? Entry Labor Market Conditions, Field of Study, and Career Success”, Journal of Labor Economics, Vol. 34/S1, pp. S361-S401, https://doi.org/10.1086/682938.

[66] Amossé, T. et al. (2021), “Les métiers “de deuxième ligne” de la crise Covid-19 : quelles conditions de travail et d’emploi dans le secteur privé ?”, DOCUMENT D’ÉTUDES, No. 246, DARES, Paris.

[33] Andrews, D., A. Charlton and A. Moore (2021), “COVID-19, productivity and reallocation: Timely evidence from three OECD countries”, OECD Economics Department Working Papers, No. 1676, OECD Publishing, Paris, https://doi.org/10.1787/d2c4b89c-en.

[37] Andrews, D. et al. (2020), “The career effects of labour market conditions at entry”, OECD Productivity Working Papers, No. 20, OECD Publishing, Paris, https://doi.org/10.1787/29c11c75-en.

[53] Baert, S. (ed.) (2022), “Firing discrimination: Selective labor market responses of firms during the COVID-19 economic crisis”, PLOS ONE, Vol. 17/1, p. e0262337, https://doi.org/10.1371/journal.pone.0262337.

[56] Balestra, C. and L. Fleischer (2018), “Diversity statistics in the OECD: How do OECD countries collect data on ethnic, racial and indigenous identity?”, OECD Statistics Working Papers, No. 2018/09, OECD Publishing, Paris, https://doi.org/10.1787/89bae654-en.

[62] Basso, G. et al. (2022), “Unsafe Jobs, Labour Market Risk and Social Protection”, Economic Policy, https://doi.org/10.1093/epolic/eiac004.

[31] Basso, G., A. Grompone and F. Modena (2021), “The (little) Reallocation Potential of Workers Most Hit by the Covid-19 Crisis”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3827512.

[21] Bergman, N., D. Matsa and M. Weber (2022), Inclusive Monetary Policy: How Tight Labor Markets Facilitate Broad-Based Employment Growth, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w29651.

[63] Blau, F., J. Koebe and P. Meyerhofer (2021), “Who are the essential and frontline workers?”, Business Economics 2021 56:3, Vol. 56/3, pp. 168-178, https://doi.org/10.1057/S11369-021-00230-7.

[30] Brewer, M., C. McCurdy and H. Slaughter (2021), Begin again? - Assessing the permanent implications of Covid-19 for the UK’s labour market, Resolution Foundation , London, https://economy2030.resolutionfoundation.org/reports/begin-again/ (accessed on 22 February 2022).

[71] Bryan, M., A. Bryce and J. Roberts (2020), “Presenteeism in the UK: Effects of physical and mental health on worker productivity”, Sheffield Economic Research Paper Series, No. 2020005, University of Sheffield, Sheffield.

[59] Cajner, T. et al. (2017), “Racial Gaps in Labor Market Outcomes in the Last Four Decades and over the Business Cycle”, Finance and Economics Discussion Series, Vol. 2017/071, https://doi.org/10.17016/FEDS.2017.071.

[34] Carcillo, S. et al. (2015), “NEET Youth in the Aftermath of the Crisis: Challenges and Policies”, OECD Social, Employment and Migration Working Papers, No. 164, OECD Publishing, Paris, https://doi.org/10.1787/5js6363503f6-en.

[9] Cook, I. (2021), Who Is Driving the Great Resignation?, Harvard Business Review, https://hbr.org/2021/09/who-is-driving-the-great-resignation (accessed on 9 March 2022).

[17] Coombs, K. et al. (2022), “Early Withdrawal of Pandemic Unemployment Insurance: Effects on Earnings, Employment and Consumption”, Working Paper, No. 22-046, Harvard Business School.

[58] Cortes, G., E. Forsythe and E. Forsythe (2021), “The Heterogeneous Labor Market Impacts of the Covid-19 Pandemic”, Industrial and Labour Relations Review, Vol. Forthcoming, https://doi.org/10.17848/wp20-327.

[32] DARES (2021), Hébergement restauration : quelle évolution des effectifs avec la crise ?, DARES Blog, https://dares.travail-emploi.gouv.fr/publication/hebergement-restauration-quelle-evolution-des-effectifs-avec-la-crise (accessed on 10 March 2022).

[49] Del Boca, D. et al. (2020), “Women’s and men’s work, housework and childcare, before and during COVID-19”, Review of Economics of the Household, Vol. 18/4, pp. 1001-1017, https://doi.org/10.1007/s11150-020-09502-1.

[42] Dorn, E. et al. (2021), COVID-19 and education: The lingering effects of unfinished learning | McKinsey, McKinsey & Company Blog, https://www.mckinsey.com/industries/education/our-insights/covid-19-and-education-the-lingering-effects-of-unfinished-learning (accessed on 25 February 2022).

[26] Dube, A. (2019), Impacts of minimum wages: review of the international evidence, Low Pay Commission, London.

[14] Duval, R. et al. (2022), “Labor Market Tightness in Advanced Economies”, Staff Discussion Notes, No. 2022/001, IMF.

[24] EPI (2022), Minimum Wage Tracker, https://www.epi.org/minimum-wage-tracker/ (accessed on 23 June 2022).

[23] Eurofound (2022), “Minimum wages in 2022: Annual review”, Minimum wages in the EU series, Publications Office of the European Union, Luxembourg.

[65] Eurofound (2021), Living, working and COVID-19 (Update April 2021): Mental health and trust decline across EU as pandemic enters another year, Eurofound, Luxembourg, https://www.eurofound.europa.eu/publications/report/2021/living-working-and-covid-19-update-april-2021-mental-health-and-trust-decline-across-eu-as-pandemic (accessed on 24 November 2021).

[19] Eurofound (2021), Tackling labour shortages in EU Member States, Publications Office of the European Union, Luxembourg.

[68] Eurofound (2021), Working time in 2019–2020, Publications Office of the European Union, Luxembourg.

[15] European Central Bank (2022), Economic Bulletin, Issue 1/2022., European Central Bank, Frankfurt.

[8] European Central Bank (2019), Economic Bulletin Issue 4, 2019, European Central Bank, Frankfurt.

[28] Eurostat (2022), EU labour force survey - correction for breaks in time series, https://ec.europa.eu/eurostat/statistics-explained/index.php?title=EU_labour_force_survey_-_correction_for_breaks_in_time_series#Comparability_over_time_-_correction_of_breaks_in_time_series (accessed on 15 July 2022).

[46] Furman, J., M. Kearney and W. Powell (2021), The Role of Childcare Challenges in the US Jobs Market Recovery During the COVID-19 Pandemic, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w28934.

[40] Gaini, M., A. Leduc and A. Vicard (2013), “Peut-on parler de «générations sacrifiées»? Entrer sur le marché du travail dans une période de mauvaise conjoncture économique”, Economie et statistique, Vol. 462-463, pp. 5-23, https://www.persee.fr/doc/estat_0336-1454_2013_num_462_1_10214 (accessed on 30 May 2022).

[39] Garrouste, C. and M. Godard (2016), “The Lasting Health Impact of Leaving School in a Bad Economy: Britons in the 1970s Recession”, Health Economics, Vol. 25, pp. 70-92, https://doi.org/10.1002/hec.3391.

[47] Goldin, C. (2022), Understanding the Economic Impact of COVID-19 on Women, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w29974.

[11] Gould, E. (2022), JOLTS | Economic Policy Institute, Economic Policy Institute Blog, https://www.epi.org/indicators/jolts/ (accessed on 10 March 2022).

[54] Hijzen, A. and A. Salvatori (2022), “The impact of the COVID-19 crisis across different socio-economic groups and the role of job retention schemes - The case of Switzerland”, OECD Social, Employment and Migration Working Papers, No. 268, OECD Publishing, Paris, https://doi.org/10.1787/38fc6bad-en.

[16] Holzer, H., R. Hubbard and M. Strain (2021), Did Pandemic Unemployment Benefits Reduce Employment? Evidence from Early State-Level Expirations in June 2021, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w29575.

[48] Hupkau, C. and B. Petrongolo (2020), “Work, Care and Gender during the COVID‐19 Crisis*”, Fiscal Studies, Vol. 41/3, pp. 623-651, https://doi.org/10.1111/1475-5890.12245.

[6] ISTAT (2022), Il Mercato del Lavoro: I trimestre 2022, ISTAT, https://www.istat.it/it/files/2022/06/Mercato-del-lavoro-I-trim_2022.pdf (accessed on 30 June 2022).

[73] Jelliffe, E. et al. (2021), “Awareness and use of (emergency) sick leave: US employees’ unaddressed sick leave needs in a global pandemic”, Proceedings of the National Academy of Sciences, Vol. 118/29, p. e2107670118, https://doi.org/10.1073/pnas.2107670118.

[64] Ker, D., P. Montagnier and V. Spiezia (2021), “Measuring telework in the COVID-19 pandemic”, OECD Digital Economy Papers, No. 314, OECD Publishing, Paris, https://doi.org/10.1787/0a76109f-en.

[38] Kroft, K., F. Lange and M. Notowidigdo (2013), “Duration Dependence and Labor Market Conditions: Evidence from a Field Experiment*”, The Quarterly Journal of Economics, Vol. 128/3, pp. 1123-1167, https://doi.org/10.1093/qje/qjt015.

[7] McGregor, T., N. Suphaphiphat and F. Toscani (2022), Europe’s Consumers are Sitting on 1 Trillion Euros in Pandemic Savings – IMF Blog, IMF Blog, https://blogs.imf.org/2022/02/10/europes-consumers-are-sitting-on-1-trillion-euros-in-pandemic-savings/?utm_medium=email&utm_source=govdelivery (accessed on 22 February 2022).

[61] Mindaroo Foundation (2021), The Impacts of COVID-19 on the Indigenous Workforce, Mindaroo Foundation, https://cdn.minderoo.org/content/uploads/2021/03/19112119/The-impacts-of-COVID-19-on-the-Indigenous-Workforce.pdf (accessed on 14 March 2022).

[4] OECD (2022), OECD Economic Outlook, Volume 2022 Issue 1, OECD Publishing, Paris, https://doi.org/10.1787/62d0ca31-en.

[3] OECD (2022), Rights and Support for Ukrainian Refugees in Receiving Countries, OECD Publishing, Paris, https://doi.org/10.1787/09beb886-en.

[52] OECD (2022), “The unequal impact of COVID-19: A spotlight on frontline workers, migrants and racial/ethnic minorities”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris, https://doi.org/10.1787/f36e931e-en.

[44] OECD (2021), “Caregiving in Crisis: Gender inequality in paid and unpaid work during COVID-19”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris, https://doi.org/10.1787/3555d164-en.

[57] OECD (2021), COVID-19 and Well-being: Life in the Pandemic, OECD Publishing, Paris, https://doi.org/10.1787/1e1ecb53-en.

[51] OECD (2021), International Migration Outlook 2021, OECD Publishing, Paris, https://doi.org/10.1787/29f23e9d-en.

[45] OECD (2021), Main Findings from the 2020 Risks that Matter Survey, OECD Publishing, Paris, https://doi.org/10.1787/b9e85cf5-en.

[1] OECD (2021), OECD Economic Outlook, Volume 2021 Issue 2, OECD Publishing, Paris, https://doi.org/10.1787/66c5ac2c-en.

[5] OECD (2021), OECD Employment Outlook 2021: Navigating the COVID-19 Crisis and Recovery, OECD Publishing, Paris, https://doi.org/10.1787/5a700c4b-en.

[43] OECD (2021), “What have countries done to support young people in the COVID-19 crisis?”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris, https://doi.org/10.1787/ac9f056c-en.

[55] OECD (2020), International Migration Outlook 2020, OECD Publishing, Paris, https://doi.org/10.1787/ec98f531-en.

[67] OECD (2020), Who Cares? Attracting and Retaining Care Workers for the Elderly, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/92c0ef68-en.

[72] OECD (2019), Health for Everyone?: Social Inequalities in Health and Health Systems, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/3c8385d0-en.

[50] OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2018-en.

[2] OECD (2016), Making Integration Work: Refugees and others in need of protection, Making Integration Work, OECD Publishing, Paris, https://doi.org/10.1787/9789264251236-en.

[27] OECD (2015), “Recent labour market developments with a focus on minimum wages”, in OECD Employment Outlook 2015, OECD Publishing, Paris, https://doi.org/10.1787/empl_outlook-2015-5-en.

[74] OECD (forthcoming), Disability, work and inclusion in Korea: Towards equitable and adequate paid sick leave, OECD Publishing, Paris.

[60] ONS (2021), A09: Labour market status by ethnic group - Office for National Statistics, https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemployeetypes/datasets/labourmarketstatusbyethnicgroupa09 (accessed on 26 November 2021).

[36] Oreopoulos, P., T. von Wachter and A. Heisz (2012), “The Short- and Long-Term Career Effects of Graduating in a Recession”, American Economic Journal: Applied Economics, Vol. 4/1, pp. 1-29, https://doi.org/10.1257/app.4.1.1.

[10] Parker, K. and J. Horowitz (2022), The Great Resignation: Why workers say they quit jobs in 2021 | Pew Research Center, PEW Research blog, https://www.pewresearch.org/fact-tank/2022/03/09/majority-of-workers-who-quit-a-job-in-2021-cite-low-pay-no-opportunities-for-advancement-feeling-disrespected/ (accessed on 13 March 2022).

[18] Petrosky-Nadeau, N. and R. Valletta (2021), “UI Generosity and Job Acceptance: Effects of the 2020 CARES Act”, Federal Reserve Bank of San Francisco Working Paper 2021-13, https://doi.org/10.24148/wp2021-13.

[12] Rottger, C. and E. Weber (2021), No Big Quit in Germany | LSE COVID-19, https://blogs.lse.ac.uk/covid19/2021/11/24/no-big-quit-in-germany/ (accessed on 19 February 2022).

[20] Samek Lodovici, M. et al. (2022), Revaluation of working conditions and wages for essential workers, Policy Department for Economic, Scientific and Quality of Life Policies, European Parliament.

[13] Shibata, I. and C. Pizzinelli (2022), “Has COVID-19 Induced Labor Market Mismatch? Evidence from the US and the UK”, IMF Working Papers, Vol. 2022/005, p. 1, https://doi.org/10.5089/9781616359027.001.

[70] Song, H. et al. (2021), “The impact of the non-essential business closure policy on Covid-19 infection rates”, International Journal of Health Economics and Management, Vol. 21/4, pp. 387-426, https://doi.org/10.1007/S10754-021-09302-9/TABLES/14.

[41] Thorn, W. and S. Vincent-Lancrin (2021), Schooling During a Pandemic: The Experience and Outcomes of Schoolchildren During the First Round of COVID-19 Lockdowns, OECD Publishing, Paris, https://doi.org/10.1787/1c78681e-en.

[25] Vacas-Soriano, C. and C. Aumayr-Pintar (2022), How to ensure adequate minimum wages in an age of inflation |, https://www.eurofound.europa.eu/publications/blog/how-to-ensure-adequate-minimum-wages-in-an-age-of-inflation#footnote-FFkhl-1 (accessed on 5 July 2022).

[69] Windsor-Shellard, B. and R. Nasir (2021), Coronavirus (COVID-19) related deaths by occupation, England and Wales - Office for National Statistics, ONS, London, https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/causesofdeath/bulletins/coronaviruscovid19relateddeathsbyoccupationenglandandwales/deathsregisteredbetween9marchand28december2020#overview-of-coronavirus-related-deaths-by-occupation (accessed on 24 November 2021).

In 2021, Eurostat implemented a number of methodological changes to the way European Labour Force Survey data are collected and managed as well as some changes to the labour market status definitions. These changes have produced a break in the series for employment and unemployment provided by Eurostat in the first quarter 2021. In the spring of 2022, Eurostat released break-adjusted series for employment and unemployment and some demographic breakdowns. The methodology employed is described in Eurostat (2022[28]). Whenever possible, this chapter uses the recently released break-adjusted series. This is the case, for example, for aggregate employment and unemployment rates, as well as for the series by education, gender and age.

However, for some of the series used in this chapter, Eurostat has not provided the break-adjusted version. This is the case, for example, for employment by industry, employment and unemployment by country of birth and employment by industry and various demographic characteristics. In all these cases, the chapter uses adjusted series using a correction factor calculated exploiting the availability of both break-adjusted and non-break-adjusted series at a higher level of aggregation.

To illustrate the procedure, consider the case of employment by industry. In this case, a correction factor (for each country and quarter) is calculated by taking the ratio between the break-corrected aggregate employment and the uncorrected aggregate employment. The same correction factor is then multiplied by the (uncorrected) employment level of each industry in the relevant quarter. For example, to correct the employment level of a given industry in Q1 2019, the level of employment for that industry reported by Eurostat is multiplied by the ratio between the adjusted total employment in Q1 2019 and the unadjusted total employment in the same quarter.

A similar procedure is adopted for the other series used in this chapter. When the series of interest is expressed as a ratio, the correction factors are also computed from the uncorrected and corrected ratios. For example, for the series of the proportion of a given demographic population employed in a given industry (for example, the proportion of all women who work in Finance and Insurance), the correction factor is computed using the ratio between the corrected and uncorrected employment rate for that demographic group (continuing the example: the ratio between the adjusted and unadjusted fraction of women in employment).

Eurostat did not provide corrected series for employment by country of birth. The correction factor for the proportion of the foreign-born population in employment is computed as the ratio between the corrected and uncorrected employment rate for the whole population. The same correction factor is then applied to correct the series for employment by country of birth by industry.

The main limitation of this approach is the underlying assumption that the outcomes of the various groups to which the correction factor is applied were indeed affected in the same way by the break in the series. For example, in the case of the employment of women by industry, the procedure assumes that the proportional change in employment produced in the aggregate for women by the break also occurred in every single industry.


← 1. This chapter has benefited from statistical support from Isac Olave Cruz and Agnès Puymoyen. Earlier versions of the material covered in Section 1.4 and Section 1.5 also benefitted from statistical analysis by Inbar Amit.

← 2. The difference in employment and unemployment figures across countries partly reflect the fact that people on temporary layoff are classified as unemployed in countries like Canada and the United States even when they expect to go back to the same job – while in most countries, workers on zero hours while on job retention schemes are still classified as employed. See Chapter 1 in (OECD, 2021[5]) for more details.

← 3. This section draws from OECD (2022[4]).

← 4. The information on the number of refugees from Ukraine recorded across Europe was retrieved from https://data.unhcr.org/en/situations/ukraine on 26 August 2022.

← 5. To document how hours have recovered from the COVID-19 crisis, Figure 1.6 uses data from Q1 2022, the most recent data point available for the largest number of OECD countries. Since seasonally adjusted data are not readily available unadjusted data are used and Q1 2019 is taken as the pre-crisis reference point. Although this method may overstate the recovery by netting out most of the hours growth in 2019, the results still show that hours recovery is still incomplete in a majority of countries for which seasonally adjusted data are available.

← 6. For the countries covered by Eurostat, all the employment series are affected by a break in Q1 2021 (see Eurostat (2022[28])). Whenever available, break-adjusted series provided by Eurostat are used in the analysis. In the other cases, a correction described in Annex 1.B has been applied.

← 7. Between Q4 2019 and Q4 2021, Mexico also saw a large proportional increase but from a rather low starting point, as its long-term unemployment rate increased from less than 0.1% to 0.24%.

← 8. https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1410032501&cubeTimeFrame.startMonth=10&cubeTimeFrame.startYear=2019&cubeTimeFrame.endMonth=10&cubeTimeFrame.endYear=2021&referencePeriods=20191001%2C20211001.

← 9. https://www.mbie.govt.nz/business-and-employment/employment-and-skills/labour-market-reports-data-and-analysis/jobs-online/using-the-all-vacancies-index-avi-as-main-indicator/.

← 10. www.dati.istat.it.

← 11. Source: OECD Short Term Labour Market Statistics Database.

← 12. https://www.nzier.org.nz/news/nziers-qsbo-shows-weaker-demand-and-confidence.

← 13. https://www150.statcan.gc.ca/n1/daily-quotidien/220225/dq220225b-eng.htm.

← 14. IAB Labour Market Barometer | IAB.

← 15. https://www.ons.gov.uk/businessindustryandtrade/business/businessservices/bulletins/businessinsightsandimpactontheukeconomy/27january2022#worker-shortages https://www.ons.gov.uk/businessindustryandtrade/business/businessservices/bulletins/businessinsightsandimpactontheukeconomy/21april2022#workforce.

← 16. https://www150.statcan.gc.ca/n1/daily-quotidien/220225/dq220225b-eng.htm.

← 17. In March 2022, seasonally adjusted quits rates were 75% above their level of Dec 2019 in manufacturing (https://fred.stlouisfed.org/series/JTS3000QUR). In retail trade, the figure was 45% (https://fred.stlouisfed.org/series/JTS4400QUR) and in Finance and Insurance 36%.

← 18. https://fred.stlouisfed.org/graph/?g=OZ23.

← 19. U.S. Bureau of Labor Statistics, Employment-Population Ratio [EMRATIO], retrieved from FRED, Federal Reserve Bank of St. Louis; https://fred.stlouisfed.org/series/EMRATIO, 10 March 2022.

← 20. They were also 40% higher than in Q1 2020 – which however was already partially affected by the beginning of the COVID-19 crisis. See https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemployeetypes/datasets/labourforcesurveyflowsestimatesx02.

← 21. https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemployeetypes/bulletins/uklabourmarket/february2022.

← 22. https://dares.travail-emploi.gouv.fr/publication/mi-2021-un-niveau-eleve-de-demissions-de-cdi.

← 23. https://dares.travail-emploi.gouv.fr/donnees/les-mouvements-de-main-doeuvre.

← 24. https://www.abs.gov.au/statistics/labour/employment-and-unemployment/job-vacancies-australia/latest-release.

← 25. Duval et. (2022[14]) report some preliminary evidence indicating that wages were more responsive to increases in labour market tightness in low pay sectors and that this in turn contributed significantly to aggregate wage growth over the course of 2021.

← 26. https://dares.travail-emploi.gouv.fr/donnees/les-indices-de-salaire-de-base.

← 27. Similar to the result for the United States reported above, Duval et. (2022[14]) use a regression-based approach to present preliminary evidence that wages were more responsive to increases in labour market tightness in low pay sectors in the United Kingdom and that this in turn contributed significantly to aggregate wage growth over the course of 2021.

← 28. The data reported in Figure 1.16 are provided by Eurostat based on EU LFS. Data from the Swiss Federal Office for Statistics show a much smaller contraction of employment in manufacturing of around 2% between Q4 2019 and Q4 2021 (see https://www.bfs.admin.ch/bfs/fr/home/statistiques/travail-remuneration/activite-professionnelle-temps-travail/caracteristiques-main-oeuvre/section-economique.assetdetail.21825634.html).

← 29. Importantly, the conclusions of any study of employment reallocation across sectors might hinge crucially on the specific time interval considered. Indeed, the significance of cross-sector transitions might well have changed over the course of the crisis as the uncertainty surrounding the prospects of the different sectors has evolved non-linearly due to the recurrence of pandemic waves of different intensity, the progress of the vaccination campaigns, and the variation in the nature of the restrictions adopted.

← 30. For the countries covered by Eurostat, all the employment series are affected by a break in Q1 2021 (see Eurostat (2022[28])). Whenever available, break-adjusted series provided by Eurostat are used in the analysis. In the other cases, a correction described in Annex 1.B has been applied.

← 31. Across the same countries considered here, the share of fixed-term among young people declined on average by more than 2 percentage points in Q2 2020 relative to Q2 2019 – with drops observed in 18 of the 28 countries.

← 32. According to European Labour Force Survey data, between Q4 2019 and Q4 2021, the total population of migrants declined by at least 10% in Poland, Ireland, the United Kingdom, Portugal and Greece. By contrast, the Czech Republic, the Netherlands, Hungary, Finland and Iceland – all saw increase in the total migrant population in excess of 10%. By comparison, In all these countries the total population of the native-born remain substantially stable (See https://ec.europa.eu/eurostat/databrowser/bookmark/adc41851-d0c0-48e6-809a-a081f5282e4e?lang=en). In the United States and Canada, the migrant population recorded in the CPS and Labour Force Survey increased by less than 2%.

← 33. The race/ethnic groups included are Indian, Pakistani, Bangladeshi, Chinese, Black/African/Caribbean/Black British, and people reporting mixed/multiple ethnic groups.

← 34. https://data.stat.gov.lv/pxweb/en/OSP_PUB/START__EMP__NB__NBLB/NBL030/ and https://andmed.stat.ee/en/stat/sotsiaalelu__tooturg__tooturu-uldandmed__aastastatistika/TT332/table/tableViewLayout2.

← 35. https://www.stats.govt.nz/information-releases/labour-market-statistics-december-2021-quarter.

← 36. Information provided by Canada in response to OECD Questionnaire on Policy Responses to the COVID-19 Crisis (see Chapter 2 for more details on the Questionnaire).

← 37. To characterise the workers in these jobs, this chapter replicates the work by Basso et al. (2022[62]) who kindly shared their code. The authors identify “at-risk” occupations as those that, based on pre-pandemic information, could not be performed remotely and involved considerable interaction with other people and therefore a heightened risk of COVID-19 infection on the job, see Basso et al. (2022[62]) for details on methodology. The same classification is also used in Chapter 3.

← 38. In practice, frontline workers are defined as those answering “Always”, “Most of the time” or “Sometimes” to the question: “In your work, are you currently in direct physical contact with people (colleagues, customers, passengers, pupils, patients, etc.)?” and who do not report “home” as a location of work during the pandemic. The data used from the analysis are from the second wave from June 2020 and cover European countries only. Data from the first wave (April 2020) do not contain information on close contact at work. While the set of workers returning to their workplace is likely to have increased between the two waves given the different stage of the pandemic and the nature of the restrictions in place, the demographic characteristics of workers who continue to work in their physical workplace across the two waves are the same.

← 39. The results do not necessarily imply that these occupations cause lower health or mental well-being. In fact, these differences can at least partly be driven by selection effects if workers with poorer health or mental well-being struggle to find better jobs. Whatever the precise causal mechanism, the result still points to a disadvantage for workers who are employed in these jobs.

← 40. All these studies refer to 2020 and early 2021 – the relative impact of the virus on different occupational categories might have changed as more transmissible variants – such as Delta and Omicron – became dominant. In addition, at the same time as these variants spread, the restrictions in place were generally less strict as vaccination rates reached high levels in most countries. The combination of these factors means that the relative risk of exposure across different occupations might well have changed over the course of the pandemic while remaining – in all likelihood – higher for jobs which involve direct contact with a large number of people.

← 41. https://www.oecd.org/housing/data/affordable-housing-database/housing-conditions.htm.

← 42. In Korea workers have no statutory right to paid or unpaid sick leave (OECD, forthcoming[74]), while in the United States only 31% of workers in the bottom decile of the wage distribution had access to paid sick leave in March 2019, a figure that had increased to 35% by March 2021 (https://www.bls.gov/ncs/ebs/xlsx/employee-benefits-in-the-united-states-dataset.xlsx). Recent evidence indicates low general awareness of the introduction of the a federal COVID-19 sick leave provision in the United States in March 2020, with particularly low levels of awareness and take-up among foreign-born – a group over-represented in frontline jobs (Jelliffe et al., 2021[73]).


This work is published under the responsibility of the Secretary-General of the OECD. The opinions expressed and arguments employed herein do not necessarily reflect the official views of the Member countries of the OECD.

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area.

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

Note by Republic of Türkiye
The information in this document with reference to “Cyprus” relates to the southern part of the Island. There is no single authority representing both Turkish and Greek Cypriot people on the Island. Türkiye recognises the Turkish Republic of Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Türkiye shall preserve its position concerning the “Cyprus issue”.

Note by all the European Union Member States of the OECD and the European Union
The Republic of Cyprus is recognised by all members of the United Nations with the exception of Türkiye. The information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.

Photo credits: Cover © León del Monte.

Corrigenda to publications may be found on line at: www.oecd.org/about/publishing/corrigenda.htm.

© OECD 2022

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at https://www.oecd.org/termsandconditions.