OECD Employment Outlook 2020: Worker Security and COVID-19
Summary
OECD Employment Outlook 2020: Worker Security and the COVID-19 Crisis, published by OECD Publishing, Paris, under the responsibility of the OECD Secretary-General. The foreword by Secretary-General Angel Gurría states that OECD-wide GDP is projected to have fallen by almost 15% between the last quarter of 2019 and the second quarter of 2020, and that the OECD-wide unemployment rate is projected at 9.4% at the end of 2020. The report has five chapters: the labour market consequences of COVID-19 and the policy response, unemployment benefits for non-standard workers, employment protection legislation, middle-skill workers, and vocational education and training graduates. It also contains a statistical annex. The document closes with a discussion of pathways from medium-level VET into higher education in Austria and Norway. The report runs 370 pages.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Full text
Full text shows the first 300 of 370 pages; the complete document is the PDF above.
OECD Employment Outlook
2020
WORKER SECURITY AND THE COVID-19 CRISIS
OECD Employment Outlook
2020
WORKER SECURITY AND THE COVID‑19 CRISIS
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 OECD member countries.
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 Turkey
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. Turkey recognises the Turkish Republic of
Northern Cyprus (TRNC). Until a lasting and equitable solution is found within the context of the United Nations, Turkey
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 Turkey. The
information in this document relates to the area under the effective control of the Government of the Republic of Cyprus.
Please cite this publication as:
OECD (2020), OECD Employment Outlook 2020: Worker Security and the COVID-19 Crisis, OECD Publishing, Paris,
https://doi.org/10.1787/1686c758-en.
ISBN 978-92-64-99828-5 (print)
ISBN 978-92-64-35320-6 (pdf)
OECD Employment Outlook
ISSN 1013-0241 (print)
ISSN 1999-1266 (online)
Photo credits: Cover : © León del Monte
Corrigenda to publications may be found on line at: www.oecd.org/about/publishing/corrigenda.htm.
© OECD 2020
The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.
3
Foreword
This edition of the Employment Outlook is released in the midst of a global health emergency that is turning
into an economic and social crisis that evokes the Great Depression. The epidemiological model developed
by the OECD shows that the severe restrictions to social and economic life that most OECD countries (and
many others) have had to take to slow the spread of the virus have prevented the collapse of health care
systems and helped to avoid hundreds of thousands, if not millions, of deaths. Yet, there is no question
that these measures have had very serious economic and social consequences. Entire sectors of the
economy were essentially closed down for weeks on end. Between the last quarter of 2019 and the second
quarter of 2020, OECD-wide GDP is projected to have fallen by almost 15%. In the first three months of
the COVID-19 crisis, in OECD countries for which data are available, hours worked fell ten times more
than in the first three months of the 2008-09 global financial crisis.
In response, governments have implemented packages of measures to support people and companies
and to cushion the impact of the crisis, which have often been impressive in their scale and speed. Some
countries expanded the support provided by unemployment benefits and made them more accessible.
Some countries expanded access to, or the generosity of, paid sick leave. Many countries have eased
companies’ access to short-time work schemes, making them more widely available (in particular to small
and medium-sized enterprises) and generous while lowering conditionality requirements. Many countries
have also stepped up means-tested assistance of last resort, introduced new ad hoc cash transfers, and
provided direct support to those who lost their livelihoods.
Despite these substantial efforts, the numbers are stark and our projections are bleak. Even if a second
wave of infections is avoided, the June 2020 OECD Economic Outlook projects a 6% annual decline in
global GDP for 2020. The OECD-wide unemployment rate is projected to be at 9.4% at the end of 2020,
above any previous historical peak, and still 7.7% the year after. The crisis will cast a long shadow over
the world and OECD economies. By 2021, it will have taken real income per capita in the majority of OECD
economies back to 2016 levels even in the absence of a widespread second wave of infections. In the
“double-hit” scenario where a second wave strikes all OECD economies in late 2020, real per capita
income in the median OECD economy in 2021 would be back to 2013 levels.
As many countries gradually move out of strict containment measures and the economy re-starts, it is
essential to sustain the recovery with a combination of macroeconomic policies and sectoral policies to
boost growth and job creation while providing support to the many still in need.
Policies need to sustain public and private investment, especially on green and other essential
infrastructure and more generally to foster job creation. Moreover, policy makers will need to modify and
adjust the composition and characteristics of their support packages, targeting support where it is most
needed and encouraging a return to work where possible. If they get these decisions right, we will be able
to look back on 2020 as a year of crisis, successfully navigated. Get them wrong, and the consequences
will be felt by many people for a long time.
The Employment Outlook 2020 outlines some of the critical decisions that countries will have to make.
Decisions on how, and at what speed, to manage a return to economic and social activity, while keeping
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
4
workers safe. Decisions on how to scale back job retention schemes without prematurely removing support
where it is still needed. Decisions on how to adapt emergency support programmes for self-employed
workers and businesses, especially small ones, as economic activity picks up, given that some viable
businesses in the most impacted sectors may continue to face restrictions and/or low demand. Decisions
on how to provide adequate income support by adapting some of the support mechanisms exceptionally
put in place during the pandemic. Decisions on how to support job creation effectively with targeted
subsidies, and how to help jobseekers with public and private employment services. Last, but certainly not
least, decisions on how to provide a comprehensive support package to the cohort of young people whose
education and early labour market experience have been blighted by the COVID-19 crisis. The crisis
cannot be allowed to result in a lost generation of young people whose careers are permanently diminished
by the disruption to the labour market.
More generally, in taking all these decisions, it is essential that the measures adopted leave no one behind.
The impact of COVID-19 is particularly severe for the elderly, low-income earners, women, migrants,
children and youth, and those with disabilities and with chronic health conditions. By accompanying labour
market and social protection measures with a broad and coordinated policy response, countries can
promote a recovery that ensures more inclusive growth. We need strengthened education and the potential
of long-distance learning, more resilient and people-centred health care, housing support and specific
interventions to enhance personal safety of women and children, as well as support for communities and
regions left behind.
This edition of the Employment Outlook is – with the June 2020 Economic Outlook and the OECD Digital
Hub on Tackling the Coronavirus – part of the OECD’s response to the crisis, providing member and
partner countries with evidence and policy advice to weather the pandemic and to foster more resilient,
inclusive and sustainable growth.
COVID-19 has exposed weaknesses in our economies and societies that will hold people back unless they
are addressed. In times of crisis, ‘normality’ sounds very appealing. However, our normal was not good
enough for the many people with no or precarious jobs, bad working conditions, income insecurity, and
limits on their ambitions. We need to capitalise on the momentum created by the strong initial national
responses to the crisis, and build better policies for better lives in the post-COVID world.
Angel Gurría
OECD Secretary-General
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
5
Acknowledgements
The OECD Employment Outlook provides an annual assessment of key labour market developments and
prospects in OECD member countries. Each edition also contains several chapters focusing on specific
aspects of how labour markets function and the implications for policy in order to promote more and better
jobs. The 2020 edition is devoted to worker security and the COVID-19 crisis. Chapter 1 focuses on the
labour market consequences of the COVID-19 outbreak and the resulting economic crisis as well as on
the labour market and social policy response. Chapter 2 investigates the uneven access to unemployment
benefits for workers in part-time and less stable jobs, which often accentuates the hardship they face in
times of crisis. Chapter 3 provides a comparative review of employment protection legislation (EPL) across
OECD countries by developing a new version of the OECD EPL indicators. Chapter 4 takes a fresh look
at job polarisation, and in particular the hollowing out of jobs in middle-skill occupations. Finally, Chapter 5
examines the changing labour market outcomes for middle-educated vocational education and training
graduates.
The OECD Employment Outlook 2020 is the joint work of staff of the Directorate for Employment, Labour
and Social Affairs. The whole Outlook has also greatly benefited from comments from other OECD
directorates and contributions from national government delegates. However, its assessments of each
country’s labour market prospects do not necessarily correspond to those made by the national authorities
concerned.
This report was edited by Andrea Bassanini, and is based on contributions from Andrea Garnero and
Sebastian Königs (Chapter 1), Horacio Levy (Chapter 2), Oliver Denk and Alexandre Georgieff
(Chapter 3), Andrew Green (Chapter 4), and Glenda Quintini, Marieke Vandeweyer and Annelore
Verhagen (Chapter 5). Chapter 5 was produced with the financial assistance of the European Union. The
views expressed herein can in no way be taken to reflect the official opinion of the European Union. The
highlights (published separately) and the infographics are based on contributions from Emily Farchy and
Alastair Wood, respectively. Pascal Marianna was responsible for the statistical annex. Specific
contributions were provided by Willem Adema, Sandrine Cazes, Michele Cecchini, Chris Clarke, Jonas
Fluchtmann, Paul Grass, Alexander Hijzen, Raphaela Hyee, Herwig Immervoll, Kristine Langenbucher,
Anne Lauringson, Duncan MacDonald, Thomas Manfredi, Luca Marcolin, Maria Mecenero, Veerle
Miranda, Marissa Plouin, Christopher Prinz, Olga Rastrigina, Andrea Salvatori, Angelica Salvi del Pero,
Maëlle Stricot, Shunta Takino, Stefan Thewissen, Chloé Touzet, Theodora Xenogiani and Natasha
Yokoyama. Statistical assistance was provided by Sébastien Martin, Agnès Puymoyen and Dana Blumin.
Editorial assistance was provided by Natalie Corry, Liv Gudmundson, Lucy Hulett, and Niamh Kinane.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
6
Table of contents
Foreword 3
Acknowledgements 5
Editorial: From recovery to resilience after COVID-19 12
Executive summary 19
1 COVID-19: From a health to a jobs crisis 21
In Brief 22
Introduction 24
1.1. The outbreak of the coronavirus 25
1.2. The pandemic took an immediate and heavy toll on the economy and labour markets 28
1.3. An unprecedented policy response by countries 48
1.4. The way ahead – What is the right policy mix for post-confinement? 77
1.5. Concluding remarks 89
References 93
Annex 1.A. Additional material 105
Notes 110
2 Unemployment benefits and non-standard dependent employment: Striking the
balance between income security and work incentives 116
In Brief 117
Introduction 119
2.1. Trajectories of non-standard dependent employment 122
2.2. Are unemployment benefit rules adapted to non-standard employment? 130
2.3. Unemployment benefits: Income protection and financial work incentives 140
2.4. Policy issues: striking the right balance 155
2.5. Concluding remarks 158
References 163
Notes 166
3 Recent trends in employment protection legislation 168
In Brief 169
Introduction 170
3.1. How job protection matters for labour market and economic outcomes 171
3.2. The design of the 2019 OECD Employment Protection Legislation indicators 173
3.3. Employment protection legislation in OECD countries in 2019 178
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
7
3.4. Recent reforms in employment protection legislation 195
3.5. Concluding remarks 206
References 207
Annex 3.A. Methodology 212
Notes 217
4 What is happening to middle-skill workers? 221
In Brief 222
Introduction 223
4.1. How are middle-skill jobs changing? 224
4.2. What drives the fall in the share of middle-skill jobs? 226
4.3. Who were middle-skill workers? 233
4.4. Where are middle-skill workers going? 236
4.5. Concluding remarks 244
References 245
Annex 4.A. Data Sources 248
Annex 4.B. Attrition Transition Decomposition 249
Annex 4.C. Additional Figures 252
Notes 260
5 Smooth transitions but in a changing market: The prospects of vocational
education and training graduates 262
In Brief 263
Introduction 265
5.1. The importance of medium-level VET in the education system 266
5.2. Job quality and quantity 271
5.3. The labour market outlook for VET graduates 290
5.4. Improving the future-readiness of VET systems 297
5.5. Concluding remarks 303
References 305
Annex 5.A. Additional tables and figures 309
Annex 5.B. Overview of education programmes 313
Annex 5.C. Simulation model – Methodology 323
Notes 333
Annex A. Statistical annex 338
FIGURES
Figure 1.1. Restrictions to individual mobility and economic activities in OECD countries 26
Figure 1.2. Individual mobility fell in all OECD countries, even where restrictions were relatively milder 27
Figure 1.3. GDP fell substantially in the first half of 2020 29
Figure 1.4. US stay-at-home orders and workplace mobility 31
Figure 1.5. US stay-at-home orders and initial unemployment insurance claims 31
Figure 1.6. Unemployment, as measured in national surveys, has increased sharply only in some countries 33
Figure 1.7. The number of unemployment insurance claims increased substantially in some OECD countries 35
Figure 1.8. Participation in job retention schemes has been massive in some countries 36
Figure 1.9. The cumulated impact of the COVID-19 crisis on employment and hours of work is ten times
greater than during the global financial crisis 37
Figure 1.10. Online job postings have declined massively 38
Figure 1.11. The decline in online job postings by industry and skill group 39
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
8
Figure 1.12. While higher-earning workers often worked from home, lower-earning workers often had to stop
working 41
Figure 1.13. In the United States, people living in higher-income neighbourhoods sheltered at home earlier
and longer than people living in poorer neighbourhoods 42
Figure 1.14. Labour market expectations deteriorated among businesses and consumers alike 46
Figure 1.15. Unemployment is projected to increase three times more than during the global financial crisis 47
Figure 1.16. OECD countries introduced bold new measures or considerably expanded existing ones in
response to COVID-19 49
Figure 1.17. Between 30% and 60% of workers worked from home in mid-April 2020 51
Figure 1.18. Paid sick leave replaces large parts of eligible employees’ wages, with significant recent changes
in regulations in a number of OECD countries 55
Figure 1.19. Workers in non-standard jobs are often less well covered by income support 67
Figure 1.20. Schematic overview of the OECD SPHeP-COVID model 79
Figure 1.21. Typical impact of COVID-19 across OECD countries under different scenarios: the example of
Italy 81
Figure 1.22. Around half of workers are employed in jobs that entail some risk of infection 82
Figure 2.1. Distribution of employment trajectories across employees and unemployed 123
Figure 2.2. More workers experience trajectories of non-standard dependent employment 125
Figure 2.3. Most workers with unstable employment have temporary contracts 127
Figure 2.4. Part-time work is more frequent among women 128
Figure 2.5. Poverty rate by type of dependent employment 130
Figure 2.6. Employment requirements in unemployment benefits 132
Figure 2.7. Main characteristics of unemployment benefits, 2020 135
Figure 2.8. Simulation scenarios 143
Figure 2.9. Unemployment benefits in Australia 145
Figure 2.10. Unemployment benefits in France 146
Figure 2.11. Unemployment benefits in Latvia 147
Figure 2.12. Unemployment benefits in Spain 148
Figure 2.13. Non-standard dependent workers receive fewer unemployment benefits 149
Figure 2.14. The distribution of working hours affects unemployment benefit entitlements in France 150
Figure 2.15. Scenarios for computing adjusted replacement rates 153
Figure 2.16. Work incentives depending on current and previous employment trajectories 154
Figure 3.1. Procedural requirements for individual dismissals of regular workers 179
Figure 3.2. Notice period and severance pay for individual dismissals of regular workers 180
Figure 3.3. The role of job tenure for notice period and severance pay 181
Figure 3.4. Regulatory framework for unfair individual dismissals of regular workers 182
Figure 3.5. Enforcement of unfair dismissal regulation for individual dismissals of regular workers 184
Figure 3.6. Changes in the assessed strictness of regulation of individual dismissals 187
Figure 3.7. The OECD indicators: Strictness of regulation of collective dismissals (defined as dismissals of
several regular workers in one month) 189
Figure 3.8. Strictness of regulation of mass dismissals (defined as dismissals of at least 120 regular workers in
one month) 190
Figure 3.9. The OECD indicators: Strictness of regulation of dismissing regular workers 191
Figure 3.10. How taking into account collective agreements affects job protection 192
Figure 3.11. The OECD indicators: Strictness of regulation of hiring temporary workers 194
Figure 3.12. Dismissal regulation for regular workers and hiring regulation for temporary workers are positively
correlated 195
Figure 3.13. Quantifying recent reforms in employment protection legislation 196
Figure 3.14. Regulatory changes in recent reforms of dismissal regulation for regular workers 197
Figure 3.15. Regulatory changes in recent reforms of hiring regulation for temporary workers 198
Figure 3.16. The new version of the indicators tends to better capture reforms of dismissal regulations 200
Figure 4.1. Employment shares of middle-skill occupations declined sharply 225
Figure 4.2. Changes in separation rates were large during the global financial crisis, but have mostly returned
to pre-crisis levels 228
Figure 4.3. Outcomes for workers separating from middle-skill jobs remain stable 229
Figure 4.4. Propensities to work in different skill groups have been driven by differential entry of younger
cohorts 232
Figure 4.5. Middle-skill workers were workers without a tertiary degree 234
Figure 4.6. Male and female middle-skill workers were concentrated in manufacturing 235
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
9
Figure 4.7. Shrinking share of middle-skill employment due (slightly more) to diminished propensity to work
more than changing composition 238
Figure 4.8. The propensity to work in middle-skill occupations fell for all workers without a tertiary degree 239
Figure 4.9. Women without a tertiary degree more likely to work in low-skill occupations 240
Figure 4.10. Men without a tertiary degree are also increasingly in low-skill occupations 241
Figure 4.11. Middle-educated men and women are only slightly more likely to work in high-skill occupations 242
Figure 4.12. Propensities to work in different skill groups mostly driven by differential entry of younger cohorts
without a tertiary degree 243
Figure 5.1. VET is an important part of many education systems 267
Figure 5.2. Tertiary education is on the rise 269
Figure 5.3. The skill composition of graduates from the last 15 years has remained stable 270
Figure 5.4. Young VET graduates have relatively strong employment outcomes 272
Figure 5.5. VET graduates spend less time than general education graduates between the end of their studies
and their first job 273
Figure 5.6. The job quantity advantage for VET versus general education is smaller for older age groups 275
Figure 5.7. The job quantity advantage of young VET graduates relative to general education graduates has
remained stable in recent years 276
Figure 5.8. A small share of VET graduates work in high-skill occupations 277
Figure 5.9. The occupational composition of young graduate employment is changing 278
Figure 5.10. Craft jobs remain important for young VET graduates despite an overall declining importance of
those jobs in the labour market for young graduates 279
Figure 5.11. Young VET graduates have slightly higher wages than general education graduates 282
Figure 5.12. The wage advantage for VET vs. general education disappears for older age groups 283
Figure 5.13. Young VET graduates are less likely to be employed on a temporary contract than general
education graduates 284
Figure 5.14. The incidence of temporary employment is on the rise for all education groups 285
Figure 5.15. Young VET graduates are less likely to have a temporary contract than general education
graduates only at the start of their career 286
Figure 5.16. Young VET graduates are more likely than general education graduates to have supervisory
responsibilities at the start of their career, but this advantage disappears later on 287
Figure 5.17. Young VET graduates work physically for long periods more often than general education
graduates 288
Figure 5.18. Young VET graduates use their literacy, numeracy and problem-solving skills less intensively at
work than tertiary education graduates with similar skill levels 290
Figure 5.19. Occupations employing mostly below upper-secondary, general education or VET graduates are
more likely to face excess supply 291
Figure 5.20. Job creation in middle-skill occupations is projected to be negative in the EU in the period
2016-30 293
Figure 5.21. The use of certain ICT tasks increased significantly in the period 2012-17 in the United States 295
Figure 5.22. Young graduates with tertiary education degrees face the lowest risk of automation of their jobs 296
Figure 5.23. The impact of a burst of automation on the skill composition of VET graduates’ employment is
relatively limited 297
Figure 5.24. VET graduates have slightly lower literacy skill than general education graduates 300
Figure 5.25. Young tertiary education graduates are more likely to participate in training than graduates from
lower education levels 301
Annex Figure 1.A.1. Individual mobility fell in all OECD countries, even where restrictions were relatively
milder 106
Annex Figure 4.B.1. Attrition accounts for the majority of the decline in the share of middle-skill employment 249
Annex Figure 4.C.1. Education shares of the working-age population 252
Annex Figure 4.C.2. Alternative shift-share analysis 253
Annex Figure 4.C.3. Shares of low-education, prime-age workers in middle-skill employment 254
Annex Figure 4.C.4. Share of middle-educated, prime-age workers in middle-skill employment 255
Annex Figure 4.C.5. Share of low-educated, prime-age workers in low-skill employment 256
Annex Figure 4.C.6. Share of middle-educated, prime-age workers in low-skill employment 257
Annex Figure 4.C.7. Share of low-educated, prime-age workers in non-employment 258
Annex Figure 4.C.8. Share of middle-educated, prime-age workers in non-employment employment 259
Annex Figure 5.A.1. Occupational composition of graduates’ employment (below upper-secondary and
tertiary) 311
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
10
Annex Figure 5.A.2. Change in skills demand after automation, by occupation 312
INFOGRAPHIC
Infographic 1. Key facts and figures 18
TABLES
Table 1.1. Countries have adjusted existing job retention schemes or adopted new ones 61
Table 1.2. Nearly two in three OECD countries expanded unemployment benefits 69
Table 1.3. Countries across the OECD have taken measures to improve support for workers and households
not covered by unemployment benefits or job retention schemes 72
Table 1.4. Many countries introduced emergency housing measures in response to COVID-19 76
Table 2.1. Incidence of unstable employment, by socio-demographic characteristics and countries 126
Table 2.2. Summary of unemployment benefits rules that impact differently on non-standard employees 137
Table 2.3. Striking the right balance by adjusting unemployment benefit’s tools 157
Table 2.4. COVID-19 extraordinary unemployment benefit measures with potential impact on non-standard
employees 160
Table 3.1. The OECD Employment Protection Legislation indicators for dismissing regular workers 175
Table 3.2. The OECD Employment Protection Legislation indicators for hiring temporary workers 177
Table 3.3. The OECD indicators: Strictness of regulation of individual dismissals of regular workers 186
Table 4.1. Middle-skill employment held steady until the global financial crisis 227
Table 4.2. Middle-skill employment adjustment led by younger workers 230
Table 4.3. Middle-skill occupations varied greatly by gender 236
Table 5.1. Among occupation changers, VET graduates are less likely than general education graduates to
make positive occupation changes 281
Annex Table 1.A.1. Projected labour market developments in OECD countries 109
Annex Table 3.A.1. Structure of Version 4 of the OECD EPL indicators for dismissing regular workers 212
Annex Table 3.A.2. Weighting in the OECD EPL indicators (Version 4) for dismissing regular workers 215
Annex Table 3.A.3. Structure of Version 3 of the OECD EPL indicator for hiring temporary workers 215
Annex Table 3.A.4. Weighting in the OECD EPL indicators (Version 3) for hiring temporary workers 216
Annex Table 5.A.1. Personal characteristics of graduates, 2004 and 2018 (European countries) 309
Annex Table 5.A.2. Probability of participating in training 310
Annex Table 5.B.1. Education programmes, by education level and orientation 313
Annex Table 5.C.1. O*NET to ISCED-11 educational crosswalk 330
Annex Table 5.C.2. O*NET experience requirement categories 330
Annex Table 5.C.3. Conversion factors for O*NET database frequency scores 331
Annex Table 5.C.4. ISIC-Rev4 to CEDEFOP Sector crosswalk 331
Annex Table 5.C.5. O*NET Knowledge and Skill categories 332
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
11
Follow OECD Publications on:
http://twitter.com/OECD_Pubs
http://www.facebook.com/OECDPublications
http://www.linkedin.com/groups/OECD-Publications-4645871
http://www.youtube.com/oecdilibrary
Alerts
http://www.oecd.org/oecddirect/
This book has... A service that delivers Excel® files from the printed page!
Look for the StatLinks2at the bottom of the tables or graphs in this book.
To download the matching Excel® spreadsheet, just type the link into your Internet
browser, starting with the https://doi.org prefix, or click on the link from the e-book
edition.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
12
Editorial: From recovery to
resilience after COVID-19
What took more than a decade to achieve has unravelled within a matter of months. In early 2020 the
employment rate in the OECD reached a record-high of 68.9%, 2.6 percentage points above the previous
record just before the global financial and economic crisis of 2008. Then the pandemic struck. Within
months, COVID-19 spread around the globe triggering the worst public health emergency in a century. It
has sparked an economic crisis not seen since the Great Depression of the 1930s. More than 10 million
people have been infected with the virus, more than half a million people have died and trillions of dollars
have been pumped into the world economy to protect lives and livelihoods. In the face of this challenge, a
four Rs strategy, which progresses from response and rehabilitation to reciprocity and resilience, is needed
to re-build a better, more robust, and inclusive labour market.
The immediate response to the pandemic has been unprecedented in scale and scope. As countries move
out of lockdown, rehabilitation will be critical to protect many jobs. Reciprocity, with everyone contributing
to rehabilitation with a sense of responsibility, will also be key to the recovery. Last but not least, the
COVID-19 crisis has exposed gaps in the labour market that must be closed to boost resilience. With the
low-paid, the young, women, the self-employed and temporary workers among the hardest hit by the crisis,
the burden of the pandemic has been shouldered disproportionately by the most vulnerable.
Countries around the world have taken major steps to deal quickly with the crisis. On the public health
side, the primary objective has been to “flatten the curve” of the virus, contain the otherwise overwhelming
pressure on hospitals and ultimately save millions of lives. Intervention was swift. Many countries adopted
drastic containment measures, which resulted in an unprecedented – at least in peacetime – shutdown of
most non-essential activities, from kindergartens, to schools, factories and most shops and recreational
activities.
The combination of fear of infection, public guidelines and mandatory lockdowns and great uncertainty,
produced a sharp contraction in economic activity with a deep and widespread shock to the labour market.
An unprecedented number of workers (39% on average) shifted to telework, pushing the boundaries of the
potential for this alternative way of work organisation. Despite this, in all countries the number of those
effectively working collapsed much more than during any recent economic and financial crisis, as
companies in non-essential sectors laid-off workers, froze hiring and put most of their workforce on hold
through subsidised job retention schemes. By May 2020, companies had claimed job-retention subsidies
for more than 30% of their employees in countries such as Germany or the United Kingdom and up to 50%
in countries such as France and New Zealand. In the meantime, the OECD-wide unemployment rate rose
from 5.3% in January to 8.4% in May.
While the virus respects no borders or socio-economic groups, its spread has disproportionally affected
the most vulnerable, either directly because of greater difficulty in protecting themselves, or indirectly via
the impact of the lockdown on their jobs. Low-income workers are paying the highest price. As shown in
this OECD Employment Outlook, during the lockdown top-earning workers were on average 50% more
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
13
likely to work from home than those in the bottom quartile; the latter were more often employed in essential
services during the lockdowns and at risk of exposing themselves to the virus while working. At the same
time, low-income workers were twice as likely to have to stop working completely as their higher-income
peers were.
Workers in non-standard jobs – i.e. self-employed workers and those on temporary or part-time contracts
– have been particularly exposed to job and income losses. In contrast to the global financial crisis, women
have also been hit harder than men, as they are over-represented in the most affected sectors and
disproportionally hold precarious jobs, while more is being asked from them in the home. And the “Class
of Corona”, this year’s graduates, are leaving schools and universities with poor chances of finding
employment or work experience this summer or in the autumn.
RESPONSE: The “emergency” response to the pandemic has been unprecedented in scale and
scope. As the health and economic shock was unprecedented in terms of speed and virulence, so was
the policy response, with several trillion dollars quickly committed globally to sustain individuals,
households, and companies. Beyond providing direct and indirect financial support to companies, the vast
majority of OECD countries have strengthened and/or extended income support to workers unable to work
or who are jobless. Many extended or introduced job retention schemes at firms suffering from a temporary
reduction in business activity, thereby avoiding severing labour contracts, which would have resulted in
the destruction of valuable competences and viable investment. Many countries also introduced or
strengthened sick pay, including for quarantined workers, and took measures to address unforeseen care
needs for working parents.
Despite the massive measures taken around the globe, uncertainty about future labour market
developments is large, as the risk of new outbreaks is high. Much of what will happen depends on the
evolution of the pandemic. The results of an epidemiological model the OECD developed during the crisis
suggests that the strict confinement measures introduced in many countries were successful in containing
the number of fatalities. Moreover, model simulations indicate that a second wave can be avoided even in
the absence of a vaccine. This requires putting in place a package of comprehensive public health
interventions, ranging from massive upscaling of testing, tracking and tracing (TTT), to enhancing personal
hygiene measures, to ensuring wide use of masks and the continuous enforcement of some physical-
distancing policies such as banning large gatherings and encouraging people to work from home.
Given the uncertainty about the evolution of the pandemic, the latest OECD Economic Outlook presents
two possible, equally probable, scenarios: one where the virus outbreak continues to recede and remains
under control, and one where a second wave of rapid contagion erupts later in 2020. Even under the single-
hit scenario, world economic output is forecast to plummet by 6% this year, before climbing back by 5.2%
in 2021. The outlook would be much worse with a double-hit scenario. In the most optimistic scenario, the
OECD-wide unemployment rate is forecast to be 9.4% in the fourth quarter of 2020, exceeding all the
peaks since the Great Depression, while average employment is projected to fall by 4.1% to 5% with
respect to 2019, depending on whether a second outbreak materialises.
Responding swiftly to the huge challenges imposed by the sudden lockdown required a Herculean effort
on the side of governments across OECD countries and beyond. As the economy re-opens, policy must
lead the labour market and society along the road to rehabilitation. But, adapting this package of measures
to the new situation of a gradual and managed re-opening is not any easier, and will require reciprocity
and responsibility from all stakeholders.
REHABILITATION: In the short-term continued support for some sectors remains vital to protect
jobs and wellbeing, but labour market mechanisms must re-start operating. Accompanying the
labour market during the gradual scaling back of confinement measures requires a two-pronged approach.
First, labour market policy must support the effort of preventing a second severe pandemic wave and
preparing for that in case it materialises. Teleworking remains, for many, an effective way to work while
limiting risks of contracting the virus. Evidence shown in this Employment Outlook suggests that, on
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
14
average, about one third of jobs can be done from home under normal conditions. Enhancing the use of
teleworking requires not only facilitating employer-employee arrangements but also investing to make sure
that workers have the instruments to work from home under good conditions (computer or tablet,
broadband connection, room to work undisturbed etc…). It will also require planning work organisation, in
particular in the case of a second pandemic wave, and training the workforce to make the most of
teleworking.
Almost two-thirds of jobs cannot, or can hardly, be performed from home. Some of them have a limited
risk of infection as they involve no or infrequent physical interactions (e.g. plumbers, truck drivers, or
archivists). However, almost one-half of all jobs require frequent interactions and, in the absence of
precautions, carry some risk for workers being infected at work (as exemplified by the large number of hot
spots that have developed in meatpacking plants). Therefore, developing and adapting rigorous
occupational safety and health standards remains a policy priority. Moreover, continuing to guarantee
extensive paid sick leave will remain crucial, so that potentially infected workers do not spread the virus at
work.
Second, as the re-opening of the economy unfolds and activity restarts, labour market and social policy
should be adapted to reflect the varying conditions of workers, households, and companies. During the
lockdown, a broad one-size-fits-all support strategy was justified, as most activities were simply prevented
from operating and companies and jobs would not have survived without immediate support. Now, policy
makers are facing the difficult task of moving the economy from emergency action, with massive,
generalised support, to recovery, where support needs to be differentiated according to the conditions of
firms, sectors, and workers.
Firms and workers in sectors that are still prevented from operating – such as parts of the entertainment
industry – should continue to be supported, at least temporarily, to increase their chance of resuming work.
However, where activities can resume the market mechanism should re-start operating, allowing for
workers and resources to move from unviable to promising activities.
Measures should be targeted better to ensure that those in need really get help, while fostering the
incentives to go back to work for those who can. This is necessary to avoid the scars of prolonged
joblessness and inactivity, on the one hand, and to ensure the sustainability of policy interventions, on the
other hand. A clear example of the need to adapt the policy intervention is provided by job-retention
schemes. For sectors where activity have resumed, firms should be required to carry part of the cost of
the job retention scheme. To avoid reinforcing financial difficulties of firms, employers’ participation can
take the form of a delayed-payment or zero-interest loan. In addition, stricter limits on the duration of
subsidies and incentives to look for work, combine temporary secondary jobs and short-term subsidies,
and take up training are among the policy levers that policy-makers and social partners should consider in
coming months.
As prospects of quickly finding new work will remain poor for many, some countries should extend
unemployment benefit durations to prevent jobseekers from sliding too quickly into much less generous
minimum-income benefits. This will be even more necessary in the case of a second wave of infections
and renewed restrictions to economic activity. Emergency support for the self-employed should also be re-
assessed, in order to improve targeting, restore incentives and ensure fairness. More generally, the
duration, targeting and generosity of all the income support programmes put in place in the early months
of the crisis should be re-examined to ensure that they are sustainable, their effects on work incentives are
minimised, and they guarantee that support goes to the most needy. Public and private employment
services will also face the daunting challenge of serving a high number of jobseekers with differing
conditions. Their capacity will have to be scaled up to avoid permanently neglecting functions that may
have been of secondary importance during the emergency phase of the crisis (e.g. career advice,
counselling).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
15
Implementation and delivery of this complex package will be crucial, however. During the crisis, many
people have waited for too long to receive the help they need and were entitled to. New programmes found
themselves entangled in a mass of administrative yarn and took too long to reach beneficiaries.
Newspapers have been filled with examples of companies going bankrupt before receiving promised
subsidies, displaced workers applying for unemployment benefits but not having received them after
several weeks, and even children not receiving lunches replacing those of locked down school canteens.
RECIPROCITY AND RESPONSIBILITY: In both the short and the long-term, all parts of society need
to contribute to this rehabilitation with sense of responsibility, in particular those who have
received, or still receive, public support.
All actors in the economy should play their role in rebuilding a better labour market. Reciprocity is needed
between public support for struggling firms and industries and private sector support for efforts to help the
unemployed return to work, boost employees’ skills and ensure no one is left behind in a recovery. This
particularly applies to those firms that receive or have received job retention and other subsidies, but all
firms must strive for the reconstruction of a dynamic labour market. Hiring and re-hiring, investment in new
technologies and in training for the workforce, and/or continued participation in apprenticeship programmes
should take a central role in corporate decisions. Time-limited hiring subsidies have proven quite effective
at supporting job creation, notably in bad times, while minimising the administrative costs of monitoring
eligibility requirements on take-up (e.g. by allowing recapturing credits when job creation goals are not met
or considering refundable hiring credits, as done by certain US states during the global financial crisis).
A similar argument applies to individuals receiving income support. For example, a priority will be restoring
the “mutual obligations” approach, in which governments commit to providing jobseekers with benefits and
effective employment services and, in turn, beneficiaries have to take active steps to search for work or
improve their employability. This is key to mobilise jobseekers to find viable jobs.
RESILIENCE: The COVID-19 crisis has shown more than ever the need to strengthen resilience and
inclusivity in the labour market. In the medium term, countries should address the structural problems
that the crisis has put under the spotlight. As stressed in the OECD Jobs Strategy, effective economic
resilience requires counter-cyclical macroeconomic policies, adequate income support for all workers,
rapid expansion of job-retention schemes during crisis, and effective social dialogue.
The COVID-19 crisis has laid bare pre-existing gaps in social protection provisions. In many countries, the
insurance function of social protection works well for employees with stable work histories. But, as shown
in this Employment Outlook, even if entitlement rules are usually the same for all dependent employees,
conditions on minimum employment duration or earnings before the unemployment spell are often harder
to meet for those who lose a part-time job or have unstable or short employment histories. The self-
employed and other non-standard workers are often poorly protected or not protected at all. At the same
time, the assistance function of social protection systems – providing last-resort minimum-income benefits
for those with little or no other resources – has been put to a severe test. The emergency has prompted
decisive actions to reduce these gaps in social protection. The challenge now is to build on these initiatives,
and transform temporary fixes into structural changes.
Workers in non-standard forms of employment need to be able to build up rights to the types of out-of-work
support that are already available to standard employees. While including self-employed in earnings-
related social-protection schemes can be fraught with moral hazard and other logistical and administrative
concerns, several countries have already been successful in establishing well-designed policies that work
for their circumstances. For instance, a number of OECD countries do include the self-employed in their
unemployment and sickness insurance schemes. A more equitable treatment of different forms of
employment can help minimise future needs for makeshift programmes – that are necessarily less targeted
and cost-effective and can be prone to leakage.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
16
Even with well-designed social insurance schemes in place, providing a minimum level of assistance
to those in need is a basic function of social protection systems. Yet, even in normal times, the
accessibility, the reactivity, and the generosity of these programmes differ markedly acr oss countries.
In many cases, complex criteria and claim procedures result in low take -up and receipt rates, long
waiting periods, and sometimes inadequate levels of support. One-off or temporary lump-sum
transfers, as introduced by many countries during the COVID-19 crisis, have played a role in providing
fast support to those in needs. But beyond the short-term, as fiscal pressures mount, sustainable and
effectively targeted programmes will be needed. Making minimum -income protection more responsive,
through timely reassessment of entitlements in the face of rapidly changing circumstances remains
an urgent policy priority.
Strengthening labour market resilience also requires stronger institutional capacity to scale up key
measures quickly, while maintaining service quality. This implies that when a crisis hit, the policy
infrastructure should already be in place and can be scaled up quickly. Evidence suggests that
implementation and delivery failures during the COVID-19 crisis were more common where emergency
solutions had to be created from scratch.
Reconstructing a better and more resilient labour market is an investment in the future and future
generations. We cannot afford losing the Corona Class generation. In the aftermath of the global
financial crisis, governments acted far too late to address the labour market difficulties of youth, which
left them with long-lasting scars that were still visible before the COVID-19 outbreak. There is no time
to waste to put in place a comprehensive policy package ensuring that no young worker is left behind.
Everybody should have a route to follow (such as, e.g. the EU Youth Guarantee). Every actor must,
again, play its role with responsibility and reciprocity: companies, for example, should be encouraged
to provide opportunities for work experiences by hiring new graduates or offering apprenticeships,
internships or work-related training, while governments should accompany them with specific financial
incentives.
A comprehensive recovery plan should include, the expansion of cost-effective active labour market
measures – such as counselling, job-search assistance, entrepreneurship programmes. Extending support
for vocational education and training (VET) would also be crucial. As shown in this Employment Outlook,
the transition from school to work of non-tertiary VET graduates remains much easier than that of their
general-education peers. Yet, it is important to make sure that these programmes remain responsive to
changing labour market needs.
Social dialogue and collective bargaining have a key role to play in enhancing the resilience of the
labour market. When social partners work co-operatively, this flexibility and granularity could allow
adapting and deploying more rapidly the required responses through tailor-made agreements and
work re-organisations that are adjusted to meet each specific situation. In many countries, for
example, collective bargaining and social dialogue have recently proved instrumental in ensuring safer
workplaces. The guidelines and codes of good conduct established by social partners and the
agreements signed between employers and trade unions in this area in various countries
(e.g. Denmark, France, Italy and Spain) are excellent examples of how social dialogue and col lective
bargaining can be mobilised to complement public action and find flexible and tailored solutions for
both companies and workers.
Countries should harness the lessons of this crisis and plan for a thorough assessment of labour
market resilience, drawing on the OECD Jobs Strategy framework. This complex exercise will have
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
17
to involve all stakeholders and lead to the identification of country-specific policy packages to enhance
resilience within a more inclusive labour market.
It is not the time to rebuild the old. It is time to build better.
Stefano Scarpetta
Director for Employment, Labour and Social Affairs
OECD
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
18
Infographic 1. Key facts and figures
Unemployment rat es soared in just Job ret ent ion schemes have played a
a few mont hs massive role in some count ries
0 % 3 % 6 % 9 % 12 % 15 %
55%
Sp ai n
Can ad a 4 5%
Un it ed St at es
Sw ed en
30 %
OECD 25%
Fran ce
18 %
16%
Ital y
11%
Germany
<1%
Fran ce Italy G erm a ny OECD Spai n C anad a Sw ede n U nit e d
St at es
The number of hours w orked
has plummeted
%
2 out of 5 w orkers w ere able to w ork f rom home
in April 20 20 across t he OECD .
Stopped
w o rking
14%
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
19
Executive summary
The world is facing one of the worst public health and economic crises in a
century
The most serious pandemic in a century has triggered one of the worst economic crises since the Great
Depression. Countries reacted with often strict containment and mitigation policies, which effectively limited
the spread of the virus and avoided the collapse of health care systems and most importantly limited the
number of fatalities. The combination of great uncertainty, fear of infection, individual restraints following
public guidelines and mandatory lockdowns, however, immediately produced a sharp contraction in
economic activity. In the first months of the crisis, new unemployment claims have soared in many
countries and projections suggest that in the OECD area the unemployment rate will be much higher than
at the peak of the global financial crisis. But the extent of the shock on the labour market is much larger:
despite a massive shift towards telework, in all countries the number of those effectively working collapsed
as companies have frozen hiring and put part of their workforce on hold through subsidised job-retention
schemes. Available evidence also suggests that vulnerable groups – the low skilled, youth and migrants –
as well as women are paying the heaviest toll of the crisis.
The labour market and social policy response has been unprecedented
OECD countries have taken massive steps to improve access to, and the generosity of, sick leave and
out-of-work income support as well as job retention schemes, whose take-up has been unprecedented in
many countries. These policy responses were aimed at containing damage and supporting workers and
companies as well as at avoiding destruction of viable activities and competences, thereby preparing the
recovery. Many countries also took steps to facilitate a massive transition towards teleworking for workers
who do not have to be physically present at their workplace. Keeping workers safe as the economy reopens
and ensuring adequate income protection and employment support for a crisis that may not yet be over
must remain a priority. As economic activity picks up, however, policy must accompany the recovery by
striking the right balance between providing continuous support to workers, households and companies
still affected by persisting restrictions and encouraging business activity as well as permitting necessary
restructuring.
Unemployment benefit rules must account for the specific trajectories of
employees in non-standard jobs
Unemployment benefits are among the key instruments providing protection against earnings falls resulting
from job losses. Yet a number of workers do not meet the criteria to receive adequate support. Even if
entitlement rules are usually the same for all dependent employees, conditions on minimum employment
duration or earnings before the unemployment spell are often harder to meet for those who lose a part-time
job or have employment trajectories involving frequent transitions between employment and
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
20
unemployment. Consequently, the risk of falling into poverty is often greater for non-standard employees.
Addressing inadequacy of benefit entitlements so as to provide greater income security for those in
non-standard forms of employment may be challenging, but several policy instruments are available to
create a policy mix that strikes the right balance between work incentives and income security for
non-standard dependent workers.
Employment protection legislation is key for worker security
Dismissal and hiring regulations are important determinants of people’s job security, career path and well-
being. They influence the extent to which firms take into account the social cost of dismissals and how
they react to technological and demand shocks. The OECD has updated and improved its detailed
indicators of employment protection legislation. Comparisons of countries highlight the differences
between employment protection systems. English-speaking countries have fewer restrictions on
dismissals than many European Union countries, for example. This puts employees in English-speaking
countries at a higher risk of job loss, but also gives them a greater chance of finding a job again if laid off.
To limit labour market duality and segmentation, countries with strict job protection for regular workers
usually have strict hiring laws for workers on temporary contracts. Several countries with apparently strict
dismissal regulations reduce their effective stringency by offering unemployment benefits even when the
worker agrees with the firm on the separation. Similarly, advance validations initially impose higher hurdles
for dismissals, but can serve to avoid disputes later.
Job polarisation is mostly due to fewer younger workers entering middle-skill
jobs than to older workers leaving them
In contrast to popular perceptions, the decline in the share of middle-skill employment is due primarily to
fewer younger workers entering middle-skill occupations than to mid-career workers being displaced and
leaving them. Since the 1990s, successive cohorts of young workers have been increasingly less likely to
enter the labour market in middle-skill jobs – e.g. truck drivers and machine operators for men, cashiers
and secretaries for women. Meanwhile, labour market trajectories of older cohorts after labour market entry
have remained essentially unchanged. These career patterns for younger cohorts can be partially
explained by the changing education and cohort demographic make-up. Individuals who would have been
once regarded as “typical” middle-skill workers are now less likely to be working in middle-skill jobs, and
more likely to be in low-skill employment. This trend has been marked for workers with a middle level of
education.
Graduates from vocational education and training have strong labour market
outcomes at the start of their career, but challenges are in sight
Many vocational education and training (VET) programmes organised at the upper secondary or
post-secondary non-tertiary education level prepare students for middle-skill jobs that have been exposed
to structural changes and face a significant risk of automation. Despite these challenges, VET graduates
usually have higher employment rates and better working conditions in the first years after graduation than
their general education peers. To ensure that VET continues to have a positive impact on students’ labour
market outcomes in a changing world of work, however, VET systems need to adapt to the rapidly evolving
skills demand. Close co-operation with social partners is crucial, as is the investment in transversal skills
in VET programmes and the development of smooth pathways between mid-level VET and higher
education.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
21
1 COVID-19: From a health to a jobs
crisis
The COVID-19 outbreak and its rapid diffusion across the globe have
turned into the worst public health crisis in living memory. The pandemic
forced countries to impose strict containment and mitigation policies and
severely affected social and economic activities, driving the global economy
to a major recession. Most countries responded quickly and put in place,
from the very first stages of the crisis, an unprecedented package of labour
market and social policies aiming at reducing the economic shock and
supporting workers, their families and companies. This chapter provides a
first assessment of the initial labour market impact of the COVID-19 crisis
and a review of countries’ wide set of policy responses. It also provides
some reflections on how countries could adapt the measures taken in the
first months of the crisis as they start softening mitigation policies.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
22
In Brief
Key findings
In a matter of a few months, the COVID-19 pandemic has turned from a public health crisis with no
parallel in living memory into a major economic and jobs crisis whose full extent is still unfolding. The
containment and mitigation strategies rapidly put in place to slow down contagion and avoid the collapse
of health care systems succeeded in limiting the spread of the virus and the associated fatalities. Even
where such confinement measures were not adopted, citizens largely assumed similar practices;
working from home where possible, while avoiding large gatherings, public transport and in-store
shopping. The unfolding pandemic led to a major “supply shock” as international supply chains were
interrupted, workers got sick, were quarantined or subject to lockdowns and companies found
themselves unable and, in some cases, forbidden to operate. Despite an unprecedented policy response
by governments and central banks, increased uncertainty, the decline in household incomes and
mandated or self-imposed physical-distancing measures led to a drop in investment and consumption.
This quickly turned what was initially a “supply shock” into a “demand shock”, putting further pressure
on companies. While economic and labour market conditions are evolving rapidly, the available
evidence on the initial impact of COVID-19 on OECD labour markets at the time of writing shows that:
Individual mobility to places of work as well as to public-transport hubs decreased drastically in
March and April, not only in countries that enforced strict mandatory measures but also where
governments relied more on public information and communication to drive changes in individual
behaviour (e.g. physical distancing and enhanced hygiene). Hence, the dramatic fall in
economic activity is the result not only of mandatory restrictions but also reflects people’s
reactions to non-binding recommendations and their greater awareness of the seriousness of
the pandemic.
The initial impact of the COVID-19 crisis on OECD labour markets where data are available has
been ten times larger than that observed in the first months of the 2008 global financial crisis:
taking into account both the drop in employment and the reduction in hours worked among those
who remained in work, total hours worked fell by 12.2% in the initial three months compared to
1.2% in 2008. This reflects the special nature of the COVID-19 crisis with many countries having
put entire sectors of their economy “on hold” to contain the spread of the virus.
Countries’ initial unemployment response to the COVID-19 crisis has varied starkly. In a few
countries, unemployment immediately jumped to record levels, while in others, it increased only
modestly so far, or not at all. This striking heterogeneity largely reflects differences in policy
responses. Few countries rely almost uniquely on unemployment benefits to secure the incomes
of job losers. Others have made massive use of job retention schemes, i.e. public support for
cutting the hours of work or furloughing workers, while keeping them employed. Cross-country
differences in the classification of “workers not at work” (because of short-time work) and those
on “temporary layoff” also contribute to disparities in measured unemployment.
Even in countries with comprehensive job retention schemes and those that banned or restricted
dismissals, however, the number of jobseekers increased, as temporary contracts were not
renewed and firms’ hiring activities collapsed: online job postings fell by 35% between February
and May in the United States and in European OECD countries.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
23
Given the exceptional uncertainties characterising the near-term outlook, the OECD considers two
epidemiological scenarios for the coming 18 months: one where the virus continues to recede and
remains under control, and one where a second wave of rapid contagion erupts later in 2020.
According to OECD projections, unemployment is set to increase to 9.4% on average across the
OECD by the end of 2020 (up from 5.3% at the end of 2019). In the event of a second pandemic
wave in late 2020, the unemployment rate would increase even further to 12.6%. Moreover,
projections point to only a gradual recovery: the unemployment rate is set to remain at or above
the peak level observed during the global financial crisis, reaching 7.7% by the end 2021 without
a second wave (and 8.9% in case of a second wave), with substantial differences across countries.
Vulnerable workers are bearing the brunt of the crisis. Low-paid workers have been key to
ensure the continuation of essential services during lockdowns, often at a substantial risk of
exposing themselves to the virus while working. They have also suffered greater job or income
losses. Workers who are not in standard (i.e. permanent, full-time dependent) employment,
notably the self-employed, have been particularly exposed to the shock. Young people risk being
once more among the big losers of the current crisis. This year’s graduates face bleak prospects,
with poor chances to secure a job, or even an internship, in the short run; their older peers are
experiencing the second heavy crisis in their still young careers. Women have so far experienced
greater declines in employment than men, unlike in the previous crisis. Meanwhile, widespread
childcare facility and school closures likely amplified their unpaid work burden at home.
OECD countries have responded with unprecedented measures to contain damages and support
workers, their families and companies. Beyond providing direct and indirect financial support to
companies, the vast majority of OECD countries have strengthened income support to workers losing
their jobs or income. Many extended or introduced job retention schemes to preserve jobs at firms
suffering from a temporary reduction in business activity. Some also tightened dismissal regulation or
facilitated hiring or renewal of workers on temporary contracts. Given workers’ risk of exposure to
COVID-19 at the workplace or their need for greater flexibility to work from home as schools and care
facilities were closed, many OECD countries also took steps to facilitate teleworking for workers who do
not have to be physically present at their workplace. Most of them also strengthened paid sick leave,
including, in some cases, to quarantined workers and took measures to address unforeseen care needs
for working parents affected by childcare facility or school closures.
As the pandemic started to recede, countries have started loosening containment policies, but solving
the health crisis remains the essential precondition for solving the economic and jobs crisis. A vaccine
may take time to be developed, produced and distributed. However, a second wave can also be avoided
by upscaling testing, tracking and tracing (TTT), enhancing personal hygienic measures and continuing
to enforce physical-distancing policies such as banning large gatherings and encouraging people to
work from home. During this post-confinement phase, some of the policies taken in the early months of
the crisis will need to be adapted and, in some cases, differentiated to account for the large
heterogeneity in conditions across sectors, firms and workers. In particular:
About half of all workers are employed in a job that requires significant physical interactions and
therefore face a risk of contagion. Strong occupational safety and health standards, defined and
enforced by public authorities and/or by social partners, remain a top priority.
Paid sick leave can continue to perform an important role in containing and mitigating the spread
of the virus and protecting the incomes, jobs and health of sick workers and their families. Moving
forward, countries should consider closing long-known gaps in paid sick-leave regulations while
reinforcing work incentives and employment support to facilitate a return to work.
Job retention schemes should be targeted only to those jobs that are at risk of being terminated
but viable in the medium/longer term. Requiring firms to carry part of the cost, stricter limits on
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
24
duration and incentives to look for work and take up training are some policy levers that policy
makers and social partners can mobilise in the coming months to this goal.
The coverage and adequacy of income support will need to be reviewed as the crisis evolves. If
weak labour market conditions persist, there can be good reasons for extending unemployment
benefit durations to prevent jobseekers from sliding too quickly into much less generous
minimum-income benefits. The duration, targeting and generosity of emergency income support
programmes put in place in the early months of the crisis should be re-assessed to ensure that
support goes to the most needy. “Mutual obligations” requirements, which commit jobless benefit
recipients to active efforts to find employment, should be progressively re-established where
they have been temporarily suspended.
Jobseekers also need assistance in finding new work. Public and private employment services
need to scale up their capacities and make larger use of digital services without giving up
standard in-person meetings with people with weaker digital skills. Online and offline training
can help jobseekers, as well as workers in job retention schemes, find jobs in sectors and
occupations more in demand and counter the risk of long-term unemployment. Hiring subsidies,
in particular if targeted to low-pay workers, can promote job creation.
Countries need to act quickly and help young people maintain their links with the labour market.
Support for companies who offer jobs or work experience to young people have proven effective
at promoting job creation. Effective outreach is crucial to re-establish contact with young people
who lost their jobs or left school without finding employment. Youth Guarantees, which entitle all
young people to a timely employment or training offer, can provide a good framework for
ensuring that no jobless young person goes without support in the current crisis.
Introduction
The outbreak in late 2019 of a novel form of coronavirus responsible for the severe respiratory disease
COVID-19, and its rapid diffusion across the entire globe, has turned into a public health crisis with no
parallel in living memory and has driven the global economy into the deepest recession since the Great
Depression. To contain the spread of the virus and its deadly effects, many countries around the world
introduced unparalleled – at least in peacetime – limitations to individual mobility and economic activities
in first half of 2020. These measures appear to have succeeded at limiting the contagion across OECD
countries. However, the combination of great uncertainty, fear of infection, individual restraints following
public guidelines and mandatory lockdowns immediately produced a sharp contraction in economic activity
and tested the resilience of labour markets, social-protection systems and societies at large.
Unlike during the global financial crisis of 2008, OECD countries reacted quickly to put in place, from the
very first stages of the crisis, an unprecedented set of fiscal and monetary policies. These measures were
necessary to contain the employment and social effects of the crisis, but also to provide people and
companies with the right incentives, and support, to comply with the restrictions that governments
mandated or recommended.
Despite these measures, the immediate impact on OECD labour markets has so far been multiple times
greater than during the first months of the global financial crisis and much more severe than what
unemployment statistics in some countries may suggest so far. Its effects are unlikely to fade away rapidly,
as the supply shock has quickly turned into a demand shock, and as economic activity in many sectors
remains subdued. Moreover, now that countries started loosening containment policies and moving to a
“new normal”, policy makers face the daunting task of moving the economy from “intensive care”, with
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
25
massive support, to “long-term care”, where support has to be differentiated according to the conditions of
sectors, firms and workers.
This chapter provides a first assessment of the initial labour market impact of COVID-19 as well as of
OECD countries’ unprecedented policy responses. It also attempts to provide some first reflections on how
countries could adapt the measures taken during the first months of the crisis to the gradual
post-confinement phase.
The chapter is organised as follows. Section 1.1 briefly describes the outbreak of the virus and the series
of restrictions that countries put in place to limit individual mobility and economic activity. Section 1.2
provides a first assessment of the impact of COVID-19 on the labour market in OECD countries, as well
as an outlook ahead based on the latest OECD projections. Section 1.3 describes OECD countries’ initial
policy response, while Section 1.4 attempts to provide – in a highly uncertain context – a discussion of how
the policy mix could evolve during a period of gradual post-confinement. This chapter heavily draws on a
series of policy briefs on labour market, social-policy and health issues released since the start of the
pandemic. They can be found on the OECD Digital Hub on Tackling the Coronavirus (COVID-19), under
http://www.oecd.org/coronavirus/en/.
1.1. The outbreak of the coronavirus
In late 2019, the city of Wuhan, located in the Hubei province of China, experienced an outbreak of
pneumonia from a novel coronavirus – the severe acute respiratory syndrome coronavirus 2
(SARS-CoV-2), which causes the infectious Coronavirus disease 2019 (COVID-19). Since these initial
cases, the number of confirmed COVID-19 cases has grown rapidly and spread to most countries and
territories across the world. Globally, there are now more than 10 million confirmed cases, and more than
500 000 deaths have been registered.1 On 11 March 2020, the World Health Organization declared a
pandemic2 and countries started to put in place an unprecedented set of measures restricting mobility and
economic activity to “flatten the curve”, avoid the collapse of their health care systems and ultimately
contain the number of fatalities (OECD, 2020[1]). The containment and mitigation strategies3 ranged from
stronger efforts to detect cases early on and trace contact with other people to severe physical-distancing
measures, including full national lockdowns and the shutdown of the economy, except for a number of
“essential activities”.
By the first half of April 2020, 90% of OECD countries had imposed some form of non-pharmaceutical
interventions (i.e. restrictions to individual mobility as well as economic activities) to contain the spread of
the virus (Figure 1.1): most OECD countries closed schools, restricted travel across but also within
countries and banned public gatherings. The exact nature and scope of these measures varied
substantially (Hale et al., 2020[2]). In some countries, such as Italy, New Zealand and Spain, the restrictions
were mandatory and applied to the entire national territory. In others, such as Mexico and Sweden,
restrictions were recommended but not imposed and they were limited to specific areas/groups.
Restrictions to economic activity also varied: in a few countries all non-essential firms were closed while in
others the restrictions applied only to activities or sectors bringing many people together such as
entertainment and accommodation.
Containment and mitigation policies have had an immediate effect on mobility patterns in all countries. As
governments issued mandatory restrictions and/or invited their citizens to reduce physical contacts,
individual mobility began to decline as people started sheltering at home. Figure 1.2 depicts data based
on smartphone locations4 in a group of selected OECD countries (all other OECD countries can be found
in Annex Figure 1.A.1). It shows that even in countries where restrictions were more limited, such as
Sweden, movements to places of work and to public-transport hubs decreased markedly between the
beginning and the end of March. The only notable exception is Korea, which, since the beginning, put in
place a strategy of rapid and massive testing, tracking and tracing (TTT) to contain the spread of the virus
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
26
without stopping economic activity (OECD, 2020[3]). As Figure 1.2 suggests, mandatory measures explain
only in part the decline in mobility observed across countries. According to Maloney and Taskin (2020[4]),
most of the decrease in mobility reflects the local and national COVID-19 case incidence and the resulting
higher awareness, or fear, or social responsibility.
Figure 1.1. Restrictions to individual mobility and economic activities in OECD countries
Percentage share of the total number of OECD countries with countries grouped by their mitigation strategy
A. Confinement and lockdown
None
Heavy restrictions on movement and activities in limited areas
Broad areas or population groups are locked down
Restrictions in many areas or population groups, with broad exceptions e.g. based on health or behavioural conditions
Nationwide lockdown, affecting many but not all groups or regions, or with many exceptions
Nationwide lockdown, with very limited exceptions
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
B. Shutdown of economic activities
None (no specific restrictions)
Most activities permitted apart from those that bring many people together
Most manufacturing and construction permitted plus selected discretionary services if they comply with enhanced sanitary standards
Businesses closed except for 'essential' production and service activities
Closure of all physical sites except for very limited 'essential' production and service activities
Complete closure of all but very limited 'essential' firms
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Note: The categories in legend are mutually exclusive, i.e. each country falls into just one category.
Source: OECD COVID-19 country policy tracker, https://www.oecd.org/coronavirus/en/#country-tracker (accessed on 8 June 2020).
StatLink 2 https://stat.link/hwzxmr
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
27
Figure 1.2. Individual mobility fell in all OECD countries, even where restrictions were relatively
milder
Percentage change in mobility relative to the median value during the 5-week period 3 Jan – 6 Feb 2020
Places of work Public transportation
Germany Italy
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Korea New Zealand
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Sweden United States
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Source: Google LLC “Google COVID-19 Community Mobility Reports”, https://www.google.com/covid19/mobility/ (accessed on 8 June 2020).
StatLink 2 https://stat.link/zs6e49
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
28
1.2. The pandemic took an immediate and heavy toll on the economy and labour
markets
The “physical distancing” resulting from voluntary restraint on mobility and/or mandatory containment and
mitigation strategies was effective in reducing the spread of the virus and avoiding a collapse of health
care systems that, in turn, would have resulted in a much higher death toll (Deb et al., 2020[5]). However,
the associated shutdown of entire sectors of the economy or, in some cases, even just the great uncertainty
and people’s fear of infection have had an immediate and dramatic effect on OECD economies and labour
markets.
At first, the COVID-19 pandemic caused a “supply shock”. The spread of the virus interrupted international
supply chains, first with China and then across most countries and regions, and reduced workers’ hours
worked as they were quarantined, sick or subject to lockdowns. Companies found themselves forced to
suspend or scale down operations, because of mandated shutdowns, because demand dropped as people
started sheltering at home or because they could not ensure safety and health conditions for their
employees. Many firms started facing liquidity constraints, and some lost capacity to continue paying their
employees’ wages. Despite unprecedented government interventions, the uncertainty about the spread of
the virus as well as, in many cases, the reduction in households’ disposable income led people and
companies to reduce investment and consumption and save more. The “supply shock” quickly turned into
a “demand shock”.
The impact on economic growth was immediate and heavy: GDP substantially dropped in the first quarter
of 2020 even though most OECD countries put in place their containment measures only in the second
half of March. Figure 1.3 shows that between the last quarter of 2019 and the first quarter of 2020, GDP
fell by 7% in Iceland, 5.3% in France and Italy, 5.2% in Spain, 3.7% in the Euro area, 2.2% in Germany,
2.1% in Canada and 1.3% in Korea and the United States and 0.6% in Japan. Chile is the only OECD
country where GDP increased significantly in the first quarter of 2020 compared to the last quarter of 2019.
This likely represents an economic rebound in the first couple of months of the year after the social unrest
that had taken place in late 2019. The projections for the second quarter point to a further dramatic fall in
all OECD countries for which quarterly estimates are available. On average across the OECD, GDP is
projected to have fallen by 13.2% in the second quarter of 2020, with values of -19% in Spain and the
United Kingdom and -18% in France and Ireland.
Early evidence available for a number of OECD countries shows a massive economic shock not only in
countries that introduced strict mandatory measures. Economic activity also dropped substantially where
governments relied more on social conformity and/or social capital. This likely reflects people’s reactions
to non-binding recommendations and their greater awareness of the seriousness of the epidemic.
An analysis of the labour market effects of state-wide stay-at-home orders on initial unemployment claims
in the United States (Box 1.1) shows that the timing and the extent of state-specific lockdowns in the
United States did not play much direct role in limiting or amplifying the extent of the nationwide shock on
the labour market and only affected within-firm work organisation (e.g. shifting to teleworking). Moreover,
Figure 1.4 and Figure 1.5 show that the largest swing in mobility and unemployment insurance claims took
place before all state lockdowns, with the partial exception of California. This suggests that spill-overs
across states, in form of reduced product demand or disruption of supply chains, were of secondary
importance, at least at the beginning.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
29
Figure 1.3. GDP fell substantially in the first half of 2020
Quarterly percentage change in real GDP, Q1 2020 and Q2 2020 projections
Q1-2020 (⭧) Q2 2020 projection
%
5
0
-5
-10
-15
-20
-25
Note: Quarterly GDP growth projections (Q2 2020) are not available for Latvia, Mexico, Poland and Turkey.
Source: OECD Quarterly National Accounts, OECD (2020), “Gross domestic product (GDP)” (indicator), https://doi.org/10.1787/dc2f7aec-en
(accessed on 30 June 2020) and OECD (2020[6]), “OECD Economic Outlook – All editions”, OECD Economic Outlook: Statistics and Projections
(database), https://doi.org/10.1787/826234be-en (accessed on 10 June 2020).
StatLink 2 https://stat.link/61emg4
This is in line with the evidence found by other recent analyses for the United States and other OECD
countries. Chen et al. (2020[7]), for example, show that both the fall in electricity consumption across
32 European countries as well as the increase in initial unemployment insurance claims across states in
the United States are associated with people’s observed mobility. Mandatory restrictions, such as school
and business closures and shelter-in-place orders, do not appear to have had much additional impact.
Maloney and Taskin (2020[4]) also show that the majority of the fall in restaurant reservations in the
United States and of movie spending in Sweden occurred before the imposition of any non-pharmaceutical
interventions. Even in May, as US states started reopening, credit card spending did not pick up faster in
states that opened quickly than in those that kept confinement measures in place (Chetty et al., 2020[8]).
Andersen et al. (2020[9]) also analyse credit card transactions and show that aggregate spending in
Denmark, where, significant restrictions on social and economic activities were taken, fell only marginally
more than in Sweden where no such measures were taken. Hensvik et al. (2020[10]) show that job postings
in Sweden fell as much as in the United States. Finally, Aum et al (2020[11]) compare the employment
effects of COVID-19 in Korea, which did not implement a lockdown but relied on testing and contact tracing,
with the effects in the United Kingdom and the United States. They find that at most half of the job losses
in the United States and the United Kingdom can be attributed to lockdowns.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
30
Box 1.1. The labour market effect of state-wide stay-at-home orders in the United States
Several OECD countries, especially those with federal structures, took non-pharmaceutical interventions
(NPIs) both at national and at regional or local level to contain the pandemic. For example, the Australian
Commonwealth Government issued a federal order on 29 March limiting gatherings to a maximum of two
people, but several states went much further while others opted not to follow the new recommendations.
Italy first imposed a strict lockdown on a number of regions and provinces in the northern part of the
country, and later extended it to the other regions. In Germany, containment policies differed across
Länder.
In the United States, the federal government issued an emergency declaration on 13 March and
non-binding physical-distancing guidelines on 16 March. Governors in most states issued binding
stay-at-home orders in the following three weeks. However, while a few states and local administrations
issued them very early – in California, for example, a few counties already issued stay-at-home orders on
16 March and a state-wide lockdown was established on 19 March – others waited longer out of concerns
for the economy. While state lockdowns were effective in curbing the diffusion of the epidemic, their direct
impact on the labour market is less obvious – particularly in the context of a rapidly spreading epidemic,
federal guidelines that already affected people’s behaviour and a connected economy. The possible
impact of state lockdowns has indeed been part of an intense political and scientific debate – see
e.g. Gupta et al. (2020[12]); Painter and Qiu (2020[13]); Kahn, Lange and Wiczer (2020[14]); Chetty
et al. (2020[8]); Friedson et al. (2020[15]); and Coibion, Gorodnichenko and Weber (2020[16]).
This box assesses the impact of state lockdowns on the labour market with the help of event study
techniques, using alternatively a measure of mobility towards workplaces and initial unemployment
insurance (UI) claims as dependent variable. The following simple model is estimated
𝑇
log 𝑌𝑠𝑡 = ∑ 𝛼𝑜 𝐷𝑜𝑠𝑡 + 𝜇𝑠 + 𝜇𝑡 + 𝜀𝑠𝑡
𝑜=−𝜏
where 𝑌 stands for workplace mobility (resp. initial UI claims), the 𝜇 stand for state 𝑠 and time 𝑡 fixed
effects, 𝐷𝑜𝑠𝑡 are state-specific time dummies indexed by 𝑜, which denotes time distance from the date of
the stay-at-home order in state s (with 𝑜 = 0 indicating the day or week of the order) and 𝜀 is a standard
error term. The 𝛼𝑜 s are the parameters of interest, which capture the effect of orders on the dependent
variable, on the day (resp. week) of implementation and each of the subsequent days (resp. weeks),
taking other states as a control group. To ensure that “treated” and control states are comparable, the
estimated 𝛼𝑜 s must be insignificantly different from 0 for negative values of 𝑜 (i.e. between −𝜏 and -1). As
standard in the case of absolute numbers, the dependent variable is in logarithm. Estimates are then
presented in terms of percentage effects in the figures below.
Model estimates shows that daily cross-state average workplace mobility slumped on the very same day
federal guidelines were announced and continued to decline the following days (Figure 1.4, Panel A). Yet,
a significant negative trend is visible also the days before. The estimated effect of state lockdowns is
smaller, at 5-10% (Panel B) but significant and persistent, remaining approximately constant for about
one month (not shown in the chart). Treated and control states appear comparable. This suggests that
state lockdowns had a significant effect on the organisation of work over and above the nationwide trend.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
31
Figure 1.4. US stay-at-home orders and workplace mobility
A. Cross-state average B. Additional effect of stay-at-home orders
Day before federal guidelines =100 Day before statewide stay-at-home order = 100
110 110
100 100
Workplace mobility Workplace mobility
90 90
80 80
70 70
60 60
-11-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
Days since federal guidelines Days since stay-at-home orders
Note: The charts show estimated time fixed effects (Panel A) and coefficients of time-distance-to-state-order dummies (Panel B) from a
log-linear regression model with the logarithm of workplace mobility as dependent variable. The sample period are the days from 15 February
to 18 April 2020. In Panel A, time fixed effects are shifted to coincide at 0 with the publication date of federal guidelines. Panel A takes the day
before the federal guidelines were announced as reference. Panel B takes the day before the stay-at-home orders were implemented as
reference. The level of workplace mobility in reference days is set to 100. The bars represent 95% confidence intervals, with errors clustered
at state level.
Source: OECD calculations using state announcements and Google COVID-19 Community Mobility Reports,
https://www.google.com/covid19/mobility/.
StatLink 2 https://stat.link/wd2m8s
Did the changes in work organisation induced by state-wide stay-at-home orders also lead to an additional
burst of unemployment? On the week of the announcement of the federal guidelines, initial UI claims
soared by about 1 000% on average relative to the week before and reached about 3 000% after two
additional weeks (Figure 1.5, Panel A).1 A moderate increase (of about 15%) is also visible in the week
before the announcement, which is the week of the federal declaration of emergency. By contrast, no
additional effect of state lockdowns on initial UI claims 2 is observable (Panel B).3 This result, which is
consistent with those of Kahn, Lange and Wiczer (2020[14]) and Chetty et al. (2020[8]), suggests that
mandatory state-wide stay-at-home orders had no discernible additional effect on unemployment over
and above aggregate trends.
Figure 1.5. US stay-at-home orders and initial unemployment insurance claims
A. Cross-state average B. Additional effect of stay-at-home orders
Week before federal guidelines =100 Week before statewide stay-at-home order = 100
6400 6400
3200 3200
Initial UI claims Initial UI claims
1600 1600
800 800
400 400
200 200
100 100
50 50
-6 -5 -4 -3 -2 -1 0 1 2 3 -6 -5 -4 -3 -2 -1 0 1 2 3
Weeks since federal guidelines Weeks since stay-at-home orders
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
32
Note: The charts shows estimated time fixed effects (Panel A) and coefficients of time-distance-to-state-order dummies (Panel B) from a
log-linear regression model with the logarithm of initial unemployment insurance claims as dependent variable. The sample period are the
weeks ending on 1 February to that ending on 18 April 2020. In Panel A, time fixed effects are shifted to coincide at 0 with the publication week
of federal guidelines. Panel A takes the week before the federal guidelines were announced as reference. Panel B takes the week before the
stay-at-home orders were implemented as reference. Y-axes are presented in logarithmic scale, with the level of initial claims in reference
weeks set to 100. The bars represent 95% confidence intervals, with errors clustered at state level.
Source: OECD calculations using state announcements and U.S. Department of Labor data, https://oui.doleta.gov/unemploy/claims.asp.
StatLink 2 https://stat.link/okmnqw
The results in this box suggest that that the timing and the extent of state-specific lockdowns in the
United States did not play much role in limiting or amplifying the extent of the nationwide shock on
unemployment insurance claims and mainly affected within-firm work organisation (e.g. shifting to teleworking).
1. It must be underlined, however, that that week is also the week in which the greatest worldwide contraction in mobility is observed (see above).
2. A state-wide stay-at-home order is attributed to a week if it took effect at the latest the day before the end of that week. Changing this
parameter to zero, two or three days yields qualitatively similar estimates.
3. Treated and control states appear comparable. Results are robust to: i) excluding the state that issued only a stay-at-home advisory
(Massachusetts); ii) excluding states which had stay-at-home orders in selected counties before state-wide orders; iii) excluding states that
never issued stay-at-home orders; iv) controlling for either overall mobility or workplace mobility; and v) replacing the dependent variable with
the level of initial claims at time t divided by the state labour force in the week before stay-at-home orders. No significant difference is observed
between early- and late-lockdown states, where early stay-at-home orders are identified as those issued no later than a week after the
publication of the federal guidelines.
1.2.1. The initial labour market impact
While the COVID-19 crisis has passed the first phase of lockdowns and business closures, the impact on
the labour market, while already unprecedented, is likely to deepen significantly going forward. This section
provides a first assessment of the initial impact of the crisis mobilising a mix of administrative and survey
data available at the time of writing. This first assessment of the initial impact, while already very dramatic
and without comparison in the post-war period, has to be considered as partial and preliminary.
The unemployment rate offers only partial guidance on the extent of the jobs crisis
Changes in countries’ unemployment rates since the onset of the COVID-19 crisis have varied starkly (see
Figure 1.6) reflecting fundamental differences in policy responses but also the complexity of collecting and
comparing labour market statistics in times of a pandemic.
In the United States, the unemployment rate jumped from its 50-year low of 3.5% in February to 14.7% in
April 2020, the highest level in the history of the series (i.e. since January 1948). It then fell to 13.3% in
May and 11.1% in June. However, 73% of the unemployed in May in the United States were on temporary
layoff, and still 59% in June.5 The large share of temporary layoffs suggests that part of the initial increase
in unemployment may be reabsorbed if the pandemic is kept under control and the economy restarts at
good speed. However, the notable decline in the share from May to June also shows that part of the initial
job losses are becoming permanent as some business are not reopening after the lockdown. In Canada,
the unemployment rate increased by 7.4 percentage points from 5.6% to 13% between February and April
and rose a bit further to 13.7% in May. As in the United States, the initial surge was driven by temporary
layoffs, with the vast majority of the newly unemployed expecting to return to their previous job within
six months.
In most other OECD countries, where data are available only up to April or May at the time of writing of
this chapter, labour market statistics do not identify a notable crisis effect yet. While in Colombia
unemployment increased by 10.3 percentage points between February and May, the increases in all other
countries are much milder: Lithuania registered the largest increases (+3 percentage points up to May)
followed by Latvia (+2.9 percentage points up to May) and Chile (+2.8 percentage points up to April). In
Italy and Portugal, decreases in the unemployment rate up to May (by, respectively, -1.2 percentage points
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
33
and 0.9 percentage points) do not reflect an improvement in the labour market, but a shift towards inactivity
as jobless people stopped searching for a job during the pandemic.
Figure 1.6. Unemployment, as measured in national surveys, has increased sharply only in some
countries
A. Unemployment rate
Percentage of the labour force, May 2020 or latest available date
%
25
20
15
10
5
0
B. Change in the unemployment rate since February 2020
Percentage-point change
p.p
12
10
8
6
4
2
0
-2
Note: April 2020 for Chile, Estonia, Hungary and Norway; and June 2020 for the United States. Greece, New Zealand, Switzerland, Turkey and
the United Kingdom have not yet released figures for April and are therefore not shown in this Chart. Countries are ordered in descending order
of the unemployment rate (Panel A). Figures for Sweden refer to the seasonally adjusted series which differ from the trend component data
published by Eurostat in its latest press release. Due to the introduction of the new German system of integrated household surveys, including
the LFS, the monthly unemployment rate for May 2020 is an estimation based on the figures recorded in previous periods, taking into account
current developments. The May 2020 figure for Mexico is an OECD estimate based on the INEGI ETOE phone survey and it is not directly
comparable with the results for earlier months. The classification of people not working because on a job retention scheme or a temporary layoff
differs across countries (see main text).
Source: OECD (2020), "Unemployment rate" (indicator), https://doi.org/10.1787/52570002-en (accessed on 7 July 2020).
StatLink 2 https://stat.link/5gdloz
The striking heterogeneity in the unemployment response across OECD countries reflects fundamental
differences in countries’ policy mix to cushion the economic and social effects of the crisis (see Section 1.3)
and the way these are reflected in labour market statistics. The United States are strongly relying on
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
34
unemployment insurance benefits to secure the income of workers who lose their jobs, even in the case
of a temporary crisis. Meanwhile, other OECD countries, not just in Europe, are making heavy use of job
retention schemes, which allow companies to cut hours of work, or even halt work entirely, while keeping
their workers attached.6
There are also other, more technical but important, reasons why unemployment rates at this stage offer
only partial guidance on the extent of the labour market crisis across OECD countries and should be read
with some caution:
Survey data are not necessarily best suited to account for sudden shocks, such as a pandemic, in
terms of their granularity and timing. The specific timing when data are collected may not allow
capturing the full shock. The COVID-19 crisis also brought very practical challenges to the
production of labour market statistics around the world. Call centres operated at a lower capacity
and carrying out face-to-face interviews was not possible. In Italy, for example, the labour force
survey sample in March was 20% smaller than usual because of the restrictions imposed to fight
the pandemic. In the United States, the household survey response rate in May, at 67%, was about
15 percentage points lower than in months prior to the pandemic.
The unemployment statistics reflect the fact that the lockdowns affected people’s job search
behaviour. To be considered “unemployed”, an out-of-work person must actively look for a job. As
the restrictions imposed by governments and the fear of infection likely severely hindered job
search behaviour, some out-of-work people may in fact be counted as inactive. This will depress
the measured number of jobseekers and the unemployment rate. In Canada, for example,
1.1 million people were not in the labour force during the week of 12 April, but had worked in March
or April and wanted to work. But because they did not actively look for work, they were not counted
as unemployed. The April unemployment rate climbs from 13% to 17.8% when including workers
who were not counted as unemployed for reasons specific to the COVID-19 economic shutdown
(Statistics Canada, 2020[17]). Also in the United States, the number of people not in the labour force
who wanted a job nearly doubled between March and April 2020, from 5.5 to 9.9 million people
(U.S. Bureau of Labor Statistics, 2020[18]). Comparable data on total employment and inactivity in
OECD countries over the recent months are not yet available.
Unemployment statistics may also be less comparable across countries because countries classify
short-time work or temporary layoffs differently in their statistics – see the detailed note in OECD
(2020[19]). In European countries, people who report being temporarily absent from work are
nevertheless counted as “employed” based on a specific question probing their formal job
attachment: respondents are classified as employed if they indicate that (i) the recall date falls
within three months from layoff (or more than that, if the return to employment in the same
economic unit is guaranteed), or that (ii) workers continue to receive remuneration from their
employer, including partial pay, even if they also receive support from other sources, including
government schemes. In the United States and Canada, people on temporary layoffs are deemed
to have weaker job attachment and they are classified as “unemployed” even if they expect to be
recalled to their job within six months.7 Typically, these differences have only a limited impact on
broad comparability of employment and unemployment statistics. However, in times of crisis, the
cross-country comparability of unemployment statistics can be significantly affected. As an
example, since the beginning of the crisis, Ireland’s Central Statistics Office has been publishing
an alternative unemployment estimate that includes workers on temporary layoff and in receipt of
a new Pandemic Unemployment Payment paid to all: doing so raises the measured unemployment
rate in May from 5.6 to 26.1%.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
35
Initial unemployment insurance claims have reached historically high levels
Administrative data on unemployment insurance claims/recipients and social-security contributions can
provide more granular and real-time evidence, at least at the onset of a crisis. They can be a
complementary and more timely source of information. And yet, they also do not permit an entirely
comparable assessment of the labour market situation across countries, as they also reflect institutional
differences across countries and the very different role that unemployment insurance plays in cushioning
the immediate effect of an economic shock.
The number of unemployment insurance claims soared in many countries as the COVID-19 crisis hit
(Figure 1.7), dwarfing the increases observed during the global financial crisis. This reflects an economic
shock that is initially much wider and more abrupt than in 2008 and that may well continue to evolve very
differently.
Figure 1.7. The number of unemployment insurance claims increased substantially in some OECD
countries
Registered unemployment (not seasonally adjusted data) in April 2020, base 100 in February 2020
Registered unemployment Extended unemployment insurance
A. Countries B. Countries experiencing moderate to strong increase
experiencing very
strong increase
1000 200
900 190
800 180
170
700
160
600
150
500
140
400
130
300 120
200 110
100 100
Note: Registered unemployed are jobseekers registered with labour offices and/or public employment services. Registered unemployment
includes workers on unpaid leave in Israel and on temporary layoffs in Norway and the United States. Figure for the Netherlands refers to the
number of unemployment insurance beneficiaries (WW-uitkeringen) and to continued UI claims for the United States. Extended unemployment
insurance refers to COVID-19 Pandemic Unemployment Payment in Ireland and Pandemic Unemployment Assistance in the United States.
Data are subject to national legislations. Consequently registered unemployment data may not be comparable across countries.
Source: OECD calculations based on registered unemployment from OECD (2020), “Labour: Registered unemployed and job vacancies”, Main
Economic Indicators (database), https://doi.org/10.1787/e9ade9e2-en (accessed on 15 June 2020), Labour Market Information Portal (Australia),
STAR (Denmark), Live Register (Ireland), Israeli Employment Agency (Israel), KOSIS (Korea), NVA (Latvia), UZT (Lithuania), ADEM (Luxembourg),
CBS (Netherlands), Ministry of Social Development (New Zealand), ESS (Slovenia), ISKUR (Turkey) and Department of Labor (United States).
StatLink 2 https://stat.link/y091fk
Across the United States, more than 40 million workers had filed unemployment insurance claims by the
end of May, two months after the beginning of lockdowns. During the global financial crisis, it took 1.5 years
to reach that number after the Lehman Brothers bankruptcy. While these numbers made the news around
the world, similar increases were registered in other OECD countries when measured relative to the size
of workforce. In Israel, the share of workers in the labour force who had filed an unemployment insurance
claim by the end of April was seven times higher than before the crisis, at 27.8%. Other OECD countries
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
36
also registered substantial increases in unemployment claim numbers, but job retention schemes often
contributed to cushion the effect of the jobs crisis.
In addition to the observed job losses, the crisis also seems to have led to strong adjustments on the intensive
margin. Where available for the recent months, data on hours of work and part-time work for economic
reasons (i.e. people who would have preferred full-time employment) show substantial adjustments also in
terms of how long employed people worked. In the United States, for example, the number of people who
work part time for economic reasons nearly doubled to 10.9 million between March and April.
Job retention schemes are cushioning the impact on open unemployment in a number of
OECD countries
Companies also made massive use of job retention schemes to receive public support for cutting the hours of
work for their workers, or putting them “on furlough”. About 60 million workers across the OECD have been
included in company claims for job retention schemes, such as the German Kurzarbeit or the French Activité
partielle. Such schemes allow preserving jobs at firms experiencing a temporary drop in business activity,
while providing income support to workers whose hours are reduced due to a shortened workweek or
temporary layoffs (see Section 1.3.2 for an in-depth discussion). The use of these instruments plays a major
role in explaining why most other OECD countries did not experience the massive surges in open
unemployment that were registered in Canada and the United States. In May, companies’ requests for support
from job retention schemes summed to 66% of dependent employees in New Zealand, over 50% in France,
over 40% in Italy and Switzerland, around 30% in Austria, Germany, Portugal and the United Kingdom
(Figure 1.8). The actual use of these schemes may be considerably lower than the initial requests. In France
and Germany, for example, the estimated actual use at the time of writing is around 60% of the initial requests.8
Figure 1.8. Participation in job retention schemes has been massive in some countries
Approved applications and actual participants in job retention schemes as a share of dependent employees
Approved applications Actual use
%
70
60
50
40
30
20
10
0
NZL FRA CHE ITA AUT PRT GBR DEU LUX NLD AUS BEL IRL CZE ESP CAN SWE DNK NOR FIN LVA USA
Note: Data refer to end May except for Luxembourg and Switzerland (end April). Data for Austria, Finland and Norway refer to the number of
registered persons in job retention schemes at the end of May. Data for Belgium refer to the estimated number of approved applications in May.
Data for Ireland and Spain refer to the number of recipients in May. Data for Canada cover the period from 10 May to 6 June. Data for France and
Germany on actual use are the estimated number of persons in job retention schemes in May. United States: data refer to participation in short-time
compensation schemes. Australia, Canada, Ireland, the Netherlands and New Zealand operate wage subsidy schemes, which are not conditional
on the reduction in working hours (see Section 1.3.2). Take-up rates are calculated as a percentage of dependent employees in 2019 Q4.
Source: National sources, for details see OECD (forthcoming[20]), “Job retention schemes during the COVID19 crisis and beyond”.
StatLink 2 https://stat.link/9hdgb3
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
37
When accounting for both the extensive margin of adjustment (fewer employed workers) and the intensive
margin (fewer hours worked among remaining workers because of part-time or short-time work), the impact
of the COVID-19 crisis on OECD labour markets has been, on average, ten times bigger than that observed
in the first months of the global financial crisis in 2008 (Figure 1.9): on average across the countries for
which data are available, total hours worked fell by 12.2% in the initial three months of the crisis compared
to 1.2% of the first three months of the global financial crisis.
Figure 1.9. The cumulated impact of the COVID-19 crisis on employment and hours of work is
ten times greater than during the global financial crisis
Percentage change in total hours worked with respect to hours worked in the month of the onset of the crisis
COVID-19 crisis For reference: Global Financial Crisis (2008-09)
of which : contribution of total employment Contribution of average hours worked
%
10
0
-10
-20
-30
-40
-50
1 2 3 1 2 1 2 3 1 2 1 2 3 1 2 3 1 2 3 1 2 3 1 2 3 1 2 1 2 3
AUS AUT CAN ISR ITA JPN KOR MEX SWE GBR USA
Months elapsed since the onset of the crisis
Note: The starting point of the global financial crisis is October 2008. No comparable data available in 2008-09 for Austria, Israel, Italy and
Mexico. The starting point of the COVID-19 crisis is January 2020 for Japan and February 2020 for all other countries. Total hours worked refers
to the total hours actually worked per month for Australia, Canada and Japan, to the total hours actually worked per week for Sweden and the
United Kingdom, to the average actual hours worked per week multiplied by total employment for Austria, Israel, Italy, Korea and Mexico, and
to the average weekly actual hours worked (not including employed not at work) multiplied by the number of employed persons at work for the
United States. The recent data for Mexico are highly uncertain because a new survey tool was introduced in April which may affect the
comparability of the results with earlier months.
Source: OECD calculations based on results from the Labour Force Survey for Australia, monthly estimates of the Microcensus Labour Force
Survey (LFS) for Austria, the Canadian Labour Force Survey for Canada, the Labour Force Survey for Israel, monthly estimates of the Labour
Force Survey for Italy, the Labour Force Survey for Japan, the Economically Active Population Survey for Korea, monthly estimates of the
Encuesta Nacional de Ocupación y Empleo (ENOE) and the Encuesta Telefónica de Ocupación y Empleo (ETOE) for Mexico, the Labour Force
Survey for Sweden, Single-month estimates for employment status and total weekly hours provided by the Office for National Statistics for the
United Kingdom and the Current Population Survey (CPS) for the United States.
StatLink 2 https://stat.link/hg3r5m
Job postings and hirings were frozen
Besides layoffs, a reduction in companies’ hiring activity played an important role in rising unemployment.
Even countries with comprehensive job retention schemes and those that banned or strictly regulated
dismissals (such as Italy and Spain, see Section 1.3.2) saw their jobseeker numbers increase, though at
a much lower scale than in Canada or the United States. Temporary contracts were not renewed, and new
jobs were not opened. Recessions are usually characterised by both large increases in the inflow rate into
unemployment (i.e. more layoffs) and large reductions in the unemployment outflow rate (i.e. fewer hirings
and longer unemployment spells (OECD, 2009[21]), and this crisis is no exception.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
38
Comparable data on hirings or job vacancies for all OECD countries are not yet available. However,
high-frequency data on online job postings can provide real-time information on labour demand, and often
with a high level of granularity in sectoral, occupational and regional information. By relying only on
information posted online, these data necessarily portray a partial picture of the overall economy, with
different degrees of representativeness across countries. Within country, they generally over-represent
high-skilled occupations and industries.
These shortcomings notwithstanding, online job postings data 9 bear witness of the recent labour market
collapse in a similar way as other figures proposed above. The number of job advertisements posted online
on a given day decreased by 35% from 1 February to 1 May, on average across the 18 OECD countries
for which data are available (Figure 1.10). Some countries experienced larger falls, such as Canada (43%),
Ireland (45%), and the United Kingdom (52%), while others experienced more moderate declines, such as
Germany (16%), Belgium, Japan and Switzerland (all 20%). The freeze in vacancy postings did not
materialise until March in most countries, and had increased 3.5 times in size by the end of April on average
across countries.10 The job posting freeze continued until 1 June despite the partial re-opening of the
economic activities in several OECD countries.
Figure 1.10. Online job postings have declined massively
Average changes in daily data
2 March 30 March 4 May 1 June (↗)
%
10
0
-10
-20
-30
-40
-50
-60
-70
MEX GBR NZL IRL AUS CAN ESP SWE FRA POL ITA BEL NLD USA JPN CHE DEU
Note: Change in the number of new job postings online between 1 February 2020 and the average day in the week beginning with the date
specified.
Source: Data sourced and elaborated by Indeed, June 2020.
StatLink 2 https://stat.link/bnm23a
Aggregate figures hide significant heterogeneity in the impact of the COVID-19 crisis on online job
openings across sectors and occupations. Some services considered “essential” were operating even at
the peak of the health crisis, while non-essential businesses had to suspend activities. Furthermore, some
sectors are naturally more exposed to contagion, either because production cannot occur off company
premises, or because they rely more heavily on inter-personal contacts among workers or between
workers and customers (Barbieri, Basso and Scicchitano, 2020[22]). Lastly, some sectors have suffered
and will continue to suffer more of the reduction in demand driven by job displacements and lower incomes,
and by disruptions in supply chains (Barrot, Grassi and Sauvagnat, 2020[23]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
39
On average across the five OECD countries for which detailed data are available, the largest contributions
to the aggregate decline in job postings is attributable to what are here defined as “public services” 11, and
business services, followed by trade and transportation, and the accommodation and food industries
(Figure 1.11, Panel A). These sectors need not correspond to those where postings fell the most in
percentage terms, as long as the latter accounted for a relatively small share of country-wide online
postings before the crisis. Unreported figures from the same five countries show that the arts and
entertainment, accommodation and food, transport and storage and private sector administration industries
experienced the largest declines in percentage terms in outstanding job postings between February and
April 2020 on average across countries (-60 to -80%), while health and social work, manufacturing, and
information services experienced minor declines.
Figure 1.11. The decline in online job postings by industry and skill group
A. Industry-composition of online job postings decline B. Decline in online job postings by skill group
% of total decline in Feb-Apr 2020 % change between Feb-Apr 2020
Public services and utilities ICT and business services Low-skilled occupations Middle-skilled occupations
Accommodation and food Primary, manufacturing, construction
High-skilled occupations
% Trade and transportation Other services %
100 0
90
-10
80
-20
70
-30
60
50 -40
40
-50
30
-60
20
-70
10
0 -80
AUS CAN GBR NZL USA AUS CAN GBR NZL USA
Note: Panel A displays the contribution of each industry to the change in the country-wide postings displaying information on the sector of
affiliation. At this level of aggregation, all industries contracted. Postings missing information on the sector of affiliation were discarded. Primary
activities refer to agriculture and mining. Panel B displays the growth rate in the count of new job advertisements posted online in the month, in
the country specified, averaged over all occupations within the skill group. Weighted averages use the share of new postings in the occupation
in total new postings in the country, for February, averaged over 2018-20. For the definitions of low-, middle-, and high-skilled occupations, see
Chapter 4.
Source: OECD calculations based on data from Burning Glass Technologies, May 2020.
StatLink 2 https://stat.link/ie3ja0
Lastly, the crisis had heterogeneous effects on the hiring activity for different occupations. Demand for
so-called essential workers, such as hospital workers, employees of food retailers, and warehouse
personnel held up or even increased during the lockdown. While many of these occupations are usually
classified as low-skilled, workers in high-skilled occupations were also relatively less affected by the labour
market shock, insofar as they could keep on working safely from home through distance work. 12
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
40
Between February and April 2020, middle-skill occupations experienced a significantly larger fall in online
job advertisements than high- or low-skill occupations, on average across the five countries for which data
are available (Figure 1.11, Panel B). In the United Kingdom, where this phenomenon is especially
pronounced, new online job postings for middle-skill occupations contracted twice as much as for low-skill
occupations, and 40% more than for high-skill occupations. While the persistence over time of such
patterns will need further investigation, these results point to the possibility that the COVID-19 shock will
reinforce the existing trend of employment polarisation in OECD countries (see Chapter 4).
Within countries, some regions were hit more than others
Despite the shock’s symmetric origin and its global extent, the impact of COVID-19 within countries differed
across regions. In many OECD countries, the outbreak has been worse in cities than in rural areas and
has affected some regions more than others. In Italy, for example, the country’s north was hardest hit, and
Lombardy, where the first outbreak of COVID-19 took place, registered the highest number of cases. In
France, the regions of Île-de-France and Grand Est were the most affected. In the United States, in early
June, the state of New York alone accounted for 20% of the country’s confirmed COVID-19 cases. Regions
or states where the outbreak was more sizeable experience significantly more severe economic losses
(Chen et al., 2020[7]). The economic impact across regions will also vary according to their sectoral
specialisation: some sectors are more exposed to confinement measures or to disruptions in the supply
chain or are structurally more volatile as they rely more on temporary and seasonal work.
Official estimates of job losses by sector, regions and groups of workers are not yet available in a consistent
manner. However, an analysis of the sectors most directly affected by containment measures, such as
those that involve travelling and direct contact between consumers and service providers (OECD, 2020[24]),
can provide a first estimate of the heterogeneous effects across regions and group of workers. Differences
in the share of regional employment at risk are very wide, ranging from less than 15% to more than 35%
across 314 regions 13 in 34 OECD countries (OECD, 2020[25]). In Greece, for example, they range from
55% of jobs at risk in South Aegean Islands to 22% in Central Greece. Regional differences are particularly
stark also in Slovak Republic and France. Touristic places often show the highest shares of jobs at risk of
disruption. In Europe, several major tourist destinations, such as Crete, the South Aegean and Ionian
islands (Greece), Balearic and Canary Islands (Spain) as well as the Algarve region in Portugal may lose
40% or more of jobs. In Korea, the highest risk of job loss is in Jeju-do, where tourism is an important pillar
of the economy. In North America, Nevada (with the tourist hub Las Vegas as its largest city) stands out
as the most affected state, followed by Hawaii. Regions in Northern and Eastern European countries
appear less affected, on average, than those in Southern Europe and North America.
More vulnerable workers are bearing the immediate brunt of this crisis
Already disadvantaged groups of workers often suffer most from economic crises, as they are the first out
when the shock hits and last in when the recovery starts. While it is still very early to assess the impact of
COVID-19 on different labour market groups, first evidence indeed suggests that the crisis has – at least
initially – exacerbated pre-existing labour market inequalities, and that vulnerable workers have so far been
paying the brunt of the costs.14
Low-paid, often low-educated workers have been particularly affected during the initial phase of the crisis.
On the one hand, many of them ensured the continuation of essential services during the lockdowns, often
at a substantial risk of exposing themselves to the virus while working. Granular evidence using
smartphone location shows that, in the United States, people living in higher-income neighbourhoods could
shelter at home earlier and for longer than people living in lower-income neighbourhoods (see Box 1.2).
The so-called “frontline workers”, who work in essential services in jobs that cannot be carried out remotely,
are on average less well educated than the overall workforce and more likely to earn low wages (Blau,
Koebe and Meyerhofer, 2020[26]; Fana et al., 2020[27]). This includes health care workers, but also cashiers,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
41
production and food processing workers, janitors and maintenance workers, agricultural workers, and truck
drivers. Low earners are also much more likely to be working in sectors affected by shutdowns and more
likely to have suffered job or earnings loss. In the United Kingdom, employees in the bottom decile of
weekly earnings are about seven times as likely to work in shutdown sectors as those in the top earnings
decile (Joyce and Xu, 2020[28]). Low-income workers are less able to work from home, are more likely to
report having lost their job because of COVID-19, and are more pessimistic about their earnings prospects
for the next few months. Real-time survey data for a number of OECD countries (Figure 1.12 based on
Foucault and Galasso (forthcoming[29])) show that those in the top earnings quartile were on average 50%
more likely to work from home in April than those in the bottom quartile. Meanwhile, low-earning workers
appear to be have stopped working twice as often. In Canada, labour force survey data show that
employment losses between February and April 2020 have been more than twice as high for low-wage
employees as for all paid employees (Statistics Canada, 2020[17]).15
Figure 1.12. While higher-earning workers often worked from home, lower-earning workers often
had to stop working
Share of total workers usually employed before the onset of the crisis by earnings quartile, selected OECD
countries, mid-April 2020
Working from home Working in the usual workplace Stopped working
%
100
90
80
70
60
50
40
30
20
10
0
1st 4th 1st 4th 1st 4th 1st 4th 1st 4th 1st 4th 1st 4th 1st 4th 1st 4th 1st 4th 1st 4th
quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile quartile
AUS AUT CAN FRA DEU ITA NZL POL SWE GBR USA
Source: Foucault and Galasso (forthcoming[29]) based on the REPEAT (REpresentations, PErceptions and ATtitudes on the COVID-19) survey.
StatLink 2 https://stat.link/x6m4pj
Box 1.2. Staying at home during the pandemic is harder for poorer people
While containment and mitigation policies formally applied to the entire population, not all population
groups adapted (or could adapt) with the same speed and depth. Very fine-grained data collected from
the smartphone location company Cuebiq show that before the COVID-19 outbreak in the United States,
people living in higher-income neighbourhoods used to move more (Figure 1.13, Panel A) and stay at
home less (Figure 1.13, Panel B). When the White House released its guidelines asking people to shelter
at home, mobility decreased and the share of people staying home increased. Also the mobility patterns
across higher- and lower-income neighbourhoods inverted: people living in poorer neighbourhoods
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
42
started sheltering at home later and less (especially during the workweek) than people living in
higher-income neighbourhoods. This may reflect differences in awareness and access to information.
However, they are also indicative of structural divides in the access to jobs (people in higher-income
neighbourhoods may be more likely to be employed in jobs that can be done at home), in the ability to
weather a sudden shock (people in higher-income neighbourhoods more likely have enough savings to
pass a period of unemployment) and in housing conditions (poor housing conditions may make it more
difficult for people to self-isolate and can make effective teleworking impossible).
Figure 1.13. In the United States, people living in higher-income neighbourhoods sheltered at
home earlier and longer than people living in poorer neighbourhoods
Higher income Lower income
A. Cuebiq Mobility Index B. Shelter in place
Base 10 of the logarithm of the distance Percentage of users staying at home
4.5 70
4
60
3.5
50
3
2.5 40
2 30
1.5
20
1
0.5 10
0 0
Note: The Cuebiq Mobility Index in Panel A quantifies how far users move each day. It is calculated as the base 10 of the logarithm of the
distance between opposite corners of a box drawn around the locations observed for users on each day. Shelter-in-Place in Panel B represents
the percentage of users staying at home. It is calculated daily by measuring how many users moved less than 330 feet (100 metres) from
home.
The charts show the gap in mobility and the share of people sheltering at home between higher- and lower-income block groups by Core-Based
Statistical Area (CBSA). A CBSA is a group of one or more counties with an urban core plus adjacent territory that has a high degree of social
and economic integration with the core measured by commuting ties. The Cuebiq Mobility Index for each CBSA is the median of these
per-person indexes. The higher and lower incomes for each metropolitan and micropolitan (an area centred on an urban cluster with a
population of at least 10 000 but fewer than 50 000 people) area are based on median household income data from the U.S. Census Bureau,
2013-17 American Community Survey 5-Year Estimates for census tracts aggregated by CBSA.
Source: Cuebiq, a location intelligence and measurement platform. Through its Data for Good programme, Cuebiq provides access to
aggregated mobility data for academic research and humanitarian initiatives. These first-party data are collected from anonymised users who
have opted in to provide access to their location data anonymously, through a GDPR-compliant framework. It is then aggregated to provide
insights on changes in human mobility over time (accessed: on 10 May 2020).
StatLink 2 https://stat.link/y348bp
Also workers in non-standard jobs – i.e. self-employed workers and those in temporary or part-time
dependent employment – were highly exposed to job and income losses. They may represent up to 40%
of total employment in sectors most affected by containment measures across European OECD countries
(OECD, 2020[30]). Some self-employed workers are overrepresented in some of the industries that have
been restricted or shut down because of quarantine, e.g. in the hospitality and culture sectors, but also in
personal services such as hairdressers. Early surveys carried out after the start of lockdowns document
this effect: 48% of self-employed workers in the Netherlands experienced an hours reduction, compared
to only 27% of employees (Von Gaudecker et al., 2020[31]); 75% of the self-employed in the
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
43
United Kingdom report having experienced a drop in earnings in the previous week, compared to less than
25% of salaried workers (Adams-Prassl et al., 2020[32]). Meanwhile, workers on temporary contracts were
among the first to lose their job during the crisis as contracts are not being renewed when coming to an
end. Canada saw sharp declines in employment among workers with a temporary job and those with a job
tenure of one year or less – -30% for each group (Statistics Canada, 2020[17]). Administrative data from
France and Italy confirm these patterns. In France, the increase in new unemployment claims in March
and April 2020 was entirely driven by temporary agency workers and workers with temporary jobs which
saw their contracts not renewed (DARES, 2020[33]). Administrative data on job flows in Italy show that the
decrease in the number of jobs between the end of February and the end of April compared to the same
period in 2019 was largely driven by reduced hiring on temporary contracts (Bovini et al., 2020[34]; Baronio
and Linfante, 2020[35]; Veneto Lavoro, 2020[36]). People who were counting on getting a new job could not
find one. The heavy job or income losses of workers in non-standard forms of employment are particularly
concerning as these workers often do not have access to job retention schemes and unemployment
benefits – see Sections 1.3.2 and 1.3.3 and OECD (2020[30]; 2020[37]).
The same applies for many informal workers, including undocumented migrants. Many of them are likely
employed in sectors severely hit by confinement measures, such as in accommodation and food services
but also as domestic workers, and they often have no access to any income support. Workers in “partial
informality”, whose employment is registered but who receive some of their remuneration in cash
(“envelope wages”), may receive compensation only for part of their lost earnings from job retention
schemes or unemployment benefits – see Section 1.3.3 and OECD (2020[37]).
The COVID-19 crisis has also exposed the vulnerabilities of many platform jobs. While, some platform jobs
offered opportunities to workers and business to reinvent themselves during the confinement and respond
to arising needs (for example by delivering food, pharmaceuticals and other goods), they were also among
those most exposed to the shock. According to a survey carried out by AppJobs (AppJobs Institute,
2020[38]) – an online platform to search for app-based jobs around the world – over half of gig workers said
they had lost their jobs; more than a quarter had seen their hours cut. Yet, at the same time, these workers
often do not benefit from employment protection legislation; they often have no access to short-time work
schemes, unemployment benefits or paid sick leave; and, in some countries, they may not even have
health insurance (OECD, 2019[39]).
Young people risk being once more among the big losers of the current crisis, much like they suffered
heavily during the global financial crisis (Carcillo et al., 2015[40]; OECD, 2016[41]). This year’s graduates,
sometimes referred to as the “Class of Corona”, are leaving schools and universities with often very poor
chances of finding employment or work experience in the short run. Meanwhile, their older peers are
already experiencing the second heavy economic crisis in their still young careers. The initial labour market
experience has a profound influence on the later working life, and a crisis can have long-lasting scarring
effects on employment and earnings perspectives (Bell and Blanchflower, 2011[42]; Schmillen and
Umkehrer, 2017[43]). First evidence of labour market data from the current crisis suggest that young workers
have been heavily affected, as they generally hold less secure jobs and are overrepresented among
workers in hard-hit industries such as accommodation and food services. In the United Kingdom,
below-25-year-olds were about 2.5 times as likely as other employees to work in shut-down sectors, a
figure that still excludes students in part-time jobs (Joyce and Xu, 2020[28]). Youth employment numbers
quickly took a dive: in Canada, the number of employed youth dropped by 33% from February to May
2020. In the United States, the teenage unemployment rate more than tripled from 7.7 to 25.2% in between
February and May. During the global financial crisis, across the OECD, almost one-in-ten jobs held by
under-30-year-olds had been destroyed, and the recovery was very slow, particularly for the
disadvantaged. It took a whole decade, until 2017, before the youth unemployment rate had gone back to
its pre-2008 level. Even so, young people have seen a general decline in their labour market fortunes, with
increases in the incidence of non-employment, low-pay and underemployment (OECD, 2019[39]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
44
Evidence on the differential employment impact of the current crisis on women and men is still weaker.
However, this crisis, unlike the previous one, appears so far to have affected the labour market prospects
of women more strongly than men. In Canada women accounted for a disproportionate share of job losses
in March, though men experienced larger employment losses in April. However, there remains a small
gender gap in employment losses (-16.9% for women vs. -14.6% for men between February and April).
Also in the United States, unemployment rates increased more sharply for women than men. In the
European Union, the unemployment rate in March 2020 increased by 4.5% for women against 1.6% for
men. Women’s labour market attachment tends to be weaker than men’s, leaving them more exposed and
easier to lay off. Moreover, many of the industries most directly affected by COVID-19 are major employers
of women, while the global financial crisis had been characterised by greater job losses in male-dominated
sectors (notably construction and manufacturing) and an increase in hours worked by women, especially
in the early years (Sahin, Song and Hobijn, 2010[44]; OECD, 2012[45]). The widespread school and childcare
facility closures during the current crisis likely also amplified women’s unpaid work burden at home (see
Box 1.3).
Box 1.3. Women on all fronts during the COVID-19 crisis
Women have been serving on the frontlines in the fight against COVID-19, and the crisis impact on
women has been stark. Leading much of the wider social and health response, they have been facing
compounding burdens:
Women are playing a key role in the health care response to the pandemic. They make up two-thirds
of the health workforce worldwide, including 85% of nurses and midwives (Boniol et al., 2019[46]), and
account for 90% of long-term care workers across OECD countries (OECD, 2020[47]). Health and social
care workers have been facing exceptional demands, and considerable risks, through the crisis. The
strain has often been particularly acute for mothers, who also had to cope with the implications of school
and childcare facility closures during confinement.
The crisis likely amplified women’s unpaid work burden, as women picked up much of the additional
unpaid work caused by widespread school and childcare facility closures. Women provided most unpaid
work at home before the crisis, spending around two hours per day more on it than men across the
OECD on average (OECD, 2020[48]). The crisis increased the amount of time that parents spent on care
and child supervision and home schooling, with much of this additional burden likely having fallen on
women. According to a German online survey carried out in March/April 2020, in about half of all
households with children the female partner alone cared for the children (Möhring et al., 2020[49]). Online
survey data collected in the United Kingdom in April/May 2020 indicate widening disparities in paid work
patterns between mothers and fathers (Andrew et al., 2020[50]).
Women face higher risks of economic insecurity. Despite the remarkable progress made over the past
half-century or so, women’s labour market attachment remains weaker than men’s, especially around
parenthood. Gender gaps in hours worked, seniority and pay, leave women more vulnerable than men
and easier to lay off. The short-term economic fallout from COVID-19 particularly affected sectors that
rely on physical customer interaction, many of which are major employers of women. On average across
OECD countries, women make up about 53% of employment in food and beverage services (e.g. cafés,
restaurants and catering), 60% in accommodation services (e.g. hotels) and 62% in the retail sector
(ILO, 2020[51]). For some women workers, the public sector may offer some protection, at least in the
short term, as women make up a disproportionate share of public-sector employees across the OECD
(OECD, 2019[52]).
Women are often more vulnerable than men to any sharp income loss. Across OECD countries,
women’s incomes are, on average, lower than men’s, and their poverty rates are higher (OECD,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
45
2020[53]). Women also often hold less wealth than men that could help cushion temporary income losses
(Sierminska, Frick and Grabka, 2010[54]; Schneebaum et al., 2018[55]). And because women tend to hold
greater care and domestic responsibilities than men, it is often more difficult for them to find alternative
employment and income streams (such as piecemeal work) following layoff. Single parents, most of
whom are women, are particularly vulnerable. They were hit much harder by the closure of childcare
facilities and schools during confinement. Reliance on a single income also means that job loss can be
critical for single parents, especially where public income support is weak or slow to react.
Early reports emerging from China and some OECD countries suggest escalating risks of domestic
violence against women during the pandemic, confirming a pattern seen in past lockdown and
confinement situations (UNDP, 2015[56]).
Source: OECD (2020[57]), “Women at the core of the fight against COVID-19 crisis”, https://www.oecd.org/coronavirus/policy-
responses/women-at-the-core-of-the-fight-against-covid-19-crisis-553a8269/.
How strongly the crisis affects different groups of workers and their families ultimately depends not only on
their exposure to job or income loss, but also on how well they are able to temporarily absorb such shocks.
Unfortunately, the exposure to labour market shocks and the capacity to deal with them are often closely
related: analyses for the United Kingdom (Benzeval et al., 2020[58]) and Norway (Alstadsæter et al.,
2020[59]) show that the largest job and income losses have fallen on already financially vulnerable workers
or parents with younger children. In particular, the UK analysis illustrates that in between those workers
experiencing little or no labour market shock and those experiencing a shock but being reasonably well
covered by social safety nets, there is a “vulnerable middle” who are hit hard and have little capacity to
mitigate those shocks: single parents, the low-educed and ethnic minorities.16
1.2.2. The outlook ahead
The economic outlook is exceptionally uncertain. With the easing of the health emergency, confinement
measures have been scaled back gradually and mobility is picking up. The restarting of activities
automatically adds to output, even though some containment measures, such as the closure of many
international borders, will remain for some time. The recovery is likely to be hesitant, and could be
interrupted by renewed outbreaks if targeted containment measures, notably test, track and trace (TTT)
programmes, are not put in place or prove ineffective.
Business and consumer confidence surveys indicate substantial pessimism about labour market
prospects.17 Across OECD countries, businesses’ employment expectations for the months ahead
plummeted in April 2020, while consumers’ unemployment expectations over the next 12 months jumped
up (Figure 1.14). These are the strongest monthly changes on record since 1985. In May the indicators
partially improved but remained far below (for employment expectations) or above (for unemployment
expectations) their long-term averages and very close to the levels registered during the global financial
crisis in March 2009. Employment expectations declined for all sectors, but the fall is much larger for
services while the outlook for manufacturing was already on the negative side before the COVID-19 hit.
Consumers’ unemployment expectations increased to a similar extent for all groups, including for
respondents from higher-income households and those with a tertiary degree.
Reflecting the unusual degree of uncertainty, the OECD Economic Outlook (2020[60]), published on
10 June, presented two equally likely scenarios for the months ahead:
A single-hit scenario in which countries successfully overcome the current outbreak due to the
containment measures put in place in the first half of 2020, with the effective reproduction rate
assumed to decline and stay persistently below unity. Higher hospital capacity and the widespread
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
46
roll-out of effective TTT are assumed to be sufficient to prevent a resurgence in infections and
intensive cases later in the year and until a vaccine becomes available.
A double-hit scenario in which the current easing of containment measures is assumed to be
followed by a second, but less intensive, virus outbreak taking place in October/November. This
could be because of seasonal factors in some countries, particularly in the Northern Hemisphere,
or because containment, TTT and isolating are not as efficient as expected. Further outbreaks in
2021 are assumed to be avoided due to pharmaceutical breakthroughs, but these remain a
significant downside risk.
Figure 1.14. Labour market expectations deteriorated among businesses and consumers alike
Percentage-point difference between the proportion of positive and negative responses, seasonally adjusted data
A. Business’ employment expectations B. Consumers’ unemployment expectations
over next 12 months over next 12 months
Manufacturing Construction Aged 16-29 Aged 50-64
First income quartile Fourth income quartile
Retail trade Services Low educated High educated
15 70
10 60
5
50
0
-5 40
-10 30
-15 20
-20
10
-25
-30 0
-35 -10
Note: A positive balance for future tendency of unemployment (Panel B) means that unemployment is expected to rise. Non respondents to
each question item are excluded from the sample that is used to calculate the balance.
Source: OECD, Business Tendency and Consumer Confidence Database, https://stats.oecd.org/Index.aspx?DataSetCode=MEI_BTS_COS.
StatLink 2 https://stat.link/4njmg8
In the “double-hit” scenario, OECD GDP is projected to decline by 9.3% this year; in the “single-hit”
scenario, OECD GDP is projected to decline by 7.5% this year. In both scenarios, the recovery will likely
be slow and gradual and, despite a rebound, total output by the end of 2021 is expected to still be well
short of its pre-crisis level. In many advanced economies, the crisis could destroy the equivalent of
five years or more of per capita real income growth by the end of 2021.
In both scenarios, unemployment rates are projected to increase significantly in all OECD countries.
According to OECD projections, unemployment in the OECD economies, which had declined to a 50-year
low of 5.3% at the end of 2019, is projected to have more than doubled by the end of June 2020 to
almost 11.4%. This is well above the level seen during the global financial crisis (Figure 1.15, Panel A). As
economies begin to re-open, unemployment is projected to fall gradually but remain above or close to its
peak level during the global financial crisis until well into 2021 even in the single-hit scenario. This reflects
the scale of immediate job losses in some countries, and the likely declines in employment in others as
temporary wage and employment support schemes end in the second half of 2020.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
47
Figure 1.15. Unemployment is projected to increase three times more than during the global
financial crisis
A. Unemployment rate
Percentage of the labour force, OECD area
Historical trend Single-hit scenario Double-hit scenario
%
14
12
10
8
6
4
2
0
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
B. Employment growth
Percentage change in 2020
Single-hit scenario (⭧) Double-hit scenario
%
2
0
-2
-4
-6
-8
-10
-12
-14
Note: Employment growth is not available for Chile (Panel B).
Source: OECD (2020[6]), “OECD Economic Outlook – All editions”, OECD Economic Outlook: Statistics and Projections (database),
https://doi.org/10.1787/826234be-en (accessed on 10 June 2020).
StatLink 2 https://stat.link/yr9eso
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
48
In the double-hit scenario, unemployment remains high for even longer in OECD economies, raising the
risk of hysteresis as long-term unemployment becomes entrenched and labour force participation falls as
workers get discouraged. The OECD-wide unemployment rate is projected to be 8.9% at the end of 2021
in this scenario, near the peak seen during the global financial crisis and 3.6 percentage points above the
rate at the end of 2019. In the single-hit scenario, unemployment would reach 7.7% by the end of 2021.
Country-specific projections are presented in Annex Table 1.A.1.
Employment is projected to decline significantly in most OECD countries (Figure 1.15, Panel B), with the
largest fall in Colombia, the United States and Ireland. The smallest changes are projected in Luxembourg
(where employment is projected to increase slightly), Korea, Austria, Mexico, Germany and Japan. The
cross-country heterogeneity is explained differences in the GDP shock, but also by institutional factors
(e.g. stricter employment protection legislation – see Chapter 3 – and the use of job retention schemes in
continental European countries).
1.3. An unprecedented policy response by countries
OECD countries have responded in an unprecedented manner, in speed, breadth and depth, to contain
the fallout from the crisis and support workers, their families and companies. While a precise and
comparable estimate of the fiscal size of these various measure is not available at this stage 18, Figure 1.16
illustrates the wide range of measures taken across the 37 OECD countries.
At the onset of the crisis, OECD countries have taken a range of measures to reduce workers’ exposure
to COVID-19 by encouraging teleworking or introducing stronger occupational safety and health standards.
Countries strengthened or extended paid sick leave, including to quarantined workers and took measures
to help working parents better deal with unforeseen care needs (Section 1.3.1) and to help ensure that
workers and their families could remain in their dwellings (see Box 1.8 below). A large majority of OECD
countries has introduced or extended job retention schemes to preserve jobs at firms experiencing a
temporary reduction in business activity. Few have also introduced changes to employment protection
legislation to either better protect workers with a permanent contract or facilitate hiring or renewal of
workers with a temporary contract. A number of measures have been taken to ensure the continuation of
essential services during the pandemic (Section 1.3.2). Moreover, almost all OECD countries have
strengthened and/or extended the income support to workers who lose their job or income (Section 1.3.3).
Finally, all countries have provided some form of financial support to boost companies’ financial liquidity,
whether through grants, loans or tax and social-security deferrals, but those measures are not covered in
detail in this chapter. On top of such national-level measures, the European Union has taken strong
initiatives to provide financial support to companies and member states, in particular to promote the use of
short-time work schemes (Box 1.4).
This section provides an overview of the main measures taken, highlights and discusses differences in
their design, and offers a first assessment of the benefits and challenges of different approaches, including
likely difficulties in their implementation.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
49
Figure 1.16. OECD countries introduced bold new measures or considerably expanded existing
ones in response to COVID-19
Percentage share of the total number of OECD countries
Financial support to firms 100%
Income support to people losing job/income 97%
Income support to quarantined workers 92%
Helping with unforeseen care needs 89%
Job retention schemes 89%
Reducing workers’ exposure to COVID-19 86%
Extensions to paid sick leave 81%
Helping workers stay in their homes 73%
Changes to dismissal regulation 24%
0 10 20 30 40 50 60 70 80 90 100
%
Source: OECD COVID-19 Employment and social policy responses by country, http://oe.cd/covid19tablesocial (accessed on 30 June 2020).
StatLink 2 https://stat.link/hm2kx3
Box 1.4. The European Union’s employment and social-policy response to COVID-19
In parallel to national policy responses, the European Union (EU) reacted to the economic, employment
and social emergency with a multi-tiered initiative to support the workers and firms in its member states:
The EU is providing financial support to enhance member states’ policy responses to the
social and employment crisis. Through the “Coronavirus Response Investment Initiative”
(CRII) and “Coronavirus Response Investment Initiative Plus” (CRII+) the European
Commission (EC) has accelerated the deployment of EUR 37 billion of cohesion funds. These
funds can be flexibly redirected towards spending on health care, support to short-time work
(STW) schemes and support to small and medium enterprises, in particular in the most affected
European regions. EU member states also agreed to create SURE (“Support to Mitigate
Unemployment Risks in an Emergency”), a temporary loan instrument to help finance STW
schemes and other similar measures supporting the self-employed across the EU. The
instrument is backed by EUR 25 billion in member states’ guarantees committed to the EU
budget. These guarantees allow the EC to borrow up to EUR 100 billion on financial markets,
to be lent to member states on favourable terms. EU member states with existing STW schemes
and/ or schemes supporting the self-employed can apply for a loan through SURE to cover the
needed expenditures. The EC will verify the application, before the Council approves the loan.
The European Central Bank (ECB) and European Investment Bank (EIB) geared up to prevent
a pro-cyclical tightening of financing conditions in the public and private sector and avoid
liquidity shortages and credit contraction. The ECB expanded its asset purchase programmes
of private and public sector securities by EUR 1 470 billion (including the EUR 1 350 billion
Pandemic Emergency Purchase Programme). It also eased the conditions for its targeted and
non-targeted longer term refinancing operations, and launched a new series of pandemic
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
50
emergency longer-term refinancing operations. After initial steps to mobilise up to
EUR 40 billion in support to European firms early in the outbreak, the EIB group created a
EUR 25 billion guarantee fund to scale up financing for companies by up to additional
EUR 200 billion, targeting in particular small and medium enterprises.
The EC amended regulations to give more flexibility to particularly affected sectors
(e.g. airline companies) or to member states. This includes using the full flexibility foreseen
under state aid rules, and activating the general escape clause of the Stability and Growth Pact
to allow countries to depart from the agreed budgetary requirements.
In late May, the EC put forward a proposal for a major recovery plan, Next Generation EU. If
approved by EU member states, Next Generation EU would raise money by temporarily lifting
the own resources ceiling to 2% of EU Gross National Income, allowing the Commission to
borrow EUR 750 billion on the financial markets to be repaid over a long period of time
throughout future EU budgets. The EUR 750 billion would be channelled through EU
programmes to support member states with investments and reforms, incentivise private
investments and reinforce EU health and civil protection programmes.
Beyond these measures designed to mitigate the economic, employment and social consequences of
the crisis, the EU also supported countries’ health care responses. Euro area countries can use
Pandemic Crisis Support credit lines (created using the framework of the European Stability
Mechanism) to borrow up to 2% of their 2019 GDP, as a benchmark, to finance direct and indirect
expenses linked to health care, cure and accident prevention-related costs (such expenses could
include costs relating to workplace safety and occupational safety and health). The EU Occupational
Safety and Health Agency, in partnership with European social partners, has produced elaborated
guidelines on how to maintain workplace safety during the pandemic.
1.3.1. Protecting workers from COVID-19 and helping them with unforeseen family care
needs
Minimising workers’ exposure to COVID-19
Workplaces and public transport gather large numbers of people and thereby often expose workers to the
risk of contracting and spreading the COVID-19 virus. A primary concern for governments, companies and
workers alike at the onset of the crisis was therefore to limit physical interaction in the workplace and during
the daily commute. Evidence from previous epidemics – see OECD (2020[1]) for a detailed review – shows
that workplace physical distancing is the most effective measure for both reducing the share of the
population who contract the disease (the “attack rate”) as well as for delaying the disease peak.19 OECD
countries therefore extensively promoted teleworking or working from home and continued to encourage
its use even when the strictest confinement measures began to be lifted in May 2020.
Most OECD countries had pre-existing teleworking regulations, in law or collective agreements; sometimes
relatively restrictive or requiring an ex ante agreement by social partners. However, take-up had remained
quite limited and, contrary to widespread belief, without much increase over the years. Across the
European Union, only 3% of workers regularly worked from home in 2015, a further 5% were highly mobile,
working regularly from several locations (including home) while another 10% teleworked occasionally from
various locations but much less often than the highly mobile workers (Eurofound, 2018[61]). Such low take
up reflects in part the nature of people’s work (i.e. not every job can be done from home), but also
resistance from employers and workers alike.20 During the COVID-19 crisis, it was suddenly in both
employers’ and employees’ direct interest to reduce the exposure to the virus to limit sickness and maintain
operations.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
51
In order to promote a rapid move to telework for all operations that allow it, countries took a series of
measures to simplify its use, including through financial and non-financial support to companies. Italy, for
example, simplified the procedure for teleworking by allowing companies and employees to arrange
teleworking without a prior agreement with unions, without a written agreement and at the employees’
place of choice. In Hungary, employers were given the possibility to introduce teleworking without their
employees’ consent. Japan made available a 50% subsidy (up to JPY 1 million) towards the cost of
introducing telework. Korea simplified the application procedures for a subsidy for introducing flexible work
arrangements. Belgium gave employers the possibility to grant their teleworking employees a tax- and
social-security-free allowance of EUR 170 per month to cover telework-related costs, such as for a desk
and office materials. Spain expedited existing programmes to support the digitisation of small and medium-
sized enterprises (SMEs). Some large tech companies also stepped in to provide assistance and
temporary free-of-charge access to some of their communication and sharing tools to companies and
workers.
Surveys conducted in mid-April show a massive surge in the share of workers working from home
compared to the pre-crisis levels (Figure 1.17). The share of workers working from home in April ranges
from little less than 30% in Sweden, Canada and Poland to around 50% in Australia, the United Kingdom
and the United States and 60% in New Zealand.
To minimise the risks of contagion for (the majority of) workers who could not work from home, several
OECD countries restricted the continuation of business operations to “essential” services only (see
discussion in Section 1.1). They issued stricter sanitary guidelines that ranged from requiring the use of
personal protective equipment, such as masks, gloves and other protective clothing, to restricting the
maximum number of workers allowed to be physically present on companies’ premises. Israel, for example,
limited the share of the workforce who were allowed to physically go to work to 15% at the beginning of
the crisis, and then raised the limit to 30% in April.
Figure 1.17. Between 30% and 60% of workers worked from home in mid-April 2020
Share of total workers usually employed before the onset of the crisis, selected OECD countries
Work from home Work in the usual workplace Stopped working
%
100
90
80
70
60
50
40
30
20
10
0
SWE CAN POL DEU FRA AUT ITA AUS GBR USA NZL
Source: Foucault and Galasso (forthcoming[29]) based on the REPEAT (REpresentations, PErceptions and ATtitudes on the COVID-19) survey.
StatLink 2 https://stat.link/lnogza
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
52
In several countries, comprehensive occupational safety and health (OSH) standards have been defined
in co-operation with social partners or autonomously by employers and unions (see Box 1.5 for a short
overview of social dialogue in times of COVID-19). In Italy, for example, the government, employers’
associations and trade unions jointly signed a protocol on OSH measures in the early phases of the crisis
and subsequently renewed and updated this protocol. This protocol was then turned into a government
decree, making it mandatory for all companies. Italian employers’ associations and trade unions also
contributed to define the list of “essential sectors” that were allowed to continue operating. Many
company-level agreements (e.g. at Fiat Chrysler Automobiles, Ferrari, etc.) were signed before the
reopening of factories in May. In Spain, several sectoral agreements were signed to better protect workers
in supermarkets, health care, hotels, restaurants and the tourism sector. The international union UNI
Global and the Spanish telecommunication company Telefónica signed a global agreement in May to
ensure the safe return to work for the company’s employees across the world.
Box 1.5. Social dialogue in times of COVID-19
Trade unions and employers’ organisations in several OECD countries responded swiftly to the
challenges raised by COVID-19 (Business at OECD, 2020[62]; TUAC, 2020[63]). Their initiatives during
the first months of the COVID-19 crisis have revolved around four main pillars:
Voicing concerns and demanding rapid government action to mitigate the economic and social
impact of COVID-19. Some social partners issued joint statements expressing a commitment
to collaborate, such as in Germany or in the French metal industry. At European level, the World
Employment Confederation-Europe and UNI-Europa issued joint recommendations on the
employment and social aspects of COVID-19 and the recovery. In many countries, unions, but
also employers’ organisations, called for stronger action to protect workers, including those in
non-standard jobs, students and parents staying at home to care for their children. For example,
the Belgian Confédération des Syndicats Chrétiens (CSC) asked for the better protection of
workers in temporary unemployment and a benefit increase regardless of their employment
status. The Canadian Labour Congress (CLC) called for an increase of the Emergency Support
Benefit and Emergency Care Benefit. The US federation of trade unions AFL-CIO called for
better protection of frontline workers, 14 days of paid sick leave and no out-of-pocket medical
expenses for all workers. The Australian Council of Trade Unions (ACTU) launched a petition
calling for additional paid leave for all workers affected by COVID-19. In Sweden and Germany,
private employment agencies engaged with the government to ensure agency workers – like all
workers – gained access to short-time working schemes.
Informing and advising their members: most employers’ organisations and unions in OECD
countries quickly set up websites and hotlines to inform their affiliates about occupational safety
and provide advice on the new policies measures. Many have issued guidelines, codes of good
conduct and protocols on good practices on teleworking and safety and health at work. The
Greek Confederation of employers (SEV) published a series of guides to help companies
implementing telework, supporting business continuity and developing e-commerce. The
Danish Confederation of Industry (DI), for example, set up a dedicated COVID-19 website with
guidance on the measures made available by the government and on infection prevention for
office workplaces. The French confederation of employers, the Mouvement des Entreprises de
France (MEDEF), published a best practices guide for businesses.
Negotiating new collective agreements: social partners in several European countries signed
far-reaching agreements on short-time work, see Section 1.3.2 and Müller and Schulten
(2020[64]). In Austria, Denmark, Norway and Sweden, short-time work schemes used during the
COVID-19 crisis derive their main features from national-level collective agreements. In Nordic
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
53
countries, also their implementation is left to company-level agreements. In Germany, sectoral
agreements have been key to raise replacement rates. In other countries, social partners have
been instrumental in simplifying procedural requirements. Other collective agreements have
focused on measures to ensure the safety and health of workers in the workplace (see
Section 1.3.1) or, in the United States, on paid sick leave. In Italy, a collective agreement was
signed in the agency work sector to allocate EUR 75 million from a solidarity bipartite fund to
protect the continuity of employment and pay for agency workers for the month of March. In the
Netherlands, the bipartite training fund of the agency work sector has put in place a EUR 500
training voucher for workers displaced because of COVID-19.
Ensuring monitoring and compliance: The Danish trade union confederation (FH), for example,
has developed recommendations to step up labour inspectorates’ activities to monitor and
sanction any violations of authorities’ provisions, non-compliance with workplace safety rules,
or misuse of the force-majeure provision on rest time and days off. The Spanish Confederación
Sindical de Comisiones Obreras (CCOO) monitors compliance with safety and health standards
and reports potential abuse, including cases where companies declare short-time work but work
regular hours.
Such initiatives – many more of which can be found in a recent Global Deal brief (Global Deal, 2020[65]) –
illustrate how social dialogue and collective bargaining can be mobilised to complement public action,
identify flexible and balanced solutions for both companies and workers and strengthen labour market
resilience (OECD, 2019[66]).
Providing paid leave to sick or quarantined workers
Paid sick leave plays a threefold role in supporting workers during a sickness spell, in protecting their
incomes, their jobs and their health (OECD, 2020[67]). Almost all OECD countries provide financial
compensation during sick leave to employees with a permanent or temporary employment contract. Often,
employers cover an initial period in the form of continued wage payment – for a period of 5-15 days in most
countries, but up to several weeks or months, e.g. in Austria, Germany, Italy and Switzerland and even for
two years in the Netherlands. In addition, most OECD countries provide publicly paid sickness benefits for
employees temporarily unable to work that can extend far beyond employers’ liabilities, for up to one year
in many OECD countries and even longer than this in some (OECD, 2018[68]). Many countries also provide
sickness benefits to those who are self-employed, often with rules that differ considerably from the
regulations governing employees (OECD, 2019[39]). However, certain groups of employees, such as casual
workers and those with zero-hour contracts, are often not entitled to paid sick leave or only during the times
when they actually work. Total spending on paid sick leave prior to the crisis, including employer payments
and public sickness benefit, sums to 3% of total employee compensation or more in countries with the
most generous systems (OECD, 2020[69]).
During a pandemic, paid sick leave can play several additional important roles in:
Permitting workers exposed to the virus to self-isolate. Providing financial compensation is of major
importance in order for workers to self-isolate. Survey data for Israel collected in the lead-up to the
COVID-19 outbreak indicate that 97% of adults report they would comply with a
government-mandated quarantine if their wage losses were compensated, whilst compliance
would drop to 57% without such compensation (Bodas and Peleg, 2020[70]).
Helping contain and mitigate the spread of the virus. Paid sick leave allows workers who are
(potentially) infected to stay at home rather than infect others at or on their way to work (OECD,
2020[71]). Access to paid sick leave for employees in the United States reduced influenza-type
disease rates by 10% and aggregate work absence by 18% (Pichler and Ziebarth, 2017[72]; Pichler,
Wen and Ziebarth, 2020[73]; Stearns and White, 2018[74]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
54
Absorbing the economic shock. Paid sick leave preserves the jobs of a potentially large number of
sick and quarantined workers, who are temporarily not available to work but who are valuable to
their employers and society at large in the longer term. By doing so, it can reduce pressure on
unemployment benefit systems and short-time work schemes and contribute to stabilising the
economy. Job losses in the United States between 8 March and 25 April 2020, measured by the
number of initial unemployment insurance claims, were larger in states that did not have statutory
paid sick leave policies in place (Chen et al., 2020[7]).
Many OECD countries have resorted to, substantially expanded or even initiated paid sick leave policies
during the last weeks and months. Crisis response policies included: first, expanding access to groups of
sick workers previously not covered; second, improving the adequacy of paid sick leave support by waiving
waiting periods, increasing benefit levels or extending benefit durations; and third, extending paid sick
leave to support workers in quarantine – an unprecedented policy in most countries. However, most of
these measures remained limited to people actually suffering from COVID-19 or in mandatory quarantine.
Those excluded from sick-leave payments because of the nature of their contracts – such as contracts that
do not specify a fixed amount of work or hours – remained by and large ineligible.
Better protecting sick employees…
Paid sick leave replaces large parts of earnings of eligible employees in many countries, though
cross-country variation is large.21 In the hypothetical case of a four-week sick leave caused by a COVID-19
infection, paid sick leave in most countries replaces around 60-80% of the last wage of a private-sector
employee earning an average wage and working with the same employer for one year (Figure 1.18). The
replacement rate even reaches 100% in many countries in Northern and Central Europe. In a minority of
countries, people would receive less than half of their last wage over this four-week period. Payment rates
usually decline over time. Over a sick leave lasting as long as three months the payment rate would fall to
on average around 60%, with larger cross-country variation.
In response to the crisis, 16 of the 38 OECD countries increased sick-leave entitlements for people with
COVID-19, as illustrated by the vertical distance between the stacked bar and the dash in Figure 1.18.
Several of them carried out rather large increases – including Finland, France, Australia, Spain,
New Zealand, the United States, Ireland and Korea – often through the introduction of new pandemic-
related payments or top-ups. Notably, Korea, which has no mandatory paid sick leave scheme in place,
provides exceptional paid sick leave through its 2015 Epidemic Act to workers who are hospitalised or
quarantined because of COVID-19. The United States introduced two weeks of mandatory sick pay for
workers with COVID-19-related symptoms for companies with up to 500 employees, paid by the employer
but fully reimbursed by the federal government. While not all employees benefit, this measure should
temporarily raise coverage significantly. 22 Seven countries (Estonia, France, Ireland, Latvia, Portugal,
Sweden and the United Kingdom) temporarily abolished existing waiting periods, thereby achieving small
increases in replacement rates. While most of these waiting periods were only a few days long, waiving
them can be an important tool to prevent the spread of the COVID-19 virus, as viral load seems to peak
quickly after the onset of symptoms (He et al., 2020[75]). Many OECD countries took additional steps to
facilitate access to benefits for all or some workers. More than ten countries eased reporting requirements,
by delaying or waiving the need for medical certification or by allowing online benefit applications. This
also lessened the burden on and the risk for health workers. Eight countries improved the protection for
health workers by recognising, only for this group, COVID-19 as occupational disease, with more generous
entitlements from occupational-accident insurance. Only Spain recognises COVID-19 as occupational
disease more generally for all employees.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
55
Figure 1.18. Paid sick leave replaces large parts of eligible employees’ wages, with significant
recent changes in regulations in a number of OECD countries
Cumulated gross sick-leave payments in the first four weeks of sick leave as a percentage of previous earnings for a
person who fell sick with COVID-19, rules valid in mid-May 2020
Non-mandatory employer sick pay in case of COVID-19 Mandatory paid sick leave in case of COVID-19 Baseline mandatory paid sick leave
%
100
90
80
70
60
50
40
30
20
10
0
Note: The results refer to an eligible full-time private-sector employee who is married with no kids, age 40, earning an average wage and working
with the same employer for one year. “Mandatory paid sick leave in case of COVID-19” refer to mandatory payments directly paid by the
government or by employers (often partly subsidised by the government) to eligible employees who contracted COVID-19. “Non-mandatory
employer sick pay” includes employer sick pay commonly agreed via collective agreements or other arrangements; these payments are included
for those countries were the majority of eligible employees would receive such payments. Baseline leave entitlements refer to regulations in
place in 2019, except for Australia, Israel, Japan, Korea, New Zealand and Turkey (all 2018). Countries emphasised with a dashed fill (Australia
and Spain) are those where employees are entitled to a benefit other than a dedicated sickness benefit.
Source: OECD (2020[67]) “Paid sick leave to protect income, health and jobs through the COVID-19 crisis”,
http://www.oecd.org/coronavirus/policy-responses/paid-sick-leave-to-protect-income-health-and-jobs-through-the-covid-19-crisis-a9e1a154/.
StatLink 2 https://stat.link/dyhgkp
… and employees in quarantine
Since the start of the pandemic, many workers across the OECD were required to temporarily shelter at
home for a variety of reasons. This may be because they had non-diagnosed mild symptoms; had close
contact with people who showed symptoms or had a diagnosis; or were at a higher risk of serious illness
in case of contracting COVID-19 because of existing health conditions.
The legal situation of eligible employees in mandated quarantine differs across countries, but they can
receive paid sick leave in almost all countries if they have mild symptoms and cannot continue to work
from home. Some, like Austria and Germany, have automatic mechanisms in place through their Epidemic
Acts that pre-date the COVID-19 pandemic. They treat quarantined employees who cannot work from
home as being on sick leave. The situation is similar in Finland and Sweden where quarantined employees
are entitled to paid sick leave following the countries’ regulations on infectious diseases. Other countries
took deliberate steps to broaden benefit coverage to quarantined workers (the Baltics, most Central
European countries, Denmark, Norway, Ireland and the United Kingdom) or introduced new crisis
payments for both sick and quarantined employees (Canada, New Zealand and the United States). In
Belgium and France, quarantined employees who cannot work from home can draw on short-time work
benefits.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
56
Better protecting sick self-employed workers
Paid sick leave can only be an effective tool during the containment, mitigation and post-confinement
periods if it is widely available to large parts of the labour force. This was by no means the case in all
countries prior to the crisis.
The self-employed stand out as a group of workers in non-standard employment who, prior to the current
public health crisis, often had poor or no access to sickness benefits (OECD, 2019[39]).23 However, rules
for this group differ substantially across the OECD. Only in a minority of countries, self-employed workers
have similar access to sickness benefits as dependent workers (Belgium, Denmark, Estonia, Finland,
France, Hungary, Iceland, Latvia, Lithuania, Norway, Slovak Republic, Spain and Sweden). A handful of
countries provide partial access to paid sick leave for self-employed workers, because of less
advantageous eligibility conditions, benefit amounts or receipt durations (Austria, Germany, Ireland,
Portugal and the United Kingdom, OECD (2020[69])).24 Also, waiting periods in many countries are
significantly longer for the self-employed than for dependent employees, with the aim to reduce costs
and/or moral hazard. In the course of this pandemic, many OECD countries have temporarily expanded
access to sickness benefits for self-employed workers who are sick with COVID-19 or quarantined (OECD,
2020[67]). A number of countries have improved access for self-employed workers to (immediate) paid sick
leave by lowering or eliminating the (often much longer) waiting periods and providing them with
entitlements in case of mandatory quarantine (Denmark, Latvia, Norway, Portugal and Sweden). Several
countries temporarily reformed sickness benefits and extended entitlements that affect dependent and self-
employed workers alike. For example, Estonia, France, Ireland and the United Kingdom temporarily
waived the waiting period that was of similar length for both dependent and self-employed workers, and
Ireland and the United Kingdom increased sickness benefit generosity. Some countries have introduced
new payments or provided entitlements to a benefit other than a dedicated sickness benefit, which self-
employed workers can access much like dependent workers (Australia, Canada, Finland, Korea,
New Zealand, Spain, Switzerland and the United States). In almost all cases, however, measures are time-
bound and limited to COVID-19 cases only.
Paid sick leave entitlements do not give a full picture of the level of income support available to the
self-employed in case of sickness or quarantine in every country. Some countries have chosen to use
different tools and benefits to support self-employed workers unable to support themselves during this
crisis irrespective of whether they had to scale down or suspend their business operations because of
sickness, quarantine or other reasons relating to lockdowns. Indeed, various countries provided hardship
funds or other payments to self-employed workers (see Section 1.3.3).
… and easing the costs for employers
Many countries provide strong incentives to employers to prevent sickness and assist sick workers in their
return to work by making them financially responsible for sick pay during an initial period of several days,
weeks or months (Palme and Persson, 2020[76]). Arguments for employer funding of sick pay, however, do
not seem to apply or even risk being counterproductive in the outbreak phase of a pandemic. In the case
of a very contagious disease, prevention requires keeping workers at home rather than encouraging them
to come to work. Reintegration is not directly relevant in a confinement. It is also not obvious that employers
should pay for extensions of existing legislation, especially if they are facing major financial stress already.
In such situations, temporarily lifting or reducing employer costs (through direct payments or tax credits)
seems justified.
Countries have reacted very swiftly to this new challenge, and many have introduced measures to support
employer costs for sick pay. More than half of the OECD countries for which information is available and in
which employers have sick-pay obligations have changed their regulation accordingly, or, like Austria and
Germany, enforced their Epidemic Acts, which include an automatic adjustment to reduce employer costs.
In some countries, employers can seek reimbursement for their sick-pay costs while in others workers sick
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
57
with COVID-19 can receive a public sickness benefit from the first day. Public funding of the costs of sick
pay for quarantined workers is even more common, for good reasons, and more often from the first day of
an entitlement. Through the reforms in the funding mechanism taken between March and May 2020, the
employer contribution during a four-week sick leave fell from around 50% to around 20% of an employee’s
gross wage at average earnings level on average across OECD countries. For a worker entitled to sick
leave in two-week mandatory quarantine, the employer contribution is on average less than 10%.
How important has paid sick leave been in the current crisis?
Real-time data on the take-up of sick leave are available less regularly than data for unemployment or
other social benefits, partly because of limited reporting requirements in many countries during the period
covered by employers. However, preliminary data available for about a dozen OECD countries suggest
that take-up of sick leave rose significantly during March and early April, often by between 30% and 100%,
with typically between 4% and 6% of the workforce receiving paid leave. 25 Data on the change in the
composition of paid-leave receivers are even scarcer, but they suggest that a sizable share of the increase
could be due to quarantined rather than sick workers. Latest data for late April and May also suggest a
sudden decline in sick-leave numbers in a few countries, such as Austria, Germany, Italy and Sweden,
largely explained by a lower likelihood of those who are teleworking to take sick leave.
Overall, the increasing use of sick leave to no more than 6% of the workforce may seem small, compared
to the massive inflow into short-time work and/or unemployment schemes in many OECD countries. In
part, use of sick leave remains low exactly because in this early phase of the crisis, short-time work has
become so prevalent. Also, a doubling in sick-leave rates is a very significant change given that at any
moment in time, infection rates of COVID-19 are a small proportion of the total population in most countries,
and a relatively small share of people with COVID-19 symptoms require sick leave for very long periods.
Helping workers with additional family care needs
The COVID-19 crisis has increased the demands on many workers to provide family care. Some workers
have had to provide care for relatives who contracted COVID-19, or who are in quarantine or self-isolation.
Many others were affected by the scaling back or closure of childcare facilities, schools and other social
care services, including for the elderly and those with disabilities.
The widespread closure of schools and childcare facilities has had a particularly stark effect. Worldwide,
more than 190 countries closed their schools at some point since the start of the crisis, affecting, at peak,
more than one and a half billion students (UNESCO, 2020[77]). Among OECD countries, only Australia,
Iceland, Sweden and the United States have decided against countrywide closures (UNESCO, 2020[77]).
School and childcare facility closures have caused considerable difficulties for working parents: many have
had to lead or supervise home schooling, and most have had to (arrange) care for their children during the
working day.
Working full hours is often very difficult, if not impossible, under such circumstances, notably for single
parents and couples where only one partner can telework. Parents with younger children, who require
closer attention, report particular difficulties balancing work and family (Eurofound, 2020[78]). Couples
where both parents have to be physically present at their workplace have faced an even greater challenge.
This includes many workers in essential occupations.
Pre-existing rights to special leave
In most OECD countries, workers have a well-established right to leave to care for sick or injured children
(OECD, 2020[79]). In several countries, these (or separate) family care leave rights also extend to other
sick dependents and adult relatives (e.g. Australia, Austria, Canada,26 the Czech Republic, Estonia, Israel,
the Netherlands, New Zealand, Norway, Poland, Portugal, the Slovak Republic and Slovenia). Family care
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
58
leave is usually paid, often at or near full earnings-replacement, either by the government or public social
insurance, or through continued payment of salaries. However, except in cases of serious illness, the
duration is often limited: in many countries, these rights stretch to only a few days per episode (e.g. Finland,
France, Switzerland) or a week or two per year (e.g. Australia, Austria, Germany, Israel, the Netherlands,
New Zealand, Norway and the Slovak Republic), which may not be sufficient for a period of quarantine or
prolonged infection. Entitlements to longer care leaves are usually only available for critical or terminal
illness. In a few countries (e.g. Israel and New Zealand), any days used for family care are deducted from
the worker’s own sick leave entitlement.
Parents’ rights to special leave in cases of school or childcare facility closure are less well established.
Only a small number of OECD countries provide parents with a pre-existing right to leave in case of school
or childcare closure (e.g. the Czech Republic, Lithuania, Poland and the Slovak Republic) or other
“unforeseen emergencies” (e.g. Australia and the United Kingdom) or force majeure (e.g. Ireland).
Moreover, in some countries (e.g. the United Kingdom), these rights extend only as far as unpaid leave,
with the decision to continue payment of salaries typically left to the employer or collective agreement.
Many working parents are unable to afford to take unpaid leave for a prolonged period. In several others,
the right to paid leave lasts only for a couple of weeks (e.g. Australia, the Czech Republic, Lithuania and
the Slovak Republic) or less (e.g. Ireland). These rights would be quickly exhausted in the face of closures
spanning months, as seen through this crisis.
In response to the limitations of existing rights, many countries have stepped up support for working
parents and those with additional family care needs, usually through temporary emergency measures.
Targeted childcare options for parents in “essential” occupations
Several countries that closed childcare facilities and schools (such as Austria, Denmark, France, Germany,
Latvia, the Netherlands, Norway and the United Kingdom) kept some facilities open, with a skeleton staff, to
look after children of essential service workers, notably in health, social care and teaching. In France, for
example, childcare facilities for such families could host up to ten children, and childminders working out of
their homes could exceptionally receive up to six rather than three children. In New Zealand, essential service
workers with children aged 5 to 14 could benefit from state funded in-home childcare while schools were
closed. Korea introduced a similar in-home childcare scheme that covers all two-earner families with children
under age 12. In Australia, the Commonwealth Government is subsidising childcare facilities that remain open
during the crisis.
Special paid leave for workers affected by school or childcare closures
A number of countries have introduced or extended special paid leave (or special income support for those
on unpaid leave) for working parents who provide care at home while schools or childcare facilities are
closed. This includes Austria, Belgium, Canada, the Czech Republic, Finland, France, Germany, Greece,
Italy, Japan,27 Korea, Lithuania, Luxembourg, Norway, Poland, Portugal, the Slovak Republic, Slovenia,
Sweden, Switzerland, the United Kingdom and the United States. In most of these countries, the right to
special paid leave or income support lasts for a fixed number of days or weeks, ranging from 10 days (per
parent) in Korea to up to 12 weeks in the United States and four months in Canada. However, in some
(e.g. Belgium, the Czech Republic, Finland, France, Luxembourg, Switzerland), it extends for as long as
schools and childcare facilities are closed. In almost all countries, the right to paid leave is conditional on
no alternative care arrangement being available.
In several countries, workers taking special leave receive either a flat-rate payment (e.g. Belgium, Canada,
Finland and Korea) or a fixed part of their salary (e.g. the Czech Republic, France, Germany, Italy,
Portugal, Switzerland, the United Kingdom and the United States); in a few (e.g. Austria, Norway), leave-
takers continue to receive their salary in full. In a minority of countries (e.g. Austria, Greece, Norway and
Portugal), financing is shared between employers, general taxation and/or public social insurance.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
59
However, as continued payment of salaries is likely to be difficult for many employers, most countries have
looked either to minimise employer contributions or to fund special leave entirely through general taxation
or social insurance.
Self-employed workers with care responsibilities can find themselves in a particularly vulnerable position.
Most countries exclude them from existing family care leave regulations, and self-employed workers may
face substantial income losses if they cannot arrange care or schooling for their children. Some countries 28
have therefore extended special paid care leave (or income support) to self-employed workers. However,
the financial compensation they receive can be lower.
1.3.2. Securing jobs, saving companies and maintaining essential service provision
Liquidity relief for firms
Mandatory business restrictions, quarantines and limitations on individual mobility have put companies
under severe strain. With sales plummeting, even productive, well-managed firms faced major liquidity
shortages in responding to their financial commitments to suppliers, employees, lenders, investors and the
state. The large number of simultaneously affected firms limited access to trade credits, an otherwise
important source of short-term financing.
OECD (2020[60]) estimates that without public support, 20% of firms would have faced a liquidity crisis after
the first month of lockdown, and 40% after three months. Failure to immediately address such liquidity
constraints may lead to a corporate solvency crisis as companies with reduced or no revenues for an
extended period of time go bankrupt. A series of corporate bankruptcies would severely disrupt not only
value chains but also the banking and financial system. Well aware of these risks, all OECD countries
adopted a vast range of emergency measures aimed at supporting firms’ liquidity (see Figure 1.16) in
addition to the monetary measures taken by central banks. These ranged from deferrals in tax and social-
security contributions to liquidity injections through equity participation, direct subsidies based on past
sales, subsidies for maintaining employment, grants. Many countries also took specific measures to
support SMEs, which usually face stronger liquidity constraints (OECD, 2020[80]).
Job retention schemes
As demand collapsed and supply chains broke, companies also found themselves with excess capacity.
This put jobs at risk on a large scale. Job retention schemes (JRS) have been one of the main policy tools
for many OECD governments to contain the employment and social fallout of the COVID-19 crisis and
avoid massive layoffs (see Section 1.2). They seek to preserve jobs at firms experiencing a temporary
reduction in business activity by alleviating firms’ labour costs while supporting the incomes of workers
whose hours are reduced. They can take the form of short-time work (STW) or temporary layoff schemes
that directly subsidise hours not worked, such as the German Kurzarbeit or the French Activité partielle.
They can also take the form of wage subsidy schemes that subsidise hours worked but that can also top
up the earnings of workers on reduced hours, such as the Dutch Emergency Bridging Measure
(Noodmatregel Overbrugging Werkgelegenheid, NOW) or the Job Keeper Payment in Australia. They differ
in their generosity for firms and workers and the requirements that they impose for eligibility (e.g. economic
need, agreement by social partners) and on the behaviour of participating firms and workers
(e.g. restrictions on economic dismissals, job search by workers) (Hijzen and Venn, 2011[81]).
A crucial aspect of all JRS is that employees keep their contracts with the firm even if their work is
suspended. This allows firms to hold on to workers’ talent and experience and quickly ramp up operations
once economic activity recovers. They provide the necessary liquidity to permit firms to continue paying at
least part of workers’ salaries and to prevent the termination of jobs that have temporarily become
unprofitable but that are likely to remain viable in the medium term. Consequently, they prevent layoffs that
are inefficient for the firm itself and costly for workers and society at large. Indeed, one of the lessons
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
60
learned from the global financial crisis was that STW schemes can play an important role in mitigating the
economic and social costs of major demand shocks (OECD, 2010[82]; 2018[83]; Hijzen and Martin, 2013[84];
Cahuc and Carcillo, 2011[85]; Hijzen and Venn, 2011[81]). In the current context of a “self-imposed” supply
shock, in which governments shut down many activities and imposed severe restrictions, the widespread
use of JRS seems even more sensible. Some countries explicitly prohibit companies participating in JRS
from dismissing workers while they participate in the scheme (the Netherlands during the first three months
of the programme, New Zealand, Poland), and in some cases including a short period after (Austria,
France, Hungary and Spain).
In the early stages of the COVID-19 crisis, the overriding concern for governments has been to help firms
and workers deal with the sudden and unpredictable decline in business activity resulting from the health
crisis and government-imposed restrictions. To this end, many governments have modified existing JRS,
or introduced new ones, to maximise take-up (see Box 1.6 for a presentation of four country cases).
Concerns over the potential negative effects of JRS, which arise in ordinary times, were initially of second
order. In particular, the risk of devoting public resources to support jobs that employers would have retained
anyway seemed limited. Restrictions in business activity during confinement heavily reduced sales and
hence financial resources in many firms across almost all sectors. In ordinary times, JRS can also impede
the reallocation of workers to more productive firms. But also this risk seemed limited during the early
phases of the current crisis, given the virtual standstill in hiring and since the government-imposed
restrictions and physical-distancing measures affected many firms independently of their pre-crisis
performance.
However, as countries move out of the strict confinement phase, policy makers have to strike the right
balance between ensuring adequate support for jobs that are temporarily unviable and limiting the extent
to which subsidies reach jobs that would be preserved anyway or that are unviable even in the long term
– see Section 1.4 and (OECD, 2020[86]). Institutional differences in JRS across countries typically reflect
different approaches to addressing this challenge – see OECD (forthcoming[20]), Hijzen and Venn (2011[81])
and Müller and Schulten (2020[64]) for in-depth discussions of the main features of existing and new JRS.
Many countries rapidly expanded their STW schemes in the early weeks of the crisis or
introduced new ones…
Twenty-two OECD countries had a STW scheme in place before the crisis erupted (Table 1.1), and a
further ten countries introduced new schemes in response to the crisis. All countries with pre-existing
schemes rapidly adjusted them to cope with the COVID-19 crisis.29 Countries’ measures to expand existing
STW schemes fall into three broad categories:
Simplifying access and extending coverage: Nineteen countries took measures to facilitate and
expedite access to STW and boost take-up among the affected firms. Several countries where
firms are required to provide an economic justification have adjusted the parameters to allow firms
to claim STW if they have experienced a decline in business activity since the start of crisis
(e.g. Japan, Korea and Poland). In others, firms can invoke the health crisis as a force majeure by
a simple declaration (e.g. Belgium, Czech Republic, France, Italy and Spain). Germany and
Norway lowered the minimum permissible reduction in working time for firms to gain access to their
STW schemes. Italy, where STW was limited to large firms and certain sectors, extended its
scheme to firms of all sizes in all sectors. France and Italy removed the condition that employers
must consult with workers’ representatives before applying for the scheme. Countries also
simplified and streamlined procedures with widespread use of online applications. The
United Kingdom facilitated the fast adoption of the newly introduced Coronavirus Job Retention
Scheme by a simple online application procedure that allows retroactive claims.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
61
Table 1.1. Countries have adjusted existing job retention schemes or adopted new ones
Pre-existing Increased Increased Increased New short-time New wage
short-time work access and benefit access for work scheme subsidy scheme
scheme coverage generosity workers in non-
standard jobs
Australia
Austria
Belgium
Canada
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Iceland
Ireland*
Italy
Japan
Korea
Latvia
Lithuania
Luxembourg
Netherlands*
New Zealand
Norway
Poland
Portugal
Slovak Republic
Slovenia
Spain
Sweden
Switzerland
Turkey
United Kingdom
United States
Note: Ireland and the Netherlands replaced their existing STW schemes with temporary wage subsidy schemes.
Extending coverage to non-permanent workers: Nine countries extended eligibility beyond workers
in standard forms of employment to include temporary, temporary agency and even certain
categories of self-employed workers. In principle, this should reduce the risk that STW schemes
reinforce labour market duality (Hijzen and Venn, 2011[81]). However, firms may have weak
incentives to hold on to workers in non-standard forms of work during periods of STW, especially
if the scheme imposes a direct cost on employers.
Raising generosity: Several countries have increased the generosity of STW schemes by raising the
replacement rates for workers and reducing the costs for firms. Sixteen countries increased the
effective replacement rate for hours not worked. In several countries where employers were required
to pay part of the wages or social-security contributions for the hours not worked these costs were
reduced to zero (e.g. France, Germany and Italy). In about half of all countries, this cost was already
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
62
zero before the crisis. The higher replacement rates and lower employer cost in the early stage of
the crisis indicate that countries gave more weight to the need to provide support for workers and
businesses than to concerns about the possible disincentive effects of the measures adopted.
While most of these changes are temporary, governments have generally made clear that the schemes
will remain in place for as long as necessary to reduce uncertainty.
Box 1.6. Job retention schemes in Germany, Italy, Japan and the United States
Germany
Germany simplified access to Kurzarbeit. Since March 2020, firms can request support if 10% of their
workforce are affected by cuts in working hours, compared to 30% before. Employers initially continue
to pay their employees any actual hours worked plus 60% of their net earnings losses because of
reduced hours (67% for employees with children). The public employment service reimburses
employers for these payments as well as for 100% of social-insurance contributions for the lost work
hours (compared to a 50% reimbursement of social-insurance contributions during the global financial
crisis). The subsidy is normally also available to workers on temporary contracts and apprentices and
it was extended to agency workers at the start of the crisis. In April, the government increased the
statutory replacement rates for lost earnings to 70% from the fourth month and 80% from the seventh
month onwards (and respectively to 77% and 87% for employees with children). In addition, restrictions
on taking another job while on STW have been lifted. Workers are allowed to cumulate additional
earnings and STW benefits as long as total income does not exceed previous earnings. In some
sectors, unions and employers agreed on higher replacement rates of up to 90%.
Italy
Italy greatly extended the reach of its STW scheme by allowing firms of any size and from all sectors to
apply. Firms can simply declare that they have been negatively affected by the COVID-19 crisis without
having to provide detailed evidence. They can apply within four months of the start of the reduction in
activity and the benefits can be paid retroactively from the end of February 2020. Nevertheless, some
of the intended new beneficiaries have experienced difficulties in accessing the scheme and receiving
prompt support. Employers’ participation in the cost of the scheme has been suspended, while benefit
levels for workers remain unchanged. Benefits pay 80% of gross wages and they are capped at
EUR 998 for wages up to EUR 2 159 and at EUR 1 199 for wages above that level. For a worker with
an average wage this translates into an effective replacement rate of about 45% when hours are
reduced to zero (OECD, forthcoming[20]).
Japan
Japan expanded the coverage and eased the requirements for access to the Employment Adjustment
Subsidy. Up until the crisis, access to the Employment Adjustment Subsidy required a 10% reduction
in production for more than three months. This has been reduced to 5% over one month. Japan
increased the subsidy rates for hours not worked to a maximum of 100% for SMEs and to 75% for larger
firms. In May 2020 the government announced an increase in the maximum benefit by 80% (from
JPY 8 330 to JPY 15 000 a day per employee). The programme has been extended to cover
non-regular workers who are not covered by employment insurance. The government further
announced a new scheme to cover workers who have remained without support because their small
and medium-sized employers have not applied for the subsidy despite reducing hours. These workers
will be able to apply to the new scheme directly and will have 80% of their usual earnings covered.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
63
United States
In the United States, 26 States (accounting for about 70% of the population) operate Short-Time
Compensation (STC) programmes. Through the Coronavirus Aid, Relief, and Economic Security
(CARES) Act, the Federal Government now funds 100% of STC payments in States with an existing
programme and 50% in States that introduce a new one. Also, STC recipients qualify for the same
weekly USD 600 increase in benefit payments that is being made to all unemployment benefit recipients
for the a period of four months (see Section 1.3.3). However, the use of STC remains very limited for a
variety of reasons, including administrative bottlenecks, lack of employer awareness, weak financial
incentives for employers (employers are liable for their part of social-security contributions for hours not
worked) and limits to the maximum reduction in working hours. To bypass such problems, the
United States introduced the Paycheck Protection Program (PPP) to provide small businesses 1 with
loans to pay their employees during the COVID-19 crisis, which are forgiven if employment and
compensation levels are maintained. While providing direct and immediate support, PPP is not a STW
scheme as it is not conditional on having financial difficulties (a reduction in turnover and/or working
time).
1. Publicly listed companies may apply but must satisfy, in good faith, that the “current economic uncertainty makes the loan necessary to
support ongoing operations”.
… while several others introduced temporary wage subsidy schemes
A number of – mostly English-speaking – countries have introduced new JRS that combine elements of
standard wage subsidies (i.e. subsidies for hours actually worked) with elements of STW schemes. These
schemes can also provide income support to workers who are temporarily not working or, more generally,
top up earnings of workers on reduced hours. Australia, Canada, Ireland and New Zealand introduced
temporary wage subsidies to cover part of normal earnings. In Canada, the subsidy covers 75% of gross
normal earnings (subject to a cap), whereas in Australia and New Zealand, schemes provide lump-sum
transfers to firms. In Ireland, the subsidy level varies with employees’ income reaching a maximum of 85%
of net normal earnings for the lowest incomes. The Netherlands replaced its pre-existing STW scheme by
a temporary wage subsidy, which is proportional to the reduction in sales and not the reduction in working
hours as in traditional STW schemes. The subsidy ranges from 22.5% of earnings in case of a 25%
reduction of sales to 90% of earnings when sales fall to zero entirely. Employees continue to receive 100%
of their usual earnings.
Wage subsidy schemes may be more flexible than traditional STW schemes, but tend to
be less well targeted
Wage subsidy schemes tend to be easier to implement than STW schemes and provide more flexibility to
firms, while being less well targeted to firms experiencing financial difficulties. They grant subsidies for
workers present in the firm at the start of the programme if these experience a significant decline in
business activity, typically in the range of 20-30%. Firms can decide themselves to what extent the subsidy
is used to support hours worked (i.e. as a pure wage subsidy) or hours not worked (i.e. as a STW scheme).
Since the subsidy is not dependent on the reduction in hours worked, firms do not need to report how the
working time reduction is distributed across workers and how it evolves over time. 30 Besides, while all
wage subsidy schemes considered here target firms experiencing significant declines in business activity,
the size of the subsidy per worker is independent of the decline in business activity in Australia, Canada,
Ireland and New Zealand. It is more strongly targeted in the Netherlands, where the subsidy per worker is
proportional to the decline in firms’ sales.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
64
There may be various reasons for why these countries have opted for introducing temporary wage subsidy,
rather than STW, schemes (OECD, forthcoming[20]). First, with the exception of the Netherlands, these
countries had no or limited earlier experience with STW schemes: Australia never had a STW scheme;
Canada, Ireland and New Zealand operated STW schemes during the global financial crisis, but these
were not very widely used. Second, firms in these countries typically face relatively low layoff costs and
therefore would have little incentive to participate in a STW scheme that generally involves considerable
procedural costs. Finally, by reducing the cost of hours worked, wage subsidies may provide incentives for
firms to maintain higher hours and increase them more quickly when conditions improved.
Limiting economic dismissals and protecting workers against unfair dismissal
A number of OECD countries introduced restrictions to collective and individual dismissals during the
current crisis (see Chapter 3 for a detailed discussion of employment protection legislation in OECD
countries) to limit an immediate rise in layoffs and ensure high take-up of STW schemes. These measures
include:
An explicit ban on economic dismissals: Italy made invalid collective or individual dismissals based
on economic grounds that were initiated after start of the confinement measures. This includes
dismissals for reasons connected to the reduction or the transformation of activities, to the
reorganisation of work, and to the closing of the business for total cessation of activities. At the
moment of writing, the ban is in force until 17 August 2020. Greece also introduced some limitations
to economic dismissals but limited them to companies that benefitted from the COVID-19 support
measures.
Increased scrutiny and costs: In Spain, any dismissal related to COVID-19 would be qualified by a
judge as either null, resulting with the employee being reinstated, or unfair in which case the
employee receives a compensation of 33 days of pay per year of tenure. France announced
increased scrutiny of collective dismissals in companies with more than 50 employees by the
authority to which these companies must notify the intention to dismiss a worker.
Limiting dismissals of employees with a permanent contract can contribute to maintaining incomes and
demand of workers during a period of already strong anxiety, limit opportunistic behaviour of few employers
who may use the crisis as an excuse to dismiss “difficult” workers and protect workers from the social
stigma of being fired.
However, in particular in case of economic dismissals, a strict ban may also provoke additional company
bankruptcies if access to JRS and other liquidity support programmes turns out to be incomplete,
impractical, delayed or too costly. A ban on dismissals also risks further shifting the burden of the
adjustment on temporary contracts, which can be terminated by simply not renewing them. Also limits on
the number of renewals and maximum durations of fixed-term contracts (see Chapter 3) may further limit
the possibility of renewal during the pandemic. To limit such risk, Spain allowed the continuation of
temporary contracts reaching the legal maximum duration during the crisis. Facing a surge in non-renewal
of temporary contracts, Italy relaxed in May the valid cases for renewal of fixed-term contracts beyond the
first year.
During post-confinement, and when combined with generous JRS, strict limitations to economic dismissals
may inhibit restructuring processes and slow down the recovery (see Section 1.4). Some workers may
remain locked in unviable companies instead of being taken care of by public employment services, which
could offer re-training and other support. They can also hold back necessary structural change in the labour
market, inhibiting mobility from sectors whose activity may remain subdued for some time (such as aviation,
tourism and entertainment) to those that may be growing again more quick (such as health care and online
and delivery services).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
65
An economic crisis that results from a pandemic also raises important questions on the boundaries of what
may or may not qualify as dismissals on personal grounds. As discussed in Section 1.3.1, the current crisis
greatly increased the number of work absences by employees who were sick, had to deal with family care
needs or could neither come to the office nor effectively work from home. Sick workers are protected
against dismissal by sick-leave policies (where such policies exist). However, employers may dismiss their
staff for personal or economic reasons during a medical leave provided that the sickness is not the reason
for dismissal. Unauthorised absences may also be a reason for fair dismissal in cases where employees
have used the totality of leave days and are still unable to return to work. This is an issue in times where
schools are closed and family members may be sick. Finally, dismissals on personal grounds may affect
employees unable to perform efficiently their work duties from home, and those refusing to come to work
because of sanitary concerns at the workplace or on the commute.31
To avoid such risks, Italy and the Slovak Republic also introduced some provisions to limit dismissal on
personal grounds. In Italy, parents living with a disabled child cannot be dismissed from work if they are
absent to care for their child provided that the absence has been previously communicated and motivated.
Parents of children between 12 and 16 years have the right to abstain from work during the period of school
closures, and their absence cannot be a cause for dismissal. In the Slovak Republic, employees who have
to take care of sick family members or young children following school’ closings are considered to be
temporarily unfit for work and therefore protected from dismissal.
Supporting essential production and services
At the peak of the pandemic, when large parts of the economy were shut down in many OECD countries,
certain sectors, such as health and long-term care, agriculture, food processing and retail and logistics,
had to continue operating smoothly. Absences of workers who were sick, quarantined or blocked at home
caring for their children, as well as the inability of seasonal workers to travel to their workplaces from
abroad, put pressure on some of these sectors. To avoid the risk of disruptions, OECD countries took a
number of measures to promote labour mobility. This includes:
Incentives to take up a job while unemployed or on STW. In most OECD countries, workers who
receive unemployment or STW benefits cannot – or have little financial incentive – to complement
these benefits with other types of income. This may be a source of concern when a large share of
the population is not working, and some sectors face labour shortages. To address this issue,
Greece, Italy and Spain, for example, have temporarily allowed unemployed people to complement
unemployment benefits (as well as minimum income benefits in Italy) with earnings from a job in
agriculture. Germany lifted restrictions on taking on part-time work for workers on STW. Additional
earnings are not credited against STW benefits as long as total income does not exceed previous
earnings. In Belgium and Italy, workers on STW are exceptionally allowed to take up a job in
agriculture without losing their benefits. The possibility of complementing income from STW with a
new job already existed in France, but it was simplified by giving workers a seven-day notice to be
called back in the old job and quit the new one. The new JRS introduced in the United Kingdom
also explicitly allows workers to take up another job and cumulate the earnings.
Promoting “loans” of workers across companies. France has actively promoted “loans” of workers
across companies (la mise à disposition). With the agreement of the worker and the two
companies, an employee can temporarily be loaned to another company while keeping the original
employment contract and wage. Loans of employees across companies exist in other countries
(e.g. Belgium and Italy) but, in general, they do not appear to be extensively used.
Adjusting working-time regulation. Several countries introduced some flexibility in their
working-time regulation. France, for example, granted essential services derogations from the
regulation on maximum working hours and weekend work. Between April and May 2020, workers
in essential services could work up to 12 hours a day (up from ten hours a day) and 60 hours per
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
66
week (up from 48 hours). The Slovak Republic also loosened working-time regulation by allowing
employers to announce schedules at a shorter notice (two days instead of one week in normal
times).
Loosening the use of temporary contracts for essential services. Belgium, for example, extended
the maximum permitted number of consecutive temporary contracts for essential-service workers.
1.3.3. Providing income security and employment support to affected workers
Income support for those losing their job or their self-employment income
In spite of governments’ bold efforts to protect jobs by expanding or newly introducing job retention schemes
and providing emergency liquidity support to firms (see Section 1.3.2), millions of workers across the OECD
have lost their jobs. In the United States alone, over 40 million workers have filed new claims for
unemployment insurance benefits between March and May 2020. Many more lost work but did not register
as unemployed, or had their hours of work considerably reduced (see Section 1.2). Meanwhile, many self-
employed workers saw their incomes collapse because they had to suspend, or substantially downscale,
their business operations during lockdown. Unemployment benefits and other out-of-work support
programmes cushion income losses for households affected by job loss or by a large fall in self-employment
income. They are crucial for reducing economic hardship and contribute to stabilising the economy by
bolstering aggregate demand, as experienced during the global financial crisis (OECD, 2014[87]).
Workers in standard jobs can often count on timely income support in case of job loss
In many OECD countries, workers in “standard” (i.e. open-ended, full-time) dependent employment are
comparatively well covered in case of job and income loss. Unemployment insurance benefits are typically
the first support layer in the initial phase of unemployment, replacing a certain share of previous earnings
for a limited time. Some countries also operate unemployment assistance systems, which provide less
generous support to jobseekers who lack the required employment or contribution histories, or who have
exhausted their benefit entitlements. Jobseekers in low-income households may also qualify for
non-contributory, means-tested minimum-income benefits, such as social assistance. The 2019 OECD
Employment Outlook (OECD, 2019[88]) showed that in most of the 17 European OECD studied, a large
majority of job losers with past continuous full-time employment had access to some sort of income support
in 2014/5 (Figure 1.19). However, it also documented substantial coverage gaps: only around 50% of
workers in standard jobs in Greece and Italy and around 60% in Poland received any income support
following job loss.32 Moreover, even workers who are covered can experience a very substantial drop in
income. For example, in about one in two OECD countries, out-of-work support for workers on modest pay
amounts to less than two-thirds of past net earnings during the initial phase of unemployment. 33
… but substantial coverage gaps exist for those in non-standard and informal employment
Workers in non-standard forms of employment are, on average, significantly less well covered by existing
social-protection schemes. The 2019 OECD Employment Outlook (OECD, 2019[88]) illustrated that those
engaged in self-employment, short-duration or part-time employment are often less likely to receive any
form of income support during an out-of-work spell than those in standard employment (Figure 1.19). In
some countries, such as the Czech Republic, Estonia, Latvia, Portugal and the Slovak Republic, the
coverage gap relative to workers in standard jobs reaches 40-50%. While gaps tend to be larger for the
self-employed, part-time workers and those with frequent transitions between employment and
unemployment also find it difficult to access out-of-work support in some countries (see Chapter 2). Already
before the COVID-19 crisis, many countries were therefore exploring how to shore up access to out-of-work
benefits in the context of a changing world of work.34
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
67
Figure 1.19. Workers in non-standard jobs are often less well covered by income support
Probability of receiving income support benefits for out-of-work individuals, by past employment, 2014-15
Baseline: past standard work Past non-standard (significant gap) Past non-standard (non-significant gap)
%
100
80
60
40
20
0
GRC ITA POL EST GBR AUT LVA PRT CZE SVK FRA HUN LUX ISL ESP SVN BEL
Note: Predicted benefit receipt during an entire year comparing: i) an able-bodied working-age adult who is out of work, had uninterrupted full-
time dependent employment with median earnings in the preceding two years, and lives in a two-adult low-income household without children
(“baseline: past-standard work”, triangle-shaped markers); and ii) an otherwise similar individual whose past work history is “non-standard”:
mostly in part-time work, mostly self-employed, or interrupted work patterns during the two years preceding the reference year (“past non-
standard”, light and dark diamond-shaped markers). Additional results for different categories of non-standard work are available for some
countries.
Statistical significance refers to the gaps between baseline and comparator cases (90% confidence interval). Full-time students and retirees are
excluded from the sample. The data source, the European Union Statistics on Income and Living Conditions (EU-SILC) covers additional
countries but they are excluded here because effective sample sizes were small (e.g. Ireland, Lithuania), because the required micro-data were
entirely unavailable (Germany), because key employment-status variables are recorded only for one individual per household (Denmark, Finland,
the Netherlands and Sweden), or because of partial or partly conflicting information on income or benefit receipt (Norway). For further details,
see Fernández, Immervoll and Pacifico (forthcoming[89]).
Source: OECD (2019[39]), OECD Employment Outlook 2019: The Future of Work, https://dx.doi.org/10.1787/9ee00155-en.
StatLink 2 https://stat.link/4728tj
Informal workers, including undocumented migrants, remain beyond the scope of contributory income
support schemes (OECD, 2020[37]). This includes employees who are not registered for mandatory social
security, who are paid less than the legal minimum wage, who are employed without a written contract
(where this is a legal requirement), and the self-employed who fail to declare some or all of their income
for tax purposes (e.g. working “cash in hand”). Workers in “partial informality”, whose employment is
registered but who receive some of their remuneration in cash (“envelope wages”) may have some
entitlements but will not receive compensation for all of their lost earnings. As a result, the effects of the
combined health and economic crisis for households are far more dramatic in emerging and developing
countries where informality rates are much higher and the vast majority of the population does not have
access to formal social-protection arrangements and cannot afford to shelter at home.
Many OECD countries extended unemployment benefits during the COVID-19 crisis
Nearly two in three OECD countries took steps in the early weeks of the crisis to improve the accessibility
and/or the generosity of “first-tier” unemployment insurance or “second-tier” unemployment assistance
benefits (see Table 1.2). While those measures primarily benefited workers in standard employment who
lost their jobs, some countries temporarily opened up benefits to groups who otherwise would not have
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
68
qualified, such as workers in non-standard employment. Countries’ early measures can be grouped into
three broad categories:
Improving access to and coverage of unemployment benefits. 16 OECD countries widened access
to unemployment insurance benefits by reducing or entirely waiving minimum-contribution
requirements (Finland, Israel, Norway, Spain and Sweden), extending the qualification period for
the employment requirement (France, Switzerland) or covering groups that had previously not been
entitled. This includes self-employed workers (in Finland and the United States), workers whose
contract was terminated during the trial period (Spain), workers on unpaid leave (Israel) and
workers who quit their job for a new job offer that fell through when the crisis hit (Belgium, France,
Spain). Canada, Latvia, Ireland, New Zealand and Slovenia introduced new unemployment
assistance benefits, Colombia made extraordinary payments to jobseekers who had not received
any unemployment benefits in the last three years. Australia temporarily relaxed the means-testing
of its unemployment benefit. In addition, a number of countries suspended or relaxed “active job
search” conditions and related activity requirements for benefit claimants.
Extending unemployment benefit durations. 12 OECD countries have lengthened the maximum
possible duration of unemployment benefit payments. Some automatically extended all expiring
benefit claims up to a certain time (until the end of the health crisis / state of emergency in
Luxembourg, Portugal and Spain; until the end of June in Norway), others for a specific period of
time (by two months in Greece, Italy and the Slovak Republic, by three months in Germany and
Switzerland, for the duration of the health crisis in Denmark and France). In the United States, the
Federal Government extended the maximum unemployment benefit duration to nine months.35 In
addition, a number of countries suspended benefit waiting periods to make support available from
the first day of unemployment.
Raising unemployment benefit generosity. Ten OECD countries temporarily increased benefit
levels: Australia introduced a coronavirus supplement of AUD 550 per fortnight for recipients of the
main out-of-work benefits and for a duration of six months. The United States raised benefits by
USD 600 per week for all recipients for a maximum period of four months. As a result of this
lump-sum increase, an estimated two-thirds of eligible unemployed workers will receive
unemployment insurance benefits that exceed their lost earnings (Ganong, Noel and Vavra,
2020[90]). Norway increased replacement rates to 80% or 62.5% depending on previous income.
Sweden raised the unemployment benefit floor (by about 30%) and ceiling (by about 40%) for
100 days. Austria, New Zealand and the United Kingdom raised benefit levels of their
unemployment assistance programmes, Colombia made extraordinary benefit payments. Belgium
froze the automatic decline in replacement rates over benefit spells for three months. Finland
raised earnings disregards for unemployment benefit recipients. France postponed part of a reform
that changes the calculation of unemployment benefit levels.
A major concern in the early weeks of the crisis in March and April 2020 was that public employment
services in some countries lacked capacity to deal with the soaring jobseeker numbers (Edwards, 2020[91])
and in some cases failed to ensure the timely pay-out of unemployment benefits. For example, an online
survey conducted in the United States in mid-April suggests that for every ten people who had filed for
unemployment benefits in the previous four weeks, three to four additional people applied but could not
get through the system while two more chose not to apply because they perceived doing so as too difficult
(Zipperer and Gould, 2020[92]). According to US media reports (Rugaber, 2020[93]), newly covered
self-employed and gig workers experienced long delays in receiving unemployment benefits, as most
States first had to establish a system to process these new claims.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
69
Table 1.2. Nearly two in three OECD countries expanded unemployment benefits
Improved Extended Raised Details
access benefit benefit
and durations generosity
coverage
Australia Coronavirus Supplement of AUD 550 per fortnight paid for the next six months.
Relaxation of the partner income test for JobSeeker Payment and waiving of asset
tests for new claims.
Austria Unemployment assistance (2nd tier) raised to the same benefit level as
unemployment insurance benefit (1st tier).
Belgium Exceptional access to unemployment benefits for workers who voluntarily quit their
job to take up a new job but whose job offer fell through. Freezing of the automatic
decline in replacement rates over benefit spells for 3 months.
Canada Introduction of new Canada Emergency Response Benefit of CAD 500 per week for
up to 24 weeks for workers who have lost their income during the COVID-19 crisis.
Colombia Extraordinary payment of 2 minimum wages (stretched over 3 months) for job losers
who have contributed to the UB system for at least 1 out of the last 5 years.
Payment of COL 160 000 per month for 3 months for job losers who have not
received any UB during the last 3 years.
Denmark Freezing of eligibility period for three months.
Finland Temporary extension of unemployment benefits to entrepreneurs and freelancers.
Shortening of the minimum contribution period from 26 to 13 weeks for employees.
Temporary layoffs are not counted towards the maximum period of eligibility.
Increase in earnings disregards for benefit recipients.
France Exceptional access to unemployment benefits for who voluntarily quit their job to
take up a new job but whose job offer fell through. The confinement period will not
be considered for calculating the unemployment benefit entitlements of new
claimants. Freezing of eligibility period by the duration of the confinement.
Germany Benefit extension by 3 months for all recipients whose entitlements end between Mai
and December.
Greece Benefit extension by 2 months for all recipients whose entitlements end in the first
quarter of 2020.
Ireland New COVID-19 Pandemic Unemployment Payment for people who lost their jobs
during the crisis at EUR 350 per week for a period of up to 12 weeks.
Israel Reduction in the required employment history from 12 to 6 months. Extension of
unemployment benefits to employees on involuntary unpaid leave.
Italy Benefit extension by 2 months for all recipients whose entitlement ended in March or
April. Unemployment benefit recipients can work in the agricultural sector without
losing their benefits (up to EUR 2 000).
Latvia Introduction of a new temporary unemployment assistance benefit for unemployment
insurance benefit recipients whose entitlements expire. Benefits are payable for up
to four months, at the level equal to the unemployment insurance benefits paid
during the 8th and 9th month (max. EUR 180 per month).
Luxembourg Benefit extension for recipients whose entitlements expire during the state of
emergency.
New Zealand New COVID-19 Income Relief Payment of up to NZD 490 per week for up to
12 weeks for workers who lose their job or self-employment due to COVID-19
between March and October. Permanent increase of all core benefits, including
Jobseeker Support, by NZD 25 per week.
Norway Reduction in the minimum-income threshold for eligibility to unemployment benefits.
Benefit extension until the end of June for recipients with at most 18 weeks of benefit
entitlements remaining on 29 February. Increase in replacement rates to 80% or
62.5% depending on previous income.
Portugal Benefit extension until the end of the containment measures.
Slovak Republic Benefit extension by at least 2 months for recipients whose entitlements expire
during the crisis.
Slovenia New temporary unemployment benefits of EUR 514 per month for workers who lost
their jobs because of COVID-19 or whose fixed-term contracts were not extended.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
70
Improved Extended Raised Details
access benefit benefit
and durations generosity
coverage
Spain Suspension of the minimum contribution period, including for temporary workers.
Access to unemployment benefits for workers who lost their job during the trial
period and workers who voluntarily quit their jobs to take on new employment and
whose job offer fell through. Benefit extension until the end of the health crisis.
Sweden Shortening of the required membership period in the unemployment insurance fund
from 12 to 3 months. Abolishment of the 6-day waiting period; increase in the
minimum benefit amount from SEK 365 to 510 per day and in the maximum benefit
amount from SEK 910 to 1 200 per day for the first 100 days.
Switzerland Doubling of the reference period to assess the employment condition to 48 months.
Benefit extension by 120 days.
United Kingdom Increase in the Universal Credit (2nd tier) by GBP 20 per week.
United States* Extension of unemployment benefits to self-employed workers. Extension of
maximum benefit duration to 9 months. Benefit increase by USD 600 per week for
up to 4 months.
Note: * Information for the United States refers to the federal level.
Source: OECD COVID-19 employment and social policy responses by country, http://oe.cd/covid19tablesocial; OECD COVID-19 country policy
tracker, https://www.oecd.org/coronavirus/en/#country-tracker.
The crisis accentuated the problem of social-protection gaps for workers in non-standard
employment
The crisis also created an immediate urgency to shore up support for workers and households not covered
by earnings-replacement programmes such as unemployment benefits or job retention schemes (OECD,
2020[37]; 2020[60]). It laid bare – or accentuated – existing social-protection gaps for workers in non-standard
and informal employment, who are among those most affected by earnings losses so far (see Section 1.2).
Without adequate support, many of them face severe and possibly long-lasting income shortfalls and – in
the absence of savings – a risk of economic hardship. Limited or irregular working hours may exclude
these workers from qualifying for job retention schemes or unemployment benefits. Low-income workers,
including many part-time employees who face earnings losses may also lose entitlements to earnings top-
ups through in-work benefits, such as Finland, France, the United Kingdom and the United States.
To respond to these challenges, OECD countries have taken measures to improve access to
non-contributory income support for vulnerable workers and low-income households, and/or to raise
support levels (Table 1.3).
A number of OECD countries reinforced minimum-income benefit schemes
Fourteen countries have facilitated access to existing minimum-income schemes, such as social
assistance, as a way of quickly channelling additional support to low-income households.36 Some of them
(e.g. Australia, Germany, Italy and the Netherlands) suspended or relaxed income and/or asset tests, both
to deliver support more quickly and to widen the circle of potential recipients. Germany, for example,
temporarily suspended all asset tests for Unemployment Benefit II, eased the income test, and permitted
the reimbursement of all housing costs (as opposed to “reasonable” housing costs before the crisis). This
will especially benefit the self-employed. In the Netherlands, recipients of social assistance for
self-employed workers no longer have to repay this allowance, unlike before the crisis. As in the case of
unemployment benefits, a number of countries have also suspended job search and other activation
requirements to account for distancing requirements and to avoid delays in payments.
Spain approved a new means-tested minimum living income (ingreso minimo vital) aimed at alleviating
risks of poverty and social exclusion. Since 15 June, this minimum-income benefit applies nationally across
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
71
Spain and complements existing regional programmes. It is expected to reach 850 000 households, with
a maximum monthly payment between EUR 462 and 1 015, depending on family type.
Data on the development benefit receipt numbers for these minimum-income support programmes are still
very scarce, also because such programmes are often not centrally administered. In the United Kingdom,
daily new claims for Universal Credit increased tenfold in the first weeks of the crisis, but quickly declined
again since peaking in late March 2020. At the time of writing, the latest available data indicate that the
number of new claims in mid-May remains about twice as high as it was in early March (Office for National
Statistics, 2020[94]). In Italy, the total number of households claiming the minimum-income Reddito di
Cittadinanza (“Citizenship Income”) has risen by 12% between January and April 2020. 37
Most countries also provided targeted cash transfers to self-employed workers and other
vulnerable groups
Most OECD countries introduced new, often time-limited, cash support programmes for people in sudden
and urgent need. Such schemes can be suitable in emergency situations to help groups who do not have
access to existing minimum-income benefits, or where claiming such benefits is time-consuming and
unlikely to provide immediate relief.
Several countries introduced new cash transfers for self-employed workers. Often, these transfers are
dependent on previous earnings or on income losses incurred during the crisis. In Austria, for example,
self-employed workers will receive a benefit replacing 80% of their net income loss compared to the same
month in the previous years, up to a limit of EUR 2000 a month.38 In the United Kingdom, the self-employed
receive a taxable grant of up to 80% of their previous earnings over the last three years. Entitlements are
capped at GBP 2 500 a month and can be claimed by self-employed workers with average annual profits
below GBP 50 000. Similar schemes exist in Denmark, Latvia and Switzerland. As determining previous
earnings of self-employed workers is complex without a structure in place to do so, several other countries
have introduced flat-rate payments (such as Belgium, Canada, Ireland, Italy, Korea, Lithuania, the
Netherlands, Poland, Portugal and Slovenia) or lump-sum transfers (Colombia, the Czech Republic,
France, Greece and Israel). To speed up payments Italy introduced a tax-free, flat-rate payment of
EUR 600 payable to self-employed workers. Germany rolled out a Corona supplement for self-employed
workers, providing cash support of up to EUR 15 000 for small firms with up to ten employees.39
New programmes are sometimes specifically targeting informal workers and undocumented migrants, who
are among the most difficult to reach in the current situation (Alfers, Moussié and Harvey, 2020[95]).
Colombia, for example, is making three transfers of COL 160 000 each for 3 million households who do
not benefit from existing programmes. The payment is delivered through bank transfers, for those who
have accounts, or by electronic transfers via mobile phone. The State of California in the United States –
where undocumented migrants account for 10% of the workforce – has announced that it will support these
workers with transfers of USD 500 to 1 000.40
… while a few have offered universal transfers
Three OECD countries have announced cash payments to (nearly) the entire population to help people
make ends meet. The appeal of such payments is their simplicity: since universal transfers do not depend
upon income, assets, or prior contributions, they avoid costly and time-consuming means tests and can be
rolled out quickly. The United States pays a transfer of USD 1 200 to all citizens earning up to USD 75 000
a year (USD 150 000 for couples). Families receive an additional USD 500 per child under 17, households
above the income threshold may receive a reduced payment. 41 Japan has begun sending a flat-rate
payment of JPY 100 000 to all its residents. Korea will make an emergency relief payment to all of its about
22 million households. The payment level depends on household size and amounts to KRW 400 000 for a
single person and an additional KRW 200 000 for each further household member (up to a four-person
household).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
72
Table 1.3. Countries across the OECD have taken measures to improve support for workers and
households not covered by unemployment benefits or job retention schemes
Extensions to New targeted cash New universal transfers Additional direct help with
means-tested programmes transfers to specific
groups household expenses
Australia
Austria
Belgium
Canada
Chile
Columbia
Czech Republic
Denmark
Estonia
Finland
France
Germany
Greece
Hungary
Iceland
Ireland
Israel
Italy
Japan
Korea
Latvia
Lithuania
Luxembourg
Mexico
Netherlands
New Zealand
Norway
Poland
Portugal
Slovak Republic
Slovenia
Spain
Sweden
Switzerland
Turkey
United Kingdom
United States*
Note: * Information for the United States refers to the federal level.
Source: OECD (2020[37]), “Supporting livelihoods during the COVID-19 crisis: closing the gaps in safety nets”,
http://www.oecd.org/coronavirus/policy-responses/supporting-livelihoods-during-the-covid-19-crisis-closing-the-gaps-in-safety-nets-17cbb92d/.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
73
While such temporary universal transfers are appealing in the current context to ensure that no-one falls
through the cracks of the social protection system, they are – by design – poorly targeted (see Box 1.7).
Many households receiving such support will not be in the greatest need. Meanwhile, such unconditional
payments should reach a meaningful level to ensure that vulnerable households who have lost most or all
of their income in the current crisis can make ends meet. Depending on other, more targeted, benefits that
may be available in addition, this may create very large budgetary costs at a time of huge pressures on
government spending.
Most countries also provided direct help with household expenses
Most OECD countries have also stepped in to help vulnerable households make ends meet by permitting
them to postpone paying bills or by providing in-kind support. A number of them have allowed for delays
in big-ticket regular expenditures such as tax and rent, e.g. by extending the deadlines for tax filing (such
as in Canada, Finland, Japan, the United Kingdom and the United States) or social-security contributions
(Japan, Spain). Several have introduced temporary deferments of mortgage payments, temporarily
suspended foreclosures or evictions (see Box 1.8). Colombia has decided to refund Value Added Tax for
the most vulnerable households. Other countries have provided direct support with pandemic-related
expenditures, notably health care. In the United States, for example, where health insurance is often
employer-provided, many workers who lost their job suddenly also found themselves without any health
insurance during the pandemic. The Federal Government therefore announced that it will meet the hospital
and testing charges incurred by uninsured COVID-19 patients. Various OECD countries have also
extended in-kind support, partly to offset the closure of food banks and suspension of schools meals during
lockdown. The United Kingdom, for example, launched a national voucher scheme to ensure that the
1.3 million eligible school-aged children will continue to have access to meals during school closures.
Spain designated EUR 25 million to provide income support through transfers and vouchers to children
who are affected by school closures. France has made available EUR 25 million of funding to support food
aid associations, plus a further EUR 14 million to be distributed in emergency food checks.
As countries have grappled to minimise the impact of the containment measures on the livelihoods of their
citizens, the usual trade-offs between support and incentives, between generosity and fiscal sustainability,
have often, temporarily, been laid aside. Indeed, concerns about undermining incentives to work appear
secondary as workers have been asked to stay at home, and worries of fiscal sustainability have been put
on pause as policy makers had to move fast in attempts to protect livelihoods and avert a deeper economic
and social crisis. These trade-offs will change as economic activity picks up over the next months, such
that making corrections to the recent measures will become inevitable (Section 1.4).
Employment services and training for jobseekers and workers
The unprecedented rise in jobseeker numbers in some countries, and companies’ massive use of job
retention schemes in others, pose an enormous challenge to benefit administrations and employment
services (OECD, 2020[96]). The vast volume of incoming support claims during the first weeks and months
of the crisis as well as the management of job retention schemes pushed public and private employment
services (PES) to the limits of their capacity. Some countries had to build rapidly the necessary
infrastructure and procedures to administer new claims. Meanwhile, liquidity-constrained businesses
depended on a fast processing of their claims to be able to cover operating costs, while many workers
anxiously awaited their benefit payments to be able to pay for their rent and other living expenses. Many
OECD countries therefore took rapid steps to streamline and re-prioritise PES operations, while
simultaneously adjusting them to physical-distancing requirements.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
74
Box 1.7. Universal support during a crisis: ad-hoc lump sum transfers vs. a genuine universal
basic income?
In some OECD countries, the policy responses to the COVID-19 crisis – in particular, the use of
universal cash transfers – have revived discussions on the desirability of a universal basic income (UBI),
an ongoing flat-rate transfer to the entire population, irrespective of employment status, income or
means. Calls for a UBI in the face of COVID-19 largely relate to two objectives: (i) the need for
instantaneous relief for those whose livelihood has suffered during the crisis, and (ii) the need to ensure
that no one is left without support.
Two of the key concerns often presented in opposition to providing a UBI as a principal pillar of social
protection – that it would undermine work incentives, and that a meaningful cash transfer to the entire
population would come at an unrealistically high budgetary cost (and/ or imply eliminating most if not
all other well-targeted cash transfers) – may have seemed less urgent during the initial phase of the
current crisis. Preserving work incentives, and conditioning benefit receipt on active job search and
participation in employment support, have been second-order priorities during the immediate lockdown
period, though they will become more important as hiring picks up in some sectors. The fiscal cost of
social-spending programmes is currently also not at the forefront of the policy discussion as countries
vow to do whatever it takes to protect people and the economy from an unprecedented crisis (Furman,
2020[97]).
However, as countries move beyond the immediate lockdown period, and perhaps into a more
protracted economic crisis, budgetary cost will invariably become a major concern. The question then
becomes whether UBI is a cost effective way to provide timely and adequate support. OECD analysis
(OECD, 2017[98]; Browne and Immervoll, 2017[99]) shows that financing a budgetary neutral UBI
(replacing most working-age benefits by a flat-rate payment to all such that cost remains constant)
would require very large tax increases and eliminate most if not all targeted transfers that effectively
reduce the risk of falling into poverty. The distributional effects of such a hypothetical UBI would be
complex: low-income groups who would normally receive other targeted benefits, and higher-income
groups paying most of the tax, would typically lose. Those who currently do not receive any benefits
would gain – this includes higher-income groups and those who fall through the gaps of existing social-
protection system. The bottom line, however, is that a fiscally realistic UBI would be too low to provide
reliable poverty alleviation on its own.
In countries with well-developed social protection in place, replacing existing support measures with a
“no questions asked” basic income for everybody would be a highly risky strategy that would provide
limited income security and be very expensive. Yet, less comprehensive types of universal transfers,
restricted to certain population groups or with some form of mild conditionality, can be valuable
complements to more targeted support measures. These include universal child benefits or basic
old-age pensions that exist in a number of OECD countries, as well as time-limited emergency
measures for groups that are known to be poorly served by the main income protection programmes.
Source: OECD (2020[37]), “Supporting livelihoods during the COVID-19 crisis: closing the gaps in safety nets”,
http://www.oecd.org/coronavirus/policy-responses/supporting-livelihoods-during-the-covid-19-crisis-closing-the-gaps-in-safety-nets-
17cbb92d/; OECD (2017[98]), Basic income as a policy option: Can it add up?, Policy Brief on the Future of Work,
https://www.oecd.org/social/Basic-Income-Policy-Option-2017.pdf.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
75
Box 1.8. Housing and COVID-19: Helping workers stay in their homes
The COVID-19 pandemic has brought to the fore a number of housing challenges and vulnerabilities
facing workers in OECD countries.
First, the heightened economic vulnerability generated by the crisis threatens workers’ housing stability.
Without assistance, workers who have been laid off, are forced to work reduced hours, or are
temporarily unable to work may struggle to cover their monthly rent, mortgage or utilities payments.
This is a particular risk for households that are already overburdened by housing costs. Across the
OECD, renters and low-income households are, on average, more likely to spend over 40% of their
disposable income on housing. More than one in ten renters are overburdened by housing costs in the
OECD, compared to less than 5% of mortgage holders; meanwhile, over half of renters in the bottom
income quintile are overburdened by housing costs in Chile, Israel, New Zealand and the
United Kingdom (OECD, 2019[100]). Moreover, even before the pandemic, many households struggled
to pay monthly housing costs: according to Eurostat data, around one in five low-income households
(below 60% of the median equivalised income) in the European Union fell behind on their mortgage,
rent or utility bills in 2018.
Second, living environments have also facilitated – or hindered – the continuity of employment during
the pandemic. The widespread shift to teleworking is not feasible for households who do not have a
computer or access to the internet at home, or difficult due to space constraints or because devices
need to be shared among household members. On average across the OECD, around 87% of
households have access to the internet at home, though the share is less than half in Colombia and
Mexico; meanwhile, nearly 81% of households in the OECD have access to a computer at home, with
less than 50% of households in Colombia, Mexico and Turkey. 1
Third, the pandemic, along with the shelter-in-place orders implemented to manage the crisis, has
elevated health and safety risks among workers living in poor quality housing or unsafe living conditions.
Overcrowding, which can increase the risk of infectious diseases (World Health Organization, 2018[101]),
is a reality for more than a quarter of all households in Latvia, Mexico, Poland and the Slovak Republic
(OECD, 2019[100]). Overcrowding makes it harder to effectively self-isolate, putting workers living in
overcrowded conditions at greater risk of contracting and spreading the disease.2 Preliminary evidence
from England and Wales finds a correlation between the number of COVID-19-related deaths and levels
of housing overcrowding in local areas (Barker, 2020[102]).
In response to COVID-19, many OECD countries have introduced emergency housing measures to
keep workers in their homes (Table 1.4).3 Emergency measures have largely addressed concerns
around housing instability, rather than housing quality gaps, which are hard to overcome in the short-
term. Eviction bans are the most common measure to support tenants (in 16 countries), followed by the
deferment of rent payments (5 countries), reforms to financial support schemes for renters (5 countries),
rent freezes (3 countries), and temporary reductions or suspensions of rent payments (3 countries). For
homeowners, 20 countries have introduced more generalised mortgage forbearance in response to
COVID-19, and two countries have banned foreclosures due to missed payments for at least some
households. In eight countries, at least some households may defer utility payments, and/or are ensured
continued service if payments are missed.
While it is too soon to assess the full impacts of the COVID-19 pandemic on housing outcomes,
emerging research suggests potential disparities between homeowners and renters – in part due to the
employment characteristics of renters in some countries. Researchers in the United Kingdom and the
United States suggest that renters face heightened economic vulnerability relative to homeowners, in
part because renters are more likely to work in industries most affected by the pandemic (Judge and
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
76
Pacitti, 2020[103]; Kneebone and Murray, 2020[104]). Further, some countries, such as France, are
anticipating that new support measures could be needed in the event of a second COVID-19 wave,
particularly during the colder winter months, where issues around housing quality, the affordability of
utilities, and homelessness would become even more pertinent.
Table 1.4. Many countries introduced emergency housing measures in response to COVID-19
Types of emergency, temporary housing measures introduced in OECD countries in response to COVID-19
Type of measure or support Country
For tenants:
Eviction ban due to missed payments Australia*, Austria*, Belgium*, Canada*, France, Germany, Hungary, Ireland,
Israel*, Luxembourg, the Netherlands, New Zealand, Portugal, Spain,
United Kingdom, United States*
Deferment of rent payments Austria, Germany, Mexico, Portugal*, Spain*
Temporary reduction or suspension of rent payments for Greece, Portugal*, Spain*
some households
Rent freeze Ireland, New Zealand, Spain*
Reforms to financial support schemes for renters Japan*, Ireland, Luxembourg, Portugal*, Spain
For homeowners:
Mortgage forbearance Australia*, Austria, Belgium, Canada*, Colombia, Czech Republic, Germany,
Greece, Hungary, Ireland, Israel, Italy, Lithuania, Mexico*, Poland, Portugal,
Slovak Republic, Spain, United Kingdom, United States*
Foreclosure ban due to missed payments United States*, the Netherlands
For all households:
Deferment of utility payments and/or assured continuity Austria, Belgium*, Colombia, Germany, Japan, Korea, Spain, United States*
of service even if payment missed
Reforms to housing subsidy schemes France (planned reform postponed), Spain
For the homeless:
Emergency support to provide shelter and/or services to Australia, Austria, Canada, France, Ireland*, New Zealand, Spain,
the homeless United Kingdom, United States*
Note: List of measures as of 15 June 2020. * indicates that the measure applies only to some jurisdictions and/or to qualifying households.
Source: OECD COVID-19 employment and social policy responses by country, http://oe.cd/covid19tablesocial.
The COVID-19 pandemic has provided a window into the disparities in workers’ access to quality,
affordable housing in the OECD. It will be essential to monitor the housing impacts of COVID-19 on
different types of households (homeowners vs. renters) as well as across workers in different sectors
in order to assess the extent to which emergency measures were appropriately targeted. Canada, for
example, plans to integrate household survey responses on housing quality and affordability during
COVID-19 with neighbourhood-level information on population density, dwelling types, and household
income to assess the relationships between housing and COVID-19 (Statistics Canada, 2020[105]).
1. In households with school-aged children, the digital divide risks deepening educational disparities during a period of extended school
closures where many institutions have transitioned to distance learning (OECD, 2020[106]).
2. Overcrowding as defined by Eurostat, measures the number of rooms per household member, taking into account different factors of
household composition. For a full explanation, see the OECD Affordable Housing Database: http://www.oecd.org/els/family/HC2-1-Living-
space.pdf. Shelter-in-place orders have led to an increase in reports of intimate-partner violence or inquiries about emergency shelters for
abuse victims in many countries (OECD, 2020[57]).
3. Table 1.4 discuses demand-side measures; for a discussion of supply-side measures to support to banks, construction companies, or
housing providers, see (OECD, forthcoming[107]).
To secure a timely pay-out of income support benefits and a rapid processing of companies’ job retention
scheme claims, several countries simplified claim procedures or prioritised claim processing. Switzerland,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
77
for example, doubled the renewal period for its STW scheme from three to six months, hence reducing the
number of applications and speeding up the approval process. Belgium and the United Kingdom facilitated
online applications. In Germany, where one-in-three companies had applied for STW by the end of April
2020, the PES increased the number of staff processing STW claims 14-fold relative to normal times. PES
in several countries also relaxed application procedures for out-of-work support or freed up resources by
temporarily scaling down and suspending other, less essential services. Some automatically renewed
benefits during the confinement period (e.g. incapacity benefits in Estonia and New Zealand, jobseeker
benefits in Greece and Spain, and housing and child allowances in the Czech Republic); others lifted
deadlines for registering as unemployed (e.g. Slovenia). Most PES temporarily suspended in-person
training, job fairs and caseworkers’ networking activities.
Soaring caseload numbers, physical-distancing requirements and the inability to look for a job during the
pandemic also forced PES to adapt their ways of supporting jobseekers and their capacity to monitor job
search behaviour. Most OECD countries have explicit job search reporting procedures (Immervoll and
Knotz, 2018[108]), aiming to direct jobseekers to look for work more intensively and earlier on. While PES
in a number of countries maintained job search and reporting requirements during the crisis, some eased
and adjusted these requirements for jobseekers with children at home because of childcare facility or
school closures, or for those in quarantine (e.g. Austria, Brussels (Belgium), the Netherlands, and
United Kingdom). Many PES temporarily suspended job search requirements and lifted sanctions
(e.g. France, Germany, Portugal, Slovenia and Sweden). Others did not apply sanctions, but encouraged
jobseekers to continue actively searching for jobs (e.g. Australia, Denmark, Estonia and Latvia).
The current crisis also represents an opportunity for upskilling and reskilling, both for jobseekers and for
workers who are idle because they their workplaces are shut down and who cannot work from home. While
most OECD PES had to suspend face-to-face training provision to respect physical distancing, many offer
training via digital channels. Pre-existing online training solutions enabled many countries to maintain
training provision with minimal investment, at least for the type of skills that can be easily taught online
(e.g. in Austria, Belgium, Denmark, Estonia, the Netherlands and some regions of Italy). Some countries
also quickly boosted online training options. Denmark, for example, amended legislation such as to allow
municipalities to offer new digital qualification courses. France made available over 150 new online training
courses on the Emploi Store. Sweden will use part of the extra funding allocated to the PES and other key
players to strengthen distance learning and internet-based education.
Also other governmental or non-governmental actors in several countries quickly developed training courses
to address immediate demand pressures (OECD, forthcoming[109]). This includes resources to support
health professionals’ upskill for the pandemic response. Health Education England, for example, offered
free e-learning programmes for the UK health workforce on infection prevention and control and the use of
personal-protection or ventilator equipment. Other programmes aimed to reskill displaced workers to help
temporarily fill roles in essential services, often in the health or social care sectors, but also in manufacturing,
logistics and distribution, or retail. In Massachusetts (United States), Partners in Health, a non-profit health
care organisation, is training one thousand workers as contact tracers, an occupation now in shortage. The
Swedish Sophiahemmet University developed a course for the medical training of laid-off staff in the airline
industry, and another one for elderly care training of hospitality workers. Several countries created,
strengthened or further advertised their online tools (matching platforms or skill assessment tools) to
connect displaced workers from recent business closures and businesses in sectors currently in demand.
1.4. The way ahead – What is the right policy mix for post-confinement?
As the first wave of the pandemic began to subside across many OECD countries in May 2020, restrictions
of people’s mobility were eased, economic activities in many sectors re-started and countries began to
move to a “new normal”. In the absence of a vaccine and effective treatments, countries are now trying to
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
78
strike the difficult balance between re-opening for business and social life whilst avoiding a new spike in
infections. Some mitigation measures will remain in place, and for people and businesses alike, the
challenge will be to ensure the application of high hygienic standards and maintain physical distancing in
order to avoid the need for renewed mandatary restrictions. Solving the health crisis is an essential
precondition to solving the economic and jobs crisis.
As the understanding of the epidemiological characteristics of COVID-19 remains limited, it is still uncertain
how the pandemic will evolve in different parts of the world. The seasonality of the virus is yet to be
confirmed, but cannot be excluded. Herd immunity42 is still far on the horizon, not least since successful
containment measures have brought the reproduction number around or below one in many countries.
Also the timing of a discovery of a vaccine remains highly uncertain. Based on the most optimistic
estimates, it will take at least 12 to 18 months for an effective vaccine for SARS‑CoV‑2 to become widely
available. However, this assumes that one of the candidates currently in clinical trials turns out to be
successful; if none are, the wait will be longer (OECD, 2020[110]).
In the absence of a vaccine, countries can avoid a second wave by identifying and putting in place a
package of comprehensive public health interventions. They range from a massive upscaling of testing,
tracking and tracing (TTT), to enhancing personal hygienic measures and the continuous enforcement of
physical-distancing policies such as banning large gatherings and encouraging people to work from home
(OECD, 2020[3]). To support countries in their planning, the OECD has developed a microsimulation
epidemiological model to assess rigorously the likely effectiveness of different containment measures. The
model shows that upscaling TTT, enhancing hygienic measures and ensuring wide use of masks would
allow a broader reopening of the economy without a new outbreak (see Box 1.9 for a short presentation of
the model’s features and the results for Italy).
Given the exceptional uncertainties characterising the near-term outlook, the OECD considers two
epidemiological scenarios for the coming 18 months – see OECD (2020[60]) and the summary in
Section 1.2 of this chapter – though a wide range of other outcomes remain possible. In the first scenario,
the containment measures taken during the spring 2020 will manage to limit the diffusion of the virus
without a second outbreak and the need to re-introduce more drastic lockdown measures. In the second
one, these containment measures do not manage to contain the spread of the virus leading to a second
infection wave in October/November 2020. In both scenarios, many service sector companies will likely
have to continue operating well below full capacity, notably in food services, accommodation, transport
and culture. This could cause a wave of company insolvencies with a further round of job and income
losses. Even in the more optimistic “single-peak” scenario, the economic recovery will likely be slow and
gradual, and the OECD projects unemployment to remain at the level around that observed at the peak of
the global financial crisis until well into 2021 (see Section 1.2). In the more pessimistic “double-peak”
scenario, countries may have to return to restricting people’s mobility and economic activity. Most
businesses will again have to suspend or scale down operations or – where possible – ask their employees
to work from their homes. Such a second round of restrictions may even hit businesses and households
harder than the first shock as many of them will have run down their savings to absorb the income losses
suffered during the first wave. Unemployment will rise further.
Irrespective of which scenario turns out being closer to reality, OECD governments will need to adapt their
labour market and social policies in the coming months to respond to the evolving pandemic and the
economic developments. During the initial weeks and months of the crisis, countries have rightly focused
primarily on providing rapid emergency relief to keep households and companies afloat and prevent the
economy from collapsing. In the upcoming months, they will likely need to modify, and adjust the
composition and characteristics of their support packages. As countries gradually open up their economies,
policies will have to better account for the large existing heterogeneity in workers and companies. Given
the cost of the policies put in place, countries will also face difficult decisions about how to target
expenditures without risking to prematurely end support for companies or households who still need it.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
79
Box 1.9. Avoiding a second pandemic wave while re-opening activities: key insights from the
OECD epidemiological model on COVID-19
To support countries in their planning, the OECD has developed a microsimulation model (Figure 1.20)
projecting the key parameters of the COVID-19 pandemic for a set of countries. The key features of the
model include:
The model is developed within the OECD SPHeP (Strategic Public Health Planning) framework
(OECD, 2019[111]) to ensure consistency with the other OECD epidemiological models
The model is an evolution of the SIR (susceptible-infected-recovered) approach, which is the basic
standard to model communicable diseases such as viral infections. More specifically, the model
also includes an ‘exposed’ compartment to also account for the incubation period as well as
compartments for hospitalisations and access to intensive care units;
The model uses country-specific epidemiological data (e.g. number of deaths) and use of health
services as well as evidence on the effectiveness of different containment measures, to generate
estimates on the number of infected people and to project the number of hospitalised people and
deaths;
All estimates can be produced under different scenarios, hence the model can be used to
understand the most effective set of policies to delay, or possibly, to avoid, future lockdowns.
Figure 1.20. Schematic overview of the OECD SPHeP-COVID model
Susceptible Exposed Infected Community
Transition state
Absorbing state
Hospital
Flow of individuals
Recovered
Age-specific transition rates
Death ICU
In a first phase, the model is fed with plausible ranges of epidemiological inputs, retrieved from the
literature, on measures such as incubation period and length of the disease. The model’s outputs are
calibrated to closely match national historical statistics on hospitalisations and deaths due to COVID-19.
Cross-country differences in the resulting parameters reflect how different health systems managed the
epidemic – for example, depending on whether only seriously affected patients, or the majority of cases,
were hospitalised. The model is also cross-validated by comparing projected outputs with other major
modelling initiatives. The resulting set of parameters is then used to carry out the scenario analysis, under
the following main assumptions:
The country-specific health care system response to COVID-19 and its effectiveness is maintained
constant for the rest of the simulation, for example implicitly assuming that there will be no
therapeutic breakthrough;
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
80
Individuals recovered from the infection acquire immunity to the virus and cannot be infected again
for the rest of the simulation (in reality, there is some uncertainty over the extent to which this is
true);
Because of inconsistent evidence on how the temperature and humidity may affect the spread of
the virus, the model assumes no seasonal effect on the epidemiology of SARS-CoV-2; and
The ‘no-lockdown’ and ‘all containment policies lifted’ scenarios represent the worst case
scenarios, with individuals’ behaviours and mobility that are maintained the same as the pre-
COVID-19 period, leading to the uncontrolled spread of the virus until herd immunity is achieved.
Findings from the model support countries’ policy decisions in the early phases of the pandemic and can
inform the next steps of the policy-making process.
More specifically, the model provides new insights on two key policy issues (see Figure 1.21):
First, by implementing restrictions to social and economic life, countries have prevented the collapse of
their health care systems and have avoided hundreds of thousands of deaths. The model simulations
suggest that in the absence of any confinement measures the death toll of COVID-19 could have been in
the order of 500 000 in Italy, and millions across OECD countries. In addition, under the same scenario,
the number of patients requiring advanced care would have been tens of times more numerous, probably
causing a full collapse of health care services and, therefore, even a much higher number of deaths than
estimated by the model.
Second, countries can prevent a second pandemic wave, and a consequent lockdown, if they implement
comprehensive packages of public health interventions to contain the spread of the infection until a vaccine
or effective treatment become available. More specifically, they can achieve this objective by implementing
a strategy based on the following three pillars:
1. Massive upscaling of testing, tracking and tracing (TTT) to quickly identify and quarantine new
cases as well as their contacts that are at a high risk of developing the infection. The risk of new
outbreaks is high, but effective TTT keeps these outbreaks at a small and local scale and prevents
the further propagation of the infection. In addition, TTT contributes to maintaining a robust
surveillance system and helps monitor key dimensions and thresholds for post-confinement;
2. Enhancement of hygienic measures such as frequent handwashing and deep cleaning to decrease
the probability of being infected by contaminated objects, and use of masks (to prevent people who
may be pre- or asymptomatic from unknowingly spreading the disease);
3. Continuous enforcement of some physical distancing policies such as banning large gatherings,
encouraging people to work from home and closing, or regulating access to, some gathering
places.
Greater effectiveness in upscaling TTT, enhancing hygienic measures and ensuring wide use of masks,
will permit a further loosening of more physical distancing measures and a broader reopening of the
economy. For example, other modelling-based studies suggest that in a scenario where half of the
population wears masks and a TTT programme successfully tracks about 40% of infected people within
four days, countries could reduce physical distancing policies by almost two thirds compared to a full
lockdown.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
81
Figure 1.21. Typical impact of COVID-19 across OECD countries under different scenarios: the
example of Italy
Number of hospitalisations and total deaths over time
Historical data Business as usual No containment
Complete loosening Some loosening
A. Hospitalisations B. Deaths
Thousands Thousands
600 600
500 500
400 400
300 300
200 200
100 100
0 0
Note: Italy implemented the lockdown approximately three weeks after the first registered death caused by COVID-19 and relaxed lockdown
measures approximately eight weeks after their implementation. Similar timelines have been observed in several OECD countries. The
“Historical” scenario shows observed numbers until 7 June 2020. The “No containment” scenario provides a counterfactual simulation under the
assumption that countries did not implement any containment policy. The “Complete loosening” scenario gives a model-based projection of
numbers under the assumption that all containment policies were lifted on the 1 September 2020. Finally, the “Business as usual” scenario and
the “Some loosening” scenario provide a model-based projection of future numbers under two scenarios respectively entailing the continuation
of containment policies as they were in place on the 7 June, or a limited loosening of these policies on 1 September 2020 along the lines of what
described in the text.
Source: OECD analysis on the OECD SPHeP-COVID model, OECD (2020[1]), “Flattening the COVID-19 peak: Containment and mitigation
policies“, http://www.oecd.org/coronavirus/policy-responses/flattening-the-covid-19-peak-containment-and-mitigation-policies-e96a4226/ and
Tian et al. (2020[112]), “Calibrated Intervention and Containment of the COVID-19 Pandemic”, http://arxiv.org/abs/2003.07353.
StatLink 2 https://stat.link/n4ra98
This section describes some of these policy challenges and discusses potential solutions. These solutions
would have to be tailored at country and, sometimes, local and/or sectoral level to account for the specific
situation as well as the national institutional settings and traditions, in particular with respect to the
involvement of social partners in the definition of labour market policies.
1.4.1. Ensuring workers’ safety
Ensuring workers’ safety is the prime objective in the near term to limit the spread of the virus, avoid a
surge in sickness absences and ensure that workers feel secure enough to work effectively.
For workers who do not need to be physically present at the workplace, working from home remains the
easiest way to ensure the continuation of work without incurring the risk of contracting an infection while
commuting and working (Section 1.3.1). Several studies43 have tried to quantify the proportion of jobs that
could be potentially performed from home, and thus be shielded from contagion. However, beyond those
that can be done from home, a number of other jobs come with only a limited risk of infection. This may be
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
82
because they imply no, little, or infrequently sustained physical contact with customers or colleagues
(e.g. for mechanics, plumbers, archivists, or truck drivers).
Estimates by Basso et al. (forthcoming[113]) suggest that, on average across 24 OECD countries, 52% of
the workforce is employed in jobs that, without taking into consideration work re-organisation during the
current crisis, are relatively safe. About 31% of workers can potentially work from home, while the
remaining 21% have at most some physical contact with others to perform their job.
However, these estimates also imply that nearly half of the workforce is employed in jobs that do entail
some risks of infection in the current situation, as they require a higher degree of physical proximity with
colleagues or more frequent physical interactions with the public. The share of workers employed in jobs
“at risk” varies from 39% in Luxembourg to 56% in Spain, reflecting cross-country differences in
occupational composition. Women (except in Greece) and younger workers are relatively more likely to
work in jobs “at risk” across all OECD countries (Figure 1.22). The same is true for low-income workers,
who more frequently take up jobs that, under normal conditions, expose them to physical contact and a
higher risk of infection. The estimated share of jobs “at risk” does not vary much with population density at
the workers’ place of residence: while urban areas have a higher share of jobs that can be done from
home, non-urban areas have a higher share of jobs that cannot be done from home but entail a low level
of physical proximity – such as in agriculture (Basso et al., forthcoming[113]).
Figure 1.22. Around half of workers are employed in jobs that entail some risk of infection
Share of total jobs by country
At risk Can be done from home Limited physical interactions required Some physical interactions required
%
100
90
80
70
60
50
40
30
20
10
0
Note: The estimations are based on 2018 data. Jobs that “require limited interactions” cannot be done from home but entail limited proximity
and interactions with colleagues, customers or the public. Jobs that “require some physical interactions” cannot be done from home, entail
limited physical proximity but also some interactions with customers or the public. “Average” is an unweighted average across all countries. See
Basso et al. (forthcoming[113]) for more details on the methodology.
Source: Basso et al. (forthcoming[113]) based on United States BLS Employment Projections and EU LFS data.
StatLink 2 https://stat.link/po8wjd
Therefore, beyond continuing encouraging telework which does not come without cost,44 occupational
safety and health practices that limit the spread of contagion are a top priority in the post-confinement
phase. This requires not only defining the appropriate practices – see, for example, the guide by the
European Agency for Safety and Health at Work (EU-OSHA, 2020[114]) – but also supporting firms, in
particular SMEs, in implementing them (for example, via tax credits). Legal and regulatory enforcement is
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
83
needed in ensuring the adaptation of practices that limit contagion. In the United States, for example, the
Occupational Safety and Health Administration (OSHA) invites workers who believe that their working
conditions are unsafe or unhealthful to file a confidential complaint and request an inspection. Whistle-
blowers are protected from being fired, demoted, transferred or from suffering other forms of retaliation.45
Beyond what can be defined in government guidance, laws, and regulation, social dialogue and collective
bargaining can be mobilised to complement public action in this area. The protocols and agreements
recently signed between employers and trade unions in various OECD countries (see Section 1.3.1) are
an excellent example of how to find flexible and tailored solutions for both companies and workers.
1.4.2. Maintaining adequate paid sick leave
One way to limit the spread of contagion via workplace exposure, and to make post-confinement safer for
everyone, is to allow sick workers to stay away from the workplace. Paid sick leave will continue to perform
an important role in containing and mitigating the spread of the virus and protecting the incomes, jobs and
health of workers and their families during post-confinement (OECD, 2020[67]). It can prove its value also
as part of an effective TTT strategy (OECD, 2020[3]), by allowing (potentially) infected workers to quickly
self-isolate. The cost to society of providing paid sick leave to these workers to ensure that they are not
financially penalised for isolating themselves is small in comparison to that of them not isolating and
spreading the virus further.
To effectively contribute to an orderly post-confinement, countries should consider keeping in place their
extraordinary paid sick leave entitlements and extending them to groups of workers who are still not
covered, including those with zero-hour contracts. Where applicable, temporary measures to support the
cost of sick pay for employers are also justified to the extent that large parts of the economy are still
confined or otherwise constrained.
Moving forward, structural considerations and adjustments to paid sick leave will likely gain in prominence
on the policy agenda to build more resilient labour markets and societies. The crisis has accentuated
long-known gaps in paid sick-leave regulations in a number of OECD countries. These countries, some of
which have introduced new mandatory regulations for the first time in history, should consider closing these
gaps more permanently and for all groups of workers. Particularly for workers on quarantine, automatic
extensions of sick-leave rules through epidemic laws have proven effective in countries where such laws
exist; other countries may wish to consider introducing such laws or mechanisms.
At the same time, when workers who have been on paid sick leave can safely return to work, governments
will have to reinforce work incentives and employment support for workers and financial incentives for
employers in order to facilitate return to work. In particular, governments should prevent paid sick-leave
systems from becoming a pathway into disability benefits for the long-term unemployed, as has happened
in many OECD countries in the past after a recession (OECD, 2010[115]). This is particularly important now,
as some workers currently on sick leave or quarantine may not be able to return to their job, as companies
may fail to remain in business when job retention schemes phase out. Connecting these workers quickly
with occupational rehabilitation or employment services, as appropriate, will be critical to prevent long-term
labour market exit of those among them unable to find new jobs.
1.4.3. Upholding support for workers with caring needs
In most countries, schools and care facilities are reopening gradually and in accordance with the local
capacity of municipalities and schools to implement public health instructions and ensure the safety of
students and staff. However, the challenge of juggling paid work with additional family care responsibilities
may continue for many parents and other within-household caregivers in some form, and potentially for
several months.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
84
With this in mind, countries may want to avoid a sharp withdrawal of temporary family care support, and
rather consider a gradual scaling back of these measures, fine-tuned to the evolution of the situation. In
countries where children are returning to school part-time, working parents may need part-time leave
support. Countries should also look at reinforcing workers’ rights to flexible working arrangements,
including remote working, but also covering flexible start and finish times, “time-banking”, and the ability to
work condensed weeks. Looking further ahead, there may be a need to (re)introduce emergency family
care leave during a potential second wave of infection.46 With more time to plan, countries should draw up
contingency plans for delivering alternative care services, should further facility closures be needed. One
option is to establish plans for delivering temporary in-home or small-group childcare and supervision
services, as New Zealand has done for essential service workers, and Korea has done for two-earner
families more generally (see Section 1.3.1). Priority could go to essential service workers and those with
no access to flexible working. To help staff in these activities, countries could explore options for
temporarily redirecting skilled staff from schools and centre-based care facilities, as and where needed.
1.4.4. Adapting job retention schemes
Job retention schemes (JRS), i.e. government-financed STW and wage subsidy schemes, seem to have
averted an initial surge in unemployment in a number of countries (Section 1.3.2). However, designed
mainly to provide immediate support, they need to be adapted to ensure sufficiently strong incentives for
firms to move off JRS support or for workers to move on to more viable jobs. This is particularly important
for schemes that provide generous support to firms and workers for relatively extended periods. This would
reduce the pressure on public budgets and also the risk that JRS become an obstacle to the recovery by
curbing job reallocation towards more viable and productive firms. Concerns about potential abuse, which
were already raised in the early phase of the crisis, may also become more prominent as some firms
continue to claim support for shortened hours even after workers have resumed their normal schedules.
However, adapting JRS is challenging given the large variation in the continued role of containment
measures across sectors and the level of uncertainty about the strength of the recovery and the risk of a
second pandemic wave. Indeed, a key question at this point is whether JRS should be differentiated across
sectors. While in some, economic activity may pick up again quickly, others will continue to face legally
imposed restrictions to their activities or have to deal with long-lasting changes in consumer patterns.
Sectors whose activity remains legally curtailed may require continued job retention support in the post-
confinement phase.47 In sectors where business can resume, JRS could be adjusted to avoid the risk that
JRS support jobs that have become permanently unviable. Moreover, JRS should be adapted with caution
and not be withdrawn too quickly to avoid a sudden surge of layoffs. Countries will also need to account
for the risk of a second infection wave in the coming months that may result in new restrictions and require
another scaling up of job retention support.
The main challenge going forward is to target JRS to jobs at risk of being terminated, but likely to remain
viable in the longer term. However, any changes to the schemes must also take account of the evolving
economic and health crisis and its varied consequences across sectors. Governments have a number of
policy levers that they can use:
Require firms to bear part of the costs of short-time work schemes, depending on the continued
impact of containment measures. Requiring firms to participate in the costs of hours not worked
increases incentives to limit requests only for jobs that they believe can re-start after the crisis. To
avoid reinforcing firms’ financial difficulties, their participation can take the form of delayed-payment
or zero-interest loans. This would be similar to experience-rating employer social-security
contributions, i.e. making contributions dependent on firms’ use of STW subsidies in the recent
past, but would be simpler to implement. As part of the phase-out of the temporary JRS, the
United Kingdom is gradually increasing the cost of employers for keeping workers on furlough.
France is currently the only country that applies different rules with respect to the cost of firms for
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
85
STW between sectors that are open for business and sectors that remain subject to government-
imposed health restrictions. In open sectors, firms have to contribute 10% of the cost of hours not
worked from 1 June. Requiring firms to participate in the costs of reduced working hours is less
obvious in the context of wage subsidy schemes.48
Support should be time-bound, but limits on the maximum duration should not be set in stone.
Imposing limits on the maximum duration of JRS helps to reduce the risk of supporting jobs that
are no longer viable even in the longer term. Maximum limits signal that support is temporary and
hence cannot be a solution to permanent problems and reduce the risk of supporting permanently
unviable jobs.49 However, limits on the maximum duration should not be set in stone: the duration
for which job retention support is provided may need to adjust to the health and economic situation.
A number of countries where temporary schemes have been introduced in response to the crisis
have recently announced or are considering to extend the maximum duration of support to avoid
that it runs out too quickly (e.g. Denmark, United Kingdom). In other countries, where the maximum
duration of job retention support is relatively long, it may be appropriate to shorten the maximum
duration of job retention subsidies for new applications. In general, governments have been clear
that support will remain available as long as government-imposed health restrictions remain in
place. However, countries may want to consider providing more information on their intentions to
extend or phase out job retention measures or the criteria that they use for making such decisions.
Promote the mobility of workers from subsidised to unsubsidised jobs. This can be achieved by
requiring or allowing workers on STW to register with the PES and benefit from their support
(e.g. job search assistance, career guidance and training) (OECD, forthcoming[116]). OECD
analysis shows that early interventions – including those before displacement takes place – can
be very effective in promoting smooth job transitions (OECD, 2018[117]). However, only few
countries require workers to register with the PES and to engage in active job search while on
STW. Countries may not see this as a priority since many of the workers on reduced working hours
will stay with their current employer even after the crisis. It may also not be practical since in most
countries STW subsidies are paid to the firm rather than to the worker. Benefit receipt therefore
does not provide a natural point of contact between workers on reduced hours and providers of
employment services as in the case of unemployment benefits. Indeed, in countries where STW
subsidies are paid directly to workers, job search requirements have traditionally been more
common (Hijzen and Venn, 2011[81]). Irrespective of whether payments are made to the worker or
to the firm, countries could encourage workers to register with the PES on a voluntary basis to
allow them to benefit from their services and support their career progression, whether in their
current firm or a different one.
Promote training participation of workers on reduced hours. Participation in training while on
reduced working hours can help workers improve the viability of their current job or improve the
prospect of finding a different job. Several countries encourage training during STW by providing
financial incentives to firms or workers (e.g. France and Germany), while in a few others
participation in training is a requirement for receiving JRS subsidies. For example, in the
Netherlands, employers applying for job retention support have to declare from June 2020 that
they actively encourage training, while the government has taken additional measures to make on-
line training and development courses freely available). A key challenge is to organise training in
such a way that it can be combined with part-time work and irregular work schedules while
maintaining physical distancing. This is easiest when training courses are targeted at individuals
rather than groups, delivered in a flexible manner through online teaching tools and if their duration
is relatively short (OECD, forthcoming[109]). In the present context, training courses that promote
the return to work in a way that is consistent with new standards for occupational safety and health
could be particularly valuable. The same applies to training courses to promote worker mobility to
jobs in expanding firms and industries (e.g. online services).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
86
1.4.5. Ensuring adequate income protection
With OECD unemployment projected to rise well above the level attained during the global financial crisis,
and to decline only gradually in 2021, income support systems across OECD countries will face heavy
pressure. Income support for jobseekers and their families is provided under various headings, including
unemployment insurance and assistance, minimum-income benefits, as well as other transfers that may
or may not depend on the family’s income situation. Among these, unemployment benefits are, in principle,
best placed to provide an effective combination of income support, job search incentives and access to
re-employment services. Some countries already experienced immediate, large inflows into their
unemployment benefit systems when the crisis struck (see Section 1.2). In many others, the number of
recipients will rise with a delay as some companies will lay off their workers when JRS end, or if a slow
recovery – or even a second infection wave – should cause another series of bankruptcies. When weak
labour market conditions persist, there can be good arguments for making unemployment benefits more
accessible. For example, with reduced job finding rates, and lengthening unemployment spells, extending
benefit durations can help to ensure that unemployment compensation systems continue to facilitate a
reasonable match between jobseeker and vacancies and provide effective income support during the
jobless spell (Immervoll, 2012[118]).
A key question is whether more generous benefits may worsen labour market outcomes and delay a
recovery by reducing job search incentives. Policy changes during the aftermath of the global financial
crisis provide useful pointers for considering the advantages and drawbacks of different benefit designs in
this respect. For example, earlier studies of benefit extensions have found that any adverse effects of
benefit generosity on individual job search intensities were indeed about the same during recessions and
booms (Schmieder, von Wachter and Bender, 2012[119]). But results also suggest that the intensity of job
search makes less of a difference to employment outcomes when there are long queues of jobseekers
and a much-reduced number of vacancies. As a result, aggregate unemployment is less sensitive to
changes in benefit generosity when labour markets are weak. In countries where this is the case, the
efficiency costs of providing support would then be no greater (and perhaps smaller) in recessions
(Rothstein, 2011[120]; Lalive, Landais and Zweimüller, 2015[121]; Landais, Michaillat and Saez, 2018[122]). At
the same time, the need for benefit support is greater, so the cost/benefit ratio of unemployment support
would be more attractive when unemployment is high.
When many unemployed exhaust their benefits without finding employment, countries should review
benefit provisions, both for social and for economic reasons. Likewise, where benefit entitlement durations
are already generous, an argument that benefit provisions should be responsive to the economic cycle
may imply shortening durations once the labour market recovers. Linking automatic changes in the
duration of receipt to the overall unemployment rate may be viable in some cases. In all cases, benefit
extensions arguably need to be accompanied by changes in related policy areas. For example, extensions
can be accompanied by measures such as “soft sanctions” (e.g. requiring claimants to re-apply before any
extensions are granted, introducing waiting periods between consecutive claiming periods, or reducing
benefit amounts over time). In general, it is important to retain a strong link between benefit receipt and
active job search. Changing benefit provisions is, however, much easier and quicker than, say, changing
PES staffing levels or intake procedures (see below).
Countries may also want to assess how to adjust or phase out emergency support programmes for
self-employed workers (e.g. new earnings replacement schemes in Austria or the United Kingdom) and
small businesses (e.g. cash support for costs in Germany) introduced in the initial phase of the crisis. While
the need for such programmes will subside as economic activity picks up again, some viable businesses
may continue to face restrictions and/or low demand because of the crisis. The trade-offs in deciding
whether, and for how long, to support these businesses are similar for small and large businesses, and
resemble those for the phase-out of JRS (see above). In any case, governments may need to re-assess
programmes that were designed to deliver support quickly, and with limited concern for targeting. Where
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
87
earnings replacement schemes were set up without a past earnings test, these tests could be introduced
now. Similarly, where payments provide very high earnings replacement rates50, this could be revised.
Unlike unemployment benefits, these benefits are not balanced by prior contributions. Effective targeting
is therefore not only important out of efficiency but also out of equity concerns. More generally, this crisis
has shown the need to let self-employed workers build up rights to the types of out-of-work support
available to dependent employees. While including the self-employed in earnings-related social-protection
schemes can be fraught with moral hazard and other administrative concerns, several countries have been
successful in establishing well-designed policies that work for their circumstances – see OECD (2018[123];
2019[88]).
As the crisis lasts longer, claimant numbers for “last-resort” minimum-income benefits may rise as workers
who lost their jobs and incomes in the initial phase of the crisis exhaust their unemployment benefit
entitlements or run down their savings. Even in normal times, the accessibility, reactivity and generosity of
these programmes differ markedly across countries (Hyee, Fernández and Immervoll, forthcoming[124]).
Many countries have eased entitlement criteria and simplified application procedures to ensure broad-
based and prompt access to these schemes during government-imposed restrictions. As they consider
rolling back these concessions, countries could review and simplify entitlement criteria and application
procedures with a view to making minimum-income benefits more reactive and accessible to encourage
take-up. Effective targeting is important as fiscal pressures mount, but countries need to ensure that those
in urgent need continue to receive support. For example, countries could gradually phase back in income
tests to allow households to adjust their expenditure, while keeping asset tests relaxed (e.g. exempt the
family home or any business assets) as long as job opportunities remain scarce. Countries may also want
to expand these programmes to cover young adults, where this is not already the case.
To further ease pressure that some workers face because of pandemic-related income loss, countries
might consider extending some of the emergency housing support measures introduced during the crisis
(see Box 1.8). In the event of a second infection wave, particularly in winter, bans on evictions and
foreclosures and targeted financial support to cover utilities could help workers remain in their homes.
Even without a second infection wave, many workers likely face an extended period of economic fragility,
making it hard to cover mortgage and rent payments in the months to come. Extending mortgage
forbearance and eviction bans, which in some OECD countries are set to expire in the fall of 2020, could
help workers in the short term, but increase financial fragility in the financial system and further impose
financial burdens on landlords. The cost of extending these measures could be shared more evenly,
e.g. through partial mortgage and rental payments, and these measures should be gradually phased out
as the economic situation improves. Meanwhile, demand for housing allowances, social housing and other
forms of support are likely to increase. Unlike the global financial crisis, the current crisis may
disproportionately affect renters and call for reinforced rental supports; relative to homeowners, renters
faced greater affordability challenges prior to the pandemic and are more likely to work in the most-affected
industries. Nevertheless, most emergency support measures remain at best temporary fixes. The
pandemic has underscored the need to develop more structural responses to address persistent housing
challenges for workers across the OECD, bringing to the fore the need for increased investment in social
and affordable housing, as well as upgrades to the existing stock to improve housing quality.
1.4.6. Expanding employment services and training
In addition to adequate income support, workers who have lost their jobs during the current crisis require
assistance and encouragement to find new work, increase their long-term employability and avoid falling
into long-term unemployment. Many countries temporarily reduced job search support and suspended
“mutual obligations” requirements for jobseekers in the initial phase of the crisis to meet physical-distancing
requirements and relieve pressure from their PES (see Section 1.3.3). As the health emergency is
subsiding, countries should gradually revive their activation regimes making government support again
conditional on active job search or participation in programmes that improve their job prospects (OECD,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
88
forthcoming[116]). This can support flows into employment, even if job opportunities continue to be
depressed in some sectors and as PES have to take into account health and safety considerations when
referring jobseekers to vacancies. Some jobseekers may be able to seize up on job opportunities that arise
even in times of crisis, including in essential occupations. For others, the crisis may represent an
opportunity for up-skilling or re-training, though physical-distancing requirements will reduce the scope of
in-person training courses on offer. Young people, as one of the groups hit hardest in the initial phase of
this crisis, deserve special attention.
This will require equipping PES with additional resources. As the number of jobseekers and participation in
JRS will remain high for the near future, PES will continue to face a much greater demand for their services
than before the crisis. PES in many countries will therefore need to build up capacity not to permanently
neglect support and services that may have been of secondary importance during the initial phase of the
crisis (e.g. career advice, counselling). Countries should scale up active labour market programmes
(ALMPs) that have proven effective to ensure effective re-employment support to all unemployed
jobseekers, promote job mobility, increase the quality of job matches, reduce unemployment and prevent
long-term unemployment. In particular, this includes programmes that support a fast return to the labour
market such as job search support and counselling. Moreover, there is a case for supporting job creation
by temporarily scaling up easy-to-expand, time-limited hiring subsidies, as many OECD countries did during
the global financial crisis (OECD, 2010[82]). Hiring subsidies, in particular if targeted at low-pay workers, can
boost job growth and be cost effective after accounting for savings in social benefit payments – e.g. Cahuc,
Carcillo and Le Barbanchon (2018[125]) for France and Neumark and Grijalva (2016[126]) for the United States.
A number of countries also extended re-employment bonuses for jobseekers (e.g. in form of re-employment
allowances) during the global financial crisis to raise the incentives to take up work (OECD, 2009[21]). Past
evidence shows that ALMPs tend to have a larger impact in periods of slow growth and higher
unemployment (Card, Kluve and Weber, 2018[127]). However, most countries increased ALMP spending
only modestly during the global financial crisis. In the OECD on average, a 1% increase in the number of
unemployed was associated with a 0.4% increase in ALMP spending (OECD, 2017[128]).
The crisis may also be an occasion for countries to modernise employment services and make them more
flexible. PES with well-developed digital services (i.e. e-services for PES users and automated PES
back-office systems) and staff teleworking arrangements found themselves much better prepared to
respond to the crisis keep their service offers largely intact. In countries where these areas are still less
developed, such innovations could contribute to making services available to a large number of jobseekers
while respecting physical-distancing requirements. However, PES will also need to develop strategies to
identify (e.g. through profiling tools) and support jobseekers without digital skills and those with complex
needs in times when the scope for face-to-face interactions may remain limited.
PES, as well as other private and public training providers, have the additional role to enable and
encourage jobseekers and workers to move from sectors that operate below capacity to those that picked
up again more quickly. Experiences made with the rapid retraining and matching of workers over the last
months may prove valuable in this respect. In the short run, job transitions are easiest when the new job
either requires little or no specialised training, or has broadly similar skill requirements as the previous job.
This may include retraining displaced workforce from “non-essential” retailers to be hired by “essential”
retailers, for example. Similarly, ultra-short courses may be sufficient to support the transition of displaced
vocational and technical workers into currently in-demand occupations. Most learning activities may have
to take place online until gathering in groups is deemed safe, but this may require mastering a certain level
of digital skills. Successful programmes therefore include provisions to support participants who may lack
the digital skills or the motivation to complete the learning activity, and they prepare teachers and design
curricula for online didactics (OECD, forthcoming[109]). To the extent that cross-sectoral imbalances in
labour and skill demand persist during post-confinement, countries will also benefit from further developing
their skill assessment and anticipation, and skills profiling tools, as well as their career guidance systems,
which can guide workers to the most efficient job transition (OECD, forthcoming[129]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
89
1.4.7. Giving young people the support they need
To prevent the crisis from leaving long-lasting scars on young people’s careers, countries need to act
quickly and help young people maintain their links with the labour market and education system. School
closures raised the risk of school dropout, temporary contracts are not being renewed, internships and
apprenticeships are being cancelled, and new graduates face great uncertainties about their labour market
entry. High and persistent youth unemployment in the aftermath of the global financial crisis showed that
once young people have lost touch with the labour market, re-connecting them can be very hard (Carcillo
et al., 2015[40]; OECD, 2016[130]). The realisation that early action is key is also the basis of the European
Union’s Youth Guarantee, a commitment made by all EU Member States in 2013 to ensure that all young
people below 25 receive a good-quality employment or training offer within four months of leaving school
or becoming unemployed.
Support for companies who offer jobs or work experience to young people have proven an effective tool to
promote job creation in times of crisis. Australia and Denmark have introduced wage subsidies to help
companies maintain or expand their apprenticeship and in-firm training programmes, while Germany and
Scotland are introducing subsidies for employers who take on apprentices who have been made redundant
during the crisis (OECD, 2020[131]). Canada expanded its Summer Jobs Program that provides wage
subsidies for below-30-year-olds, and France is considering a hiring premium or a reduction in employer
contributions for young workers. In times of depressed labour demand, volunteering can be a useful
alternative for young people to gain practical experience and acquire new skills, and governments could
encourage its use through grants.
Effective outreach strategies are crucial to re-establish contact with young people who recently lost their
jobs or left school without finding employment. Particularly the more vulnerable young people often do not
get in contact with the PES, because they are not entitled to income support, lack trust in public authorities
or are simply not aware of the support they can receive. Rapid and proactive outreach – in collaboration
with schools and youth organisations and through social-media campaigns – may be particularly important
in the current crisis.
The OECD Action Plan for Youth (OECD, 2013[132]) sets out a toolkit of measures that countries and
stakeholders can take to promote better outcomes for young people. This includes cost-effective active
labour market measures, such as counselling, job search assistance, entrepreneurship programmes, and
intensive support for more disadvantaged young people. Increased use of online support and virtual-
learning platforms, including in vocational education and training, can allow the PES and education
providers to continue offering their services while meeting physical-distancing requirements (OECD,
2020[131]).
1.5. Concluding remarks
This chapter provides a first assessment of the crisis’ initial labour market impact as well as an overview
of the massive policy response that OECD countries quickly put in place. The immediate impact of the
crisis on employment and hours worked has been ten times larger than in the first months of the 2008
global financial crisis, even in countries where unemployment rates have so far not increased much. Once
more, vulnerable workers are bearing the brunt of the shock, with low-skilled workers and those in
non-standard employment having been particularly exposed. Women seem to have suffered greater initial
employment losses than men. They have also been playing a key role in the health care response to the
pandemic, and the crisis likely amplified their unpaid work burden. Young people have again been hit hard,
and some of them are experiencing already the second deep crisis in their still young careers.
OECD countries rapidly took comprehensive and far-reaching measures to contain the economic fallout
and support workers, their families and companies. The massive use of job retention schemes in many
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
90
OECD countries saved jobs and protected the survival of many companies by allowing employers to cut
the hours of work for their workers, or putting them “on hold”, without having to lay them off. Countries
also increased the coverage and adequacy of income support, including for groups previously poorly
covered, or not covered at all, hence cushioning income losses for many of those hit hardest. As
countries are now gradually re-opening their economies, they will have to adapt these initial policy
packages to better account for the large existing heterogeneity in situations across workers and
companies, while fostering incentives to resume work without running the risk to end support prematurely
where it is still needed.
Uncertainty about the future labour market developments remains large, and much depends on how the
pandemic evolves. The virus has by no means been defeated, and the risk of new outbreaks is still looming
until a vaccine is available. The big challenge for countries is, therefore, to find ways of re-starting economic
and social life and steering the economy towards recovery while keeping the pandemic in check without
having to revert back to strict containment measures. This requires putting in place comprehensive public
health interventions, which range from massively upscaling of testing, tracking and tracing (TTT), to
enhancing personal hygienic measures and the continuous enforcement of some physical-distancing
policies.
While there is no doubt that bold measures were needed to avoid health systems from collapsing and
mitigate the economic fallout from the pandemic, the evaluation of these emergency policy packages has
only just begun. There is much to be learned about how countries’ strategies and policy packages are
affecting various groups of workers and companies across sectors and regions. The heterogeneity in the
mix, timing and design of measures across countries provides a strong potential for policy evaluation and
mutual learning. Such analysis will provide crucial insights into how OECD labour markets and
social-protection systems react in times of extreme pressure, and it is an occasion to learn lessons for
strengthening their resilience.
Overview of Chapters 2 to 5: Worker security
Protecting individuals against labour market risks is a key pillar of the OECD Jobs Strategy (OECD,
2018[83]). The unprecedented health and economic crisis that the world is currently experiencing with
the COVID-19 pandemic has shone the spotlight on the crucial importance of well-designed worker
security strategies to protect workers and households against unforeseeable shocks.
Effective social safety nets are fundamental for cushioning income shocks. And unemployment
benefits are among the key instruments providing protection against earnings falls resulting from job
losses. Yet a number of workers do not meet the entitlement or eligibility criteria to receive benefits and
are therefore at greater risk of facing severe income losses. On average, only about one-quarter of
jobseekers receive unemployment benefits (OECD, 2018[133]). The additional evidence provided by
OECD (2019[39]) suggests that social safety nets are particularly weak for the self-employed, although
a number of countries have extended access to out-of-work support for this category of workers during
the current crisis (see Section 1.3.3 above).
Chapter 2 of this volume sheds further light on safety net disparities by looking at the uneven access
to unemployment benefits for different types of dependent employees. Even if entitlement rules
are usually the same for all dependent employees, conditions on minimum employment durations,
working hours or earnings before the unemployment spell, are often harder to meet for those who lose
a part-time job or have an employment trajectory involving frequent transitions between employment
and unemployment. The same applies to the rules for unused entitlements, which in a number of
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
91
countries may put workers alternating short spells of employment and unemployment at a disadvantage
compared with employees with fewer transitions and longer unemployment spells. Consequently, even
when workers are in the same family and income situation, have the same average annual wage and
have accumulated the same number of hours of work as dependent employees over a given period,
entitlements tend to be smaller for those with non-standard employment trajectories than for those who
were previously in long-term, full-time positions. In turn, the risk of falling into poverty tends to be greater
for workers in non-standard dependent employment.
Correcting the possible inadequacy of benefit entitlements to provide more income security may be
challenging, however. Avoiding trade-offs between benefit generosity and work incentives can be like
walking a tightrope, as Chapter 2 shows. Nevertheless, several policy instruments can be used to
create a policy mix that strikes the right balance between work incentives and income security:
customised extensions of employment reference periods; earnings disregards and withdrawal rates
when combining earnings from work and unemployment benefits; waiting periods and tight rules on the
retention of unused benefits; differentiated contribution rates by type of contract; integration of in-work
and out-of-work benefits; and co-ordination of active and passive labour market policies.
Protecting workers against income shocks following job losses is, however, costly. Yet, individual
employers typically do not factor in the social costs of unemployment benefits when they take their
decision to dismiss a worker, nor other social costs, such as firm- and sector-specific human capital
destruction, negative health effects (particularly psychosocial risks) and possible intergenerational
consequences. Experience rating of unemployment and other social security contributions and
employment protection legislation (EPL), in particular regulations concerning individual and
collective dismissals, are the primary instruments that policy makers can use to induce employers
to avoid socially inefficient dismissals (those which are decided without taking account of their social
impact).
Chapter 3 provides an up-to-date comparative review of where OECD countries stand as regards
employment protection legislation (EPL). To do this, the chapter develops a new version of the OECD
EPL indicators, which takes more account of regulations for collective dismissals, enforcement issues
and regulations concerning unfair dismissals.
As Chapter 3 underlines, EPL has several dimensions and its effects on worker security may
depend on the balance among them. For example, sufficiently long advance notice periods are crucial
to allow early interventions by employment services before the dismissal takes effect, thereby facilitating
the transition to another job. This suggests that countries with short notice periods and high severance
pay could consider reducing severance pay and increasing notice periods, while activating early
interventions, to smooth job transitions without increasing costs to employers. Similarly, EPL measures
against unfair dismissals play a vital role in preventing abuses, but clear enforcement rules are
necessary to avoid creating uncertainty. Moreover, excessively stringent EPL rules harm worker
security both directly, by reducing hiring and making jobless spells longer, and indirectly, by slowing
growth in productivity and, therefore, wages and incomes. A balanced employment protection
framework that provides effective adaptability for firms and adequate protection for workers is thus
required.
As stressed by the OECD Jobs Strategy (OECD, 2018[83]), however, one of the best ways of
protecting workers and promoting an inclusive labour market is by addressing problems before
they arise. This means that preventive policies are at least as important as remedial policies. Preventive
measures can enable workers to avoid many of the social and financial costs associated with labour
market risks (such as unemployment, sickness and disability) and to enjoy better jobs and careers.
Education, training and skills policies play a fundamental role in this context and countries therefore
need to develop high-quality education and training systems that enable workers to acquire and develop
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
92
the skills that are in demand in the labour market. The last two chapters of this volume focus on the
demand and supply of skills.
The share of middle-skill jobs – occupations in the middle of the wage distribution – declined in OECD
countries over the past two decades due to falling demand for these jobs. At the same time, the shares
of both high-skill and low-skill occupations have increased. The causes and consequences of this
phenomenon, termed job polarisation, have been the subject of a heated debate in the economics and
policy literature. The contribution of Chapter 4 is, first and foremost, to dispel a myth. A popular
perception is that the contraction of middle-skill occupations has occurred through firms increasingly
dismissing middle-skill workers and forcing mid-career workers to find new employment in other skill
groups. While downsizing, especially of manufacturing firms, has obviously played a prominent role in
specific situations, such transitions do not appear to explain the aggregate polarisation trend. Rather,
the gradual retirement of older middle-skill workers and the different entry patterns of younger
workers in other, growing occupations appear to drive job polarisation.
Policy makers consequently have to pay special attention to education choices and the transition
between school and work. Vocational Education and Training (VET) programmes lead to market-
relevant, vocational qualifications and typically enhance student engagement in education, reduce
school dropout rates and facilitate school-to-work transitions. However, there is a growing concern that
the increasing polarisation of the labour market may be having a negative impact on the labour market
performance of non-tertiary VET graduates, who are typically preparing for middle-skill occupations.
The results of Chapter 4 add to this concern, as young generations appear to be bearing the brunt of
job polarisation. Chapter 5 therefore re-examines the labour market performance of middle-educated
VET graduates. Reassuringly, it finds that these graduates maintain a labour market advantage over
their general education peers on labour market entry, although this advantage tends to disappear later
in the career and their performance is worse than that of higher education graduates, even at labour
market entry. Middle-educated VET graduates have managed to maintain their position in shrinking
middle-skill occupations by increasing their share in these occupations relative to other groups.
However, in some occupations that have a larger share of VET graduates among their young workers,
the supply of labour with the relevant skills exceeds the corresponding demand, and many of these
typical VET jobs are at high risk of automation. In a number of countries, to reinforce the positive
impact VET systems can have on the labour market outcomes of VET graduates, some re-
engineering of VET programmes may be necessary, including reinforcing their foundational skills
component and developing closer co-operation between VET institutions and social partners, as occurs
in a number of countries with successful VET systems.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
93
References
Adams-Prassl, A. et al. (2020), “Inequality in the Impact of the Coronavirus Shock: Evidence [32]
from Real Time Surveys”, IZA Discussion Paper, No. 12183, IZA Bonn.
Alfers, L., R. Moussié and J. Harvey (2020), The COVID-19 crisis: income support to informal [95]
workers is necessary and possible, https://oecd-development-matters.org/2020/04/22/the-
covid-19-crisis-income-support-to-informal-workers-is-necessary-and-possible/.
Alstadsæter, A. et al. (2020), The First Weeks of the Coronavirus Crisis: Who Got Hit, When and [59]
Why? Evidence from Norway, National Bureau of Economic Research, Cambridge, MA,
http://dx.doi.org/10.3386/w27131.
Andersen, A. et al. (2020), “Pandemic, Shutdown and Consumer Spending: Lessons from [9]
Scandinavian Policy Responses to COVID-19”, mimeo, University of Copenhagen.
Andrew, A. et al. (2020), How are mothers and fathers balancing work and family under [50]
lockdown?, Institute for Fiscal Studies, London, http://dx.doi.org/10.1086/703115.
Angelici, M. and P. Profeta (2020), “Smart-working: Work Flexibility Without Constraints”, [143]
Working Paper, No. 137, Dondena Centre Bocconi Unviersity.
AppJobs Institute (2020), How Does the COVID-19 Outbreak Affect the Life of Freelancers & Gig [38]
Workers, https://www.institute.appjobs.com/how-does-the-covid-19-outbreak-affect-the-life-of-
freelancers-gig-workers-march-2020 (accessed on 11 May 2020).
Aum, S., S. Lee and Y. Shin (2020), COVID-19 Doesn’t Need Lockdowns to Destroy Jobs: The [11]
Effect of Local Outbreaks in Korea, National Bureau of Economic Research, Cambridge, MA,
http://dx.doi.org/10.3386/w27264.
Baldwin, R. (2019), The Globotics Upheaval: Globalization, Robotics, and the Future of Work, [155]
Oxford University Press.
Barbieri, T., G. Basso and S. Scicchitano (2020), “Italian Workers at Risk during the COVID-19 [22]
Epidemic”, SSRN Electronic Journal, http://dx.doi.org/10.2139/ssrn.3572065.
Barker, N. (2020), The housing pandemic: four graphs showing the link between COVID-19 [102]
deaths and the housing crisis, Inside Housing, https://www.insidehousing.co.uk/insight/the-
housing-pandemic-four-graphs-showing-the-link-between-covid-19-deaths-and-the-housing-
crisis-66562 (accessed on 2 June 2020).
Baronio, G. and G. Linfante (2020), Prime evidenze degli effetti della crisi sanitaria sulla [35]
dinamica dei rapporti di lavoro, ANPAL, Rome,
https://www.anpal.gov.it/documents/552016/554592/focus-anpal-covid-14-05-1.pdf (accessed
on 22 May 2020).
Barrot, J., B. Grassi and J. Sauvagnat (2020), “Sectoral Effects of Social Distancing”, SSRN [23]
Electronic Journal, http://dx.doi.org/10.2139/ssrn.3569446.
Basso, G. et al. (forthcoming), “The new hazardous jobs and worker reallocation”, OECD Social, [113]
Employment and Migration Working Papers, OECD Publishing, Paris.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
94
Battiston, D., J. Blanes i Vidal and V. Kirchmaier (2017), “Is Distance Dead? Face-to-Face [144]
Communication and Productivity in Teams Acknowledgements We thank”, Discussion Paper,
No. 1473, Centre for Economic Performance.
Beckmann, M. (2015), “Working-time autonomy as a management practice”, IZA World of Labor, [147]
http://dx.doi.org/10.15185/izawol.230.
Beckmann, M., T. Cornelissen and M. Kräkel (2017), “Self-managed working time and employee [146]
effort: Theory and evidence”, Journal of Economic Behavior and Organization, Vol. 133,
pp. 285-302, http://dx.doi.org/10.1016/j.jebo.2016.11.013.
Bell, D. and D. Blanchflower (2011), “Young people and the Great Recession”, Oxford Review of [42]
Economic Policy, Vol. 27/2, pp. 241-267, http://dx.doi.org/10.1093/oxrep/grr011.
Benzeval, M. et al. (2020), “The Idiosyncratic Impact of an Aggregate Shock: The Distributional [58]
Consequences of COVID-19”, Working Paper Series, No. 9, Understanding Society,
http://dx.doi.org/10.5255/UKDA-SN-8644-1.
Blau, F., J. Koebe and P. Meyerhofer (2020), Essential and Frontline Workers in the COVID-19 [26]
Crisis, econofact.org, https://econofact.org/essential-and-frontline-workers-in-the-covid-19-
crisis (accessed on 21 May 2020).
Bloom, N. et al. (2014), “Does Working from Home Work? Evidence from a Chinese [158]
Experiment”, The Quarterly Journal of Economics, Vol. 130/1, pp. 165-218,
http://dx.doi.org/10.1093/qje/qju032.
Bodas, M. and K. Peleg (2020), “Self-Isolation Compliance In The COVID-19 Era Influenced By [70]
Compensation: Findings From A Recent Survey In Israel”, Health Affairs,
http://dx.doi.org/10.1377/hlthaff.2020.00382.
Boniol, M. et al. (2019), “Gender equity in the health workforce: Analysis of 104 countries”, [46]
Health Workforce Working Paper, No. 1, World Health Organization,
https://www.who.int/hrh/resources/gender_equity-health_workforce_analysis/en/ (accessed
on 24 March 2020).
Bovini, G. et al. (2020), La crisi Covid e il mercato del lavoro: alcune conferme, alcune specificità [34]
regionali, alcune ipotesi, Banca d’Italia, Rome.
Browne, J. and H. Immervoll (2017), “Mechanics of replacing benefit systems with a basic [99]
income: comparative results from a microsimulation approach”, The Journal of Economic
Inequality, Vol. 15/4, pp. 325-344, http://dx.doi.org/10.1007/s10888-017-9366-6.
Business at OECD (2020), Covid-19 special webpage, http://biac.org/coronavirus/ (accessed on [62]
21 May 2020).
Cahuc, P. and S. Carcillo (2011), Is Short-Time Work a Good Method to Keep Unemployment [85]
Down?, https://pdfs.semanticscholar.org/17bc/c2fab64fd829f868667dac2186babf38bf2c.pdf
(accessed on 25 October 2018).
Cahuc, P., S. Carcillo and T. Le Barbanchon (2018), “The Effectiveness of Hiring Credits”, The [125]
Review of Economic Studies, Vol. 86/2, pp. 593-626, http://dx.doi.org/10.1093/restud/rdy011.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
95
Carcillo, S. et al. (2015), “NEET Youth in the Aftermath of the Crisis: Challenges and Policies”, [40]
OECD Social, Employment and Migration Working Papers, No. 164, OECD Publishing, Paris,
https://dx.doi.org/10.1787/5js6363503f6-en.
Card, D., J. Kluve and A. Weber (2018), “What Works? A Meta Analysis of Recent Active Labor [127]
Market Program Evaluations”, Journal of the European Economic Association, Vol. 16/3,
pp. 894–931, http://dx.doi.org/10.1093/jeea/jvx028.
Carnevale, A., T. Jayasundera and D. Repnikov (2014), Understanding online job ads data, [136]
Georgetown University, Washington, DC.
CDC (2012), Principles of Epidemiology in Public Health Practice, U.S. Department of Health [135]
and Human Services, Atlanta, GA, https://www.cdc.gov/csels/dsepd/ss1978/SS1978.pdf
(accessed on 4 May 2020).
Chen, S. et al. (2020), “Tracking the Economic Impact of COVID-19 and Mitigation Policies in [7]
Europe and the United States”, IMF Special Series on COVID-19, International Monetary
Fund, Washington, D.C., https://www.imf.org/~/media/Files/Publications/covid19-special-
notes/enspecial-series-on-covid19tracking-the-economic-impact-of-covid19-and-mitigation-
policies-in-europe.ashx?la=en (accessed on 11 May 2020).
Chetty, R. et al. (2020), “Real-Time Economics: A New Platform to Track the Impacts of COVID- [8]
19 on People, Businesses, and Communities Using Private Sector Data”, Harvard University
and Opportunity Insight,
https://www.hks.harvard.edu/centers/mrcbg/programs/growthpolicy/real-time-economics-new-
platform-track-impacts-covid-19-people (accessed on 22 May 2020).
Coibion, O., Y. Gorodnichenko and M. Weber (2020), “The Cost of the Covid-19 Crisis: [16]
Lockdowns, Macroeconomic Expectations, and Consumer Spending”, NBER Working Paper,
No. 27141, NBER, Cambridge, MA, http://dx.doi.org/10.3386/w27141.
DARES (2020), Focus sur les demandeurs d’emploi inscrits à Pôle emploi en avril 2020, [33]
DARES, Paris.
Deb, P. et al. (2020), “The effect of containment measures on the COVID-19 pandemic”, Covid [5]
Economics 19, https://cepr.org/sites/default/files/news/CovidEconomics19.pdf (accessed on
4 June 2020).
Dingel, J. and B. Neiman (2020), “How Many Jobs Can be Done at Home?”, Becker Friedman [151]
Institute, University of Chicago.
Edwards, K. (2020), Millions Need Unemployment Benefits. Unfortunately, the Delivery System [91]
Is Broken, The RAND Blog, https://www.rand.org/blog/2020/04/millions-need-unemployment-
benefits-unfortunately-the.html (accessed on 14 May 2020).
Espinoza, R. and L. Reznikova (2020), “Who can log in? The importance of skills for the [156]
feasibility of teleworking arrangements across OECD countries”, OECD Social, Employment
and Migration Working Papers, No. 242, OECD Publishing, Paris,
https://dx.doi.org/10.1787/3f115a10-en.
EU-OSHA (2020), COVID-19: Back to the Workplace - Adapting workplaces and protecting [114]
workers, Public Office of the European Union, Luxembourg.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
96
Eurofound (2020), “Labour market change: Trends and policy approaches towards [141]
flexibilisation”, Challenges and Prospects in the EU Series,
https://www.eurofound.europa.eu/publications/flagship-report/2020/labour-market-change-
trends-and-policy-approaches-towards-flexibilisation (accessed on 19 May 2020).
Eurofound (2020), Living, working and COVID-19: First findings - April 2020, Publications Office [78]
of the European Union, Luxembourg, https://www.eurofound.europa.eu/topic/covid-19
(accessed on 9 May 2020).
Eurofound (2018), Living and working in Europe, Publications Office of the European Union, [61]
Luxembourg.
Fana, M. et al. (2020), The COVID confinement measures and EU labour markets, Publications [27]
Office of the European Union, Luxembourg.
Fernández, R., H. Immervoll and D. Pacifico (forthcoming), “Beyond repair? Anatomy of income [89]
support for standard and non-standard workers in OECD countries”, OECD Employment,
Social and Migration Working Papers, OECD Publishing, Paris.
Foucault, M. and V. Galasso (forthcoming), “Working during COVID-19: Cross-Country Evidence [29]
from Real-Time Survey Data”, OECD Social, Employment and Migration Working Papers,
OECD Publishing, Paris.
Friedson, A. et al. (2020), “Did California’s Shelter-In-Place Order Work? Early Coronavirus- [15]
Related Public Health Effects”, IZA Discussion Paper, No. 13160, IZA, Bonn.
Furceri, D. et al. (2020), “Will Covid-19 affect inequality? Evidence from past pandemics”, Covid [159]
Economics, Vol. 12, pp. 138-57, https://voxeu.org/article/covid-19-will-raise-inequality-if-past-
pandemics-are-guide.
Furman, J. (2020), “Protecting people now, helping the economy rebound later”, in Baldwin, R. [97]
and B. Weder di Mauro (eds.), Mitigating the COVID Economic Crisis: Act fast and Do
Whatever It Takes, VoxEU.org, https://voxeu.org/article/protecting-people-now-helping-
economy-rebound-later.
Ganong, P., P. Noel and J. Vavra (2020), “US Unemployment Insurance Replacement Rates [90]
During the Pandemic”, Working Paper, No. 2020-62, Becker Friedman Institute for Economics
at the University of Chicago, Chicago, https://github.com/ganong-noel/ui_calculator (accessed
on 16 June 2020).
Glenn Dutcher, E. (2012), “The effects of telecommuting on productivity: An experimental [145]
examination. The role of dull and creative tasks”, Journal of Economic Behavior &
Organization, Vol. 84/1, pp. 355-363, http://dx.doi.org/10.1016/j.jebo.2012.04.009.
Global Deal (2020), Social Partnership in the times of the COVID-19 Pandemic, Global Deal, [65]
Paris, https://www.theglobaldeal.com/resources/Social-Partnership-in-the-times-of-the-
COVID19-pandemic-Global-Deal.pdf.
Goos, M., A. Manning and A. Salomons (2014), “Explaining Job Polarization: Routine-Biased [138]
Technological Change and Offshoring”, American Economic Review, Vol. 104/8, pp. 2509-
2526, http://dx.doi.org/10.1257/aer.104.8.2509.
Gottlieb, C., J. Grobovšek and M. Poschke (2020), “Working from home across countries”, Covid [157]
Economics 8, pp. 71-91, https://cepr.org/sites/default/files/news/CovidEconomics8.pdf.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
97
Gupta, S. et al. (2020), “Tracking Public and Private Responses to the COVID-19 Epidemic: [12]
Evidence from State and Local Government Actions”, NBER Working Paper, No. 27027,
NBER, Cambridge, MA, http://dx.doi.org/10.3386/w27027.
Hale, T. et al. (2020), “Variation in government responses to COVID-19”, Working Paper, [2]
Blavatnik School of Government, http://www.bsg.ox.ac.uk/covidtracker (accessed on
11 May 2020).
Hensvik, L., T. Le Barbanchon and R. Rathelot (2020), “Job Search during the COVID-19 Crisis”, [10]
IZA Discussion Paper, No. 13237, IZA, Bonn.
Hershbein, B. and L. Kahn (2018), “Do Recessions Accelerate Routine-Biased Technological [137]
Change? Evidence from Vacancy Postings”, American Economic Review, Vol. 108/7,
pp. 1737-1772, http://dx.doi.org/10.1257/aer.20161570.
He, X. et al. (2020), “Temporal dynamics in viral shedding and transmissibility of COVID-19”, [75]
Nature Medicine, Vol. 26/5, pp. 672-675, http://dx.doi.org/10.1038/s41591-020-0869-5.
Hijzen, A. and S. Martin (2013), “The role of short-time work schemes during the global financial [84]
crisis1 and early recovery: a cross-country analysis”, IZA Journal of Labor Policy, Vol. 2/1,
p. 5, http://dx.doi.org/10.1186/2193-9004-2-5.
Hijzen, A. and D. Venn (2011), “The Role of Short-Time Work Schemes during the 2008-09 [81]
Recession”, OECD Social, Employment and Migration Working Papers, No. 115, OECD
Publishing, Paris, https://dx.doi.org/10.1787/5kgkd0bbwvxp-en.
Hyee, R., R. Fernández and H. Immervoll (forthcoming), “How reliable are ’safety-net’ benefits? [124]
Value and accessibility of minimum-income support in comparative perspective”, OECD
Social, Employment and Migration Working Papers, OECD Publishing, Paris,
https://doi.org/10.1787/1815199X.
ILO (2020), ILOSTAT Database, http://www.ilo.org/ilostat/ (accessed on 24 March 2020). [51]
Immervoll, H. (2012), “Reforming the benefit system to ’make work pay’: Options and priorities in [118]
a weak labour market”, IZA Policy Papers 50, http://ftp.iza.org/pp50.pdf.
Immervoll, H. and C. Knotz (2018), “How demanding are activation requirements for jobseekers”, [108]
OECD Social, Employment and Migration Working Papers, No. 215, OECD Publishing, Paris,
https://dx.doi.org/10.1787/2bdfecca-en.
Joyce, R. and X. Xu (2020), “Sector shutdowns during the coronavirus crisis: which workers are [28]
most exposed?”, IFS Briefing Note, No. BN278, Institute for Fiscal Studies, London.
Judge, L. and C. Pacitti (2020), Resolution Foundation Housing Outlook, Resolution Foundation, [103]
London.
Kahn, L., F. Lange and D. Wiczer (2020), “Labor Demand in the Time of COVID-19: Evidence [14]
from Vacancy Postings and UI Claims”, NBER Working Paper, No. 27061, NBER,
Cambridge, MA, http://dx.doi.org/10.3386/w27061.
Kneebone, E. and C. Murray (2020), Estimating COVID-19’s Near-Term Impact on Renters, [104]
Terner Center for Housing Innovation - UC Berkeley,
https://ternercenter.berkeley.edu/blog/estimating-covid-19-impact-renters (accessed on
21 May 2020).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
98
Knutsson, P., A. Tsvetkova and A. Lembcke (2020), “Using Burning Glass data for regional [139]
analysis: Opportunities and caveats”, OECD Publishing, forthcoming.
Kopp, D. and M. Siegenthaler (2019), “Short-Time Work and Unemployment in and after the [150]
Great Recession”, KOF Working Papers, No. 462, ETH, Zurich,
http://dx.doi.org/10.3929/ETHZ-B-000359533.
Kröll, C. and S. Nüesch (2019), “The effects of flexible work practices on employee attitudes: [148]
evidence from a large-scale panel study in Germany”, The International Journal of Human
Resource Management, Vol. 30/9, pp. 1505-1525,
http://dx.doi.org/10.1080/09585192.2017.1289548.
Lalive, R., C. Landais and J. Zweimüller (2015), “Market Externalities of Large Unemployment [121]
Insurance Extension Programs”, American Economic Review, Vol. 105/12, pp. 3564-3596,
http://dx.doi.org/10.1257/aer.20131273.
Landais, C., P. Michaillat and E. Saez (2018), “A Macroeconomic Approach to Optimal [122]
Unemployment Insurance: Theory”, American Economic Journal: Economic Policy, Vol. 10/2,
pp. 152-181, http://dx.doi.org/10.1257/pol.20150088.
Mallapaty, S. (2020), “How do children spread the coronavirus? The science still isn’t clear”, [153]
Nature, Vol. 581/7807, pp. 127-128, http://dx.doi.org/10.1038/d41586-020-01354-0.
Maloney, W. and T. Taskin (2020), “Determinants of Social Distancing and Economic Activity [4]
during COVID-19 A Global View”, Policy research working paper, No. 9242, World Bank,
Washington, DC.
Mas, A. and A. Pallais (2020), Alternative Work Arrangements, National Bureau of Economic [142]
Research, Cambridge, MA, http://dx.doi.org/10.3386/w26605.
Möhring, K. et al. (2020), Die Mannheimer Corona-Studie: Schwerpunktbericht zu [49]
Erwerbstätigkeit und Kinderbetreuung, Universität Mannheim, Mannheim, https://www.uni-
mannheim.de/gip/corona-studie. (accessed on 29 May 2020).
Mongey, S., L. Pilossoph and A. Weinberg (2020), “Which Workers Bear the Burden of Social [152]
Distancing Policies?”, SSRN Electronic Journal, http://dx.doi.org/10.2139/ssrn.3586077.
Müller, T. and T. Schulten (2020), Ensuring fair Short-Time Work - a European overview, [64]
European Trade Union Institute, Brussels.
Neumark, D. and D. Grijalva (2016), “The Employment Effects of State Hiring Credits”, ILR [126]
Review, Vol. 70/5, pp. 1111-1145, http://dx.doi.org/10.1177/0019793916683930.
OECD (2020), “Beyond Containment: Health systems responses to COVID-19 in the OECD”, [71]
OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
http://www.oecd.org/coronavirus/policy-responses/beyond-containment-health-systems-
responses-to-covid-19-in-the-oecd-6ab740c0/.
OECD (2020), “Combatting COVID-19’s effect on children”, OECD Policy Responses to [106]
Coronavirus (COVID-19), OECD Publishing, Paris, http://www.oecd.org/coronavirus/policy-
responses/combatting-covid-19-s-effect-on-children-2e1f3b2f/.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
99
OECD (2020), “Corporate sector vulnerabilities during the Covid-19 outbreak: Assessment and [80]
policy responses”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing,
Paris, http://www.oecd.org/coronavirus/policy-responses/corporate-sector-vulnerabilities-
during-the-covid-19-outbreak-assessment-and-policy-responses-a6e670ea/.
OECD (2020), “Distributional risks associated with non-standard work: Stylised facts and policy [30]
considerations”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing,
Paris, http://www.oecd.org/coronavirus/policy-responses/distributional-risks-associated-with-
non-standard-work-stylised-facts-and-policy-considerations-68fa7d61/.
OECD (2020), “Evaluating the initial impact of COVID-19 containment measures on economic [24]
activity”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/policy-responses/evaluating-the-initial-impact-of-covid-19-
containment-measures-on-economic-activity-b1f6b68b/.
OECD (2020), “Flattening the COVID-19 peak: Containment and mitigation policies”, OECD [1]
Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
http://www.oecd.org/coronavirus/policy-responses/flattening-the-covid-19-peak-containment-
and-mitigation-policies-e96a4226/.
OECD (2020), “Flattening the unemployment curve policies to support workers’ income and [86]
promote a speedy labour market recovery”, OECD Policy Responses to Coronavirus (COVID-
19), OECD Publishing, Paris, http://www.oecd.org/coronavirus/policy-responses/flattening-
the-unemployment-curve-policies-to-support-workers-income-and-promote-a-speedy-labour-
market-recovery-3dbd4087/.
OECD (2020), “From pandemic to recovery: Local employment and economic development”, [25]
OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
http://www.oecd.org/coronavirus/policy-responses/from-pandemic-to-recovery-local-
employment-and-economic-development-879d2913/.
OECD (2020), OECD Composite Leading Indicators News Release: Unprecedented collapse in [140]
CLIs in most major economies, OECD Publishing, Paris,
http://www.oecd.org/newsroom/Unprecedented-collapse-in-CLIs-in-most-major-
economies.pdf (accessed on 12 May 2020).
OECD (2020), “OECD Economic Outlook - All editions”, OECD Economic Outlook: Statistics and [6]
Projections (database), https://dx.doi.org/10.1787/826234be-en (accessed on 25 June 2020).
OECD (2020), OECD Economic Outlook, Volume 2020 Issue 1: Preliminary version, OECD [60]
Publishing, Paris, https://dx.doi.org/10.1787/0d1d1e2e-en.
OECD (2020), OECD Family Database, http://www.oecd.org/els/family/database.htm. [79]
OECD (2020), OECD Gender Data Portal, http://www.oecd.org/gender/data (accessed on [48]
13 May 2020).
OECD (2020), OECD Income Distribution Database (IDD), http://www.oecd.org/social/income- [53]
distribution-database.htm (accessed on 24 March 2020).
OECD (2020), OECD unemployment rate up sharply in March 2020 with data for April pointing to [19]
an unprecedented rise, OECD, Paris, http://www.oecd.org/sdd/labour-stats/unemployment-
rates-oecd-05-2020.pdf (accessed on 14 May 2020).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
100
OECD (2020), “Paid sick leave to protect income, health and jobs through the COVID-19 crisis”, [67]
OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
http://www.oecd.org/coronavirus/policy-responses/paid-sick-leave-to-protect-income-health-
and-jobs-through-the-covid-19-crisis-a9e1a154/.
OECD (2020), “Public employment services in the frontline for jobseekers, workers and [96]
employers”, OECD Covid-19 Policy Brief, OECD Publishing, Paris,
http://www.oecd.org/coronavirus/policy-responses/public-employment-services-in-the-
frontline-for-employees-jobseekers-and-employers-c986ff92/.
OECD (2020), “Supporting livelihoods during the COVID-19 crisis: closing the gaps in safety [37]
nets”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
http://www.oecd.org/coronavirus/policy-responses/supporting-livelihoods-during-the-covid-19-
crisis-closing-the-gaps-in-safety-nets-17cbb92d/.
OECD (2020), “Supporting people and companies to deal with the Covid-19 virus: Options for an [69]
immediate employment and social-policy response”, OECD Policy Responses to Coronavirus
(COVID-19), OECD Publishing, Paris, http://www.oecd.org/coronavirus/policy-
responses/supporting-people-and-companies-to-deal-with-the-covid-19-virus-options-for-an-
immediate-employment-and-social-policy-response-d33dffe6/.
OECD (2020), “Testing for COVID-19: A way to lift confinement restrictions”, OECD Policy [3]
Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/policy-responses/testing-for-covid-19-a-way-to-lift-
confinement-restrictions-89756248/.
OECD (2020), “Treatments and a vaccine for COVID-19: the need for coordinating policies on [110]
R&D, manufacturing and access - OECD”, OECD Policy Responses to Coronavirus (COVID-
19), OECD Publishing, Paris, http://www.oecd.org/coronavirus/policy-responses/treatments-
and-a-vaccine-for-covid-19-the-need-for-coordinating-policies-on-r-d-manufacturing-and-
access-6e7669a9/.
OECD (2020), “VET in a time of crisis: Building foundations for resilient vocational education and [131]
training systems”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing,
Paris, http://www.oecd.org/coronavirus/policy-responses/vet-in-a-time-of-crisis-building-
foundations-for-resilient-vocational-education-and-training-systems-efff194c/.
OECD (2020), Who Cares? Attracting and Retaining Care Workers for the Elderly, OECD Health [47]
Policy Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/92c0ef68-en.
OECD (2020), “Women at the core of the fight against COVID-19 crisis”, OECD Policy [57]
Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/policy-responses/women-at-the-core-of-the-fight-against-
covid-19-crisis-553a8269/.
OECD (2019), Affordable Housing Database - OECD, https://www.oecd.org/social/affordable- [100]
housing-database/ (accessed on 21 May 2020).
OECD (2019), Government at a Glance 2019, OECD Publishing, Paris, [52]
https://dx.doi.org/10.1787/8ccf5c38-en.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
101
OECD (2019), “Left on your own? Social protection when labour markets are in flux”, in OECD [88]
Employment Outlook 2019: The Future of Work, OECD publishing, Paris,
https://doi.org/10.1787/bfb2fb55-en.
OECD (2019), Negotiating Our Way Up: Collective Bargaining in a Changing World of Work, [66]
OECD Publishing, Paris, https://dx.doi.org/10.1787/1fd2da34-en.
OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, [39]
https://dx.doi.org/10.1787/9ee00155-en.
OECD (2019), SPHeP-NCDs’s documentation webpage, [111]
http://oecdpublichealthexplorer.org/ncd-doc/ (accessed on 12 June 2020).
OECD (2018), “Back to work: Lessons from nine country case studies of policies to assist [117]
displaced workers”, in OECD Employment Outlook 2018, OECD Publishing, Paris,
https://dx.doi.org/10.1787/empl_outlook-2018-8-en.
OECD (2018), Good Jobs for All in a Changing World of Work: The OECD Jobs Strategy, OECD [83]
Publishing, Paris, https://dx.doi.org/10.1787/9789264308817-en.
OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris, [133]
https://dx.doi.org/10.1787/empl_outlook-2018-en.
OECD (2018), The Future of Social Protection: What Works for Non-standard Workers?, OECD [123]
Publishing, Paris, https://doi.org/10.1787/9789264306943-en.
OECD (2018), Towards Better Social and Employment Security in Korea, Connecting People [68]
with Jobs, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264288256-en.
OECD (2017), “Basic income as a policy option: Can it add up?”, Policy Brief on the Future of [98]
Work, OECD, Paris, https://www.oecd.org/social/Basic-Income-Policy-Option-2017.pdf.
OECD (2017), OECD Employment Outlook 2017, OECD Publishing, Paris, [128]
https://dx.doi.org/10.1787/empl_outlook-2017-en.
OECD (2016), Society at a Glance 2016: OECD Social Indicators, OECD Publishing, Paris, [41]
https://dx.doi.org/10.1787/9789264261488-en.
OECD (2016), “The NEET challenge: What can be done for jobless and disengaged youth?”, in [130]
Society at a Glance 2016: OECD Social Indicators, OECD Publishing, Paris,
https://dx.doi.org/10.1787/soc_glance-2016-4-en.
OECD (2014), “The crisis and its aftermath: A stress test for societies and for social policies”, in [87]
Society at a Glance 2014: OECD Social Indicators, OECD Publishing, Paris,
https://dx.doi.org/10.1787/soc_glance-2014-5-en.
OECD (2013), The OECD Action Plan for Youth: Giving Youth a Better Start in the Labour [132]
Market, Meeting of the OECD Council at Ministerial Level, 29-30 May 2013,
http://www.oecd.org/employment/Action-plan-youth.pdf.
OECD (2012), Closing the Gender Gap: Act Now, OECD Publishing, Paris, [45]
https://dx.doi.org/10.1787/9789264179370-en.
OECD (2010), OECD Employment Outlook 2010: Moving beyond the Jobs Crisis, OECD [82]
Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2010-en.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
102
OECD (2010), Sickness, Disability and Work: Breaking the Barriers: A Synthesis of Findings [115]
across OECD Countries, OECD Publishing, Paris,
https://dx.doi.org/10.1787/9789264088856-en.
OECD (2009), OECD Employment Outlook 2009: Tackling the Jobs Crisis, OECD Publishing, [21]
Paris, https://dx.doi.org/10.1787/empl_outlook-2009-en.
OECD (forthcoming), “COVID-19 and the Housing Trilemma”, OECD Policy Responses to [107]
Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/en/policy-responses.
OECD (forthcoming), “Job retention schemes during the COVID-19 lockdown and beyond”, [20]
OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/en/policy-responses.
OECD (forthcoming), “Mitigating the rise in (long-term) unemployment: what role for active labour [116]
market policies and public employment services?”, OECD Policy Responses to Coronavirus
(COVID-19), OECD Publishing, Paris, https://www.oecd.org/coronavirus/en/policy-responses.
OECD (forthcoming), “Skill measures to mobilize the workforce during the COVID-19 crisis”, [129]
OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/en/policy-responses.
OECD (forthcoming), “The potential of online learning: Early lessons from the COVID-19 crisis”, [109]
OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris,
https://www.oecd.org/coronavirus/en/policy-responses.
Office for National Statistics (2020), Coronavirus and the latest indicators for the UK economy [94]
and society,
https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddi
seases/bulletins/coronavirustheukeconomyandsocietyfasterindicators/21may2020#universal-
credit (accessed on 27 May 2020).
Painter, M. and T. Qiu (2020), “Political Beliefs affect Compliance with COVID-19 Social [13]
Distancing Orders”, SSRN Electronic Journal, http://dx.doi.org/10.2139/ssrn.3569098.
Palme, M. and M. Persson (2020), “Sick Pay Insurance and Sickness Absence: Some European [76]
Cross-Country Observations and a Review of Previous Research”, Journal of Economic
Surveys, Vol. 34/1, pp. 85-108, http://dx.doi.org/10.1111/joes.12351.
Pichler, S., K. Wen and N. Ziebarth (2020), “Positive Health Externalities of Mandating Paid Sick [73]
Leave”, Unpublished manuscript.
Pichler, S. and N. Ziebarth (2017), “The pros and cons of sick pay schemes: Testing for [72]
contagious presenteeism and noncontagious absenteeism behavior”, Journal of Public
Economics, Vol. 156, pp. 14-33, http://dx.doi.org/10.1016/j.jpubeco.2017.07.003.
Rothstein, J. (2011), Unemployment Insurance and Job Search in the Great Recession, National [120]
Bureau of Economic Research, Cambridge, MA, http://dx.doi.org/10.3386/w17534.
Rugaber, C. (2020), “Gig workers and self-employed keep waiting for jobless aid”, Associated [93]
Press, https://apnews.com/e17fdbbd8a49169e9a5d131d961d472c (accessed on
13 May 2020).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
103
Sahin, A., J. Song and B. Hobijn (2010), “The Unemployment Gender Gap During the 2007 [44]
Recession”, SSRN Electronic Journal, http://dx.doi.org/10.2139/ssrn.1582525.
Schmieder, J., T. von Wachter and S. Bender (2012), “The Effects of Extended Unemployment [119]
Insurance Over the Business Cycle: Evidence from Regression Discontinuity Estimates Over
20 Years”, The Quarterly Journal of Economics, Vol. 127/2, pp. 701-752,
http://dx.doi.org/10.1093/qje/qjs010.
Schmillen, A. and M. Umkehrer (2017), “The scars of youth: Effects of early-career [43]
unemployment on future unemployment experience”, International Labour Review,
Vol. 156/3-4, pp. 465-494, http://dx.doi.org/10.1111/ilr.12079.
Schneebaum, A. et al. (2018), “The Gender Wealth Gap Across European Countries”, Review of [55]
Income and Wealth, doi: 10.1111/roiw.12281, pp. 295-331,
http://dx.doi.org/10.1111/roiw.12281.
Sierminska, E., J. Frick and M. Grabka (2010), “Examining the gender wealth gap”, Oxford [54]
Economic Papers, Vol. 62/4, pp. 669-690, http://dx.doi.org/10.1093/oep/gpq007.
Statistics Canada (2020), Housing characteristics and staying at home during the COVID-19 [105]
pandemic, https://www150.statcan.gc.ca/n1/pub/45-28-0001/2020001/article/00009-eng.htm
(accessed on 21 May 2020).
Statistics Canada (2020), The Daily — Labour Force Survey, April 2020, [17]
https://www150.statcan.gc.ca/n1/daily-quotidien/200508/dq200508a-eng.htm (accessed on
21 May 2020).
Stearns, J. and C. White (2018), “Can paid sick leave mandates reduce leave-taking?”, Labour [74]
Economics, Vol. 51, pp. 227-246, http://dx.doi.org/10.1016/j.labeco.2018.01.002.
Tian, L. et al. (2020), Calibrated Intervention and Containment of the COVID-19 Pandemic, [112]
http://arxiv.org/abs/2003.07353 (accessed on 21 May 2020).
TUAC (2020), Covid19 crisis: Mapping out trade union and social partners’ responses, [63]
https://tuac.org/news/covid19-crisis-mapping-out-trade-union-and-social-partners-responses/
(accessed on 21 May 2020).
U.S. Bureau of Labor Statistics (2020), Frequently asked questions: The impact of the [18]
coronavirus (COVID-19) pandemic on the Employment Situation for April 2020,
https://www.bls.gov/bls/employment-situation-covid19-faq-april-2020.htm#ques1 (accessed
on 21 May 2020).
U.S. Bureau of Labor Statistics (2019), Paid Leave Benefits, [149]
https://www.bls.gov/ncs/ebs/benefits/2019/benefits_leave.htm.
UNDP (2015), Assessing Sexual and Gender Based Violence during the Ebola Crisis in Sierra [56]
Leone, UNDP,
https://www.sl.undp.org/content/sierraleone/en/home/library/crisis_prevention_and_recovery/
assessing-sexual-and-gender-based-violence-during-the-ebola-cris.html (accessed on
24 March 2020).
UNESCO (2020), COVID-19 Education Disruption and Response, [77]
https://en.unesco.org/covid19/educationresponse (accessed on 11 May 2020).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
104
Veneto Lavoro (2020), Emergenza COVID-19: L’Impatto sul lavoro dipendente in Veneto (23 [36]
febbraio-17 maggio 2020), Veneto Lavoro, Venice,
https://www.venetolavoro.it/documents/10180/1693590/Misure%2093_Covid-19.pdf
(accessed on 22 May 2020).
Von Gaudecker, H. et al. (2020), “Labour Supply in the Early Stages of the COVID-19 Pandemic: [31]
Empirical Evidence on Hours, Home Office, and Expectations”, IZA Discussion Papers No.
13158.
Waters, T., H. Miller and S. Adam (2020), Income protection for the self-employed and [154]
employees during the coronavirus crisis, Institute for Fiscal Studies,
http://dx.doi.org/10.1920/bn.ifs.2020.bn0277.
WHO (2010), What is a pandemic?, [134]
https://www.who.int/csr/disease/swineflu/frequently_asked_questions/pandemic/en/
(accessed on 4 May 2020).
World Health Organization (2018), WHO housing and health guidelines., World Health [101]
Organization, Geneva.
Zipperer, B. and E. Gould (2020), Unemployment filing failures: New survey confirms that [92]
millions of jobless were unable to file an unemployment insurance claim, Economic Policy
Institute Working Economics Blog, https://www.epi.org/blog/unemployment-filing-failures-new-
survey-confirms-that-millions-of-jobless-were-unable-to-file-an-unemployment-insurance-
claim/ (accessed on 13 May 2020).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
105
Annex 1.A. Additional material
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
106
Annex Figure 1.A.1. Individual mobility fell in all OECD countries, even where restrictions were
relatively milder
Unweighted percentage change in mobility relative to the median value during the 5-week period 3 Jan – 6 Feb,
2020
Places of work Public transportation
% Australia % Austria
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Belgium Canada
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Chile Colombia
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Czech Republic Denmark
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Estonia Finland
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
107
Places of work Public transportation
% France % Greece
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Hunagary Ireland
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Israel Japan
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Latvia Lithunia
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Luxembourg Mexico
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
108
Places of work Public transportation
% Netherlands % Norway
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Poland Portugal
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Slovak Republic Slovenia
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Spain Switzerland
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Turkey United Kingdom
% %
40 40
20 20
0 0
-20 -20
-40 -40
-60 -60
-80 -80
-100 -100
Source: Google LLC “Google COVID-19 Community Mobility Reports”, https://www.google.com/covid19/mobility/ (accessed: 8 June 2020).
StatLink https://stat.link/koy3ih
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
109
Annex Table 1.A.1. Projected labour market developments in OECD countries
A. Real GDP growth B. Employment growth C. Unemployment rate
Percentage change from previous Percentage change from previous Percentage of total labour force
period period
2019 Projections 2019 Projections 2019 Projections
Single-hit Double-hit Single-hit Double-hit Single-hit Double-hit
scenario scenario scenario scenario scenario scenario
2020 2021 2020 2021 2020 2021 2020 2021 2020 2021 2020 2021
OECD 1.7 -7.5 4.8 -9.3 2.2 1.0 -4.1 1.6 -5.0 0.3 5.4 9.2 8.1 10.0 9.9
Euro area 1.3 -9.1 6.5 -11.5 3.5 1.2 -2.6 0.9 -3.2 -0.6 7.6 9.8 9.5 10.3 11.0
Australia 1.8 -5.0 4.1 -6.3 1.0 2.3 -2.4 0.9 -3.0 -0.6 5.2 7.4 7.6 7.6 8.8
Austria 1.5 -6.2 4.0 -7.5 3.2 0.8 -0.6 1.4 -0.9 1.1 4.5 5.8 5.2 6.0 5.7
Belgium 1.4 -8.9 6.4 -11.2 3.4 1.6 -1.4 1.6 -2.2 -1.0 5.4 7.4 6.5 8.2 9.3
Canada 1.7 -8.0 3.9 -9.4 1.5 2.1 -4.4 1.6 -5.1 0.7 5.7 8.9 8.0 9.4 9.0
Chile 1.0 -5.6 3.4 -7.1 1.9 .. .. .. .. .. 7.2 9.5 8.7 10.1 11.0
Colombia 3.3 -6.1 4.3 -7.9 2.8 -0.8 -10.6 2.2 -12.6 -0.7 10.5 18.0 16.0 19.8 19.9
Czech Rep. 2.5 -9.6 7.1 -13.2 1.7 0.2 -1.7 -0.1 -2.0 -1.0 2.0 3.5 3.8 3.8 5.0
Denmark 2.4 -5.8 3.7 -7.1 0.9 1.5 -2.4 0.5 -2.8 -2.5 5.0 6.6 6.5 7.0 9.1
Estonia 4.4 -8.4 4.3 -10.0 1.6 1.0 -4.2 1.1 -5.2 -1.5 4.4 9.2 8.1 10.1 11.2
Finland 0.9 -7.9 3.7 -9.2 2.4 1.0 -2.0 0.9 -2.6 -0.9 6.7 8.7 8.5 9.1 10.3
France 1.5 -11.4 7.7 -14.1 5.2 0.4 -2.7 1.8 -3.1 0.5 8.4 11.0 9.8 11.3 11.2
Germany 0.6 -6.6 5.8 -8.8 1.7 1.1 -0.8 0.0 -0.9 -1.0 3.2 4.5 4.3 4.6 5.3
Greece 1.9 -8.0 4.5 -9.8 2.3 2.2 -3.5 -1.0 -3.8 -1.8 17.3 19.4 19.8 19.6 20.4
Hungary 4.9 -8.0 4.6 -10.0 1.5 0.9 -3.2 1.8 -3.8 1.0 3.4 6.3 4.9 6.9 6.2
Iceland 1.9 -9.9 4.6 -11.2 3.0 1.3 -3.3 2.4 -3.8 1.0 3.5 7.4 6.0 7.8 7.7
Ireland 5.5 -6.8 4.8 -8.7 -0.2 2.9 -6.7 2.0 -8.2 -1.7 4.9 10.8 8.5 12.3 12.9
Israel 3.5 -6.2 5.7 -8.3 2.6 1.6 -3.2 2.4 -3.8 0.5 3.8 7.5 6.6 8.0 8.8
Italy 0.3 -11.3 7.7 -14.0 5.3 0.6 -2.9 -0.4 -3.2 -1.5 9.9 10.1 11.7 10.1 11.9
Japan 0.7 -6.0 2.1 -7.3 -0.5 0.9 -1.0 0.1 -1.2 -0.8 2.4 3.2 3.2 3.4 3.9
Korea 2.0 -1.2 3.1 -2.5 1.4 1.1 -0.6 0.5 -0.7 0.1 3.8 4.5 4.4 4.6 4.7
Latvia 2.2 -8.1 6.3 -10.2 2.0 0.2 -2.9 -0.7 -3.3 -3.0 6.3 9.2 9.3 9.6 11.7
Lithuania 3.9 -8.1 6.4 -10.4 3.4 0.3 -2.6 0.9 -3.2 0.6 6.3 9.1 8.2 9.5 8.7
Luxembourg 2.3 -6.5 3.9 -7.7 0.2 2.7 0.4 1.6 0.2 0.5 5.4 7.1 7.5 7.2 8.6
Mexico -0.1 -7.5 3.0 -8.6 2.0 2.4 -0.7 1.6 -1.0 1.8 3.5 6.0 5.8 6.3 6.0
Netherlands 1.8 -8.0 6.6 -10.0 3.4 2.0 -3.6 3.3 -4.4 1.1 3.4 5.9 4.9 6.5 6.6
New Zealand 2.2 -8.9 6.6 -10.0 3.6 1.1 -2.8 1.4 -3.0 -0.1 4.1 7.9 7.2 8.1 8.9
Norway 1.2 -6.0 4.7 -7.5 1.3 1.1 -2.0 1.6 -2.5 0.8 3.7 5.9 4.6 6.3 5.6
Poland 4.1 -7.4 4.8 -9.5 2.4 -0.1 -4.6 1.6 -5.3 -1.2 3.3 7.3 5.8 7.9 8.8
Portugal 2.2 -9.4 6.3 -11.3 4.8 1.0 -5.7 2.9 -7.1 2.0 6.5 11.6 9.6 13.0 11.8
Slovak Rep. 2.4 -9.3 6.4 -11.1 2.1 0.9 -3.0 1.7 -3.9 -0.1 5.8 8.9 7.0 9.6 9.2
Slovenia 2.4 -7.8 4.5 -9.1 1.5 0.2 -1.4 1.2 -2.0 -1.0 4.4 6.4 5.4 6.9 8.1
Spain 2.0 -11.1 7.5 -14.4 5.0 2.3 -5.3 1.1 -6.4 -1.6 14.1 19.2 18.7 20.1 21.9
Sweden 1.2 -6.7 1.7 -7.8 0.4 0.6 -3.2 0.3 -3.9 -0.4 6.8 10.0 10.0 10.6 11.1
Switzerland 1.0 -7.7 5.7 -10.0 2.3 0.7 -1.6 -0.3 -1.6 -0.5 4.4 5.7 6.4 5.7 6.6
Turkey 0.9 -4.8 4.3 -8.1 2.0 -2.2 -3.7 2.7 -5.0 0.8 13.7 15.6 14.2 16.8 16.1
United Kingdom 1.4 -11.5 9.0 -14.0 5.0 1.1 -4.6 2.1 -5.9 1.0 3.8 9.1 7.8 10.4 10.0
United States 2.3 -7.3 4.1 -8.5 1.9 1.1 -8.1 3.1 -9.8 1.4 3.7 11.3 8.5 12.9 11.5
Note:..: not available.
Source: OECD (2020), “OECD Economic Outlook – All editions”, OECD Economic Outlook: Statistics and Projections (database),
https://doi.org/10.1787/826234be-en (accessed on 10 June 2020).
StatLink 2 https://stat.link/vhre3u
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
110
Notes
1
The latest info can be found here https://coronavirus.jhu.edu/map.html.
2
A pandemic is linked to the geographical spread of a new disease, not its severity. According to the WHO
(2010[134]), “a pandemic is the worldwide spread of a new disease. An influenza pandemic occurs when a
new influenza virus emerges and spreads around the world, and most people do not have immunity.”
According to the United States Centers for Disease Control and Prevention (2012[135]) “a pandemic refers
to an epidemic that has spread over several countries or continents, usually affecting a large number of
people.” Since the historic “Spanish Influenza” of 1918, the world has witnessed six pandemics: the “Asian
flu” of 1957, the “Hong Kong flu” of 1968, the Severe Acute Respiratory Syndrome (SARS) in 2002, the
N1H1 influenza in 2009 (“bird flu”), the Middle East Respiratory Syndrome (MERS) in 2012, and Ebola
which peaked in 2013-14.
3
Containment strategies aim at minimising the risk of transmission from infected to non-infected people in
order to stop the outbreak – i.e. reducing the reproduction number to below one (OECD, 2020[1]). This
included actions to detect cases early on and trace an infected individual’s contact with other people, or
the confinement of affected people. Mitigation strategies, which include physical-distancing, including a full
society “lockdown”, and improved personal and environmental hygiene, aim at slowing the disease, and,
where the disease has occurred, to lessen its impact or to reduce the peak in health care demand –
i.e. getting the reproduction number as close as possible to, or below, one. In practice, containment and
mitigation actions largely overlap and are often implemented concurrently. In fact, containment and
mitigation policies may even be considered as a continuum with gradual increments of the same strategy;
with mitigation that could go to the extreme level of a full lockdown of a city, region or country.
4
Google released aggregated, anonymised data to chart movement trends over time by geography, across
different high-level categories of places such as retail and recreation, groceries and pharmacies, parks,
transit stations, workplaces, and residential areas.
5
The U.S. Bureau of Labor Statistics defines workers as unemployed on temporary layoff if they have
either been given a date to return to work by their employer or expect to be recalled to their job within
6 months (U.S. Bureau of Labor Statistics, 2020[18]). The term is sometimes used interchangeably with
being “furloughed”. The U.S. Bureau of Labor Statistics does not use the term “furlough” in its household
and establishment surveys, but includes furloughed workers among those on temporary layoffs.
6
The United States also put in place the Paycheck Protection Program (PPP) to provide small businesses
with loans to maintain employment levels (see Section 1.3.2 and Box 1.6 for a more detailed discussion).
7
If they have not been given a date to return to work by their employer and if they have no expectation to
return to work within six months, they need to fulfil the “job search” criteria to be classified as “unemployed”.
8
Based on these estimates of actual use for France and Germany, the total number of persons
participating in job retention schemes across the OECD would be about 50 million. Yet, actual use in
France and Germany is computed here on a shorter period of time (the month of May) than approved
applications.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
111
9
The analysis leverages information from online job vacancies as collected by two private companies:
Indeed (Figure 1.10) and Burning Glass Technologies (Figure 1.11). Indeed is a large job postings search
engine aggregating information from thousands of websites including firms’ career websites and job
boards. Burning Glass Technologies is an employment analytics company sourcing and coding job
postings from hundreds of millions of job postings to provide insight into labour market patterns. Burning
Glass Technologies data for the United States have been shown to align well with official data from the
U.S. Job Openings and Labor Turnover Survey, e.g. in Carnevale et al. (2014[136]), Hershbein and Kahn
(2018[137]), and Kahn, Lange and Wiczer (2020[14]). Knutsson et al. (2020[139]) further show that the regional
distribution of Burning Glass Technologies data for Australia, Canada and United States is generally well
aligned with official data for the most recent years. To the best of the authors’ knowledge, similar exercises
have not been performed on Indeed data, which should be therefore interpreted with greater caution.
Nevertheless, for the five countries for which both Burning Glass Technologies and Indeed data are
available, the aggregate trends shown by both sources of data are similar. The exact data used in this
publication were not benchmarked on official job vacancy data, where already available, and may therefore
be misaligned with those. Misalignments may result from the difference between the overall vs online-only
market for job postings, and from the data collection technology by Burning Glass Technologies and
Indeed, among other factors.
10
A first empirical exploration also finds a positive correlation between changes in mobility and in job
postings between February and April, similarly to Hensvik et al. (2020[10]). The simultaneous occurrence
of the two phenomena, however, makes it impossible to identify causal links at present.
11
The label “public services” is attributed for convenience and covers services that can be supplied by
private entities, as long as they fall in the education, health care and social work, or public administration
and defence sectors.
12
The definition of low-, medium- and high-skilled occupations is sourced from Chapter 4, which extends
Goos et al. (2014[138]).
13
Regions are defined as large subnational regions or as Territorial Level 2 (TL2) regions according to the
OECD classification. TL2 regions generally represent the first government layer after the national or federal
level.
14
If recent pandemics are any guide, the toll on poorer and vulnerable segments of society will be very
high. An analysis of the consequences of SARS (2003), H1N1 (2009), MERS (2012), Ebola (2014) and
Zika (2016) by Furceri et al. (2020[159]) shows that recent epidemics have led to an increase in income
inequality and hurt employment prospects of those with only a basic education while scarcely affecting
employment of people with advanced degrees.
15
Average hourly wages in April rose by 10.8% year on year in Canada because of the relatively larger
employment declines in low-paying industries, notably in accommodation and food services and in
wholesale and retail trade. In the United States, in April, average hourly earnings increased also well above
the recent average, reflecting the substantial job loss among lower-paid workers. More granular data will
be necessary to estimate the wage effect of the crisis conditional on job type.
16
The first group includes those in professional jobs able to work from home and those in industries with
less human-facing contact. The second group includes those with precarious employment at the bottom of
the income distribution but who are potentially well covered by Universal Credit.
17
Indicators based on expectations need to be interpreted with care in the current crisis. They are usually
a leading indicator of the outlook ahead. However, uncertainty surrounding the duration of lockdown
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
112
measures has complicated the ability of these data to provide those forward-looking signals and in such a
situation they represent more a coincident rather than a leading indicator (OECD, 2020[140]). The magnitude
of expectations decline should not be regarded as a measure of the degree of contraction in economic
activity, but rather as an indication of the signal strength.
18
Current data on spending are subject to frequent revisions and adjustments as not all countries provide
precise estimates, in particular for measures such as tax deferrals. Moreover, some policies are still being
rolled out and initial estimates may vary depending on the take-up and the actual duration. Also, in some
countries, loans to firms may eventually turn into grants, and guarantees on loans may be activated and
have a budgetary impact that cannot be foreseen at this stage.
19
Workplace physical distancing measures, such as working from home and workplace closures, can
reduce the disease attack rate by between 23-73%, with lower values for highly infectious diseases and
where there is lower compliance (OECD, 2020[1]).
20
Ample empirical evidence exists that working from home or space flexibility increase workers’ effort and
motivation (Beckmann, Cornelissen and Kräkel, 2017[146]), and job satisfaction (Bloom et al., 2014[158]; Kröll
and Nüesch, 2019[148]). These gains, however, partially rely on workers’ ability to choose whether to work
from home or from the office, which is not possible in case of a pandemic. Some workers are indeed found
to perform worse at home than in the office and to experience loneliness (Bloom et al., 2014[158]). To be
effective, they require adequate equipment and a proper space to work and no concurrent care duties such
as during the COVID-19 crisis. Moreover, workers may not be keen on the new flexible arrangements if
they are associated with a large pay cut (Mas and Pallais, 2020[142]), or if their tasks may be substituted by
software or by “telemigrants”, i.e. equivalent workers sitting abroad where labour costs are lower (Baldwin,
2019[155]). Employers, conversely, may be wary that workers reduce their effort while working from home
(a fact for which the economic literature has not found empirical backing, e.g. Beckmann (2015[147])), and
offer lower wages as a consequence. Evidence on the effects of telework on productivity is also mixed,
with some studies finding a positive effect (Angelici and Profeta, 2020[143]; Bloom et al., 2014[158]) and
others a negative or mixed one (Battiston, Blanes i Vidal and Kirchmaier, 2017[144]; Glenn Dutcher,
2012[145]).
21
Certain groups of workers in non-standard dependent employment, such as casual workers or workers
on zero-hour contracts, often have poor or no access to paid sick leave. For instance, casual workers in
Australia (about a quarter of all employees) are not eligible to sick pay and zero-hour contract workers in
the Netherlands (about 7% of all employees) only for hours they were called upon by their employer. These
two groups did not obtain better access to sick pay, although casual workers in Australia who meet the
residence requirements can temporarily access special unemployment benefits in case of sickness from
COVID-19 or mandatory quarantine.
22
There exist subnational requirements for paid sick leave in the United States. In 2019, a quarter of
US workers did not have access to paid sick leave at all (rising to one half for low-wage workers), and
two-thirds of them had less than ten days of paid sick leave per year (U.S. Bureau of Labor Statistics,
2019[149]).
23
Workers in other hybrid forms of self-employed work, such as freelancers and gig workers, lack access
to sickness benefits even more often, and informal workers are not covered by definition (Eurofound,
2020[141]). A few countries have taken initiatives to extend sickness benefits to these workers in case of
sickness due to COVID-19. For instance, gig workers in Canada and the United States are now temporarily
covered under certain conditions. Colombia introduced a COVID-19 specific flat-rate benefit for low-wage
informal workers.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
113
24
For instance, sickness benefit coverage may be mandatory for self-employed workers only if they have
incomes above a certain threshold.
25
Weekly or monthly administrative data from national social-insurance authorities for Austria, Chile, the
Czech Republic, Finland, Germany, Italy, Latvia, Portugal and Sweden, or data from special employer
surveys in France and the United Kingdom. For more detail, see OECD (2020[67]).
26
In Canada, the right to Employment Insurance Caregiving leave and benefits applies only in cases of
critical illness or injury or someone in need of end-of-life care. The right covers workers who need to provide
this type of care for family members of others who are considered to be like a family member.
27
Japan, rather than to establish a statutory right to special paid leave as such, has introduced a subsidy
for employers that allow their workers to take paid leave due to school or childcare facility closure.
Employers are compensated for the continued payment of salaries while workers are on leave, up to a limit
of JPY 8 330 per worker per day.
28
This includes Canada, France, Italy, Japan, Luxembourg, New Zealand, Norway, Portugal, Switzerland
and the United Kingdom.
29
In some countries, such as Denmark, these extensions build on a tripartite agreement between the
government, trade unions and employers.
30
This also implies that there is no risk that firms continue claiming benefits even once hours have been
restored.
31
The very large majority of OECD countries have encouraged teleworking but, in practice, the decision
has been left to employers in several countries. Employees may, at least on paper, risk disciplinary
measures or even dismissal if they do not show up for work out of sanitary concerns.
32Italy has since significantly expanded minimum-income provisions, in 2018 and 2019, and introduced a
number of changes to the unemployment benefit system in 2015.
33
“Net replacement rate” for a single, childless person with previous earnings of two-thirds of the national
average wage in the third month of unemployment (OECD tax-benefit models, http://oe.cd/taxben).
34
For instance, Austria, Canada, France and Spain have extended entitlements to unemployment benefits
to independent workers. Denmark has strengthened the portability of earned entitlements across different
jobs and forms of employment. Italy has facilitated access to means-tested safety-net benefits.
35
In addition, automatic benefit extensions can be available at the State level if unemployment in that State
exceeds the federally prescribed trigger level.
36
In the United States, a court ruling temporarily suspended the tightening of access to the federal
Supplemental Nutrition Assistance Program (SNAP, previously “food stamps”) that had initially been
foreseen for April 2020.
37
Receipt trends for the Universal Credit and the Reddito di Cittadinanza cannot be directly compared
because the former measures daily “inflows” and the latter monthly “stocks”.
38
Newly self-employed workers who only started their business in 2020 and therefore cannot prove their
income with a tax declaration will receive a flat rate payment of EUR 500 per month.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
114
39
These payments are meant to cover three months of business operating costs such as rent, wages of
employees not covered by short-time work schemes etc.; for their own living costs, self-employed workers
will have to rely on the means-tested Unemployment Benefit II, eligibility to which has been temporarily
relaxed.
40
Personal information from undocumented workers will not be required. Officials estimate that 150 000
undocumented immigrants in the state will benefit.
41
Above the income threshold of USD 75 000 per year, payments are reduced by USD 5 for every
USD 100 of additional income earned. People classified as “dependent” on another household member’s
tax return are excluded from this transfer. This includes many students over the age of 17 and some
disabled people living with family members.
42
Herd immunity is a form of indirect protection from infectious disease that occurs when a large
percentage of a population has become immune to an infection, whether through vaccination or previous
infections, thereby providing a measure of protection for individuals who are not immune.
43
See, among many, Dingel and Neiman (2020[151]), Espinoza and Reznikova (2020[156]), Gottlieb,
Grobovšek and Poschke (2020[157]); Hensvik, Le Barbanchon and Rathelot (2020[10]), Mongey, Philossoph
and Weinbger (2020[152]).
44
See brief summary of the literature in endnote 20.
45
The OSHA keeps a daily record of complaints, referrals and closed cases related to COVID-19 as well
as the number of whistle-blower complaints filed. See US OSHA
https://www.osha.gov/enforcement/COVID-19-data and https://www.whistleblowers.gov/COVID-19-data.
46
This depends largely on whether governments deem widespread childcare facility and school closures
necessary on public health grounds. Given the apparent low infection rates among children – see OECD
(2020[106]) and Mallapaty (2020[153]) – many countries will likely want to consider options for re-opening
and/or keeping schools and childcare centres open during a possible second wave, especially for younger
children, also in view of the budgetary costs of funding paid care leave and the risk of lost educational
opportunities for children. However, this needs to be done with care. The evidence on whether or not
children have a lower risk of transmitting the infection is still inconclusive (Mallapaty, 2020[153]). If, after
further study, it is established that children carry similar transmission risks as adults, re-opening schools
and childcare facilities – and potentially leaving them open during a possible second wave – could
contribute to heightened infection.
47
The type of support may nonetheless depend on the timing of expected re-opening, as activities that are
potentially viable now may turn unviable with a prolonged shutdown.
48
In such schemes, the cost of labour hoarding falls entirely on the government or workers in the form of
uncompensated reductions in working time. In principle, firms can be made to share in the cost of labour
hoarding by placing limits on the extent to which uncompensated reductions in working time are possible.
In New Zealand, total worker earnings in subsidised jobs in principle cannot decline more than 80% of
normal earnings. However, it is not clear how binding this requirement is in practice. The wage subsidy
scheme operated in the Netherlands mimics STW schemes that require firms to share the cost of labour
hoarding. While workers continue to receive 100% of their earnings, employers are not fully compensated
for the loss in revenue. This may induce some employers to request support only for workers whose jobs
are viable in the longer term.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
115
49
Evidence from Switzerland (Kopp and Siegenthaler, 2019[150]) during the global financial crisis shows
that workers in viable jobs tended to leave the scheme before the maximum duration, while those firms
who did use the scheme for the maximum duration tended to layoff some workers eventually.
50
For example, over 100% for some claimants in the Self-employment Income Support Scheme in the
United Kingdom (Waters, Miller and Adam, 2020[154]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
116
2 Unemployment benefits and
non-standard dependent
employment: Striking the balance
between income security and work
incentives
This chapter provides an in-depth discussion of the income security and
work incentives provided by unemployment benefits to jobseekers with
previous periods of part-time and unstable dependent employment. It sheds
light on the accessibility and adequacy of this key social protection tool, and
on the work incentives affecting different types of workers, two key factors
for its design and implementation. In particular, the chapter compares
entitlements to unemployment benefits for workers with a range of typical
employment trajectories, including alternating spells of dependent
employment and unemployment. Issues, such as the extension of
out-of-work support to individuals in “part-time” or “partial” unemployment,
options for accumulating entitlement rights across different spells of
employment, saving “unused” benefit entitlements for future out-of-work
spells, and strategies for integrating in-work and out-of-work support, are all
assessed.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
117
In Brief
Key findings
This chapter provides an in-depth discussion of how unemployment benefits provide income security
and work incentives for jobseekers whose recent work history included periods of non-standard
dependent employment. Non-standard dependent employment refers to wage and salary workers
working either on a part-time or on an unstable basis (i.e. involving frequent transitions between
dependent employment and unemployment over a number of years, see Box 2.1 for details). Typically,
entitlements to unemployment benefits depend on the characteristics of the job held before becoming
unemployed (e.g. number of months employed, wage, hours worked). Since such characteristics refer
to a period rather than just to the moment before becoming unemployed, this chapter assesses
non-standard dependent employment in terms of trajectories instead of the type of contract. This allows
for a more nuanced and sophisticated analysis of the interactions between unemployment protection
and non-standard employment.
Using longitudinal microdata, unemployment benefit legislation and simulations of entitlements under
several policy-relevant employment trajectories, the main findings of this chapter are:
Workers with non-standard dependent employment trajectories make up a sizeable part of
dependent employment (22% on average across 26 European OECD countries). On average
across European OECD countries, 69% of employees who have been unemployed have
histories of non-standard dependent employment (51% in the form of unstable employment and
18% as part-time employment).
Non-standard dependent employment, particularly part-time work, has increased in the last
decade, especially among the young. In the decade leading up to 2016-18, non-standard
dependent employment rose by 5 percentage points on average among employees aged 20 to
29. Most of the rise was due to an increase in part-time employment. Yet, there is considerable
variation across countries. In Spain, for example, the share of unstable dependent employment
among youth increased by 8 percentage points, in spite of the contraction of fixed-term contracts
(OECD, 2019[1]).
Women are more than two times more likely to work part time than men. On average, almost
one quarter of women – often mothers – work part-time, although figures differ widely across
OECD countries. Part-time employees tend to receive lower hourly wages in all OECD countries,
have higher job insecurity and participate less in training. Yet, in a number of countries, many
part-time employees work shorter hours by choice and as a way to achieve work-life balance.
Poverty rates for workers with non-standard dependent employment trajectories are higher than
for workers with standard employment. Across European OECD countries, on average, 10% of
workers in part-time employment and 19% of workers in unstable employment live in a
household with an annual disposable income of less than 50% of the national median. For
standard employees the figure is only 3%.
Unemployment benefit legislation directly affects the duration of benefit entitlements and the
level of benefits payable for workers with standard and non-standard dependent employment
trajectories. Required minimum employment durations, working hours or earnings are harder to
meet for those in unstable or part-time employment. Some countries operate second-tier
unemployment benefits with less demanding employment requirements, which are easier for
jobseekers with non-standard employment trajectories to access.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
118
Analysis of detailed rules of first- and second-tier unemployment benefits in 11 OECD countries
(Australia, Austria, Canada, Finland, France, Japan, Latvia, Netherlands, Poland, Spain and the
United Kingdom) finds that the same rules can affect workers with part-time and unstable
employment differently from standard employees.
Benefit claimants with a history of part-time employment usually have the amount of their
benefits calculated in line with the number of hours worked in their previous job. In Australia,
Poland and the United Kingdom, however, jobseekers previously in part-time employment
receive the same amount as those previously working full-time.
Many countries operate part-time unemployment schemes, which combine the entitlement to
unemployment benefit with some casual part-time work. In most countries for which the analysis
was undertaken, the amount of benefit is reduced in such cases, although there is a considerable
variation in the rules among countries. In Latvia and Poland, jobseekers receive no
unemployment benefit if they take up any amount of work.
Jobseekers can keep unused entitlements for future claims if they find full-time work before
benefits expire, in all analysed countries but in Latvia and Poland. Countries use a range of rules
concerning the possibility for jobseekers to keep and access unused entitlements. In Canada,
the Netherlands and the United Kingdom, strict conditions determine whether entitlements are
retained for subsequent unemployment spells. If the jobseeker with an unused entitlement
accrues a new one from more recent employment, she must choose between them in Spain,
while can utilise both in France – first using up the older entitlement and then claiming the new
one.
Simulations based on a set of typical employment trajectories for four OECD countries (Australia,
France, Latvia and Spain) show that benefit entitlements can vary substantially depending on
the type of employment and the specific sequence of in-work and out-of-work spells. Significant
differences exist even when workers have the same personal characteristics and same level of
earnings and working hours over a given period. These gaps call into question the adequacy of
benefit generosity for all types of workers.
In Australia and France, entitlements do not differ significantly between jobseekers with fairly
stable employment records (working for a full year before becoming unemployed) and
jobseekers with more unstable employment trajectories (working either every other month or
every other semester, but for the same total number of hours). In France, an upcoming reform
will reduce the benefit entitlements of workers with unstable employment records, as the
calculation will take into account the months in which she has not worked. In Latvia and Spain,
where contribution requirements are more demanding, jobseekers who previously had more
unstable employment records are entitled to fewer months of benefits.
In all four countries, entitlements are lower for workers on part-time unemployment. In Australia,
France and Spain, earnings from work reduce the benefit amount. In Latvia, benefits are
suspended as soon as the jobseeker has some earnings from employment.
Unemployment benefits also create different financial incentives to work across countries and
employment trajectories. On average, work incentives are weaker in France, followed by
Australia, Spain and Latvia. In Latvia and Spain, however, in some cases work incentives are
stronger because some employment trajectories do not lead to any entitlement to unemployment
benefits.
To address the main issues of entitlement and access, on the one hand, and work incentives,
on the other, for workers with non-standard employment trajectories, a number of policy tools
could be considered:
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
119
o Reductions in benefit payments when jobseekers have earnings from part-time unemployment
schemes help to smooth income variations and to prevent benefit misuse. However, they also
affect the neutrality of the system and reduce the incentives of using casual part-time work as
a stepping stone to a better job. Earnings disregards and withdrawal rates can help balance
these objectives. Earnings disregards are the amounts of earnings not considered when
calculating benefits during periods of partial work. The withdrawal rate is the percentage of
earnings (above any disregard) by which benefit is reduced in such cases.
o Easy access to benefits and the possibility of retaining unused benefits for future out-of-work
spells may encourage frequent periods of unstable employment or increase the overall
duration of unemployment benefits. Waiting periods, moderately long employment
requirements, reductions in case of frequent reclaiming, and limits on the accumulation of
old and new entitlements can help to prevent such distortions. However, employment
requirements, even if they provide work incentives, prevent benefit abuse and protect the
system’s financial sustainability, they also penalise workers with unstable employment
records. Customised extensions of reference periods for workers more prone to job
instability (e.g. young workers, employees with temporary contracts) could alleviate these
negative effects. Differentiated unemployment insurance contribution rates could also be
used to create financial incentives for employers and employees to choose more stable
employment contracts and discourage collusion.
o Work incentives can be enhanced through improved co-ordination of in-work and out-of-work
support. Potential measures include: extending to full-time workers with low wages the
possibility to cumulate earnings and unemployment benefits; making use of in-work benefits;
or integrating in- and out-of-work benefits.
o During periods of economic downturns, unemployment benefit instruments can be adjusted
to the changes in labour market circumstances. In the current COVID-19 crisis, many
countries extended maximum duration and generosity of unemployment benefits. A number
of countries also adopted measures to facilitate benefit access, such as reducing minimum
employment requirements or extending reference periods, which are likely to specifically
increase benefit coverage of non-standard workers during the emergency.
Introduction
Non-standard forms of dependent employment (i.e. jobs that are part time or of short duration) represent
a significant share of wage and salary workers. On average across OECD countries, part-time employment
accounts for 16.5% of all employment – 2 percentage points more than two decades ago (OECD, 2020[2]).
Temporary employment accounts for 11.7% of dependent employment and has remained somewhat
stagnant in recent years. Among young workers, however, temporary employment makes up 25.7% of
dependent employment – almost 2 percentage points more than 20 years ago (OECD, 2020[3]).
Non-standard forms of dependent employment have been associated with increasing labour market
instability, underemployment and economic vulnerability, particularly among young people and those with
less than tertiary education (OECD, 2019[1]).
As also stressed in the 2019 edition of the OECD Employment Outlook (OECD, 2019[1]), in many countries,
social protection systems in general, and unemployment benefits in particular, have not yet fully adapted
to the specific needs and circumstances of workers in non-standard forms of dependent employment.
Analyses of out-of-work social benefits in six countries (France, Greece, Hungary, Italy, Spain and the
United Kingdom) found that in Spain and Italy unstable employees were significantly less likely to receive
benefits than standard employees (i.e. those with full-time jobs and open-ended contracts). This gap was
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
120
also considerable in the United Kingdom. The social protection gap between standard and non-standard
employees was even larger when assessed in terms of benefit generosity (benefit amount as a proportion
of median income), especially in Greece, Italy and Spain (OECD, 2019[1]).
Due to their unusual work trajectories, jobseekers with previous non-standard dependent employment may
not receive the type of support that unemployment benefits offer to jobseekers with standard employment
careers. In fact, non-standard employees’ access to unemployment benefits tends to be more difficult than
to other insurance programmes such as maternity and sickness benefits (Avlijas, 2019[4]). Non-standard
dependent employment may also be one of the factors that explain why, in many OECD countries, only a
small share of jobseekers receive unemployment benefits – fewer than one-third on average across
32 countries (OECD, 2018[5])).
At the same time, depending on their design, unemployment benefits may contribute to job instability. For
example, they may provide incentives for alternating between short periods of employment and
unemployment (Boeri, Cahuc and Zylberberg, 2015[6]). In particular, this can arise in the absence of waiting
periods, overly short qualification periods for contributory benefits and ill-designed partial unemployment
insurance schemes (Kyyrä, 2010[7]; Le Barbanchon, 2016[8]; Fontaine and Malherbet, 2016[9]). Results in
OECD (2019[1]) illustrate that, in some circumstances, entitlements for those with patchy work histories
may be equally or more generous than for those with more stable employment records.
This chapter provides an in-depth review of the key policy mechanisms that affect how unemployment
benefits strike a balance between income security and financial work incentives for non-standard
employees. By providing comparable cross-country estimates of protection for jobseekers with different
employment trajectories, the chapter seeks to provide input into the policy debate on the accessibility,
adequacy and effectiveness of unemployment benefits for different types of employees.
Typically, entitlements to unemployment benefits depend on the characteristics (e.g. number of months
employed, wage, hours worked) of the job held before becoming unemployed. Since such characteristics
refer to a period rather than just to the moment before becoming unemployed, this chapter assesses non-
standard dependent employment in terms of trajectories instead of the type of contract (Box 2.1 explains
in detail how employment trajectories are defined and measured). Thus, employment is observed and
characterised over a continuous period, instead of at a single point in time. This approach provides a more
refined assessment of job instability and the possibility to measure the unemployment risk of each type of
employment. In turn, this allows for a more nuanced and sophisticated analysis of the interactions between
unemployment protection and non-standard employment.
The chapter is organised as follows. Section 2.1 assesses the scale and development of unstable and part-
time dependent employment and the characteristics of workers with such trajectories, comparing them with
workers with stable full-time dependent employment trajectories. Section 2.2 takes stock of the legal
provisions on unemployment benefits that may give rise to an uneven treatment of standard and non-
standard employees. Section 2.3 sheds light on the provision and level of unemployment benefits in a number
of non-standard dependent employment trajectories, by making use of simulation techniques. Furthermore,
it computes indicators of income security and financial work incentives for a range of policy-relevant scenarios
of standard, unstable or part-time employment. Section 2.4 discusses desirable features in the design of
unemployment benefits to provide reliable and effective support to workers with non-standard employment.
Box 2.1. Trajectories of non-standard dependent employment: definitions and measures
“Non-standard” employment is an umbrella term that typically covers all temporary, part-time and self-
employment arrangements, i.e. everything deviating from the “standard” of full-time, open-ended
employment with a single employer (OECD, 2014[10]; OECD, 2018[5]). This chapter analyses a slightly
different employment group: “non-standard dependent employment”.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
121
Definitions
Non-standard dependent employment refers to wage or salary workers who experience periods of part-
time or unstable work. Self-employed workers and persons not in the labour force are not included. The
OECD defines part-time employment as people in employment who usually work less than 30 hours per
week in their main job (OECD, 2020[2]). Unstable employment is defined here as a situation
characterised by frequent transitions between employment and unemployment over a number of years. 1
Measurement for the empirical analysis
The empirical analysis in this chapter is based on panel data of the European Union Statistics on Income
and Living Conditions (EU-SILC). Employment trajectories are measured through a continuous month-
to-month observation of the employment statuses of individuals over a period of 36 months.
The definition and measurement of unstable employment used here differs from previous OECD work,
which focussed on transitions from one year to the next (OECD, 2019[1]). In the context of the present
chapter, annual transitions may conceal one (or several) transitions into and out of employment that
occur over the course of the reference year.2 By measuring monthly transitions, the definition used here
is more likely to capture transitions that involve short periods of unemployment between jobs.
Workers in part-time dependent employment are people reporting to work for a wage or salary on a
part-time basis over most of the period of 36 months. The definition is self-reported and implausible
information is detected and corrected as far as possible (Eurostat, 2018[11]).
Workers in unstable dependent employment are people reporting to work for a wage or salary with
at least three transitions between (full-time or part-time) dependent employment and unemployment
over the 36-month period. The rationale for requiring a minimum of three transitions is to ensure that
the person experienced at least two spells of unemployment during the 36-month period. Workers
reporting at least three transitions are classified into unstable employment even if part-time work is the
most frequent employment status.3
Standard employment refers to workers whose dependent employment is stable (i.e. not unstable)
and work mainly on a full-time basis within three years.
Sample
The sample pools together rolling panel samples ending on the years 2016, 2017 and 2018 (and the
years 2006, 2007 and 2008) to minimise year-to-year variability in small subsamples. The sample is
restricted to individuals aged between 20 and 59 years old who declared to be in dependent
employment for at least one month in each of the three observed years. People reporting some months
not in the labour force (e.g. students or pensioners) are included only if they were active (i.e. employed
or unemployed) in most of the three-year period and were employed at least one month in each
calendar year. Individuals experiencing self-employment are only part of the sample if self-employment
is not the dominant status and there is no transition from self-employment to unemployment.
1. Employees with unstable jobs are closely associated with employees with temporary contracts (see Section 2.1.3). However, instead of being
defined based on the type of contract they are identified based on the actual instability of their employment trajectories. Thus, this definition may
include employees with permanent or long-term contracts who have frequent periods of unemployment and may exclude employees with temporary
contracts who remain employed either with the same or with a new employer. It is possible that workers with permanent contracts have different
underlying reasons and labour market pressures for changing jobs from workers with temporary contracts. However, regarding entitlements to
unemployment benefit, rules are more sensitive to work tenure than the type of contract. Furthermore, across OECD countries, there is wide variation
in the use of permanent and temporary contracts, partly reflecting differences in employment protection legislation (see Chapter 3).
2. On average across analysed countries, 1.4% of employees in the 2016-18 sample have employment trajectories that are both mostly part
time and unstable (i.e. with many transitions).
3. On average across analysed countries, 1.4% of employees in the 2016-18 sample have employment trajectories that are both mostly part
time and unstable (i.e. with many transitions).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
122
2.1. Trajectories of non-standard dependent employment
According to OECD data for 2018, part-time employment accounts for 16.5% of all employment –
2 percentage points more than two decades ago (OECD, 2020[2]). Temporary employment accounts for
11.7% of dependent employment and has remained somewhat stagnant. Among young workers, however,
it makes up 25.7% of dependent employment – almost 2 percentage points more than 20 years ago and
8 percentage points more than in 1980 (OECD, 2020[3])
While some non-standard jobs, particularly for high-skilled professionals, may pay high and stable earnings
and provide good working conditions, there is an association between non-standard work and poorer job
quality, particularly for workers with low and middle skills. Wages tend to be lower, employment less
protected, access to employer and social benefits reduced, safety and health risks greater, investments in
lifelong learning lower, and bargaining power weaker (OECD, 2014[10]; OECD, 2019[1]).
Non-standard jobs are also associated with higher job instability. In a context of job polarisation (see
Chapter 4), some workers (particularly those in middle-skill occupations) may not enjoy the positive
aspects of low job tenure, like upward transitions when former jobs disappear. Past OECD research
(OECD, 2013[12]) indicates that workers who lose their job involuntarily experience a fall in job quality after
re-employment. Whereas income losses affect most previously displaced workers, the magnitude of this
effect differs by country (lower in Northern European countries, higher in others), age (higher and more
persistent for older workers), skill level (higher for low-skilled workers) and gender (higher for men).
Furthermore, displaced workers are more likely to be re-employed in temporary or part-time jobs.
Evidence from the 2019 edition of the OECD Employment Outlook (OECD, 2019[1]) shows that job
instability has increased in most OECD countries once changes in the demographic composition of the
workforce are taken into account. The average job tenure decreased across the OECD by 4.9% (or around
five months) between 2006 and 2017. The largest declines in job stability occurred for low-educated
workers (i.e. those without an upper secondary qualification).
This section assesses the incidence, characteristics and trends of workers with trajectories of unstable and
part-time dependent employment (see Box 2.1 for detailed definitions), and compares them to those with
stable full-time dependent employment histories.
2.1.1. How prevalent are non-standard employment trajectories?
Workers with non-standard dependent employment trajectories make up a sizeable part of dependent
employment. On average across 26 European OECD countries for which data is available, non-standard
dependent employment accounts for 22% of employees, part-time employment comprises 16% and
unstable employment makes up 6%. Among employees who experienced spells of unemployment, past
unstable employment accounts for 51%, standard employment for 30% and part-time employment for 18%.
The shares of employees with trajectories of unstable and part-time dependent employment vary
considerably across countries. Unstable employment is lower than 5% of dependent employment in
Belgium, Czech Republic, Denmark, Estonia, Germany, Luxembourg, Norway, Slovak Republic and the
United Kingdom, while it amounts to more than 10% in Finland, Greece and Spain. Part-time employment
is low in Eastern European countries, Portugal and Finland, above 25% in Belgium, Germany and
Switzerland, and highest in the Netherlands (41%).
In most countries, previous patterns of non-standard dependent employment account for the bulk of
workers who experience unemployed, with unstable employment being the largest group. Workers with
past unstable employment make up at least half of the unemployed in Austria, Finland, Greece, Iceland,
Hungary, Spain, Sweden and the United Kingdom. The shares are lowest in Czech Republic and
Slovak Republic. The share of unemployed with previous part-time employment is generally low (less than
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
123
20%), except in Germany, Ireland and the Netherlands, where they make up at least 30%, and, to some
extent, in Belgium, France, Italy and Switzerland.
Figure 2.1. Distribution of employment trajectories across employees and unemployed
Employment trajectory types, 2016-18 (in percentage)
A. As a share of dependent employment
% A. Standard employment B. Part-time employment C. Unstable employment
100
90
80
70
60
50
40
30
20
10
0
B. As a share of unemployment 2
% A. Standard employment B. Part-time employment C. Unstable employment
100
90
80
70
60
50
40
30
20
10
0
Note: Data is from the longitudinal panel of the EU Statistics on Income and Living Conditions (EU-SILC). The sample pools together rolling
panel samples for the years 2016, 2017 and 2018 to minimise year-to-year variability in small subsamples. The sample is restricted to employees
between 20 and 59 years old. People not in the labour force (e.g. students or pensioners) are excluded. OECD: weighted average of listed
countries.
1. Employment trajectories are measured through a continuous month-to-month observation of the employment statuses of individuals over a
period. Unstable dependent employment comprises individuals with at least three transitions between employment and unemployment in
3 years. Standard employment refers to individuals whose employment is stable (i.e. not unstable) and work mainly on a full-time basis within
3 years. Part-time employment refers to individuals whose employment is stable and work mainly on a part-time basis within 3 years.
2. The unemployment subsample is defined as employees who have been unemployed for at least one month in the year. Employment trajectory
types in this case refer to employment trajectories before entering unemployment. The unemployment status is based on individual self-
classification of main activity status stated by respondent. This definition may not satisfy the ILO criteria for unemployment, which require active
job search and immediate availability.
Source: Longitudinal EU-SILC.
StatLink 2 https://stat.link/oj2rmw
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
124
2.1.2. More young workers are in non-standard dependent employment than a decade
ago
Non-standard dependent employment, particularly part-time work, has slightly increased in the last decade,
especially among the young. However, on average across European OECD countries between 2006-08
and 2016-18, the rise has been small and restricted to part-time employment (Figure 2.2, Panel A). These
patterns are partly confounded, however, by changes in the demographic composition of the labour force.
Population ageing and pension reforms (restricting early retirement and increasing the statutory retirement
age) have increased the proportion of older workers1, whose jobs are usually stable and full-time (OECD,
2019[1]).
Focusing on young employees (aged 20 to 29), non-standard dependent employment rose by
6 percentage points on average between 2006-08 and 2016-18 (Figure 2.2, Panel B). Again, most of the
rise took place among part-time employment, with unstable employment rising by just 1 percentage point.
Young unstable employment increased considerably in Denmark, Greece, Ireland, Slovenia and Spain. In
contrast, the share of young employees with unstable employment trajectories fell in the Netherlands and
Norway, where it was offset, in part, by increases in part-time employment. In fact, part-time employment
among youth increased in 20 out of 23 countries. The rise was notably high (more than 10 percentage
points) in Greece, Italy, Ireland and Spain, while only in Denmark the share of part-time jobs among young
employees fell significantly. These changes may be affected, in part, by the economic cycle, as the
reference period (2006-08) may partly reflect labour market consequences of the abrupt downturn
associated with the Great Recession.
2.1.3. Young workers on temporary jobs are more likely to have unstable employment
trajectories
Unstable employment trajectories are more prominent among workers who are single and young, have
less than upper secondary education, work in jobs with temporary contracts, are in low-skill occupations,
and earn less than the median monthly wage. The results in Table 2.1 indicate that workers on temporary
contracts are more likely to experience unstable employment. On average across countries shown, a
person with a temporary contract is 17% more likely to be in unstable employment than an employee with
a permanent contract. The association between temporary contracts and unstable employment is
particularly strong in Austria, Estonia, Finland, Greece and Spain.
Most workers with unstable employment have a temporary contract – 64% on average across European
OECD countries (Figure 2.3), but cross-country variation is considerable. Temporary contracts account for
the vast majority of workers with unstable employment in Poland and Southern European countries. On
the other hand, in the Baltic countries, Austria, Denmark, Iceland, Ireland, Norway, Switzerland and the
United Kingdom, most workers with unstable employment have permanent contracts (fewer than 30% have
temporary contracts), reflecting, in part, the low incidence of temporary contracts in employment and the
lighter protection of open-ended contracts against dismissal (see Chapter 3).
In most countries, unstable employment is also more likely among jobs in low- and middle-skilled
occupations, particularly in Austria, the Netherlands, Spain and Eastern European countries (Table 2.1).
Unstable employment is also more frequent among young people, especially in Northern European
countries as well as Hungary, France and Switzerland. In many countries, people with less than upper
secondary education are more likely to be in unstable employment (especially in Greece, Hungary, Latvia
and Poland), but in Denmark and Germany, the estimates suggest the opposite. Unstable employment is
more frequent among women than men in Belgium, Finland and Sweden; while the opposite holds in
Austria, Denmark and Germany. In most countries, people in couples (i.e. married or in partnership) are
less likely to be in unstable employment; this is particularly the case in the Denmark, Hungary and
Lithuania. The association between having children and unstable employment is ambiguous; it is positive
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
125
and significant in Germany, Luxemburg, Norway and Portugal but negative in the Czech Republic, the
Netherlands and Poland. Higher monthly wages (normalised to within-country median wage) are negatively
related to unstable employment in most countries, particularly in Austria, Denmark, Finland and
Luxembourg; however, these results must be treated with caution as they may be biased by the wage data
available.2
Figure 2.2. More workers experience trajectories of non-standard dependent employment
Percentage point change in the share of employees with standard, part-time and unstable employment trajectories,
2006-08 to 2016-18
A. Employees (aged 20 to 59)
A. Standard employment B. Part-time employment C. Unstable employment
30
20
10
0
-10
-20
-30
ESP GRC ITA SVN BEL AUT LVA LUX EST OECD HUN PRT FIN NLD GBR FRA IRL LTU SVK CZE DNK SWE POL NOR
B. Young employees (aged 20 to 29)
A. Standard employment B. Part-time employment C. Unstable employment
30
20
10
0
-10
-20
-30
ESP GRC ITA SVN BEL AUT LVA LUX EST OECD HUN PRT FIN NLD GBR FRA IRL LTU SVK CZE DNK SWE POL NOR
Note: Data is from the longitudinal panel of the EU Statistics on Income and Living Conditions (EU-SILC). To minimise year-to-year variability in
small subsamples, the sample pools together rolling panel samples. The reference sample includes panels from 2006-08 and the recent sample
panels from 2016-18. The sample is restricted to employees between 20 and 59 years old. People not in the labour force (e.g. students or
pensioners) are excluded. OECD: weighted average of listed countries. Employment trajectories: unstable dependent employment comprises
those individuals with at least three transitions between employment and unemployment in 3 years. Standard employment refers to those
individuals whose employment is stable (i.e. not unstable) and work mainly on a full-time basis within 3 years. Part-time employment refers to
those individuals whose employment is stable and work mainly on a part-time basis within 3 years.
Source: Longitudinal EU-SILC.
StatLink 2 https://stat.link/5igcfn
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
126
Table 2.1. Incidence of unstable employment, by socio-demographic characteristics and countries
ALL AUT BEL CHE CZE DEU DNK ESP EST FIN FRA GBR
Wage -0.003 -0.056*** -0.012** -0.006 -0.026*** -0.028*** -0.064*** -0.019*** -0.000 -0.070*** -0.003 0.003*
Gender (ref: male)
Female -0.002 -0.068*** 0.010** 0.002 0.003 -0.022*** -0.018** 0.001 0.006 0.019** -0.005 -0.000
Education (ref: middle)
Low 0.021*** 0.023* 0.016** -0.000 0.000 -0.012** -0.016** 0.011 0.014* 0.052** 0.006 -0.004
High -0.002 -0.000 -0.007 -0.005 0.004 0.003 0.044** -0.044*** -0.011*** -0.034*** -0.005 -0.003
Age (ref: 40s)
20s 0.020*** -0.004 0.014** 0.045*** 0.018*** 0.007 0.032*** -0.010 0.027*** 0.076*** 0.043*** 0.018***
30s 0.007*** -0.021** 0.019*** 0.037*** 0.013*** 0.013*** 0.047*** 0.008 0.006 0.023** 0.007 0.013**
50s -0.001 -0.010 0.001 -0.004 0.001 0.003 0.039*** 0.005 0.004 0.012 -0.011** 0.011**
In couple -0.014*** -0.023*** -0.012** -0.015*** 0.004 0.002 -0.032*** -0.016** -0.024*** -0.013 -0.024*** -0.024***
Dependent 0.000 0.002 0.009* 0.008* -0.006** 0.013*** 0.002 0.004 0.003 0.017* 0.003 -0.007
children
Occupation (ref: high skill)
Mid-skill 0.017*** 0.039*** -0.008* 0.014*** 0.006** 0.004 0.013 0.019*** 0.013*** 0.016 0.026*** 0.014**
Low-skill 0.028*** 0.053*** 0.008 0.017** 0.011*** -0.006 0.012 0.034*** 0.024*** -0.007 0.011** 0.015**
Contract
Temporary 0.173*** 0.264*** 0.148*** 0.140*** 0.073*** 0.051*** 0.151*** 0.271*** 0.301*** 0.303*** 0.137*** 0.057***
Observations 316 090 10 982 10 107 13 165 16 512 9 997 5 071 21 624 12 342 9 335 20 876 8 412
Pseudo-R: 0.198 0.169 0.165 0.114 0.217 0.132 0.229 0.231 0.147 0.198 0.259 0.060
GRC HUN IRL ITA LTU LUX LVA NLD NOR POL PRT SWE
Wage -0.020*** -0.013* -0.028*** -0.003 -0.012** -0.052*** -0.014 -0.017* -0.015** -0.001 -0.002 -0.020
Gender (ref: male)
Female 0.001 0.001 -0.003 0.003 -0.007 -0.004 -0.009* -0.002 0.006 0.002 -0.000 0.025***
Education (ref: middle)
Low 0.071*** 0.049*** 0.015 0.010*** -0.002 -0.006 0.027*** 0.018* 0.010 0.042*** 0.010** -0.005
High 0.013** -0.026*** -0.001 -0.006* -0.031*** -0.006 -0.015*** 0.003 -0.001 0.002 0.004 -0.004
Age (ref: 40s)
20s 0.019* 0.035*** 0.052*** 0.007 0.019** 0.024*** 0.020** -0.012 0.022*** 0.021*** 0.013* 0.045***
30s -0.002 0.018*** -0.010 -0.000 -0.001 0.011** 0.000 -0.011 0.006 -0.010*** 0.006 0.018*
50s -0.017** 0.007 0.008 -0.001 0.014** -0.002 0.002 -0.001 -0.006 -0.007 0.008 0.004
In couple -0.017** -0.026*** -0.001 -0.003 -0.029*** 0.003 -0.006 -0.015** -0.003 -0.012*** -0.012*** -0.009
Dependent 0.010 0.016*** -0.001 -0.006* 0.000 0.013*** -0.003 -0.017** 0.010** -0.011*** 0.020*** -0.012
children
Occupation (ref: high skill)
Mid-skill 0.004 0.024*** 0.002 0.006* 0.034*** -0.028*** 0.019*** 0.022*** 0.002 0.025*** 0.002 0.032***
Low-skill 0.017** 0.032*** -0.013 0.016*** 0.037*** -0.017* 0.049*** 0.049*** 0.008 0.052*** 0.009 0.014
Contract
Temporary 0.318*** 0.234*** 0.127*** 0.193*** 0.240*** 0.128*** 0.031 0.155*** 0.026** 0.104*** 0.180*** 0.178***
Observations 20 286 14 046 5 025 27 378 9 228 8 597 10 133 10 699 6 271 29 138 15 943 4 675
Pseudo-R: 0.245 0.294 0.0886 0.263 0.200 0.267 0.0746 0.209 0.135 0.238 0.242 0.218
Note: Marginal effects from probit regressions. ***, ** and * denote statistical significance at the 1%, 5% and 10% significance level.
The dependent variable “unstable employment” equals 1 if employment is unstable, i.e. there are at least 3 transitions between employment and
unemployment, and 0 if employment is not unstable. Estimated marginal effects indicate the estimated change in the probability of being in
unstable employment associated with an increase of a given continuous independent variable, and are calculated for an infinitesimal change of
that variable from the sample average. Marginal effects for categorical variables refer to a discrete change from the base level. While unstable
employment is based on month-to-month information (Box 2.1), the characteristics used are observed in the data only once a year. Data from
the longitudinal panel of the EU Statistics on Income and Living Conditions (EU-SILC). The sample pools together longitudinal samples ending
with the years 2016, 2017 and 2018 to minimise year-to-year variability in small subsamples. The sample is restricted to employees between
20 and 59 years old; people not in the labour force (e.g. students or pensioners) are excluded. All countries: sample includes observations for
all listed countries.
Source: Longitudinal EU-SILC.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
127
Figure 2.3. Most workers with unstable employment have temporary contracts
Share of temporary contracts by type of employment trajectories, 2016-18, in percentage
A. Standard employment B. Part-time employment C. Unstable employment
100
90
80
70
60
50
40
30
20
10
0
Note: Data from the longitudinal panel of the EU Statistics on Income and Living Conditions (EU-SILC). The sample pools together rolling panel
samples ending with the years 2016, 2017 and 2018 to minimise year-to-year variability in small subsamples. The sample is restricted to
employees between 20 and 59 years old; people not in the labour force (e.g. students or pensioners) are excluded. OECD: weighted average
of listed countries. Employment trajectories: measured through continuous month-to-month observation of the employment statuses of
individuals over a period of 3 years. Unstable dependent employment comprises those individuals with at least three transitions between
employment and unemployment in 3 years. Standard employment refers to those individuals whose employment is stable (i.e. not unstable) and
work mainly on a full-time basis within 3 years. Part-time employment refers to those individuals whose employment is stable and work mainly
on a part-time basis within 3 years. Temporary contract: a contract is classified as temporary if the termination of the job is determined by
objective conditions such as a specific contract end date, completion of an assignment or return of another employee who is temporarily replaced.
Source: Longitudinal EU Statistics on Income and Living Conditions (EU-SILC).
StatLink 2 https://stat.link/6gl4e2
2.1.4. Women are more likely than men to work part-time
Part-time work can be an option for employees who need to reduce their working hours on a permanent
basis, though it widely comes at the price of reduced earnings and curtailed career prospects (OECD,
2017[13]; 2018[5]). Almost one quarter of women – often mothers – work part-time in OECD countries, and
are more than two times more likely than men to work part-time (Figure 2.4).
Beyond average figures, female part-time employment differs widely among OECD countries (Figure 2.4),
ranging from less than 15% in Eastern European countries and Portugal to almost 60% in the Netherlands.
Women account for the majority of part-time jobs in all countries. In Austria, Belgium, Luxembourg and
Switzerland, about four-in-five part-time jobs are performed by women.
There is no definitive conclusion why the share of part-time work differs so widely among OECD countries.
The institutional framework may play a role to explain cross-country differences, but there seems to be no
clear relationship between the generosity of protection for part-time workers and the incidence of part-time
work (OECD, 2010[14]). Childcare policies can also play a role in incentivising women to take up part-time
jobs. In particular, several studies find a positive effect of reduced childcare costs on maternal labour
supply (Baker, Gruber and Milligan, 2015[15]; Berlinski and Galiani, 2007[16]; Carta and Rizzica, 2018[17])
Part-time employees tend to receive lower hourly wages in all OECD countries (OECD, 2010[14]; 2018[5]).
Individual characteristics explain part of the part-time pay penalty, but occupational profile differences are
more important in accounting for the wage differentials between part-time and full-time workers (Manning
and Petrongolo, 2008[18]). Job insecurity is higher among part-time employees and more so for men than
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
128
for women. Part-time workers also take part in less training and tend to be less optimistic about their career
prospects. Yet, workers, especially women, use part-time work as the primary way to achieve work-life
balance (Mas and Pallais, 2019[19]).
Despite higher job insecurity and lower pay, in most countries part-time employees work shorter hours by
choice, even though this choice is often dictated by external constraints, such as uneven family
responsibilities within couples (OECD, 2018[5]). In 2018, only 5.2% of female employment was involuntary
part-time work (Figure 2.4), i.e. working fewer hours than desired because they could not find a full-time
job. Evidence from the Netherlands, the country with the highest part-time employment share, shows that
women who work part-time display high job satisfaction rates and do not aim to work more hours (Booth
and van Ours, 2013[20]). There are exceptions, however. In Italy, Chile, France and Spain, the countries
with the greatest share of involuntary part-time work, the majority of part-time women would have liked to
work more hours (OECD, 2019[1]).
Figure 2.4. Part-time work is more frequent among women
Incidence and share of female part-time and involuntary part-time employment, 2018, in percentage
Female incidence of part-time employment Female share of part-time employment Female incidence of involuntary part-time employment
90
80
70
60
50
40
30
20
10
0
Note: Part-time employment as a proportion of total employment. “Part-time” here refers to persons who usually work for less than 30 hours per
week in their main job. For Japan and Korea, part-time employment is based on actual rather than usual weekly working hours. Population of
persons aged 15 or more, except for Italy, Spain and United Kingdom where the lower age limit is 16. For Ireland and Portugal, involuntary part-
time employment refers to 2017. OECD unweighted average for all countries but Korea and Mexico, where data on involuntary part-time
employment is missing.
Source: OECD Employment Database http://www.oecd.org/employment/emp/onlineoecdemploymentdatabase.htm.
StatLink 2 https://stat.link/il0a6b
2.1.5. Poverty rates are higher among workers with non-standard employment
Workers with trajectories of non-standard dependent employment have higher income poverty rates
(i.e. live in a household with an annual disposable income of less than 50% of the national median). On
average across 24 European OECD countries, the share of workers living in households with income below
the poverty line is 10% among those with trajectories of part-time employment and 19% among workers in
unstable employment; in contrast, the poverty rate among standard workers is 3% (Figure 2.5).
There is considerable variation between different countries. Among workers with trajectories of part-time
employment, the poverty rate is below 5% in Belgium and the Netherlands, and around or above 20% in
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
129
Greece, Latvia, Portugal and Spain. Further, in the Baltic countries and Sweden, more than one-in-four
workers with unstable employment trajectories live in a household with income below the poverty threshold.
Among standard employees, the poverty rates are lower and more homogenous across countries, ranging
from less than 1%, in Finland, Ireland and the Netherlands, to slightly above 5% in Estonia and
Luxembourg.
Non-standard dependent employment may be a driving factor increasing the risk of falling into income
poverty for several reasons. Earnings potential may be reduced due to lower work intensity, and lower
current and future wages.
Non-standard employees may work for fewer hours in part-time and unstable employment.3 Workers with
trajectories of non-standard dependent employment may also undergo a penalty in terms of hourly wages.
In comparison to standard employees, workers with part-time and temporary jobs experience not only
wage penalties in their current earnings (lower hourly wages) but are also less likely to participate in training
and be promoted, which impacts future earnings (OECD, 2010[21]; OECD, 2014[10]).4 OECD evidence also
suggests that personal characteristics (e.g. age, education and work experience) play a minor role, while
job characteristics (e.g. occupation, industry, firm size and contract) explain some of such penalties
(OECD, 2010[21]).
Household characteristics can also play an important role in determining the poverty risk of non-standard
employees. Evidence on part-time employees shows that they are more likely to be income poor if their
job is the main source of household earnings than if they cohabit with standard employees with higher
earning (OECD, 2015[22]).
Workers with part-time and unstable employment may also experience penalties in terms of access to
social benefits and protections. In many countries, social protection systems in general, and unemployment
benefits in particular, have not yet fully adapted to the specific needs and circumstances of workers in non-
standard forms of dependent employment. Analyses of out-of-work social benefits in six countries (France,
Greece, Hungary, Italy, Spain and the United Kingdom) found that in Spain and Italy unstable employees
were less likely to have access to benefits than standard employees (OECD, 2019[1]). This gap was also
considerable in the United Kingdom, although not statistically significant. The social protection gap
between standard and non-standard employees was larger when assessed in terms of benefit generosity
(benefit amount as a proportion of median income), especially in Greece, Italy and Spain (Fernández,
Immervoll and Pacifico, forthcoming[23]).
Due to their unusual work trajectories, jobseekers with previous non-standard dependent employment may
not receive the type of support that unemployment benefits typically provide to jobseekers with standard
employment. Unemployment benefits, in particular, tend to be more difficult to access for non-standard
employees than other insurance programmes, such as maternity and sickness benefits (Avlijas, 2019[4]).
In fact, low coverage by unemployment benefits – fewer than one-in-three jobseekers receive
unemployment benefits on average across countries (OECD, 2018[5]) – may be associated with non-
standard dependent employment. The condition for access to unemployment benefits in OECD countries
are examined in the next section.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
130
Figure 2.5. Poverty rate by type of dependent employment
Share of employees with household disposable income below the poverty line by type of employment trajectory,
2016-18 (in percentage)
A. Standard employment B. Part-time employment C. Unstable employment
40
35
30
25
20
15
10
5
0
Note: Data is from the longitudinal panel of the EU Statistics on Income and Living Conditions (EU-SILC). To minimise year-to-year variability in
small subsamples, the sample pools together rolling panel samples ending with the years 2016, 2017 and 2018. The sample is restricted to
employees between 20 and 59 years old. People not in the labour force (e.g. students or pensioners) are excluded. OECD: weighted average
of listed countries. Poverty: defined as household disposable income (adjusted for household size) below 50% of the median. Household
disposable income includes, from all household members, all gross personal income components (e.g. earnings, social transfers, income from
rent, regular inter-household transfers received, and income from capital) minus taxes on wealth, regular inter-household transfers paid and
income tax and social insurance contributions. Employment trajectories: unstable dependent employment comprises those individuals with at
least 3 transitions between employment and unemployment in 3 years. Standard employment refers to those individuals whose employment is
stable (i.e. not unstable) and work mainly on a full-time basis within 3 years. Part-time employment refers to those individuals whose employment
is stable and work mainly on a part-time basis within 3 years.
Source: Longitudinal EU-SILC.
StatLink 2 https://stat.link/hf4zvy
2.2. Are unemployment benefit rules adapted to non-standard employment?
Requirements related to minimum time in employment or social contributions are harder to meet for those
in unstable or part-time employment. With frequent job changes and job losses, unstable workers tend to
have comparatively short employment tenure. Depending on the country, differences between workers
with standard and non-standard employment might go beyond a mere pro-rata equivalence of entitlement
and result in less favourable entitlement to unemployment insurance such as smaller benefit amount,
shorter duration or restricted access. Furthermore, already acquired entitlements may be lost during a
change in employment status or job (e.g. following a transition from dependent employment to self-
employment when entitlements differ across employment statuses or between jobs if they are tied to a
specific employment relationship). To deal with these gaps, several countries have special measures
including exemptions for specific contractual arrangements, such as casual employment and seasonal
work (OECD, 2019[1]).
In many OECD countries, some part-time work is compatible with the receipt of unemployment benefits
(“part-time unemployment benefit”) and may or may not open new entitlements. Some countries offer
partial benefits to workers whose working hours have been reduced (“short-time unemployment benefit”)
– see e.g. Cahuc (2018[24]). In contributory systems, out-of-work support for unstable workers often
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
131
includes provisions that allow for contribution periods or unused entitlements to be carried forward to future
claim periods.
2.2.1. Main unemployment benefit rules
This section describes the main unemployment benefit rules focusing on their possible specific adaptation
– or lack of it – to the special circumstances of workers with part-time or unstable employment. Formally,
workers with part-time and unstable employment have equal access to unemployment benefits as workers
with standard employment. In practice, however, if eligibility criteria are not adapted to their special
circumstances their access and protection level may be considerably different and, in some cases,
inadequate. This is particularly the case in contributory systems, which require meeting a minimum amount
of time in employment, of hours worked and/or of earnings received.
Most unemployment benefits in OECD countries are contributory and eligibility is conditional on
employment requirements (Figure 2.6). Employment requirements of first-tier unemployment benefits vary
considerably across OECD countries. On average, the minimum time in employment is about 12 months
and ranges from three months, in Iceland and Italy, to 24 months in the Slovak Republic. The reference
period is on average about 24 months, and varies between nine months in the Netherlands and six years
in Spain. Interestingly, the country with the longest minimum time in employment, the Slovak Republic,
permits a longer assessment period for workers with temporary contracts.
Employment requirements also apply to some second-tier unemployment benefits. Usually, both minimum
time in employment and reference periods are shorter than first-tier benefits. In France, however, access
to the second-tier unemployment benefit requires at least five years of employment in the last ten years,
typically targeting the long-term unemployed rather than non-standard employees.
Some countries have restrictions on the minimum amount of contributions or of hours worked, which may
hinder access of part-time employees, who are less likely to fulfil those criteria. Poland, for example,
requires monthly contributions to be based on earnings that are equal to, or greater than, the national
minimum wage. In the United Kingdom, a minimum earnings level is also required for contributions to count
towards unemployment insurance entitlements – the minimum earnings level is approximately 15% of
average full-time earnings in the United Kingdom (OECD, 2019[1]). In Finland, contributions must be based
on at least 18 hours of work per week. In Australia, the Jobseeker Payment provides social protection for
jobseekers through means-tested payments that are independent of past employment.
The financing of contributory unemployment benefits usually stems, at least in part, from social security
contributions paid by employees and/or employers. In most countries, social security contributions include
“unemployment insurance contribution rates” which correspond to the part of contributions earmarked to
unemployment protection. Some countries differentiate unemployment insurance rates to create financial
incentives for employers and employees to choose more stable employment contracts and discourage
collusion. Box 2.2 shows unemployment insurance rules in a subsample of OECD countries.
There is considerable variation across countries in the design of unemployment benefit systems, and
hence, the extent to which they support incomes during joblessness and facilitate job search. Figure 2.7
summarises some of the key institutional details of first-tier and second-tier unemployment benefits in all
OECD countries, except Colombia. Second-tier unemployment benefits protect jobseekers who are not (or
no longer) eligible to the first-tier contributory unemployment benefits.5 Special unemployment
programmes such as training allowances, additional “lower-tier” unemployment benefits6, social assistance
benefits and other programmes that are not exclusively targeted at jobseekers (e.g. family or sickness
benefits) are not considered.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
132
Figure 2.6. Employment requirements in unemployment benefits
Minimum time in employment and reference period, number of months at 1st of January 2020
Additional reference period for temporary workers Reference period Minimum time in employment
Number of months 72 120
60
40
20
0
First-tier Second-tier
Note: Minimum time in employment indicates the minimum contribution period to be eligible for receiving first-tier unemployment benefit
payments. Reference period is the period to assess whether the respective employment condition is fulfilled. First-tier and second-tier
unemployment benefits are identified by the country acronym plus the number 1 and the number 2, respectively. Second tier benefits are only
included if claim is conditional on fulfilling an employment condition. In countries without a contributory second-tier unemployment benefits, first-
tier benefits are identified only by the country acronym. Countries where first-tier benefits are not conditional on fulfilling an employment condition
are not shown.
(1) Minimum earnings requirement of NOK 149 787 in the past 12 months or alternatively NOK 299 574 in the past three years. (2) At least
30 days of employment in the 12 months prior to the start of the unemployment spell. (3) Or 200 days in last two years. (4) Earnings condition
also applies. (5) Assuming 40-hour workweek. (6) six months in any one of the past two years. (7) The first-tier benefit requires claimants to
meet a minimum membership period in an unemployment insurance fund (A-kassa) of 12 months. (8) Or 26 weekly contributions in each of
previous two years. The claimant must also have made 104 weekly contributions in the whole career. (9) First unemployment benefits claim:
12 months within two years in general, and 6.5 months within one year for people under 25. For all subsequent UB claims, seven months within
one year or 12 months within two years. (10) The claimant must have 12 months of contributions since the previous unemployment spell or the
last time a withdrawal from the individual account was made; the last three contributions must be continuous and with same employer. (11) The
claimant must have held the labour contract for the last 120 days before its termination. (12) The last three monthly contributions before
unemployment must be continuous and with same employer.
Source: OECD tax-benefit model and policy database (http://oe.cd/TaxBEN), Avilijas (2019[4]), and information collected and compiled by the
Employment, Labour and Social Affairs Directorate of the OECD.
StatLink 2 https://stat.link/3qmgtc
Most unemployment benefits are contributory and not means-tested, especially if they are first-tier benefits.
Out of the 47 unemployment benefits, 37 are contributory and ten are non-contributory (Figure 2.7,
Panel A). Only in Australia and New Zealand, first-tier unemployment benefits are not contributory and
means-tested. Most second-tier unemployment benefits are not contributory and means tested. Only in
Austria and Chile, second-tier unemployment benefits are contributory and not means test. Conversely, in
Estonia, France, Portugal and Spain, second-tier unemployment benefits are contributory and means
tested.
Typically, second-tier unemployment benefits are subject to strict income and asset testing, which may
hamper access for jobseekers in households with other income sources, such as earnings from a spouse.
Among OECD countries with a second-tier unemployment benefit, only in Austria and Chile the benefit is
not means-tested (Figure 2.7, Panel B). In Austria, France and Greece, the second-tier unemployment
benefits are available only for people who ran out of the first-tier benefit – thus, they are not available to
jobseekers who did not qualify for the primary benefit in the first place.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
133
Box 2.2. Unemployment insurance contribution rates of employees and employers
In some countries, social security contributions for unemployment insurance vary for specific types of
employment, including part-time and temporary contracts. Total contribution rates, i.e. both employer
and employee contributions, vary from 0.9% of pre-tax wages in Japan to 8.3% for temporary
employees in Spain. In most countries with contributory unemployment insurance schemes, the lion’s
share of total contributions is borne by the employer. The employer-to-employee contribution ratio
ranges from 50:50 in Finland to 100:0 in France 1, the Netherlands and Poland. Some countries apply
different contribution rates according to occupation (Japan), type of dependent employment (Spain and
Netherlands) and wages (Austria and the United Kingdom).
Different contribution rates for temporary and permanent contracts in Spain and the Netherlands may
significantly alter employer and employee incentives to favour a certain type of contract. In France, a new
experience rating system according to firm-specific job separation rates will modify employer contributions
rates from January 2021 onwards.2 Employers will pay a reduced rate if firm separation rates are below the
median sectoral separation rates. The contribution floor of the new system is 3% and the ceiling is 5.05%.
A recent study suggests, however, that, by increasing labour cost for marginal works, taxing temporary jobs
of short duration […] increases the share of open-ended contracts but reduces the mean duration of jobs
and decreases job creation, employment and welfare of unemployed workers (Cahuc et al., 2019[25]).
Similarly, discontinuities in the contribution schedule, as in Austria and the United Kingdom, may lead
to coordinated behaviour of employers and employees to determine wages or adapt labour market
participation. Results from empirical studies exploiting discontinuities in social security contribution
ceilings for employers indicate that labour supply and wage responses are small to negligible (Saez,
Matsaganis and Tsakloglou, 2012[26]; Saez, Schoefer and Seim, 2019[27]). At the lower end of the wage
distribution, however, reduced taxes or social security contributions for employees can significantly alter
labour supply (Meyer and Rosenbaum, 2001[28]; Chetty, Friedman and Saez, 2013[29]). Considering this,
sharp discontinuities in social security contributions at low earnings levels in Austria and the
United Kingdom may affect both the decision of whether to work (extensive margin) and how many
hours to work (intensive margin).
1. In France, the contributions paid on wages are not the only source of funding. Unemployment protection is also funded by a general tax
of 1.47% on market incomes.
2. France implemented different contribution rates by contract types in 2013 but abolished them in 2017. Since January 2020, a fixed rate of
EUR 10 applies on contracts based on “CDD d’usage”– this is a specific flexible temporary contract authorised only in certain sectors of
activity. In addition, for certain activity sectors (dockers and entertainment workers), instead of this flat-rate tax, an increased rate of employer
contributions to unemployment insurance of 0.05 points is applied to “CDD d’usage” fixed-term contracts shorter than three months.
In some countries, second-tier unemployment benefits facilitate access to non-standard employees
through less demanding employment requirements. In Spain, the second-tier unemployment benefit
requires a shorter minimum time in employment (six months in general, three months in case of having
family responsibilities). In the United Kingdom and Finland, the second-tier unemployment benefits do not
depend on past employment.
Generally, the amount of first-tier contributory unemployment benefits fluctuate between a minimum and
maximum amount, while second-tier, non-contributory and means-tested unemployment benefits have
fixed amounts. Benefit amounts are usually considerably higher in first-tier than second-tier unemployment
benefits, but the difference depends on whether the entitlement to the first-tier benefit is closer to the
minimum or maximum limit. Panel A in Figure 2.7 indicates the benefit amount for a recipient previously
earning a salary equivalent to two-thirds of the average wage in the country.7 Under these circumstances,
in most countries, the amounts of first-tier unemployment benefits exceed those from second-tier benefits
by a significant margin (at least 10 percentage points of the average wage).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
134
In some countries, however, the benefit amounts of first- and second-tier unemployment benefits are
similar (Figure 2.7, Panel A). In Austria, where access to the second-tier benefit is conditional on
exhausting first-tier benefit, the amount of the second-tier benefit is just slightly lower than that of the first-
tier benefit, though it is means-tested and other household incomes reduce the entitlement. In Ireland and
the United Kingdom, the benefit amounts of first-tier and second-tier benefits are the same. This may help
explain why only 14% of the unemployment benefit claimants in the United Kingdom receive the first-tier
benefit (Bradshaw and Bennett, 2017[30]; Avlijas, 2019[4]).
In Estonia, Finland, Greece, Portugal, Spain and Sweden, the amount of the flat-rate means-tested
second-tier benefit is similar to the minimum amount of the earnings-related first-tier benefit (Figure 2.7,
Panel A). In the case of Finland, not only the amounts are the same but also the means test applied to the
second-tier benefit has limited impact on benefit amounts as many income sources are fully disregarded
(e.g. earnings of the spouse, social assistance and housing allowance). In contrast to other earnings-
related social security systems, in Finland, the first-tier benefit does not have an upper ceiling for the benefit
level. This feature aims to incentivise high-wage earners to contribute to the system.
The maximum duration of first-tier unemployment benefits tends to be shorter than of second-tier benefits.
In most countries, the maximum duration of first-tier unemployment benefits is up to 36 months, with
24 months being the most frequent limit (Figure 2.7, Panel B). In the non-contributory and means-tested
first-tier unemployment benefits of Australia and New Zealand there is no maximum duration limit.
Similarly, no maximum duration limit applies in Belgium, where first-tier unemployment benefits are
contributory and not means tested. In several countries, the duration of second-tier unemployment benefits
is unlimited. In most cases, these benefits with unlimited duration are means tested. In Austria, however,
the benefit is not means tested on household income, although it is suspended if the claimant earns above
the minimum base of social contributions.8 Among second-tier unemployment benefits with limited
duration, the maximum duration ranges from nine months in Estonia to 24 months in Portugal.
2.2.2. Detailed unemployment benefit rules
This section describes detailed rules of first- and second-tier unemployment benefits that can affect
workers with part-time and unstable employment differently from standard employees. The data is based
on a tailored questionnaire, which was submitted to 11 OECD member countries (Australia, Austria,
Canada, Finland, France, Japan, Latvia, the Netherlands, Poland, Spain and the United Kingdom). This
country selection aims to illustrate unemployment benefits in European and non-European countries,
based on contributory and non-contributory systems, using and not using means testing, and with different
levels of legal complexity and access conditions. The detailed information on unemployment benefit
legislation is summarised in Table 2.2.
In all selected countries, benefit claimants with previous part-time employment are subject to the same
access conditions as those who previously worked full-time. The calculation of benefit amounts may differ
however, particularly in the case of first-tier benefits. In most countries (Austria, Canada, France, Japan,
Latvia, the Netherlands and Spain), the amount of first-tier benefits is calculated pro-rata, i.e. in proportion
to the wage in the previous job. In Australia, Poland and the United Kingdom, where the benefit amount is
a flat rate, jobseekers previously in part-time employment receive the same amount as those previously
working full-time. Finland has a hybrid system combining a fixed component with a supplement that is
proportional to earnings. Second-tier benefit amounts are also independent of previous hours worked in
Finland, France, and the United Kingdom. In contrast, in Austria and Spain, the amounts of second-tier
benefits are computed in proportion to the previous number of hours worked.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
135
Figure 2.7. Main characteristics of unemployment benefits, 2020
A. Benefit amount
(minimum, maximum and level for jobseeker previously earning 2/3 of average wage)
Contributory .Not contributory Amount when previous earnings were 2/3 of average wage
% of average wage ∞ 251 ∞ ∞ 114
100
80
60
40
20
0
B. Benefit duration
Not means tested Means tested
No. of months
∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞ ∞
40
35
30
25
20
15
10
5
0
Note: Panel A shows the interval between minimum and maximum benefit as a percentage of the average wage and the amount the jobseeker
is entitled to when previous-job earnings were exactly two-thirds of the average wage. Panel B shows maximum duration in months. First-tier
and second-tier unemployment benefits are identified by the country acronym plus the number 1 and the number 2, respectively. In countries
without second-tier unemployment benefits, first-tier benefits are identified only by the country acronym.
Benefit rules and calculations assume a 40-year old individual in a single household without children and not eligible to any supplement. For
other household types and age groups, see annual OECD tax-benefit country reports. In some countries, benefit minimums may be lower than
stated for former part-time employees (not considered).
Unemployment benefits as of 1st January, 2020. Benefit amounts for jobseeker previously earning 2/3 of average wage estimated by combining
results from the OECD tax-benefit web calculator for 2018 and information collected and compiled by the Employment, Labour and Social Affairs
Directorate of the OECD for 2020. Estimations of the benefit amount at 2/3 of average wage additionally assume that claimants are in the second
benefit month and have a long contribution record (22 years).
Average wage: gross annual value for a full-time worker. Wages for 2020 are preliminary wage estimates based on projected average wage
data, calculated by the Secretariat. All estimations use annualised benefit amounts.
[1] Benefits may be extended up to a maximum of 48 months if taking part in certain active labour market policies.
[2] Information on the USA reflects the situation of the Michigan unemployment benefit scheme.
[3] Claimants deemed to be difficult to re-employ may receive up to 360 days of unemployment benefits.
[4] The maximum benefit duration is determined by the regional unemployment rate. The estimated maximum duration shown here is based on
the national unemployment rate for December 2019.
Source: OECD tax-benefit model and policy database (www.oecd.org/social/benefits-and-wages.htm) and information collected and compiled
by the Employment, Labour and Social Affairs Directorate of the OECD.
StatLink 2 https://stat.link/ib7fo5
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
136
Many countries use part-time unemployment benefit schemes, which enable claimants to keep part of their
unemployment benefits while earning low wages from a job. These schemes aim at people who have lost
a full-time and found a part-time job or have lost a secondary job, and are seeking a new job in order to
work more hours (Cahuc, 2018[24]). Typically, jobs are restricted to a temporary basis or subject to specific
limits regarding working hours and/or earnings.
Except for Latvia and Poland, all analysed countries have schemes of part-time unemployment, although
specific conditions apply. Most countries set an implicit earnings limit on part-time unemployment by
phasing out the amount of the benefit by reducing it in line with earned wages. Some unemployment
benefits operate, however, explicit limits based on earnings, hours or duration. Explicit earning limits are
in place in Austria (lower than the “marginal income limit”), second-tier benefit in Japan (80 000 JPY) and
the Netherlands (87.5% of the benefit amount). Limits on the number of hours worked are imposed in
Finland (80% of full-time), first-tier benefit in Japan (20 hours per week) and Spain (70% of full-time). In
France, the second-tier unemployment benefit sets a duration limit, as work is only compatible with the
benefit for up to three months.
Some countries encourage unemployment benefit recipients to take up part-time work using financial
incentives that allow recipients to “keep” some of the additional income received in wages by not reducing
the benefit amount by the same magnitude. First- and second-tier unemployment benefits in Austria and
second-tier benefit in France do not reduce the amount of benefits with wages earned, although both apply
limits, as seen above. In Australia, Canada, Finland, France (first-tier benefit) and the Netherlands, benefits
apply ‘withdrawal rates’, which reduce the magnitude of benefits at a slower pace than wages.
Furthermore, in Australia, Finland and the United Kingdom, benefits apply ‘earning disregards’, which
ignore part of the earnings amount to be deducted from the benefit. Only Spain and first-tier benefit in
Japan do not provide any financial incentive.
Putting together the limits and incentives to part-time unemployment, Table 2.2 presents the maximum
possible sum of unemployment benefits and earnings (SUBE), expressed as a proportion of the average
wage in the country. SUBE ranges from 9% in the United Kingdom to 70% in Canada. The sum of
unemployment benefits and earnings can also reach levels from 40% of the average wage in Austria,
Finland, France, Japan, the Netherlands and Spain (first-tier benefit only).
Workers experiencing unstable employment, with repeated transitions between work and unemployment,
may not satisfy all unemployment benefit conditions when applying for a subsequent time. To facilitate
their access, some countries make use of ‘recharging’ rules, with specific access conditions for jobseekers
who are not applying to unemployment benefit for the first time. In Austria, for example, people who
received unemployment benefits before are favoured by an alternative employment requirement (28 weeks
of employment in the past 12 months) besides the one available for first-time claimants (52 weeks in the
past 24 months).
Rules easing repeated benefit claims may produce incentives for workers and employers to adjust their
behaviour. Unemployment benefit entitlements readily available after the end of temporary contracts can
trigger “carousel effects” (i.e. repeated movements in and out of unemployment) because benefit claims
are not restricted to situations where the firm has an objective economic reason for layoffs (OECD,
2002[31]).
Some countries reduce the amount of first-tier unemployment benefits after some period to encourage job
search (France, Latvia, the Netherlands and Spain). 9 However, such reductions may also incentivise
workers with intermittent employment to plan strategically their employment spells in a way to maximise
the amount of unemployment benefit payments (Fontaine and Malherbet, 2016[9]; Kyyrä, 2010[7]; Le
Barbanchon, 2016[8]).
In some countries, the jobseeker can keep unused entitlements if she finds full-time work before exhausting
the benefit duration to which she is entitled. Workers with unstable employment may then access such
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
137
unused entitlements in a later unemployment spell. The rules for saving unused entitlements diverge
significantly across countries and benefit programmes. Unused entitlements to unemployment benefits are
completely lost in Poland and Latvia,10 so that workers need to rebuild their benefit rights from scratch
whenever they start a new employment spell. In Canada, the Netherlands and the United Kingdom, strict
conditions determine whether entitlements are kept for subsequent unemployment spells. 11 In Spain, “old”
(unused) entitlements are kept, but cannot be accumulated with “new” entitlements. If, due to a recent
period in employment, jobseekers are eligible to “new entitlements”, they must choose between the old
and the new entitlement, and discard the other. In France’s first-tier unemployment benefit rules,
jobseekers can accumulate old and new entitlements, as long they first finish the old one before claiming
the new one. Otherwise, jobseekers have the droit d’option, which allows them to start receiving the new
entitlement, while discarding the old one. In some countries, such as France and Japan, unused
entitlements can be (partly) paid out as an in-work benefit or re-employment allowance/bonus.
About half of the countries shown use benefit waiting periods to incentivise job search during an
unemployment spell. However, these provisions may also make support difficult to access for those with
unstable employment. Waiting periods typically last for a week and apply to all applicants, but individual
circumstances can extend them to one month or more in Australia or (for second-tier benefits) in Spain. In
some countries, waiting periods are waived for those with unstable employment, i.e. if the last benefit
payment was received a short time before (12 weeks in the United Kingdom and 12 months in France).
Table 2.2. Summary of unemployment benefits rules that impact differently on non-standard
employees
Selected countries, 2020
Part-time Part-time Unstable employment
employment before unemployment
unemployment
Recharging Amount Unused Waiting
across time entitlements period
Australia
1st tier Same requirements Compatible. Not applicable. No Not applicable. 7 days for
Jobseeker Payment [1] and amounts as full- Disregard of reduction. all.
time. AUD 104 per Additional
two weeks, transfer waiting
withdrawal rate of period may
50% between apply for
AUD 104-254 and high-paid
60% above that seasonal
threshold. work.
SUBE*: 31% of
AW.
Austria
1st tier Same requirements Fully compatible if Facilitated No Kept. None.
Arbeitslosengeld as full-time. earnings below the access: reduction. Old entitlements
Amount calculated marginal income Only 28 weeks can be used for
pro-rata to earnings limit of EUR 461 of employment 5 years after the
in the reference per month.[2] in last 2 years if last day of
period. SUBE*: 48% AW. repeated benefit receipt.
unemployment New entitlement
spells. makes the old
one void.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
138
Part-time Part-time Unstable employment
employment before unemployment
unemployment
Recharging Amount Unused Waiting
across time entitlements period
2nd tier Same requirements Fully compatible up Same No . None.
Notstandshilfe as full-time. to earnings of conditions apply reduction.
Amount calculated EUR 461. to first and
pro-rata to working SUBE*: 46% AW. subsequent
hours. claims.
Canada
1st tier Same requirements Compatible. Same conditions No Kept, if benefit 7 days for all.
Employment Insurance as full-time. Transfer withdrawal apply to first and reduction. was suspended
Amount (pro-rata) rate of 50% up to subsequent due to high
and duration 90% of reference claims. earnings from
calculated by number earnings; transfer partial work.
of contribution hours. withdrawal rate of Dropped after
100% above this benefit period
threshold. ends.
SUBE*: 70% AW.
Finland
1st tier Same requirements Compatible. Same No Kept. 7 days for
Peruspäiväraha as full-time. Transfer withdrawal conditions apply reduction. New entitlement all.
Ansiosidonnainen Weeks of work rate of 50% above to first and makes the old
työttömyyspäiväraha defined as being at disregard of subsequent one void.
least 18 hours of EUR 311; claims.
work. Sum of benefits
Amount includes a and earnings may
fixed basic not exceed
component and an reference earnings,
earnings-based working hours may
part. not exceed 80% of
full-time hours.
SUBE*: 65% AW.
2nd tier Same requirements Compatible. Same No . 7 days for
Työmarkkinatuki and amounts as full- Same rules as for conditions apply reduction. all.
time. first-tier benefit to first and
apply. subsequent
Earnings from part- claims.
time work
disregarded in
means test.
SUBE*: 46% AW.
France
1st tier Same requirements Compatible. Same conditions 30% Kept, within the Exempt from
Allocation d’aide au retour à as full-time. Transfer withdrawal apply to first and reduction limit of 5 years 7 days wait if
l’emploi (ARE) Amount calculated rate of about 70%. subsequent after from the opening already had
pro-rata to number of SUBE*: 55% of AW. claims. 6 months if of the right. it in the last
hours worked. reference Can opt for new 12 months.
wage is entitlement or old
above one.
EUR 4 500
per month.
2nd tier Same requirements Fully compatible up Same conditions No Fully paid out None.
Allocation de solidarité spécifique and amounts as full- to 3 months (in the apply to first and reduction. until next re-
(ASS) time. limit of remaining subsequent application.
entitlements until claims.
subsequent
renewal).
SUBE*: n/a.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
139
Part-time Part-time Unstable employment
employment before unemployment
unemployment
Recharging Amount Unused Waiting
across time entitlements period
Japan
1st tier Same requirements Compatible for paid Same No Partly paid out as 7 days for all.
Koyo hoken as full-time. work of up to 4 hours conditions apply reduction. in-work benefit if
Amount calculated per day and to first and at least one third
pro-rata to reference 20 hours per week. subsequent of benefit days
period income. Transfer withdrawal claims. remains (at least
rate of 100%. 45 days).
SUBE*: 40% AW.
2nd tier Same requirements Fully compatible up Same conditions No n/a. None.
Kyuusyokusyashienseido and amounts as to earnings of apply to first and reduction.
fulltime. JPY 80 000 per subsequent
month. claims
SUBE*: 41% AW
Latvia
1st tier Same requirements Not compatible. Same 25% Dropped. None.
Bezdarbnieka pabalsts as full-time. Temporarily conditions apply reduction
Amount calculated suspension for to first and in months
based on average max. 2 months, subsequent 3-4;
past wages. otherwise claims. 50%
termination. reduction
SUBE*: n/a. in months
56; 55%
reduction
in months
7-8.
Netherlands
1st tier Same requirements Compatible. Same 7% Kept only if: None.
Werkloosheidswet/Toeslagenwet as full-time. Transfer withdrawal conditions apply reduction weekly working
Amount calculated rate of 75% in the to first and after hours are about
pro-rata to gross- first 2 months and subsequent 2 months. the same as
wage. 70% subsequently. claims. before
Earnings from work becoming
may not exceed unemployed and
87.5% of the benefit earnings from
amount. the new job are
SUBE*: 61% AW. below 87.5% of
the
unemployment
benefit amount.
Poland
1st tier Same requirements Not compatible. Same No Dropped. 7 days for
Zasiłek dla bezrobotnych and amounts as full- Possibility to conditions apply reduction. Possibility to all.
time. receive activation to first and receive
Wage must be over allowance if taking subsequent activation
social up employment claims. allowance if
unemployment below minimum taking up
insurance wage. employment
contribution SUBE*: n/a. below minimum
threshold. wage.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
140
Part-time Part-time Unstable employment
employment before unemployment
unemployment
Recharging Amount Unused Waiting
across time entitlements period
Spain
1st tier Same requirements Compatible Same 28.6% Kept, but cannot None.
Prestación por desempleo as full-time.[3] between 10% and conditions apply reduction accumulate with
Amount calculated 70% of full-time to first and after new entitlement.
pro-rata to number work hours. [4] subsequent 6 months. Must opt for one
of hours worked. Transfer withdrawal claims. or the other.
rate of 100%.
SUBE*: 47% AW.
2nd tier Same requirements Compatible. Same No Kept, but cannot 1 month
Subsidio por desempleo as full-time. Transfer withdrawal conditions apply reduction. accumulate with waiting
Amount calculated rate of 100%. to first and new entitlement. period
pro-rata to working SUBE*: 18% AW. subsequent Must opt for one except if not
hours. claims. or the other. meeting
contribution
criteria for
UI.
United Kingdom
1st tier Same requirements Compatible up to Same No Kept only if 7 days
Jobseeker’s Allowance New Style and amounts as full- 16 hours per week. conditions apply reduction. break between waiting
time. Transfer withdrawal to first and unemployment period
rates of 100% subsequent spells is less exempted if
above disregard of claims. than 12 weeks last
GBP 5 per (otherwise lost). entitlement
2 weeks.[5] is less than
SUBE*: 9% AW. 12 weeks
ago.
2nd tier Same requirements Compatible. Same No n/a. None.
Universal Credit and amounts as full- Transfer withdrawal conditions apply reduction.
time. rate of 63%. to first and
SUBE*: 15% AW. subsequent
claims.
Note: All benefit rules and calculations assume a single household without children. For other household types, see annual OECD tax-benefit
country reports.
* SUBE: Sum of unemployment benefits and part-time employment earnings. This indicator is estimated assuming a single individual without
children, aged 40 years-old, eligible for the unemployment benefit and, if previous earnings are required, previously earning two-thirds of the
average wage in the country.
[1] From 20 March 2020, Newstart Allowance was replaced by the JobSeeker Payment. JobSeeker Payment has the same payment rates and
indexation arrangements as Newstart Allowance https://www.dss.gov.au/about-the-department/benefits-payments/working-age-payments.
[2] Months worked do not count for future benefit eligibility if earnings are below EUR 461.
[3] In case of multiple part-time jobs, only contributions of lost jobs are considered for eligibility and duration;
[4] Part-time work during benefit payment periods do not count for fulfilling renewed employment conditions.
[5] GBP 20 per week in some special cases.
Source: OECD tax-benefit model and policy database and tailored questionnaire to national authorities.
2.3. Unemployment benefits: Income protection and financial work incentives
Do unemployment benefits provide adequate income protection and strong work incentives to jobseekers
who were in non-standard dependent employment? How different are indicators of income protection and
work incentives for jobseekers with a history of standard employment? What are the circumstances and
policies that drive non-standard employees to be treated differently than standard employees? Are there
particular systems (e.g. contribution-based or means-tested based) or policy mechanisms that make
unemployment benefits better equipped to protect non-standard employees?
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
141
In order to address these questions, this section develops new indicators of the impact of unemployment
benefits on income protection and financial work incentives for non-standard employees. The indicators
are obtained by simulating the unemployment benefit rules on employment scenarios that characterise
workers with standard and non-standard dependent employment, using definitions that are comparable
across countries. The simulations compute unemployment benefit entitlements for people who worked the
same total number of hours and earned the same amount of wages over a period of several years, but
through different employment trajectories.
The simulations are similar in spirit to model calculations, such as those based on the OECD tax-benefit
model (TaxBEN), that are commonly used to compare benefit replacement rates and work incentives.
However, TaxBEN currently does not cover benefit provisions for unstable employment. The simulations
focus on the first-tier and second-tier unemployment benefits described in Table 2.2. Lower-tier benefits,
such as social assistance, as well as in-work support and tax provisions also shape the income
consequences of these different work patterns. Accounting for the full range of tax-benefit policy levers
would require an extension of the TaxBEN model to cover unstable employment, which is left to future
work. Box 2.3 provides a detailed description of each scenario as well as additional assumptions.
The simulations were carried out for four OECD countries out of the 11 whose unemployment benefits
were described in Table 2.2: Australia, France, Latvia and Spain. While the long-term objective would be
to include all OECD countries, the simulations for these four countries provides a pilot for future extensions.
These specific countries were selected to deliver diversity in terms of geography (European and
non-European countries), benefit system (contributory and non-contributory), legal complexity and access
conditions. Despite such diversity, given the small number of countries, the evidence obtained with
simulations aims to illustrate and highlight policy issues rather than to be representative of all OECD
countries.
2.3.1. Income protection favours some forms of non-standard employment trajectories
over others
Unemployment benefits provide varying degrees of income protection depending on the type of
employment trajectory. Such differences are observed even when workers have the same personal
characteristics, have earned the same amount of wages and have been in and out of work the same
number of hours, over a given period.
In Australia, France, Latvia and Spain, workers with a standard employment trajectory receive at least as
much income protection as any worker with a non-standard employment trajectory. Figure 2.9 to
Figure 2.12 illustrate the unemployment benefit entitlements in each month and for each scenario
described in Box 2.3. As a summary for these figures, Figure 2.13 decomposes income protection into two
indicators: receipt and level. Receipt measures the number of months that unemployment benefits are
received. Level assesses the average amount of unemployment benefit received over the analysed
seven-year period (accounting only the months in which the benefit is effectively paid).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
142
Box 2.3. Simulation scenarios
Results presented in the following sections are drawn from four different simulation scenarios, each
corresponding to a distinct employment trajectory over a seven-year period. In all scenarios, the
employment status is the same in the first two years (full-time work) and in the last year (unemployment
during the entire year, without any earnings). Employment patterns in years 3 to 6 differ across scenarios.
This relatively long period of analysis allows the assessment of a range of different circumstances. The
first three years correspond to a worker who transitions from full-time stable work to unemployment or non-
standard employment. This trajectory is akin to that of a displaced worker (OECD, 2018[5]; Farber, 2017[32]),
who is struggling to secure a stable job after several years of continuous full-time employment.1 The next
three-year period (years 4 to 6) is illustrative for workers who are caught in a long cycle of non-standard
employment. The seventh year of full unemployment permits measuring the entitlements that are left after
an extended period of non-standard employment.
In all the simulated scenarios, over the whole period, the person works a total number of hours that is
equivalent to 48 full-time months: 24 months during the first two years and 24 months during the following
four years. The person is unemployed a total number of hours that is equivalent to 36 full-time months:
24 months between the third and sixth years, and 12 months in the seventh year. In all instances, the
person works at an hourly wage that is equivalent to two-thirds (67%) of the average wage (AW)2.
The simulations assume full take-up of unemployment benefits (de jure entitlements). Entitled jobseekers
claim benefits from the first month they are eligible, thus maximising their receipt in the short-term; in some
circumstances, jobseekers may receive larger benefit amounts by strategically postponing a claim. In all
scenarios, the jobseeker is 45 years old, lives alone and has no children. Fixing these characteristics
boosts cross-country comparability, by avoiding specific age- and family related unemployment benefit
rules that change from country to country.
The simulated scenarios are illustrated in Figure 2.8. The horizontal axis represents each of the 84 months
of the seven years that are assessed. The vertical axis represents the amount of wages and unemployment
benefits, expressed as a percentage of the average wage – hence, in the first two years the wage is 67%
(two-thirds) of the average wage. UB_1 represents the amount of the first-tier unemployment benefit and
UB_2 represents the amount of the second-tier unemployment benefit.
The characteristics of each simulation scenario – particularly each employment patterns, which is different
between years 3 to 6 – are explained as follows:
A. Standard: illustrates a benchmark employment trajectory (further referred to as a “standard
employment trajectory”), which alternates full-time dependent employment and full unemployment every
12 months.
B. Unstable 6 months: illustrates an “unstable employment trajectory with changes every semester”, with
job changes between full-time dependent employment and full unemployment every 6 months.
C. Unstable 1 month, illustrates an “unstable employment trajectory with changes every month”, with job
changes between full-time dependent employment and full unemployment every month.
D. Part-time: illustrates a “part-time employment trajectory”, which consists of recurrent one-year periods
seeking for a full-time job, with the first month in full unemployment and the following 11 months at part-
time employment (i.e. working at 54.5% of full-time).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
143
Figure 2.8. Simulation scenarios
A. Standard B. Unstable 6 months
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
C. Unstable 1 month D. Part-time
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
StatLink 2 https://stat.link/1jd23u
1. The US Bureau of Labor Statistics defines displaced workers to be “persons 20 years and over who lost or left jobs because their plant or
company closed or moved, there was insufficient work for them to do, or their position or shift was abolished” (Bureau of Labor Statistics,
2020[33]). According to previous OECD work, between 1% and 7% of the workforce is displaced annually, implying a significant probability that
a typical worker will experience displacement one or more times during her working life. A considerable number of displaced workers find a
suitable new job rapidly, but the majority experience significant losses of income and potentially would benefit from re-employment assistance
and income support (OECD, 2018[5]).
2. AW is the annual average wage among full-time employees in the non-agricultural business sector – sectors B to N (inclusive) of the
International Standard Industrial Classification of All Economic Activities (ISIC rev. 4) (United Nations Statistical Division, 2008[34]). More details
on the calculation of the AW measure for OECD countries are available in the methodological section of the OECD Taxing Wages publication
(OECD, 2019[35]).
In Australia and France, in each month that the worker is fully unemployed she receives some
unemployment benefit, in all simulated scenarios. In France, workers with unstable employment
trajectories (scenarios B and C) receive unemployment benefits for the same number of months
(36 months) as those with standard employment (scenario A). This is not the case in Australia, Latvia and
Spain.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
144
In Australia, workers with a standard employment trajectory (scenario A) receive unemployment benefit for
38 months, workers who switch between employment and unemployment every six months
(scenario B) receive unemployment benefit for 40 months, and those who switch every month
(scenario C) receive it for 60 months. The entitlement, means test and payment of unemployment benefit
in Australia is assessed every fortnight based on the income from the previous fortnight. Because of that,
the benefit payment shifts by a fortnight and may be received in a month when the person is employed
(see the spikes in Figure 2.9). This explains why in scenario C the unemployment benefit is received in
more months than in scenarios A and B. If the number of payments were measured in fortnights, however,
the number of times across scenarios would be the same.
In Latvia, workers with unstable employment trajectories (scenarios B and C) receive unemployment
benefits for 6 months or less, while those with standard employment (scenario A) receive it for 24 months.
The unemployment benefit in Latvia requires a relatively long minimum time in employment (12 months),
while Latvia has the shortest reference period (16 months) out of the four countries (Figure 2.6). Meeting
these entitlement conditions is hard for those with unstable employment patterns. Latvia also has no
second-tier unemployment benefit as an additional protection to those with shorter contribution records.
In Spain, workers with unstable employment trajectories (scenarios B and C) receive the first-tier
unemployment benefit for half the time received by standard workers (8 instead of 16 months). As in Latvia,
this is partly because the first-tier benefit requires a relatively long minimum time in employment
(12 months). In addition, after six months of employment, jobseekers can claim the second-tier
unemployment benefit. After six months of unstable work, workers may prefer to claim the lower-paid
second-tier benefit than to wait until they are able to accumulate 12 months of work; months used to claim
the second-tier unemployment benefit cannot be used to claim the first-tier benefit (see Table 2.2).
Across the four analysed countries, unemployment benefits are more generous under contributory than
non-contributory systems. In Australia, where unemployment protection is non-contributory, the benefit
level for workers with standard or unstable employment trajectories ranges from 10% to 15% of AW. Benefit
levels are considerably higher in France, Latvia and Spain, where their systems are contributory.
In France, benefit levels are the same for workers with stable and unstable employment trajectories. In
Australia and Spain, benefit levels are slightly higher for workers with stable employment trajectories,
because they are less subject to reductions due to waiting periods 12 or are often eligible for new benefits
with higher amounts. Conversely, in Latvia, benefit levels are higher for unstable workers, because of
shorter benefit durations that are less affected by the reduction in the benefit rate as the unemployment
spell lengthens (Table 2.2).13
Workers with employment trajectories that include periods combining unemployment and part-time work
within a month (scenario D) generally receive unemployment benefits for fewer months than workers with
standard employment (scenario A). In Latvia, access is restricted because the unemployment benefit is
incompatible with any type of employment (see Table 2.2). In Australia, Spain and France, the
unemployment benefit is compatible with some work, but in some cases the eligibility depends on the wage
received while in employment (in the simulations presented here the wage is 33.3% AW). In Australia, the
unemployment benefit is completely depleted by the means test if earnings are above an amount
equivalent to 23% AW (see Table 2.2). In Spain, unemployment benefits are reduced in proportion to the
number of hours worked. In addition, the second-tier benefit also applies a means test that makes benefit
receipt incompatible with earnings higher than an amount equivalent to 29% AW (see Table 2.2). In
France, the first-tier benefit is reduced with a 70% withdrawal rate; for the second-tier benefit, work and
benefits are compatible for up to three months, after which the benefit is interrupted if the professional
activity continues.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
145
Figure 2.9. Unemployment benefits in Australia
Simulations under different employment scenarios
A. Standard B. Unstable 6 months
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
C. Unstable 1 month D. Part-time
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
Note: AW: average wage. UB_1: entitled amount of first-tier unemployment benefit. UB_2: amount of second-tier unemployment benefit. The
horizontal axis represents each of the 84 months of the seven years that are assessed. The vertical axis represents the amount of wages and
unemployment benefits, expressed as a percentage of the AW. All cases consider two initial years of full-time employment paid at 67% of AW.
Different cases: A. Alternating periods of full-time dependent employment and full unemployment every 12 months. B. Unstable 6 months and
C. Unstable 1 month: situations with job changes between full-time dependent employment and full unemployment every 6 months and every
other month, respectively. D. Part-time: “part-time employment trajectories”, which consists of recurrent one-year spells seeking for a full-time
job, with the first month in full unemployment and the following 11 months at partial employment (working at 54.5% of full-time hours in working
periods). See Box 2.3 for more details about assumptions and simulation scenarios.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/4eph9q
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
146
Figure 2.10. Unemployment benefits in France
Simulations under different employment scenarios
A. Standard B. Unstable 6 months
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
C. Unstable 1 month D. Part-time
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
Note: AW: average wage. UB_1: entitled amount of first-tier unemployment benefit. UB_2: amount of second-tier unemployment benefit. The
horizontal axis represents each of the 84 months of the seven years that are assessed. The vertical axis represents the amount of wages and
unemployment benefits, expressed as a percentage of the AW. All cases consider two initial years of full-time employment paid at 67% of AW.
Different cases: A. Alternating periods of full-time dependent employment and full unemployment every 12 months. B. Unstable 6 months and
C. Unstable 1 month: situations with job changes between full-time dependent employment and full unemployment every 6 months and every
other month, respectively. D. Part-time: “part-time employment trajectory”, which consists of recurrent one-year spells seeking for a full-time job,
with the first month in full unemployment and the following 11 months at partial employment (working at 54.5% of full-time hours in working
periods). See Box 2.3 for more details about assumptions and simulation scenarios.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/hqebks
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
147
Figure 2.11. Unemployment benefits in Latvia
Simulations under different employment scenarios
A. Standard B. Unstable 6 months
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
C. Unstable 1 month D. Part-time
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
Note: AW: average wage. UB_1: entitled amount of first-tier unemployment benefit. UB_2: amount of second-tier unemployment benefit. The
horizontal axis represents each of the 84 months of the seven years that are assessed. The vertical axis represents the amount of wages and
unemployment benefits, expressed as a percentage of the AW. All cases consider two initial years of full-time employment paid at 67% of AW.
Different cases: A. Alternating periods of full-time dependent employment and full unemployment every 12 months. B. Unstable 6 months and
C. Unstable 1 month: situations with job changes between full-time dependent employment and full unemployment every 6 months and every
other month, respectively. D. Part-time: “part-time employment trajectory”, which consists of recurrent one-year spells seeking for a full-time job,
with the first month in full unemployment and the following 11 months at partial employment (working at 54.5% of full-time hours in working
periods). See Box 2.3 for more details about assumptions and simulation scenarios.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/lzqfmd
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
148
Figure 2.12. Unemployment benefits in Spain
Simulations under different employment scenarios
A. Standard B. Unstable 6 months
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
C. Unstable 1 month D. Part-time
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
Note: AW: average wage. UB_1: entitled amount of first-tier unemployment benefit. UB_2: amount of second-tier unemployment benefit. The
horizontal axis represents each of the 84 months of the seven years that are assessed. The vertical axis represents the amount of wages and
unemployment benefits, expressed as a percentage of the AW. All cases consider two initial years of full-time employment paid at 67% of AW.
Different cases: A. Alternating periods of full-time dependent employment and full unemployment every 12 months. B. Unstable 6 months and
C. Unstable 1 month: situations with job changes between full-time dependent employment and full unemployment every 6 months and every
other month, respectively. D. Part-time: “part-time employment trajectory”, which consists of recurrent one-year spells seeking for a full-time job,
with the first month in full unemployment and the following 11 months at partial employment (working at 54.5% of full-time hours in working
periods). See Box 2.3 for more details about assumptions and simulation scenarios.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/zfb3ym
Unemployment benefit protection is also less generous for workers with employment trajectories that
include periods combining unemployment and part-time work (scenario D) than for a standard worker
(scenario A). This is because entitlements are reduced with the wage earned from concomitant work.
Usually, unemployment benefit entitlements are the same independently of how workers distributed their
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
149
working hours within a month (e.g. worked part-time every day or full-time some days of the month). In
France, however, the unemployment benefit differs depending on the distribution of working hours within
the month (see Box 2.4).
Figure 2.13. Non-standard dependent workers receive fewer unemployment benefits
A. Receipt (number of months receiving unemployment benefit over the seven-year period)
UB-1 UB-2
70
60
50
40
30
20
10
0
A. Standard D. Part-time A. Standard D. Part-time A. Standard D. Part-time A. Standard D. Part-time
B. Unstable 6 months C. Unstable 1 month B. Unstable 6 months C. Unstable 1 month B. Unstable 6 months C. Unstable 1 month B. Unstable 6 months C. Unstable 1 month
Australia France Latvia Spain
B. Level (average amount of unemployment benefit received over months received, in %AW)
UB-1 UB-2
50%
40%
30%
20%
10%
0%
A. Standard D. Part-time A. Standard D. Part-time A. Standard D. Part-time A. Standard D. Part-time
B. Unstable 6 months C. Unstable 1 month B. Unstable 6 months C. Unstable 1 month B. Unstable 6 months C. Unstable 1 month B. Unstable 6 months C. Unstable 1 month
Australia France Latvia Spain
Note: AW: average wage. UB_1: entitled amount of first-tier unemployment benefit. UB_2: amount of second-tier unemployment benefit. The
vertical axis represents, in the first panel, the number of months that unemployment benefits are received, and in the second panel, the amount
of unemployment benefits, expressed as a percentage of the AW. All cases consider two initial years of full-time employment paid at 67% of
AW. Different cases: A. Alternating periods of full-time dependent employment and full unemployment every 12 months. B. Unstable 6 months
and C. Unstable 1 month: situations with job changes between full-time dependent employment and full unemployment every 6 months and
every other month, respectively. D. Part-time: “part-time employment trajectory”, which consists of recurrent one-year spells seeking for a full-
time job, with the first month in full unemployment and the following 11 months at partial employment (working at 54.5% of full-time hours in
working periods). See Box 2.3 for more details about assumptions and simulation scenarios.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/bo3qef
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
150
Box 2.4. The 2020 reform of the “reference wage” in France
Until 2020, France is the only analysed country where the amounts of unemployment benefits differed
depending on the distribution of working hours within a calendar month. Despite working the same number
of hours each month, a person working full-time some days per month would receive a larger amount of
unemployment benefit than if working part-time for the entire month (Figure 2.14).
The cause of this difference was the formula used to compute the “reference wage” of the first-tier
unemployment benefit (Allocation d’aide au retour à l’emploi). The reference wage is an indicator that
summarises the wage that the person earned before becoming unemployed and to be replaced, to some
extent, by the unemployment benefit. In most unemployment insurance systems, the reference wage is
based on the average of previous monthly wages, whether the person worked every day over the reference
period or not. In France, however, the reference wage is calculated based on daily wages on working days
(i.e. days not worked do not count in the calculation). Because of that, the reference wage of a person
working part-time for a whole month (scenario ‘Part-time’) is lower than the reference wage of a person
working full-time for half of the month (scenario ‘Part-month’).
A reform of the French first-tier unemployment benefit, originally planned for April 2020 and later postponed
to September 2020, eliminates this feature. The new reference wage will be calculated using monthly
wages over a 12-month period, irrespective of the number of months and days worked.
Figure 2.14. The distribution of working hours affects unemployment benefit entitlements in France
Part-time Part-month
(person working half-time for a whole month ) (person working full-time for half of the month )
wage UB_1 UB_2 wage UB_1 UB_2
Wages and benefits (% AW) Wages and benefits (% AW)
100% 100%
75% 75%
50% 50%
25% 25%
0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7
Employment trajectory (months/years) Employment trajectory (months/years)
Note: AW: average wage. UB_1: entitled amount of first-tier unemployment benefit. UB_2: amount of second-tier unemployment benefit. The
horizontal axis represents each of the 84 months of the seven years that are assessed. The vertical axis represents the amount of wages and
unemployment benefits, expressed as a percentage of the AW. All cases consider two initial years of full-time employment paid at 67% of AW.
Different cases: Part-time and Part-month, which both consist of recurrent one-year spells seeking for a full-time job, with the first month in full
unemployment and the following 11 months in partial employment (working at 54.5% of full-time hours in working periods). The difference
between the Part-time and Part-month scenarios regards the distribution of working hours within the month. In Part-time, the person works every
day for 54.5% of the full working time. In Part-month, the person works on a full-time basis 54.5% of the days in the month.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/zstq6j
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
151
2.3.2. Financial work incentives depend on current and previous employment
trajectories
Unemployment benefits aim to provide adequate income protection to jobseekers and their families, while
also maintaining work incentives. Improving work incentives and facilitating the return to self-sufficiency is
important because the risk of long-term poverty is much higher for jobless individuals on benefits than for
fully employed people. Moreover, the cost of safety nets to budget-constrained governments reinforces
the need to induce people – who can – to get back into work (OECD, 2005[36]).
When the unemployment benefit amount received by jobseekers is high relative to the wage they can
obtain from work, jobseekers may find themselves in an “unemployment trap” – i.e. discouraged from
searching for a new job (OECD, 2005[36]). In the literature (OECD, 2007[37]; OECD, 2020[38]), the financial
incentive of a jobseeker to pick up a new job is typically measured through replacement rates (RR). This
sub-section uses a variation of the indicator of replacement rates (adjusted replacement rates, ARR), which
is adjusted to account for the specificities of the simulated scenarios (Box 2.5).
Figure 2.16 presents the ARR for the final year of the 16 scenarios depicted in Figure 2.15, in Australia,
France, Latvia and Spain. The ARR is an indicator of financial work disincentives – the higher the ARR,
the lower the financial incentives to work more. The ARR expresses the amount of total income (wages
and benefits) that a worker would get relative to the full-time wage (Box 2.5).
Since workers with non-standard employment earn wages that are equivalent to 50% of a full-time wage,
their ARR is at least 50%. ARR for workers in full unemployment provide a benchmark to assess the
incentives for a worker not to work at all.
Work incentives vary considerably across countries, current and previous employment trajectories. On
average, the ARR is higher in France, followed by Australia, Spain and Latvia. Across current non-standard
employment trajectories, the ARR is higher among workers with unstable employment. Across previous
employment trajectories, ARR are unsurprisingly higher among workers with previous standard
employment, except for Australia.
In Australia, a worker in part-time employment would get an income equivalent to 53% of a full-time wage.
Hence, her income would increase by 47% of a full-time wage if she worked full time. If she did not work
at all, her income would decrease by 18% of a full-time wage, as the ARR in full unemployment is 35%.
Conversely, the ARR for a worker with current unstable employment would be 68%, full employment would
increase income by 32% of a full-time wage, and full unemployment would reduce income by 33% of a
full-time wage.
In France, workers with unstable employment have lower incentives to take up full employment. A worker
in unstable employment would get a benefit amount equivalent to 78% of a full-time wage but working only
half of the time. If she did not work at all, her income would decrease by 21% of a full-time wage, as the
ARR in full unemployment is 57%. In contrast, the ARR for a worker in part-time employment is 64%. If
she did not work at all, her income would decrease by only 7% of a full-time wage. Previous unstable
employment trajectories produce the same work incentives as previous stable employment. For previous
part-time employment, however, adjusted replacement rates are lower – 18 percentage points lower for
full-time unemployed and around 9 percentage points lower for workers in current non-standard
employment.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
152
Box 2.5. Measuring work incentives across employment trajectories
Adjusted Replacement Rates (ARR)
The financial incentives of a jobseeker to pick up a new job are typically measured through a replacement
rate (RR). A replacement rate is the ratio between out-of-work and in-work income (OECD, 2007[37]; OECD,
2020[38]).
𝑦𝑜𝑤
𝑅𝑅 =
𝑦𝑖𝑤
where yow denotes income received while out of work and yiw denotes income received while in work.
Usually, RR is calculated to analyse the effects of a transition between employment and unemployment.
In the case of a transition from employment to unemployment, yiw would denote the income before the
transition and yow would denote the income after the transition. Generally, income is assessed based on
the pay period (e.g. week, fortnight or month).
The adjusted replacement rate (ARR) used in this sub-section uses a variation of the definition above.
Instead of analysing transitions, it compares the effects of counterfactual scenarios. Two scenarios are
compared, full employment (which would be similar to iw) and not full employment (which would be similar
to ow). “Full employment” refers to full-time full-year employment. A person “not in full employment” works
reduces the number of hours or days, or does not work at all (i.e. fully unemployed). Finally, incomes are
assessed over a year instead of the pay period. So, ARR could be express as:
𝑌𝑛𝐹𝐸
𝐴𝑅𝑅 =
𝑌𝐹𝐸
where Y denotes annual income, YnFE denotes income received if not in full employment, and YFE denotes
income received if in full employment.
Current and previous employment trajectories
In the absence of taxes or other social benefits, the amount of YFE depends exclusively on wages, while
YnFE depends on the amount of wages and unemployment benefits. The amounts of unemployment
benefits depend on current and on previous employment circumstances.
Figure 2.15 presents 16 scenarios that combine different current and previous employment trajectories.
Each scenario includes seven years. The first six years correspond to the “previous trajectory”. The
seventh year corresponds to the “current trajectory”, as this is the year in which ARR is computed in
Figure 2.16.
The employment patterns in the first six years are the same as presented in Box 2.1: “A. Standard”,
“B. Unstable 6 months”, “C. Unstable 1 month” and “D. Part-time”. The seventh year repeats the
employment pattern of the third year of each of the four scenarios: a. full unemployment, b. 6 months in
unemployment followed by 6 months in employment, c. alternating between unemployment and
employment, and d. one month unemployed followed by 11 months working at 54.4% of the time.
While this framework has the advantage of accounting for changes across time and counterfactual
scenarios, it also has some limitations:
1) Taxes and other transfers are not taken into account. The ARR is computed as a “gross rate”, rather
than a “net rate”, as often done in work incentive analysis (OECD, 2007[37]; OECD, 2020[38]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
153
2) Work incentives are measured in one specific period (the seventh year of each scenario). Therefore,
the circumstances analysed are specific and cannot be held as representative of what would happen at
different points in the trajectory.
3) An extremely low ARR may indicate the absence of unemployment protection rather than adequate
work incentives. In these cases, constraints are more likely to be on labour demand.
Figure 2.15. Scenarios for computing adjusted replacement rates
Previous: A. Standard Previous: A. Standard Previous: A. Standard Previous: A. Standard
Current: a. Full unemployment Current: b. Unstable 6 months Current: c. Unstable 1 month Current: d. Part-time
100% 100% 100% 100%
75% 75% 75% 75%
50% 50% 50% 50%
25% 25% 25% 25%
0% 0% 0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 3036 42 48 54 60 66 72 78 84 6 12 18 24 30 3642 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7
Previous: B. Unstable 6 months Previous: B. Unstable 6 months Previous: B. Unstable 6 months Previous: B. Unstable 6 months
Current: a. Full unemployment Current: b. Unstable 6 months Current: c. Unstable 1 month Current: d. Part-time
100% 100% 100% 100%
75% 75% 75% 75%
50% 50% 50% 50%
25% 25% 25% 25%
0% 0% 0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 3036 42 48 54 60 66 72 78 84 6 12 18 24 30 3642 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7
Previous: C. Unstable 1 month Previous: C. Unstable 1 month Previous: C. Unstable 1 month Previous: C. Unstable 1 month
Current: a. Full unemployment Current: b. Unstable 6 months Current: c. Unstable 1 month Current: d. Part-time
100% 100% 100% 100%
75% 75% 75% 75%
50% 50% 50% 50%
25% 25% 25% 25%
0% 0% 0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 3036 42 48 54 60 66 72 78 84 6 12 18 24 30 3642 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7
Previous: D. Part-time Previous: D. Part-time Previous: D. Part-time Previous: D. Part-time
Current: a. Full unemployment Current: b. Unstable 6 months Current: c. Unstable 1 month Current: d. Part-time
100% 100% 100% 100%
75% 75% 75% 75%
50% 50% 50% 50%
25% 25% 25% 25%
0% 0% 0% 0%
6 12 18 24 30 36 42 48 54 60 66 72 78 84 6 12 18 24 3036 42 48 54 60 66 72 78 84 6 12 18 24 30 3642 48 54 60 66 72 78 84 6 12 18 24 30 36 42 48 54 60 66 72 78 84
1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7 1 2 3 4 5 6 7
StatLink 2 https://stat.link/1dstec
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
Previous Current
154
Previous Current Previous Current Previous Current
trajectory trajectory 0% 25% 50% 75% 100% 0% 25% 50% 75% 100% 0% 25% 50% 75% 100% 0% 25% 50% 75% 100%
trajectory trajectory trajectory trajectory trajectory trajectory
a. Full unemp a. Full unemp a. Full unemp a. Full unemp
b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths
c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths
A. Standard A. Standard A. Standard A. Standard
d. Part-time d. Part-time d. Part-time d. Part-time wage
a. Full unemp a. Full unemp a. Full unemp a. Full unemp
b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths
c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths UB_1
d. Part-time d. Part-time d. Part-time d. Part-time
B. Unstable 6 months B. Unstable 6 months B. Unstable 6 months B. Unstable 6 months
Latvia France Australia
Spain
a. Full unemp a. Full unemp a. Full unemp a. Full unemp
UB_2
b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths
c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths
d. Part-time d. Part-time
C. Unstable 1 month C. Unstable 1 month
d. Part-time d. Part-time
C. Unstable 1 month C. Unstable 1 month
a. Full unemp a. Full unemp a. Full unemp a. Full unemp
Adjusted replacement rates, based on current and previous employment trajectories (in percentage)
b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths b. Uns 6 mths Full unemployment
Figure 2.16. Work incentives depending on current and previous employment trajectories
c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths c. Uns 1 mths
D. Part-time D. Part-time D. Part-time D. Part-time
d. Part-time d. Part-time d. Part-time d. Part-time
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
155
Note: The vertical axis represents adjusted replacement rates (ARR), which expresses the amount of total income (wages and benefits) that a
worker would get relative to the full-time wage. ARR is decomposed in three parts: “wage”, representing worker’s wage relative to the full-time
wage; “UB_1”, representing the entitled amount of first-tier unemployment benefit relative to the full-time wage, and “UB_2”, representing the
amount of second-tier unemployment benefit relative to the full-time wage.All scenarios combine different current and previous employment
trajectories. Each scenario includes seven years. The first six years correspond to the “previous trajectory”. The seventh year corresponds to
the “current trajectory”, as this is the year in which ARR is computed. Previous trajectories consider two initial years of full-time employment paid
at 67% of AW, followed four years with one of the following employment patterns: A. Alternating periods of full-time dependent employment and
full unemployment every 12 months. B. Unstable 6 months and C. Unstable 1 month: situations with job changes between full-time dependent
employment and full unemployment every 6 months and every other month, respectively. D. Part-time: “part-time employment trajectory”, which
consists of recurrent one-year spells seeking for a full-time job, with the first month in full unemployment and the following 11 months at partial
employment (working at 54.5% of full-time hours in working periods). The seventh year repeats the employment pattern of the third year of each
of the four scenarios: a. full unemployment, b. 6 months in unemployment followed by 6 months in employment, c. alternating between
unemployment and employment, and d. one month unemployed followed by 11 months working at 54.4% of the time. See Box 2.5 for more
details about assumptions and simulation scenarios.
Source: Secretariat calculations based on unemployment benefit legislation extracted from OECD tax-benefit model and tailored questionnaire
to national authorities.
StatLink 2 https://stat.link/raoz96
In Latvia, outcomes are to a large extent conditioned by the previous employment trajectory. In the
scenarios considered here, workers with previous unstable employment trajectories would not be entitled
to any unemployment benefit, because minimum contribution conditions requires 12 contributions in the
last 16 months (Figure 2.6). Independently of previous employment trajectories, the worker would also not
receive unemployment benefit if working on a part-time basis, as the Latvian benefit is incompatible with
any work (in scenario D, the small part of ARR related to benefits is due to the month in full unemployment).
Workers with current unstable and previous stable employment would receive benefits with sizeable
replacement rates. A worker switching every month between unemployment and employment would get
an amount equivalent to 79% of a full-time wage. If she did not work at all, her income would decrease by
52% of a full-time wage, as the ARR in full unemployment is 27%. The adjusted replacement rates are
slightly lower for workers current switching every six months between unemployment and employment,
and for workers who previously were in part-time employment.
In Spain, previous employment trajectories also play an important role in determining the amount of future
unemployment benefits. Workers without previous standard employment trajectories would not be entitled
to first-tier unemployment benefit, because minimum contribution conditions and means testing rules (as
discussed in Section 2.3.1). They would be entitled, however, to a few months of the second-tier benefit.
Workers with current and previous part-time employment trajectories would not be entitled to any
unemployment benefit. Conversely, a worker with current unstable and previous stable employment would
receive benefits with sizeable replacement rates (73%). If she did not work at all, her income would
decrease by 37% of a full-time wage, as the ARR in full unemployment is 36%. The adjusted replacement
rate is lower for a worker with current part-time and previous stable employment (64%).
2.4. Policy issues: striking the right balance
This section discusses some key features of unemployment benefit systems and considers potential policy
options to improve their reliability and effectiveness to support workers with trajectories of non-standard
employment. Following the evidence provided in previous sections, the discussion addresses benefit
neutrality, work incentives, activation and protection during an economic downturn. Table 2.3 recaps the
arguments below with a summary of unemployment benefit instruments in terms of their objectives, trade-
offs and effects on workers with non-standard dependent employment trajectories, and possible
alternatives of adjustments in order to strike the right balance between conflicting objectives.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
156
2.4.1. Benefit neutrality
In general, for the countries reviewed in this chapter, unemployment benefits are not neutral with respect
to the distribution of time between dependent employment and unemployment before the benefit is
claimed. Evidence gathered in section 2.3 reveals considerable differences concerning the extent of
protection that unemployment benefits provide to workers who, over the medium run, have worked the
same amount of time.
Strict employment requirements tend to penalise workers with unstable employment trajectories, in
particular when the minimum time in employment is relatively long and the reference period is short.
Unstable employment prolongs the period in which workers need to work in order to fulfil the minimum
requirement, thus hindering their ability to claim the benefit. While employment requirements are important
features of unemployment benefits – they provide work incentives, prevent benefit abuse and protect the
financial sustainability of the system – some adjustments can contribute to reduce such a penalty on
instability. The difference between the length of the reference period and the minimum time in employment
can be designed to be sufficiently long in order to account for unstable employment or specifically extended
for workers more prone to job instability, as in Austria (young workers) and the Slovak Republic (temporary
contracts).
The reduction of benefit amounts with wages from part-time unemployment penalises jobseekers able to
carry out some casual part-time work while searching for an adequate full-time job. In some countries, the
penalty is absolute as unemployment benefits are incompatible with any work. Reducing the amount of
benefits in line with wages is justified by the need to smooth income and avoid the misuse of unemployment
benefits as a sort of wage subsidy. On the other hand, besides neutrality, part-time unemployment may
produce desirable outcomes such as improving jobseekers’ employability through a “stepping-stone effect”
(Kyyrä, Arranz and García-Serrano, 2019[39]). Financial incentives, such as earnings disregards and lower
withdrawal rates, encourage unemployment benefit recipients to take up part-time work by allowing them
to keep some of the additional income received in wages, and can be effective instruments for balancing
objectives. Yet, part-time unemployment benefits may also produce “lock-in effects”, by reducing the exit
rate from part-time to full-time employment. Thus, like in other activation policies, it is important to use
“mixed strategies” that profile and monitor the effect of causal part-time jobs on the long-term employability
of jobseekers (OECD, 2018[40]).
2.4.2. Work incentives
The potential adverse impact of unemployment insurance systems on work incentives is a main concern
in many countries, particularly for those who have been out of work for a long time. Furthermore, ill-
designed unemployment benefits may incentivise unstable employment.
Soft access conditions may generate a carousel of shifts between periods in short-tenure jobs and periods
in unemployment while receiving generous benefit amounts. Waiting periods, moderately long
requirements of minimum time in employment and reductions in case of frequent reclaiming can contribute
to preventing such distortion. Differentiated unemployment insurance contribution rates may also be used
to create financial incentives for employers and employees to choose more stable employment contracts
and discourage collusion (OECD, 2004[41]; OECD, 2019[1]).
Keeping unused benefits for future use may encourage frequent periods of unstable employment (as
discussed above) and incentivise over-extending the duration of unemployment benefits, thus promoting
long-term unemployment. Limits on the possibility of keeping and accumulating previous and new
entitlements can regulate the trade-off between work incentives and acquired rights.
The coordination between in-work and out-of-work support can produce important back-to-work incentives.
One option could be to extend to full-time workers with low earnings the possibility to cumulate
unemployment benefits and income from work as is the case for part-time unemployment (Hijzen and
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
157
Salvatori, 2020[42]). Another possibility would be introducing in-work benefits, like the Earned Income Tax
Credit in the United States or the Prime d’activité in France (Carcillo et al., 2019[43]). Alternatively, in-work
and out-of-work benefits could be integrated in a single comprehensive strategy (e.g. the United Kingdom’s
Universal Credit).
Table 2.3. Striking the right balance by adjusting unemployment benefit’s tools
Unemployment benefit’s tools, objectives, trade-offs and adjustments
Tool Objectives Drawbacks Adjustments
Strict employment requirements. Provide work incentives. Penalty on instability. Adjust employment requirements
Prevent abuse. Non-neutrality. so that the difference between the
Protect the financial sustainability. Lack of Benefit lengths of the reference period
adequacy. and of the minimum time in
employment account for
employment instability.
Customise employment
requirements for workers more
prone to job instability (e.g. young,
temporary contracts, recurrently
unemployed).
Reduction of benefit amounts with Income smoothing. Penalty on part-time Earnings disregards.
wages from part-time Prevent misuse. unemployment. Withdrawal rates below 100%.
unemployment. Working incentives (reduce lock-in Non-neutrality.
effect). Lack of benefit
adequacy.
Working incentives
(stepping-stone effect).
Soft access conditions. Accessibility. Incentivise job Waiting periods.
Neutrality. instability. Moderately long requirement of
Benefit adequacy. Incentivise collusion minimum time in employment.
between employers and Reductions in case of frequent
employees to abuse reclaiming.
system. Differentiated unemployment
insurance contribution rates.
Keeping un-exhausted benefits for Acquired rights. Encourage frequent Limits on the possibility of keeping
future use. Accessibility. periods of unstable and accumulating previous and
Neutrality. employment. new entitlements.
Overextend the duration
of unemployment
benefits.
2.4.3. Activation
Unemployment benefit’s activation requirements also play a role in strengthening incentives to look for,
prepare for, and accept employment. Evidence suggests that unemployment benefit programmes featuring
job-search monitoring and sanctions yield positive employment effects, while overly demanding eligibility
criteria can exclude some intended recipients (Immervoll and Knotz, 2018[44]). Exceedingly strict sanctions
and excessive reliance on job-search incentives could be counterproductive and increase the
unemployment exit rate into unstable jobs. Profiling plays a crucial role. Workers in need of more intensive
activation services (such as training) should be identified and supported (OECD, 2018[40]).
According to the 2018 OECD Jobs Strategy, unemployment benefits can play a pivotal role in the success
of activation strategies through the rigorous enforcement of a “mutual obligations framework” (OECD,
2018[40]). Within a mutual obligations framework, governments have the duty to provide jobseekers with
benefits and effective services to enable them to find work and, in turn, beneficiaries have to take active
steps to find work or improve their employability. The threat of potential sanctions in terms of benefit
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
158
withdrawal significantly increases the financial incentive for seeking and taking up gainful employment as
well as seriously participating in active programmes. Thus, unemployment benefit recipients are referred
to employment services.
Unemployment benefits are key to connect jobless people to active labour market programmes, by
referring them to employment services, which provide job-search assistance or interventions to increase
their employability. In the absence of unemployment (or social assistance) benefits, it is often difficult to
reach out to those facing multiple barriers to employment, who risk being left behind. Similarly, in the
absence of effective active labour market policies, there is a risk that unemployment benefits reduce work
incentives and deepen labour market exclusion. Passive and active policies should therefore be conceived
together rather than in isolation (OECD, 2018[40]).
Increasing the access of workers in non-standard dependent employment to unemployment insurance is,
therefore, a promising avenue for promoting labour market security and inclusiveness, provided it is carried
out together with the rigorous enforcement of a “mutual-obligations” framework to preserve work
incentives.
2.4.4. Economic shock protection
Unemployment benefits play an important role as key instruments of social protection to provide income
support and employment opportunities in all phases of the economic cycle. This role is stronger during and
after an economic shock. By helping individuals and families cope with the consequences of economic
shocks, unemployment benefits contribute to prevent temporary economic problems from turning into
long-term disadvantage for workers and a social crisis for society (OECD, 2014[45]).
Unemployment benefits need to be sufficiently responsive to changes in circumstances, following an
economic shock. In countries where access to, and duration of, unemployment benefits is restricted, there
may be a case for, temporarily, loosening employment and other requirements and extending the
maximum duration of the benefit (e.g. in the Canada and the United States). In a context of fewer job
opportunities, the emphasis of active labour market policies may temporarily shift from work-first to
train-first, in particular for hard-to-place jobseekers (OECD, 2018[40]).
Complementing the overall discussion in Chapter 1 on the challenges coming from the COVID-19 crisis,
Box 2.6 discusses the measures introduced by countries that have revised unemployment benefit
elements in order to extend unemployment protection of non-standard workers during the crisis.
2.5. Concluding remarks
This chapter provides an in-depth review of the key policy mechanisms that affect how unemployment
benefits strike a balance between income security and financial work incentives for those with
non-standard dependent employment trajectories. It presents new evidence on the scale and development
of unstable and part-time dependent employment and the characteristics of workers with such trajectories.
In particular, it shows that workers who experienced non-standard dependent employment trajectories
make up for a considerable part of dependent employment and the majority of the unemployed; that their
share of employees has increased, particularly among young workers; and that their poverty rates are
several times higher than those of workers in standard employment.
The chapter also takes stock of and analysed the legal unemployment benefit provisions that apply to
non-standard employees, and how these might give rise to an uneven treatment of standard and
non-standard employees. Unemployment benefit rules can result in considerable differences between
workers with standard and non-standard dependent employment trajectories in terms of the duration of
benefit entitlements and the level of benefits paid. Some requirements are harder to meet for those in
unstable or part-time employment.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
159
Box 2.6. COVID-19 crisis: Non-standard dependent employment and extraordinary
unemployment benefit measures
The coronavirus crisis has hit hard, with unprecedented job losses around the world. Workers in non-
standard dependent employment are particularly vulnerable against the risks of job and income loss.
Unemployment benefits and related income support are crucial for cushioning these losses. However,
as discussed in the 2019 edition of the OECD Employment Outlook (OECD, 2019[1]) and in this chapter,
workers in non-standard dependent employment are less likely to receive adequate unemployment
support.
Many OECD countries have taken rapid action by adjusting their unemployment benefit systems to the
new labour market circumstances. In line with the OECD Jobs Strategy (OECD, 2018[40]), most
measures adjusted existing benefits by loosening requirements in order to facilitate access and revised
entitlements to extend income support and duration. Some countries have introduced new benefits.
Table 2.4 summarises the extraordinary unemployment benefit measures, with potential impact on non-
standard dependent employment, which were introduced during the COVID-19 crisis between March
and May 2020. Chapter 1 provides a wider analysis of employment- and social-policy responses to the
COVID-19 crisis.
Facilitating access: six OECD countries facilitated access to unemployment benefits. France,
Israel, Spain and Sweden reduced or entirely waived minimum employment requirements to
workers. In Norway, the minimum previous income to be eligibility to unemployment benefits
was reduced. The length of the reference period to assess employment requirements was
extended in Switzerland (doubled) and in France (by disregarding the confinement period for
calculating the entitlements of new claimants). All these measures are, in particular, expected
to facilitate access for workers with unstable trajectories, who often find it difficult to meet
minimum employment requirements.
Extend benefit duration: 13 OECD countries have extended the maximum possible duration
of unemployment benefit payments. Some countries triggered automatic extensions. Other
countries extended the expiring benefit claims until the end of the health crisis (Luxembourg,
Portugal and Spain) or for a specific period (by up to one month in the Slovak Republic,
two months in Greece, three months in Germany, Norway until the end of June, and four months
in Switzerland). Denmark and France froze benefit entitlements for the duration of the health
crisis. In the United States, the Federal Government extended the maximum unemployment
benefit duration to nine months. Since, in some countries, non-standard employees are entitled
to fewer months of unemployment benefits (Figure 2.13), these measures are likely to be
particularly important for them.
Raising unemployment benefit generosity: seven OECD countries temporarily increased
benefit levels. Australia introduced, for six months, a “Coronavirus Supplement” of AUD 550 per
fortnight to recipients of income support payments, including the unemployment benefit
“Jobseeker Payment”. Austria increased the benefit amount of the second-tier unemployment
benefit (Notstandshilfe) to the same level as the first-tier benefit (Arbeitslosengeld). Sweden
raised, until the end of 2020, the unemployment benefit floor (by about 30%) and ceiling (by
about 40%) for 100 days. New Zealand increased the amount of “Jobseeker Support” by
NZD 25 per week. Norway increased the unemployment benefit from 62.4% to 80% of the
calculation basis up to an income of NOK 24 966 per month. The United Kingdom raised the
allowance of the second-tier unemployment benefit (Universal Credit) by GBP 20 per week. The
United States raised the unemployment benefits by USD 600 per week for all recipients for a
maximum period of four months, for up to four months. While all these measures are not
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
160
specifically targeted to non-standard employees, to the extent that those with “naturally”
unstable jobs are likely to bear the brunt of the employment adjustment, they are among the
categories potentially more positively affected by these measures. Moreover, non-standard
employees are likely to benefit the most from the increase in generosity of second-tier benefits
to the extent that the risk of poverty is greater for non-standard employees.
New unemployment benefits: six OECD countries introduced new temporary non-contributory
benefits, which have clearly a greater impact for people with unstable trajectories. In Canada,
the “Canada Emergency Response Benefit” provides CAD 500 a week for up to 16 weeks, to
people who have involuntarily lost their job during the COVID-19 crisis, receive the first-tier
unemployment benefit (Employment Insurance) or recently exhausted its claim. In Ireland, the
new “COVID-19 Pandemic Unemployment Payment” supports individuals who lost their job
during the COVID-19 crisis. Claimants receive EUR 350 per week until at least 8 June. The
payment is fully compatible with other unemployment benefits. Latvia introduced a new
temporary unemployment assistance benefit for unemployment insurance benefit recipients
whose entitlements expire. Benefits are payable for up to four months, at the level equal to the
unemployment insurance benefits paid during the 8th and 9th month (max. EUR 180 per
month). New Zealand created the “COVID-19 Income Relief Payment” for people who lost their
jobs from 1 March 2020 to 30 October 2020 due to COVID-19. The payment amount depends
on previous hours worked, is means-tested against partner’s earnings, is incompatible with
receiving other main benefit (but people are allowed to choose) and is exempt from taxes.
Slovenia introduced a temporary unemployment benefit (Začasno denarno nadomestilo plače
zaradi izgube zaposlitve v obdobju epidemije) for those who lost their jobs in the crisis.
Claimants receive EUR 513.64 per month for the duration of the temporary restrictions; the
benefit is not compatible with the first-tier unemployment benefit.
Table 2.4. COVID-19 extraordinary unemployment benefit measures with potential impact on
non-standard employees
Existing benefits New Details
benefit
Access Entitlement
Reduced Extended Extended Raised
minimum reference benefit benefit
employment period duration amount
Australia “Coronavirus Supplement” of AUD 550 per fortnight. [1]
Austria Notstandshilfe (2nd tier) amount increased to equal level
as the Arbeitslosengeld (1st tier).
Canada Automatic extension of maximum duration according to
regional unemployment rate.
New “Canada Emergency Response Benefit” [2].
Chile Automatic activation of the 6th and 7th monthly
payments of the Seguro de cesantía solidario (2nd tier
benefit). [3]
Denmark Freezing of entitlement period for 3 months.
France Exceptional access for people who voluntarily quit their
job to take up a new job but whose job offer fell through.
The confinement period will not be considered for
calculating the entitlements of new claimants.
Freezing of entitlement period during confinement.
Germany Benefit extension by 3 months for all recipients whose
entitlements end between May and December
(announced).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
161
Greece Benefit extension by 2 months for all recipients whose
entitlements end in the first quarter of 2020.
Ireland New “COVID-19 Pandemic Unemployment Payment”
for individuals who lost their job in the COVID-19 crisis.
[4]
Israel Reduction in the required employment history from 12
to 6 months.[5]
Latvia New temporary unemployment assistance benefit for
recipients of unemployment insurance benefit whose
entitlements expire. Benefit level equal to
unemployment insurance paid during the 8th and 9th
months (max. EUR 180 per month), payable for up to
4 months.
Luxembourg Benefit extension for recipients whose entitlements
expire during the state of crisis.
New Zealand Amount of “Jobseeker Support” increased NZD 25 per
week.
Created “COVID-19 Income Relief Payment” for people
who lost their jobs from 1 March 2020 to 30 October
2020 due to COVID-19 [6].
Norway Temporarily reduced minimum income for eligibility to
unemployment benefits. [7]
Prolonged benefit duration until end of June for
recipients who had 18 weeks or less of remaining.
Benefit replacement rate increased from 62.4% to 80%
for incomes lower than NOK 24 966 per month.
Portugal Benefit extension until the end of the confinement
period.
Slovak Benefit extension by 1 month for recipients whose
Republic entitlement expire during the crisis.
Slovenia New temporary unemployment benefit for those who
lost their jobs in the crisis. [8]
Spain Suspension of the minimum contribution period.
Benefit extension until the end of the health crisis.
Sweden Reducing employment requirements. [9]
Increased the benefit floor and ceiling. [10]
Switzerland The reference period to assess employment condition
was doubled to 48 months.
Benefit extension by 120 days.
United Raised weekly Universal Credit allowances (2nd tier
Kingdom benefit) by GBP 20.
United Extension of maximum benefit duration to 9 months.
States* Benefit increase (“Pandemic Unemployment
Assistance”) by USD 600 per week for up to 4 months.
Note: * Information for the United States refers to the federal level.
[1] To recipients of income support payments, including the unemployment benefit “Jobseeker Payment”. The supplement will be paid for
6 months.
[2] For newly unemployed and people who receive Employment Insurance or recently exhausted their claim. CERB provides USD 500 a
week for up to 16 weeks. Newly unemployed must have involuntarily lost their job in the COVID-19 crisis. Claimants need to have earned
at least CAD 5 000 in 2019 and may not have earned more than CAD 1 000 for the four-week period of the benefit claim.
[3] Maximum duration extension triggered as the national unemployment rate is 1 percentage point above its 4-year mean.
[4] Claimants receive EUR 350 per week until at least June, 08. Fully compatible with 1st and 2nd tier benefits.
[5] Claimants who only fulfil the shortened employment condition receive half the number of benefit days.
[6] The Income Relief Payment is paid for up to 12 weeks. The payment is NZD 490 per week if the person was previously working 30 hours
or more a week, and NZD 250 per week if she was previously working 15 hours to 29 hours a week. The payment is means-tested against
partner’s earnings. A person cannot get the benefit if the partner earns NZD 2 000 or more per week (before tax). It is incompatible with
receiving other main benefit (but people are allowed to switch to it if they meet the criteria for who can get it). The payment is tax exempt.
[7] Minimum income reduced from NOK 149 787 to 74 894 over the past 12 months, or from NOK 299 574 to 224 680 over the past
36 months.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
162
[8] Začasno denarno nadomestilo plače zaradi izgube zaposlitve v obdobju epidemije is available for those who lost their jobs in the crisis
for economic motives or due to the end of a temporary contract. Claimants receive EUR 513.64 per month for the duration of the temporary
restrictions. The benefit is not compatible with the first-tier unemployment benefit.
[9] Shortening of the required membership period in the unemployment insurance fund from 12 to 3 months (1st tier); reduction of the
employment requirement from at least 80 hours per months in last 6 months or 480 hours in last 6 months and at least 50 hours per month
to 60 hours per months in last 6 months or 420 hours in last 6 months and at least 40 hours per month (1st and 2nd tier).
[10] minimum benefit amount from SEK 365 to 510 per day and maximum benefit amount from SEK 910 to 1 200 per day for the first
100 days (1st tier). All rules until end of 2020.
Source: Chapter 1 and OECD (2020[46]), “Supporting people and companies to deal with the COVID-19 virus: Options for an immediate
employment and social-policy response”, ELS Policy Brief on the Policy Response to the COVID-19 Crisis, OECD, Paris,
http://oe.cd/covid19briefsocial.
Simulations of some policy-relevant employment trajectories for four OECD countries (Australia, France,
Latvia and Spain) show that unemployment benefits provide different levels of support depending on the
employment trajectory. Such differences are evident even when workers have the same personal
characteristics, have earned the same amount of wages and have been in and out of work the same
number of hours over a given period. In other words, the systems are not neutral in terms of social
protection and incentives. Jobseekers with unstable employment trajectories (working either every other
month or every other semester) are entitled to the same number of months and the same amount of
unemployment benefits in Australia and France, but considerably less in Latvia and Spain, where
contribution conditions are more demanding. In all of the four countries, workers with trajectories that
include periods combining unemployment and partial work are entitled to lower unemployment benefit than
workers with standard employment trajectories.
The chapter shows that there are several policy tools that can be deployed to design unemployment benefit
systems that strikes the right balance between the sometimes conflicting objectives of income security and
work incentives for those with unstable employment histories. For example:
Customised extensions of employment reference periods could be used for groups who are more prone to
job instability, such as young workers and those on temporary contracts.
Earnings disregards and withdrawal rates applied to the benefits of jobseekers with earnings from casual
or part-time work can be calibrated to enable such work to be used as a stepping stone to a better job.
Waiting periods, moderate minimum employment duration requirements, reductions in case of frequent
reclaiming and limits on the accumulation of old and new entitlements can be used to reduce incentives to
choose an unstable employment trajectory.
Differentiated unemployment insurance contribution rates could create financial incentives for employers
and employees to choose more stable employment contracts and deter them from colluding in unstable
arrangements that reduce workers’ pay and maximise benefit entitlements.
Finally, the integration or at least improved co-ordination of in-work and out-of-work benefits can enhance
work incentives. Activation measures can also strengthen incentives to seek and accept stable
employment.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
163
References
Avlijas, S. (2019), “The Dynamism of the New Economy: Non-Standard Employment and Access [4]
to Social Security in EU-28”, SSRN Electronic Journal,
http://dx.doi.org/10.2139/ssrn.3354463.
Baker, M., J. Gruber and K. Milligan (2015), “Non-cognitive deficits and young adult outcomes: [15]
the long-run impacts of a universal child care program”, NBER Working Paper, No. 21571,
National Bureau of Economic Research, http://www.nber.org/papers/w21571.
Berlinski, S. and S. Galiani (2007), “The effect of a large expansion of pre-primary school [16]
facilities on preschool attendance and maternal employment”, Labour Economics, Vol. 14/3,
pp. 665-680, http://dx.doi.org/10.1016/j.labeco.2007.01.003.
Boeri, T., P. Cahuc and A. Zylberberg (2015), “The Costs of Flexibility-Enhancing Structural [6]
Reforms: A Literature Review”, OECD Economics Department Working Papers, No. 1264,
OECD Publishing, Paris, http://dx.doi.org/10.1787/5jrs558c5r5f-en.
Booth, A. and J. van Ours (2013), “Part-time jobs: What women want?”, Journal of Population [20]
Economics, Vol. 26/1, pp. 263-283, http://dx.doi.org/10.1007/s00148-012-0417-9.
Bradshaw, J. and F. Bennett (2017), ESPN Thematic Report on Access to social protection of [30]
people working as self-employed or on non-standard contracts United Kingdom,
https://ora.ox.ac.uk/objects/uuid:f6d86772-fba0-4088-858c-9e96a9b1fcbf.
Bureau of Labor Statistics (2020), Glossary : U.S. Bureau of Labor Statistics, [33]
https://www.bls.gov/bls/glossary.htm#D (accessed on 23 April 2020).
Cahuc, P. (2018), “Wage Insurance, Part-Time Unemployment Insurance and Short-Time Work [24]
in the XXI Century”, IZA Discussion Papers, No. 12045, IZA, http://ftp.iza.org/dp12045.pdf.
Cahuc, P. et al. (2019), “Taxation of Temporary Jobs: Good Intentions with Bad Outcomes?”, [25]
The Economic Journal, Vol. 130/626, pp. 422-445, http://dx.doi.org/10.1093/ej/uez062.
Carcillo, S. et al. (2019), “Assessing recent reforms and policy directions in [43]
France: Implementing the OECD Jobs Strategy”, OECD Social, Employment and Migration
Working Papers, No. 227, OECD Publishing, Paris, https://dx.doi.org/10.1787/657a0b54-en.
Carta, F. and L. Rizzica (2018), “Early kindergarten, maternal labor supply and children’s [17]
outcomes: Evidence from Italy”, Journal of Public Economics, Vol. 158, pp. 79-102,
http://dx.doi.org/10.1016/j.jpubeco.2017.12.012.
Chetty, R., J. Friedman and E. Saez (2013), “Using differences in knowledge across [29]
neighborhoods to uncover the impacts of the EITC on earnings”, American Economic Review,
Vol. 103/7, pp. 2683-2721, http://dx.doi.org/10.1257/aer.103.7.2683.
Eurostat (2018), Methodological Guidelines and Description of EU-SILC Target Variables. [11]
(Version August 2017).
Farber, H. (2017), “Employment, Hours, and Earnings Consequences of Job Loss: US Evidence [32]
from the Displaced Workers Survey”, Journal of Labor Economics, Vol. 35/S1, pp. S235-
S272, http://dx.doi.org/10.1086/692353.
Fernández, R., H. Immervoll and D. Pacifico (forthcoming), “Beyond repair? Anatomy of income [23]
support for standard and non-standard workers in OECD countries”, OECD Social,
Employment and Migration Working Papers, OECD Publishing, Paris.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
164
Fontaine, F. and F. Malherbet (2016), CDD vs CDI : les effets d’un dualisme contractuel, [9]
Presses de Sciences Po, Paris.
Hijzen, A. and A. Salvatori (2020), “Designing fair and work-oriented unemployment [42]
benefits: The case of Belgium”, OECD Social, Employment and Migration Working Papers,
No. 237, OECD Publishing, Paris, https://dx.doi.org/10.1787/ac17d171-en.
Immervoll, H. and C. Knotz (2018), “How demanding are activation requirements for jobseekers”, [44]
OECD Social, Employment and Migration Working Papers, No. 215, OECD Publishing, Paris,
https://dx.doi.org/10.1787/2bdfecca-en.
Kyyrä, T. (2010), “Partial unemployment insurance benefits and the transition rate to regular [7]
work”, European Economic Review, Vol. 54/7, pp. 911-930,
http://dx.doi.org/10.1016/j.euroecorev.2010.02.005.
Kyyrä, T., J. Arranz and C. García-Serrano (2019), “Does subsidized part-time employment help [39]
unemployed workers to find full-time employment?”, Labour Economics, Vol. 56, pp. 68-83,
http://dx.doi.org/10.1016/j.labeco.2018.12.001.
Le Barbanchon, T. (2016), “Optimal Partial Unemployment Insurance: Evidence from Bunching [8]
in the U.S”, undefined, https://www.semanticscholar.org/paper/Optimal-Partial-
Unemployment-Insurance%3A-Evidence-in-
Barbanchon/1ba1750d1779a5ff1c9b867c08413ac5efc393bc (accessed on 12 May 2020).
Manning, A. and B. Petrongolo (2008), “The Part‑Time Pay Penalty for Women in Britain”, The [18]
Economic Journal, Vol. 118/526, pp. F28-F51, http://dx.doi.org/10.1111/j.1468-
0297.2007.02115.x.
Mas, A. and A. Pallais (2019), “Alternative Work Arrangements”, IRS Working Papers, No. 634, [19]
http://arks.princeton.edu/ark:/88435/dsp01bk128d78n (accessed on 7 April 2020).
Meyer, B. and D. Rosenbaum (2001), “Welfare, the Earned Income Tax Credit, and the Labor [28]
Supply of Single Mothers”, The Quarterly Journal of Economics, Vol. 116/3, pp. 1063-1114,
http://dx.doi.org/10.1162/00335530152466313.
OECD (2020), Part-time employment rate (indicator), https://dx.doi.org/10.1787/f2ad596c-en [2]
(accessed on 12 March 2020).
OECD (2020), “Supporting people and companies to deal with the Covid-19 virus: Options for an [46]
immediate employment and social-policy response”, OECD Policy Responses to Coronavirus
(COVID-19), OECD Publishing, Paris, http://www.oecd.org/coronavirus/policy-
responses/supporting-people-and-companies-to-deal-with-the-covid-19-virus-options-for-an-
immediate-employment-and-social-policy-response-d33dffe6/.
OECD (2020), TaxBEN: The OECD tax-benefit simulation model - Methods and policy [38]
applications, OECD, https://taxben.oecd.org/tax-ben-
resources/OECD_TaxBEN_methodology.pdf.
OECD (2020), Temporary employment (indicator), https://dx.doi.org/10.1787/75589b8a-en [3]
(accessed on 12 March 2020).
OECD (2019), OECD Employment Outlook 2019: The Future of Work, OECD Publishing, Paris, [1]
https://dx.doi.org/10.1787/9ee00155-en.
OECD (2019), Taxing Wages 2019, OECD Publishing, Paris, [35]
https://dx.doi.org/10.1787/tax_wages-2019-en.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
165
OECD (2018), Good Jobs for All in a Changing World of Work: The OECD Jobs Strategy, OECD [40]
Publishing, Paris, https://dx.doi.org/10.1787/9789264308817-en.
OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris, [5]
https://dx.doi.org/10.1787/empl_outlook-2018-en.
OECD (2017), The Pursuit of Gender Equality: An Uphill Battle, OECD Publishing, Paris, [13]
https://dx.doi.org/10.1787/9789264281318-en.
OECD (2015), In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris, [22]
https://dx.doi.org/10.1787/9789264235120-en.
OECD (2014), OECD Employment Outlook 2014, OECD Publishing, Paris, [10]
https://dx.doi.org/10.1787/empl_outlook-2014-en.
OECD (2014), Society at a Glance 2014: OECD Social Indicators, OECD Publishing, Paris, [45]
https://dx.doi.org/10.1787/soc_glance-2014-en.
OECD (2013), “Back to work: Re-employment, earnings and skill use after job displacement”, in [12]
OECD Employment Outlook 2013, OECD Publishing, Paris,
https://dx.doi.org/10.1787/empl_outlook-2013-8-en.
OECD (2010), “How Good is Part-Time Work?”, in OECD Employment Outlook 2010: Moving [14]
beyond the Jobs Crisis, OECD Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-
2010-5-en.
OECD (2010), OECD Employment Outlook 2010: Moving beyond the Jobs Crisis, OECD [21]
Publishing, Paris, https://dx.doi.org/10.1787/empl_outlook-2010-en.
OECD (2007), Benefits and Wages 2007: OECD Indicators, OECD Publishing, Paris, [37]
https://dx.doi.org/10.1787/ben_wages-2007-en.
OECD (2005), OECD Employment Outlook 2005, OECD Publishing, Paris, [36]
https://dx.doi.org/10.1787/empl_outlook-2005-en.
OECD (2004), OECD Employment Outlook 2004, OECD Publishing, Paris, [41]
https://dx.doi.org/10.1787/empl_outlook-2004-en.
OECD (2002), OECD Employment Outlook 2002, OECD Publishing, Paris, [31]
https://dx.doi.org/10.1787/empl_outlook-2002-en.
Saez, E., M. Matsaganis and P. Tsakloglou (2012), “Earnings Determination and Taxes: [26]
Evidence From a Cohort-Based Payroll Tax Reform in Greece”, The Quarterly Journal of
Economics, Vol. 127/1, pp. 493-533, http://dx.doi.org/10.1093/qje/qjr052.
Saez, E., B. Schoefer and D. Seim (2019), “Payroll taxes, firm behavior, and rent sharing: [27]
Evidence from a young workers’ tax cut in Sweden”, American Economic Review, Vol. 109/5,
pp. 1717-1763, http://dx.doi.org/10.1257/aer.20171937.
United Nations Statistical Division (2008), International Standard Industrial Classification of All [34]
Economic Activities (ISIC), United Nations Publications.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
166
Notes
1
The impact of the demographic change on the incidence of standard and non-standard employment is
mainly due to an “age effect” rather than a “cohort effect”. Even among young workers, the incidence of
standard employment increases as a cohort ages.
2
Average monthly wages were computed dividing the annual wage of the current year by the number of
months in work. Since people experiencing many transitions are unlikely to have changed status at the
first or last calendar day of each month, each month in work is likely to correspond to fewer hours of work
than, say, a standard employee staying in the same job the whole year. This is particularly the case for
those that are more employable and therefore are less likely to spend most of the month out of work.
Therefore, the prevalence of unstable employment among below-median wages could in part driven by
the underestimation of their wages.
3
Previous OECD evidence suggests, however, that the average number of hours worked by part-timers
does not explain much of the cross-country difference in poverty penalties (OECD, 2010[21]).
4
Results from Table 2.1 also suggest that workers with unstable employment have lower wages, although
these results control for the number of months and not for the number of hours worked.
5
Jobseekers who are not entitled to first- or second-tier unemployment benefits may be protected by
safety-net programmes, such as social assistance benefits, with neither employment nor employment-
availability requirements.
6
Some countries have unemployment benefits available for jobseekers not entitled to first-year and
second-tier unemployment benefits. In Spain, for example, Active Integration Income (Renta Activa de
Inserción) and Extraordinary unemployment allowance (Subsidio extraordinario por desempleo) are aimed
at jobseekers in long-term unemployment who have extinguished their access to unemployment insurance
and unemployment assistance benefits.
7
All benefit amounts assume a single individual without children and not eligible to any supplement. For
other family circumstances, see annual OECD tax-benefit country reports.
8
In 2020, the minimum base of social contributions in Austria was 461 EUR per month.
9
In France, the benefit amount is reduced only for jobseekers with previous high wages (i.e. those who
earned more than EUR 4 500 per month on average before unemployment).
10
In Latvia, one can suspend the benefit for up to two months of work without losing entitlement.
11
In Canada, entitlements are only retained if they were suspended due to high earnings from partial work.
In the Netherlands, entitlements are only kept if the new job provides the same amount of working hours
as the job prior to unemployment or if earnings from the new job are below 87.5% of the unemployment
benefit level. In the United Kingdom, entitlements are kept for 12 weeks.
12
In Figure 2.9, Figure 2.10 and Figure 2.12 (UB_2), waiting periods can be identified as the first and lower
amount of benefit received over a period. Waiting periods shift the time of receipt, but not the potential
duration and potential total amount of payments, which remain the same. Yet, in comparison to the case
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
167
without waiting period, the effective duration and effective total amount may be lower if the jobseeker finds
a job before exhausting the benefit.
13
In Figure 2.11 and Figure 2.12 (UB_1), reductions due to the length of the receipt periods can be
identified as the lower amounts once the benefit has been received for a few months.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
168
3 Recent trends in employment
protection legislation
Dismissal and hiring regulations – or employment protection legislation in
short – are an important determinant of worker security and firm
adaptability. This chapter provides an up-to-date review of employment
protection legislation in OECD countries, building on earlier work by the
OECD in the area. Taking into account legislation and actual practices, it
describes the regulation of individual and collective dismissals of workers
on regular contracts and the regulation for hiring workers on temporary
contracts. It also discusses recent reforms in employment protection
legislation. The comparison of employment protection across countries in
this chapter brings evidence to the policy debate on the relative importance
that different systems attach to the twin aspirations of protecting workers
and promoting adaptable labour markets.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
169
In Brief
Key findings
Employment protection legislation is at the heart of the contractual working relationship between a firm
and its employees. For the firm, it influences the ability to attract new staff and dismiss workers to react
to changes in economic conditions and technology. For the employees, it is an important element of the
stability of their job. Hiring and dismissal regulations therefore affect the rights and well-being of every
employee in the labour market. As this chapter explains, one motivation for employment protection
legislation is to induce firms to at least partially internalise the social costs of their hiring and dismissal
decisions, in terms of unemployment benefit costs, psychosocial distress and income shocks.
The OECD has a long history of being at the forefront of international comparisons of employment
protection legislation, also by providing comparable indicators via the OECD Employment Protection
Legislation Database. Over the past three decades, these indicators have become one of the most
widely used sources for benchmarking labour market regulation. The chapter extends the database, last
updated in 2013, to 2019 and describes recent trends in employment protection in OECD countries.
Besides extending past series, the new indicators allow for a more systematic and accurate assessment
of different types of job protection provisions, due to methodological improvements and the inclusion of
important aspects that were not considered previously. For example, collective dismissals
(i.e. dismissals of several workers) are now evaluated using the same methodology as the one for
individual dismissals, and enforcement issues enter the quantitative comparisons in a substantial way
for the first time. The new design of the indicators ensures better comparisons between countries and
better assessments of reforms.
The main findings are as follows:
The OECD indicators measure job dismissal regulations along four dimensions: i) procedural
requirements before notice is given; ii) notice period and severance pay; iii) the regulatory
framework for unfair dismissals; and iv) enforcement of unfair dismissal regulation. The new
dimension on enforcement takes account of policies that, by explicitly limiting the scope for unfair
dismissal complaints, ease de facto dismissal regulations. This can be the case with advance
validations of dismissals and pre-termination resolution mechanisms (that allow a worker to
leave an employer by mutual agreement or resignation without losing the right to unemployment
benefits).
Job dismissal regulations exhibit large differences across OECD countries: English-speaking
countries are among those with fewer restrictions on dismissals, so that the layoff risk for workers
is higher. Many European Union (EU) countries as well as a few non-EU countries have more
restrictions on dismissals and high job security for workers on regular contracts.
Countries with strict job protection provisions for regular workers usually have strict hiring laws
for workers on fixed-term, or temporary work agency, contracts. Strict job protection provisions
for regular workers require strict hiring laws for temporary workers to limit labour market duality
and segmentation (the degree to which firms substitute less flexible regular contracts with more
flexible temporary contracts).
The granular comparisons of employment protection in this chapter highlight which elements of
job dismissal regulations play a particularly important role in the different OECD countries. The
chapter shows for example where notice periods and severance pay are the highest (in Turkey,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
170
Lithuania and Israel for workers who have been in their job for four years), where there are
relatively fewer rights relating to unfair dismissals (in the United States, the United Kingdom and
Canada) and where advance validations of dismissals and pre-termination resolution
mechanisms are strongest (in Austria and the Netherlands).
Thirty-two of the 37 OECD countries impose more restrictions on collective dismissals than on
individual dismissals, mostly because of stricter consultation requirements before notice can be
given. These higher restrictions reflect the greater challenge for the economy of dealing with a
collective dismissal. Nevertheless, pooling several individual layoffs in one collective dismissal
can, in some cases, reduce the administrative burden of the firm.
In the aftermath of the global financial crisis, a number of countries – including Greece, Portugal
and several other EU countries – had eased strict dismissal regulations for regular workers to
lower dualism in the labour market. In the following years from 2013 to 2019, the period under
study in this chapter, 12 OECD countries reformed job dismissal regulations for regular workers,
while 17 OECD countries reformed hiring regulations for temporary workers. Some reforms had
as objective to reduce the stringency of employment protection against dismissals (as in France,
Italy, Lithuania and Slovenia). A second category of reforms focused specifically on hiring
regulations for temporary workers. Belgium and the Netherlands aligned dismissal regulation for
different types of workers and dismissals. Among the countries that undertook a reform during
2013-19, more countries relaxed dismissal regulations for regular workers than strengthened
them. Countries reforming hiring regulations for temporary workers were evenly split between
those that reduced restrictions on temporary contracts and those that imposed additional
restrictions on them. In the current COVID-19 health and economic crisis, several EU countries
have taken temporary action to considerably strengthen protection against dismissals.
Introduction
Regulations on the hiring and dismissal of employees – or employment protection legislation (EPL) in
short – are one of the most discussed areas of labour market policy. They are an important policy
intervention in the labour market as they usually affect every single employee and every firm, contrary to
sector- or region-specific policies, for example. They are also important because they influence job security
for employees (the risk of being dismissed and the chances of moving into a job) and firm adaptability (the
scope for firms to respond swiftly to changing demand and new technologies). Losing a job and finding a
job are pivotal moments in many people’s lives, making job protection provisions a key determinant of
well-being.
Employment protection is the central policy pillar that supports worker security, alongside publicly funded
policies such as unemployment benefits, short-time work schemes and active labour market programmes
(see Chapters 1 and 2). Well-designed dismissal regulations are motivated by the desire to protect workers
against arbitrary dismissals and to have the company dismissing a worker bear some of the social costs
(in particular the fiscal, psychological and health costs) of the dismissal (Pissarides, 2010[1]; Scarpetta,
2014[2]). At the same time, dismissal provisions tend to preserve existing jobs, rather than help workers
move into better jobs; they therefore go against the often expressed principle that policy should “protect
workers, not jobs”. Another consideration is the interaction between regulations for regular contracts and
regulations for temporary contracts, which influences who has access to stable jobs. These different
functions and effects make the design of employment protection an important, complex matter.
This chapter provides a comprehensive overview of employment protection legislation in OECD countries
and recent reforms in employment protection legislation. The chapter builds on and refines earlier streams
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
171
of work in the area by the OECD (Grubb and Wells, 1993[3]; OECD, 1999[4]; OECD, 2004[5]; OECD, 2013[6];
Venn, 2009[7]). These lines of work have fed into the OECD Employment Protection Legislation Database
and associated research projects. Governments, organisations and researchers have used the OECD
indicators extensively for cross-country benchmarking in reports, books and academic papers. The last
vintage of the indicators dates from 2013, so the present analysis extends the database by six years to
2019.
The chapter is organised as follows. Section 3.1 provides an overview of the ways in which job protection
matters for labour market outcomes. Section 3.2 describes the design of the OECD Employment Protection
Legislation indicators and how they aim to capture the main elements of job protection. It also presents the
key aspects of the revision of the indicators, made necessary to reflect ongoing changes in the labour
market since the last revision in 2008. Section 3.3 describes employment protection legislation in all
37 OECD countries. Section 3.4 looks at employment protection reforms in the OECD during 2013-19 and
how they are reflected in the indicators. The new data are available for free download at http://oe.cd/epl.
3.1. How job protection matters for labour market and economic outcomes
The rationale for job dismissal regulations is mainly twofold: to protect workers against arbitrary dismissals
and to have the firm dismissing a worker carry some of the social costs of the dismissal (Cahuc, Carcillo
and Zylberberg, 2014[8]). The need to protect workers against arbitrary dismissals is especially relevant for
firms facing high labour supply relative to their labour demand. In such situations, some firms may undercut
labour standards and threaten their workers that, if they are unhappy, they will be replaced. Dismissal
regulations also reduce another possibly excessive motive for firms to dismiss a worker that can arise
when the firm does not take into account the consequences of the job separation for fiscal revenues (due
to lower labour income) and fiscal costs (higher expenditure on unemployment benefits). Regulation can
also induce employers to internalise the health consequences of dismissals (Bassanini and Caroli, 2015[9])
and the destruction of human capital following job loss and joblessness (Neal, 1995[10]). A further rationale
for dismissal regulation, in particular an advance notice period, is that it helps workers, potentially with the
early support of the public employment services, to smooth the transition to the next job (OECD, 2018[11]).
The intended effect of the two main motivations for job dismissal regulation – protecting workers against
arbitrary dismissals and having the firm bear some of the dismissal costs – is that layoffs are less frequent
than they would be in the absence of regulation. Economic theory predicts that job dismissal regulation
also reduces hiring, since firms anticipate the higher layoff costs already at the time of hiring and because
workers’ opportunity cost of moving to another job is higher. Different models have obtained the result that
job dismissal regulations reduce both hiring and layoffs and hence job and worker flows (Bentolila and
Bertola, 1990[12]; Garibaldi, 1998[13]; Mortensen and Pissarides, 1994[14]; Nickell, 1978[15]).1 Some reduction
in hiring and layoffs as a consequence of job dismissal regulation is desirable to avoid excessive worker
turnover; overly strict regulation can, however, reduce hiring and layoffs below their optimal level.
A large number of empirical studies, both cross-country and single-country analyses, confirm that dismissal
regulation lowers job and worker flows (Autor, Donohue and Schwab, 2006[16]; Boeri and Jimeno, 2005[17];
Gielen and Tatsiramos, 2012[18]; Haltiwanger, Scarpetta and Schweiger, 2014 [19]; Marinescu, 2009[20];
Micco and Pagés, 2006[21]; Millán et al., 2013[22]; OECD, 2010[23]; Salvanes, 1997[24]). Dismissal regulation
is a multifaceted concept, which the design of the OECD Employment Protection Legislation indicators in
this chapter will take account of. Among the various elements of job protection, the factor that has been
found to reduce labour market fluidity the most is the regulatory framework for unfair dismissals (Bassanini
and Garnero, 2013[25]). Of the features defining unfair dismissal regulation, long trial periods and strict
reinstatement rules seem to play the most important role.
Job dismissal regulation therefore reduces job creation and job destruction. It also seems to lengthen the
duration of unemployment, as most studies, including those relying on well-identified natural experiments,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
172
find that dismissal protection for regular-contract workers has no or a small negative effect on employment
and that unemployment is little affected as well. These are the conclusions of the OECD Jobs Strategy
(OECD, 2018[26]) and several literature surveys (Boeri, 2011[27]; Martin and Scarpetta, 2012[28]; OECD,
2013[6]). One case when employment may be lower is when job protection increases step-wise with job
tenure (i.e. with large hikes at specific seniority levels), as this may encourage firms to anticipate layoffs
before their cost becomes too high (Cahuc, Malherbet and Prat, 2019[29]; García Pérez and Osuna,
2014[30]).2 Moreover, in declining sectors strict employment protection may salvage jobs with no or little
effect on hiring (Messina and Vallanti, 2007[31]), and in a macroeconomic downturn reforms relaxing
dismissal regulation may depress employment, at least temporarily (Bassanini and Cingano, 2019[32];
OECD, 2016[33]).
The main downside of overly strict dismissal regulation is that, by lowering job and worker flows, it tends
to make labour markets less adaptable to economic change, with too little worker movement from declining
towards fast-growing businesses and reduced entry and exit of firms. Several empirical papers confirm
that strict dismissal regulation dampens the scope for productivity-enhancing worker reallocation from low-
to high-productivity firms (Andrews and Cingano, 2014[34]; Bottasso, Conti and Sulis, 2017[35]; Bravo-
Biosca, Criscuolo and Menon, 2016[36]). It can thus weaken labour productivity growth and slow economic
development.
Economy-wide labour productivity growth is determined not only by worker reallocation, but also by within-
firm productivity growth, which depends on investment and innovation at a given firm. In principle, dismissal
regulation can increase or decrease investment and innovation. Investment and innovation might be lower
because actual and anticipated adjustment costs for firms are greater. However, investment could also be
higher if firms substitute more flexible capital for less flexible labour. Innovation, too, could be higher to the
extent that dismissal regulation reduces the risk of being laid off, making it more likely for employees with
innovative ideas that they can reap the rewards for their initiative. The empirical evidence on the effects
on investment is mixed (Autor, Kerr and Kugler, 2007[37]; Bai, Fairhurst and Serfling, 2020[38]; Cingano
et al., 2010[39]; Cingano et al., 2016[40]), while most studies find that overly strict dismissal regulation is
linked with fewer innovative activities and weaker multifactor productivity growth (Bartelsman, Gautier and
De Wind, 2016[41]; Bassanini, Nunziata and Venn, 2009[42]; Bjuggren, 2018[43]; Griffith and Macartney,
2014[44]; Murphy, Siedschlag and McQuinn, 2017[45]).3
Summarising the effects of job dismissal regulations discussed so far: strict dismissal regulation tends to
reduce layoffs, which is a direct result of its intended effect to raise the costs of dismissals. It also tends to
reduce hiring, as firms factor in the higher costs for a potential dismissal already at the time of hiring.
Dismissal regulation therefore reduces both flows out of jobs and flows into jobs. Aggregate employment
and unemployment do not appear to be affected much. Fewer job flows means lower risk of job loss, which
to a certain extent is a good outcome as it countervails an otherwise inefficiently high dismissal rate. When
job protection is too high, however, efficient job allocation and innovation are likely to suffer. Hence, overly
strict dismissal regulation tends to reduce productivity growth and increase the duration of unemployment
spells.
One consequence of the negative effect of strict dismissal regulations on productivity growth is that strict
dismissal regulation also limits the scope for pay increases, given that, at least to some degree, wage
developments are tied to productivity developments. Higher dismissal costs may also dampen wage levels
as they add to the total expected labour costs for firms which know that they will dismiss some workers.
These negative effects of strict dismissal regulations on pay and pay increases may be countervailed by
dismissal regulations increasing the bargaining power of workers and therefore the labour share – see the
evidence from cross-country analysis and laboratory experiments (Ciminelli, Duval and Furceri, 2018[46];
Falk, Huffman and Macleod, 2015[47]). Moreover, the literature points to a distinction between newly hired
workers and already employed workers: a higher stringency of dismissal regulations has been found to
lower wages of new hires (Leonardi and Pica, 2013[48]), while raising wages of incumbent workers (Martins,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
173
2009[49]; van der Wiel, 2010[50]). One question for future research is whether stricter job dismissal regulation
influences the pace of automation, which in turn would affect productivity and wage growth.
An important difference needs to be drawn between workers on regular contracts and workers on
temporary contracts. Workers on regular contracts usually benefit from greater employment protection than
workers on temporary contracts. The evidence indicates that, the larger is the gap in employment
protection between these two contractual forms of work, the more firms use temporary contracts (Centeno
and Novo, 2012[51]; Hijzen, Mondauto and Scarpetta, 2017[52]; Kahn, 2010[53]). Youth, women and the low
skilled tend to be the population groups for which temporary work relationships are particularly common.
Large use of temporary jobs can amplify the increase in unemployment during a business cycle downturn
(OECD, 2017[54]). Simultaneously strict dismissal regulations for regular workers tend to be positive for the
resilience of the labour market initially, but can hinder job creation in the subsequent recovery.
Larger duality – in the sense of a segmented labour market between highly protected workers on regular
contracts and little protected workers on temporary contracts – has been shown to be associated with
weaker productivity levels and growth rates (Bassanini, Nunziata and Venn, 2009 [42]; Cahuc, Charlot and
Malherbet, 2016[55]; Damiani, Pompei and Ricci, 2016[56]; Dolado, Ortigueira and Stucchi, 2016[57]; Hijzen,
Mondauto and Scarpetta, 2017[52]). One reason is that the limited scope for career advancement in the firm
for people on temporary jobs tends to reduce their commitment to the job and hence their incentives to
invest in firm-specific knowledge and skills.4 Another potential reason is that duality induces an inefficiently
high share of temporary workers whose employment spell is too short to exploit all production opportunities
of the firm.
A deeper divide between workers on regular contracts and others on temporary contracts has also been
found to be associated with worse working environments, weaker job stability and greater wage inequality
(García-Pérez, Marinescu and Vall Castello, 2018[58]; OECD, 2011[59]; OECD, 2014[60]). In addition, it can
have negative effects from one generation to the next: children with fathers on a temporary contract have
been shown to be more likely to drop out of the education system and be unemployed than children with
fathers on a regular contract (Ruiz-Valenzuela, 2020[61]).
3.2. The design of the 2019 OECD Employment Protection Legislation indicators
Many of the papers referenced in the previous section rely on earlier vintages of the OECD Employment
Protection Legislation Database. The indicators have been used to investigate the effects of employment
protection on worker flows, employment, productivity growth, wages, investment, resilience to a downturn,
the extent of use of temporary work, wage inequality, subjective job security and political economy aspects.
They have also served as a control variable or descriptive tool in many other papers. Denk and Georgieff
(forthcoming[62]) survey the academic papers that have used the OECD Employment Protection Legislation
Database. They refer as well to some of the numerous policy reports that have drawn on the indicators –
by the OECD (e.g. OECD Economic Outlook, OECD Economic Surveys), national governments and
supranational bodies and institutions (e.g. European Commission and International Monetary Fund). The
wide use of the database underlines its importance for informing and influencing the setting of job dismissal
and hiring regulations.
Besides legislation, the OECD indicators quantify also actual practices, by considering court rulings and
collective bargaining agreements. The main distinction in the database is between indicators that assess
dismissal regulation for regular workers and indicators that assess hiring regulation for temporary workers.
The first part of this section looks at dismissal regulation for regular workers and the second part at hiring
regulation for temporary workers.5 Importantly, the OECD indicators quantify employers’ dismissal and
hiring costs and not the degree of protection for workers.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
174
So far, three annual time series (Versions 1-3) existed, all ending in 2013. Version 1 begins in 1985, and
the two subsequent versions sought to improve the ways in which the indicators capture job protection
provisions. Version 2 begins in 1998 and has some coverage of collective (besides individual) dismissals
of regular workers. Version 3 begins in 2008 and introduces certain aspects of enforcement for regular
workers and additional items on hiring regulation for temporary workers. The present analysis extends
Versions 1-3 to preserve the time series dimension of the database. For regular workers, however, the
chapter develops, and mostly relies on, the new Version 4, which is available from 2013 to 2019. The
design of the indicators for temporary workers is unchanged.
The indicators consider information on employment protection legislation in a detailed, but al so
pragmatic way: they take account of national and sectoral, but not firm -level, collective bargaining
agreements. They focus on the private, not public, sector and evaluate regulation applying to medium -
sized and large, not small, firms and their employees.6 Where there are differences by firm size, the
scored value is the average of the values for a firm with 35, 150 and 350 employees. Where there are
differences between categories of workers (for example blue-collar and white-collar workers), the scored
value is the average of the values corresponding to each category. These standardisations are
necessary in light of the available information. A further consideration is the design of the scoring scale
which, as from the beginning of the database, is constructed to enable quantitative comparisons of
regulatory stringency. The nature of the indicators, which convert mostly qualitative information into
numerical data, means that readers should nevertheless be cautious when interpreting small differences
in scores across countries and over time.
3.2.1. The OECD Employment Protection Legislation indicators for dismissing regular
workers
The new Version 4 of the OECD Employment Protection Legislation indicators for regular workers better
reflects the differences in job protection regulation across countries and time. This is achieved in four
ways: i) by improving the granularity of some elements of employment protection regulation that have
already entered the indicators; ii) by adding important elements of employment protection regulation that
have so far been absent from the indicators; iii) by expanding the assessment of employment protection
regulation of collective dismissals to align it with the assessment of employment protection regulation of
individual dismissals; and iv) by changing the way in which dismissal size thresholds for dismissals of
several workers enter the indicator of employment protection regulation of collective dismissals.
All employment protection legislation indicators in this chapter refer to no-fault dismissals; hence, the
stated reason for the dismissal is not related to illegitimate behaviour of the worker (such as theft,
misconduct or unauthorised absence from work). The indicators, in the case of both individual and
collective dismissals, take account of four aspects of dismissal regulations (Table 3.1): procedural
requirements, notice period and severance pay, the regulatory framework for unfair dismissals and
enforcement of unfair dismissal regulation. The first two of these categories are defined by two lower -
level elements, the last two by four elements. The four broad categories determine with equal weight the
aggregate score, and the lower-level elements determine with equal weight the scores for the four broad
categories.7 Where there are differences between dismissals for personal and for economic reasons,
the scored value is the average of the two. Annex 3.A provides further details on the methodology and
the full scoring scale.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
175
Table 3.1. The OECD Employment Protection Legislation indicators for dismissing regular workers
Category of dismissal regulation Lower-level elements of dismissal regulation
Notification procedures (substantively revised)
Procedural requirements
Time delay before notice can be given (substantively revised)
Length of notice period
Notice and severance pay
Amount of severance pay
Definition of unfair dismissal (substantively revised)
Length of trial period (i.e. the initial period during which unfair dismissal claims cannot be made)
Regulatory framework for unfair dismissals
Compensation to the worker following unfair dismissal
Possibility of reinstatement following unfair dismissal
Maximum time to make a claim of unfair dismissal
Burden of proof when the worker files a complaint for unfair dismissal (new item)
Enforcement of unfair dismissal regulation
Ex-ante validation of the dismissal by an external authority (new item)
Pre-termination resolution mechanism granting unemployment benefits (new item)
Note: The changes indicated in blue are with respect to the 2008 version (i.e. Version 3) of the indicators. The four broad categories of dismissal
regulation determine with equal weight (25%) the aggregate score; the lower-level elements determine with equal – or almost equal – weight
the scores of the four broad categories. Length of trial period is not included in the indicator for collective dismissals. Annex 3.A provides the full
scoring scale.
Procedural requirements captures the actions that the firm must take before or when issuing the dismissal
to the worker. It consists of two components: notification procedures and time delay before notice can be
given. Notification procedures assess whether the dismissal notification needs to be provided with reasons,
its legal value and whether it needs to be preceded by a warning procedure, a discussion with the worker
and the consultation or authorisation of a third party, such as the relevant administrative body. This item
on notification procedures was already included in previous versions of the indicators, but the new version
modifies it to better reflect the differences in procedure linked with the various types of notification
requirements. Also, in the new version, the scoring scale for the time delay before notice has been aligned
with the one for the length of the notice period at four years of job tenure, so that days, weeks and months
count the same in both these lower-level elements.
The second category considers the length of the notice period and the amount of severance pay, the two
elements of dismissal regulation that often come first to mind. As in earlier versions, both are evaluated as
the average of the values at three points of job tenure: 9 months, 4 years and 20 years.
The third category, the regulatory framework for unfair dismissals, is concerned with the breadth of the
definition of fair and unfair dismissals and the stringency of remedies imposed by courts when a dismissal
is judged to be unfair. It takes into account four aspects: the definition of unfair dismissal, the length of the
initial (or trial) period during which the employee is not protected against unfair dismissal, the monetary
compensation to the worker following an unfair dismissal and the possibility of reinstatement following an
unfair dismissal.8
The new version substantially expands the item on the definition of (un)fair dismissal. This now considers
unfair dismissals for economic reasons separately to those for personal reasons.9 Within unfair dismissals
for economic reasons, the item assesses the freedom that judges have in their decision, the restrictions
on the firm as to which worker is to be selected for the dismissal and the requirements in terms of alternative
employment and training opportunities that the firm needs to offer to the worker and without which the
dismissal is considered unfair. The item on the length of the trial period has been slightly refined as well.
It now takes higher values when the regulation for dismissing workers before the end of the trial period is
more stringent in terms of notice and severance pay.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
176
The new version of the employment protection indicators for regular workers places a stronger emphasis
on enforcement of unfair dismissal regulation. Three items have been newly introduced: i) whether the
worker alone has the burden of proof when filing a complaint for unfair dismissal; ii) whether an ex-ante
validation of the dismissal limits the scope of unfair dismissal complaints; and iii) whether a pre-termination
resolution mechanism exists that, by granting eligibility for unemployment benefits, offers an attractive
alternative to dismissal for the employee.
In most countries, the burden of proof in unfair dismissal cases does not lie only with the employee. While
this increases the cost and the uncertainty for the firm that it can prove that the dismissal was fair (Boeri,
Garibaldi and Moen, 2017[63]), it reflects that the firm is the party that knows the reason or motivation for
the dismissal. Validation of the dismissal as a preventive check is closely related with the item on
notification procedures in the first category, procedural requirements. It tends to make notifications more
burdensome for the firm, but has the advantage for the firm of limiting the risk that the dismissal will be
judged as unfair later on. Pre-termination resolution mechanisms (termination by mutual consent or
resignation) give the right to unemployment benefits in many OECD countries (sometimes with sanctions).
France, for example, introduced in 2008 a formalised scheme of termination by mutual agreement that
provides entitlement to unemployment benefits, and the reform has been found to have increased worker
flows (Batut and Maurin, 2019[64]).
The coverage of enforcement issues in the OECD indicators remains limited overall, as the indicators do
not take account of certain aspects of the functioning of the judicial system, such as access to labour courts
or the length of proceedings. Such aspects of the complexity of judicial procedures matter for the decision
of the firm whether to dismiss a worker (Espinosa, Desrieux and Ferracci, 2018[65]; Gianfreda and Vallanti,
2017[66]), but can also influence the incentives of the employee to file a complaint (Campolieti and Riddell,
2020[67]; Espinosa, Desrieux and Wan, 2017[68]; Fraisse, Kramarz and Prost, 2015[69]). Their effects on
employers’ costs remain therefore ambiguous, which makes it difficult to incorporate them in the indicators.
Another consideration is that integrating statistics on judicial procedures in the indicators would be
problematic given lack of data and poor cross-country comparability and raise issues concerning the
endogeneity of judicial outcomes to regulation and labour market conditions (Ichino, Polo and Rettore,
2003[70]).
The overall employment protection legislation indicators for dismissing regular workers assign a weight of
5/7 to individual dismissals and 2/7 to collective dismissals, as in previous versions. In the design of the
indicators, a dismissal is seen as collective when a firm lays off several workers at around the same time.
More precisely, the indicator for collective dismissals in Version 4 evaluates for each item the average for
dismissals of 10, 45 and 120 workers by a firm within one month, rather than including a separate item on
dismissal thresholds as before.10 In all OECD countries with specific legislation for collective dismissals
(and for the firm sizes considered by the indicator), this legislation always applies in the case of dismissals
of 120 workers or more in one month, which will be referred to as mass dismissals in the remainder of this
chapter.11 Moreover, in contrast to individual dismissals, collective dismissals can only occur for economic
reasons. Therefore, while the indicators for individual dismissals give the same weight to dismissals for
personal and economic reasons, in the aggregate indicators dismissals for economic reasons take a weight
of almost two-thirds.
3.2.2. The OECD Employment Protection Legislation indicators for hiring temporary
workers
The OECD Employment Protection Legislation indicators for temporary workers distinguish between fixed-
term contracts and temporary work agency contracts. They focus on hiring restrictions instead of dismissal
regulations, in contrast to the indicators for regular workers. This is natural to a certain extent, given that
terminations of temporary contracts tend to be rare during the duration of the contract and easy at the end,
while legislation in many countries aims to avoid an excessive use of temporary contracts. Nevertheless,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
177
dismissals of temporary employees do happen, and in some cases severance pay or other tools used in
dismissal regulation protect temporary workers when their contract expires. Future work is planned to go
beyond hiring restrictions for workers on fixed-term contracts, by also considering regulations in the context
of dismissals of fixed-term workers and expirations of fixed-term contracts.
The two categories of hiring regulations, for fixed-term and temporary work agency contracts (Table 3.2),
contribute in equal shares to the total score. This is to be kept in mind in an environment where in all OECD
countries fixed-term contracts are more common than temporary work agency contracts (OECD, 2014[60]).
The first three lower-level elements serve the same purpose for the two types of temporary contracts: to
capture constraints to the ease with which such contracts can be used in place of regular contracts. The
fourth item on authorisation and reporting obligations is in practice mainly relevant for temporary work
agency contracts. The last item on equal treatment of temporary work agency workers and regular workers
at the user firm concerns both pay and working conditions. The full scoring scale is in Annex 3.A.
Table 3.2. The OECD Employment Protection Legislation indicators for hiring temporary workers
Category of hiring regulation Lower-level elements of hiring regulation
Valid cases for use of fixed-term contracts
Fixed-term contracts Maximum number of successive fixed-term contracts
Maximum cumulated duration of successive fixed-term contracts
Types of work for which temporary work agency employment is legal
Restrictions on the number of renewals of the assignment to the user firm
Temporary work agency contracts Maximum cumulated duration of successive assignments to the user firm
Authorisation and reporting obligations
Equal treatment of regular workers and temporary work agency workers at the user firm
Note: There are no changes with respect to the 2008 version (Version 3) of the indicators. The two broad categories of hiring regulation determine
with equal weight (50%) the aggregate score; the lower-level elements determine with equal weight the scores for the two broad categories,
except for the first element in both cases (“Valid cases for use of fixed-term contracts” and “Types of work for which temporary work agency
employment is legal”) which carry a weight of 50% and 33% respectively. Annex 3.A provides the full scoring scale.
3.2.3. Comparison with other employment protection legislation databases
The last two decades have seen an increasing effort by international organisations and research centres
to create indicators of labour market institutions, including of employment protection legislation.12 Country
coverage of the OECD Employment Protection Legislation indicators themselves was extended in 2013 to
include, for example, several non-OECD Latin American and Caribbean countries, in a partnership with
the Inter-American Development Bank. Besides the OECD indicators, among the most well-known are the
EPLex by the International Labour Organization, the Labour Regulation Index by the Centre for Business
Research (CBR-LRI) at the University of Cambridge and the (suspended) Employing Workers indicator in
the Doing Business dataset by the World Bank.13 While these three databases share some similarities with
the OECD indicators,14 important differences remain. Overall, the OECD indicators continue to be the most
widely used for cross-country comparisons in policy reports and academic papers (Denk and Georgieff,
forthcoming[62]).
The three other databases cover more countries than the OECD indicators (from 100 countries for the
EPLex to 190 countries for the Doing Business). However, they provide a less comprehensive overview of
employment protection regulation.15 In particular, the three other databases do not cover enforcement of
unfair dismissal regulation, and the EPLex and the Doing Business do not cover temporary work agency
employment. In addition, the CBR-LRI and the Doing Business have a less deep evaluation of collective
dismissals and of the regulatory framework for unfair dismissals (for example they have no information on
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
178
compensation following an unfair dismissal). There are also certain aspects of job protection that the OECD
indicators do not cover, but some of the other databases do (for example prohibited grounds for dismissal
such as discrimination in the EPLex). Two other differences between the EPLex and the OECD indicators
are that the EPLex is more descriptive, rather than quantitative, and available for a shorter time span, since
the series starts in 2009, against 1985 for the OECD indicators. Also, the three other databases do not, or
little, take account of sectoral collective agreements and case law.
3.3. Employment protection legislation in OECD countries in 2019
This section compares employment protection legislation in OECD countries, based on information in the
OECD Employment Protection Legislation Database. It begins with a detailed look at the components of
job dismissal regulation. It then presents assessments of the overall regulation of individual and collective
dismissals of regular workers and the regulation for hiring temporary workers. These indicators of dismissal
and hiring regulations are based on the methodology outlined in the previous section.
3.3.1. Country details on individual elements of dismissal regulations for regular workers
This part discusses country details on individual elements of job dismissal regulations for regular workers
in the order of the four broad categories entering the indicator: procedural requirements before notice can
be given, notice period and severance pay, the regulatory framework for unfair dismissals and enforcement
of unfair dismissal regulation.
Procedural requirements
Procedural requirements for individual dismissals of regular workers vary significantly between countries
(Figure 3.1). This variation is mostly due to differences in notification procedures, rather than in the time
delay before notice is given, the other regulatory element entering this category.
The employer usually needs to inform the employee of the reason for the dismissal, at least if the worker
requests it. This notification generally takes the form of a written statement. Some countries require, at
least in some cases, a prior warning (Australia, Austria, Québec in Canada, the Czech Republic, Denmark,
Estonia, Greece, Ireland, Lithuania, New Zealand, Portugal and the Slovak Republic), an interview with
the employee (Australia, Colombia, France, Iceland, Ireland, Luxembourg, Slovenia and Turkey) or a
consultation with a third party (the Czech Republic, Finland, Israel, Norway, Poland, the Slovak Republic
and Sweden). Only in Canada (except Québec) and in most states of the United States, the firm never
needs to provide a reason for the dismissal to the worker before or at the time of dismissal. By contrast,
procedures are particularly stringent in the Netherlands, where a dismissal cannot occur without prior
authorisation by the Public Employment Service or the Sub-district Court. In Germany and Sweden, the
dismissal can be paused until the final judgement by the court if the work council (Germany) or worker
(Sweden) objects to the dismissal.
Notification procedures for collective dismissals are more homogeneous and stringent than for individual
dismissals for economic reasons (Denk and Georgieff, forthcoming[62]). All countries, except some
provinces in Canada, Chile and the United States, require a consultation or even an authorisation before
the dismissal can take place from a certain threshold number of workers dismissed. In particular, above
this number of dismissals, a dismissal can never occur without the authorisation of the administration in
Colombia and France (for firms with more than 50 employees). In Belgium, if the work council and worker
object to the dismissal, the dismissal can be paused until the employer has provided evidence of
compliance with the notification and consultation procedures. In Mexico, dismissals for economic reasons
are allowed only if they involve several workers and they require an authorisation by the Labour Court.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
179
Figure 3.1. Procedural requirements for individual dismissals of regular workers
2019
Indicator score
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. Procedural requirements consists of two components: notification procedures and time delay before notice
can be given.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/x2zf8y
Notice and severance pay
At the point at which a firm decides to dismiss a worker, it often cannot do so without informing the worker
in advance (i.e. respecting a notice period) and providing severance pay. Notice period and severance pay
represent a cost to the firm. To the worker, they reduce the economic and possibly psychological burden
of a layoff, by preventing an abrupt loss of labour income. Employees usually continue to work for the firm
during the notice period. However, in some cases, notified employees are released from work, and even
employees continuing work during the notice period are likely to be less motivated. A combined estimate
of the costs to firms and the benefits to workers from notice period and severance pay is thus provided by
the number of months a firm must give notice before dismissal and the amount of severance pay (in months
of work).
The database collects information on notice period and severance pay at three points of job tenure:
9 months, 4 years and 20 years. At 4 years of job tenure, compensation in the form of notice period and
severance pay in the case of an individual dismissal varies widely among OECD countries, from no
compensation in the United States to six months of pay in Turkey (Figure 3.2). The compensation of half
a year of pay in Turkey corresponds to one-eighth of the four years of labour income earned until then.
Costs to firms and benefits to dismissed workers are also high in Israel and Lithuania, where more than an
extra 10% of the labour income earned until then is paid as compensation to a dismissed worker. Notice
periods are more widely used overall than severance pay is: only two countries have no notice period,
while 12 countries have no severance pay. Nonetheless, countries with the highest total compensation
stand out with very high severance pay. The variation in severance pay across countries is much greater
than the variation in notice periods.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
180
Figure 3.2. Notice period and severance pay for individual dismissals of regular workers
Four years of job tenure, measured in months of pay after dismissal notice, 2019
Notice period Severance pay
7
6
5
4
3
2
1
0
Note: These values are for individual (not collective) dismissals. They take the average of dismissals for personal and economic reasons.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/brk4sa
Notice period and severance pay both tend to increase with the length of job tenure. The longer workers
stay with their firm, the more secure their job tends to be. This greater job security at higher job tenure may
also contribute to the observed lower job mobility of long-tenured, and often older, workers. In all OECD
countries, notice period and severance pay are increasing, or at least not decreasing, with longer job tenure
across the three points of tenure in the database: 9 months, 4 years and 20 years. On average in OECD
countries, together they are seven times as high at 20 years of job tenure as at 9 months of job tenure
(Figure 3.3).
Severance pay increases more steeply with job tenure than notice period. On average in OECD countries,
severance pay is smaller than notice period at 9 months of job tenure, but greater than notice period at
20 years of job tenure. At 9 months of job tenure, two-thirds of OECD countries require no severance pay.
At 20 years of tenure, workers in nine OECD countries – Belgium, Chile, France, Israel, Luxembourg,
Mexico, the Netherlands, Spain and Turkey – have the right to a severance pay that is worth at least half
a year of work. It is relatively more common for the notice period to be the same at the three points of job
tenure.
Notice periods are similar for dismissals for personal and for economic reasons. They are also similar for
individual and collective dismissals, with the exception of a few countries where they are longer for
collective dismissals. For example, in the United States dismissed workers are entitled to two months of
notice in the case of mass layoffs and large plant closures. In Canada, the notice that must be given to an
employee affected by a collective dismissal is often longer than for an individual termination of employment.
In Luxembourg and the United Kingdom, in the case of a large dismissal, a long time period must occur
between notification to the labour authority and the day the dismissal takes effect, de facto prolonging
notice periods for employees with short job tenure.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
181
Figure 3.3. The role of job tenure for notice period and severance pay
OECD average at three different points of job tenure, measured in months of pay after dismissal notice, 2019
Notice period Severance pay
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
9 months 4 years 20 years
Note: These values are for individual (not collective) dismissals. They take the average of dismissals for personal and economic reasons. OECD
average is the unweighted average for the 37 OECD countries.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/imrpqs
For severance pay, more notable differences arise between dismissals for economic and personal reasons
and between individual and collective dismissals. At 20 years of job tenure, severance pay is higher in the
case of dismissals for economic reasons in Australia, the Czech Republic, Estonia, Ireland, Poland, the
Slovak Republic and the United Kingdom. At this point of tenure, no country has higher severance pay for
dismissals for personal reasons. This on average somewhat lower level of security in the case of dismissals
for personal reasons is potentially motivated by workers themselves sharing more of the responsibility for
the layoff, for example due to insufficient performance. With regard to the comparison of individual and
collective dismissals, additional compensation can often be granted for dismissals that exceed a certain
number of workers, usually as a result of the consultation with worker representatives.
Counting months of advance notice and severance pay on equal terms as in the analysis of this section is
necessarily a simplification. For the firm and the worker, notice periods are often somewhat less costly and
more protective than severance pay: the worker is required to continue work while on notice and sufficiently
long notice periods allow the employment services to intervene before the dismissal takes place, thereby
facilitating the transition to another job. These considerations motivate OECD advice that countries, where
notice periods are short and severance pay is high, could consider extending notice periods and lowering
severance pay, while activating early interventions, to smooth job transitions without increasing employers’
costs (OECD, 2018[11]).
Regulatory framework for unfair dismissals
In almost all OECD countries, a dismissal based on a reason that is beyond the scope of allowed (or “fair”)
reasons can, if it is challenged in court, lead the employer to pay specific compensation to workers or even
reinstate workers to the positions from which they were dismissed. The category “regulatory framework for
unfair dismissals” captures the definition of unfair dismissal, the length of the trial period during which all
dismissals are fair and the compensation and reinstatement rules following an unfair dismissal (Figure 3.4).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
182
Fair reasons for dismissal generally include operational reasons (e.g. economic difficulties or technological
changes) or personal reasons related to workers themselves (e.g. insufficient performance or unsuitability).
In Canada (except Québec) and the United States, an employee can be fairly dismissed without reason,
provided that the dismissal was not based on prohibited grounds. 16 Contrasting examples are Chile, which
forbids dismissals for insufficient performance and unsuitability, and Mexico, which allows dismissals for
economic reasons only if they involve several workers.
Figure 3.4. Regulatory framework for unfair individual dismissals of regular workers
2019
Definition of unfair dismissal Length of the trial period Compensation following unfair dismissal Reinstatement following unfair dismissal
Indicator score
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. Compensation following unfair dismissal is missing for Canada, Greece, Iceland and the United States,
and length of the trial period is missing for Canada and the United States. Values for these subcomponents are set equal to the average of non-
missing subcomponents of regulatory framework for unfair dismissals for the same country.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/z5fnel
The scope of fair dismissals for economic reasons depends to a large extent on the freedom that judges
have in their decision. In about half of the OECD countries (including Finland, Germany, Poland, Spain
and the United Kingdom), dismissals for economic reasons can only be challenged if the reason for the
dismissal was false or patently irrational. By contrast, in the other half of the countries (including Australia,
Chile, Italy, Japan, the Netherlands and Norway), judges can question the operational need of the dismissal
decision. In some countries, when redundancy could concern several workers occupying similar positions,
the employer should select the employees who are to be dismissed based on objective criteria other than
performance. Job tenure may be part of these criteria, as for example in France, Latvia, Portugal and
Sweden.17 In Italy, if the dismissal concerns five employees or more within a period of 120 days, judges
cannot question the operational need for the dismissal and the firm must follow social and economic criteria
for selecting the workers to be dismissed.
About two-thirds of OECD countries require substantive conditions for a dismissal for economic reasons.
These conditions generally include attempting the transfer of the worker to another position, possibly with
retraining. They may also require priority for rehiring (e.g. Finland and France) or provision of outplacement
services (e.g. Belgium). Other countries have no conditions, at least for individual dismissals: Canada, the
Czech Republic, Denmark, Greece, Hungary, Iceland, Israel, Slovenia, Spain, Switzerland, Turkey, the
United Kingdom and the United States. However, Canada, Denmark, Greece, Slovenia, Spain,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
183
Switzerland and Turkey have additional requirements from a certain number of workers dismissed. These
requirements typically include the establishment of a social plan, i.e. a set of measures of reemployment,
retraining, outplacement and, in some cases, extra monetary compensation for the workers.
Dismissals for personal reasons exist in most countries. Employers can dismiss workers who have become
unsuitable for the position (due to medical or qualification reasons) or whose performance has become
insufficient. In some countries, however, insufficient performance, without unsuitability, is not a fair reason
for dismissal (e.g. Chile, Finland, France, Mexico, Norway, Portugal, Spain and Sweden). Attempting
substantial alternatives can be required also in the event of a dismissal for personal reasons. For example,
a transfer to a suitable position should be attempted in the case of a dismissal for medical unsuitability in
Belgium and France and for all types of unsuitability in Finland and Italy. In Japan and Spain, workers
should be trained to avoid a dismissal for insufficient qualification.
In almost all OECD countries, unfair dismissal regulation does not apply during an initial (or trial) period at
the beginning of the employment relationship. Belgium, Chile, Greece, Israel, Japan and Poland are the
only exceptions, although temporary contracts might sometimes act as a substitute for the trial period. 18
The median value of the trial period is three months. It is longest in the United Kingdom (24 months) and
Ireland (12 months).
When judges deem the dismissal (at some point after the completion of the trial period) to be unfair, they
can order the payment of a compensation or the reinstatement of the worker to the position. Compensation
following an unfair dismissal is particularly high in Italy. Reinstatement is always made possible to the
employee in Austria, the Czech Republic, Korea, Latvia and Turkey. By contrast, reinstatement, except in
the case of dismissals on prohibited grounds, cannot be imposed on the employer in Belgium, Colombia,
Estonia, Finland, France (for individual dismissals), Iceland, Lithuania, Luxembourg, Spain, Sweden,
Switzerland and the United States. In France, in the event of a dismissal of more than ten workers in a firm
with more than 50 employees, the absence (or insufficient elaboration) of a social plan can entail the nullity
of the redundancy procedure; in these cases, the judge may order the reinstatement of the employees
upon their request. Overall, reinstatements tend to be more common in countries with a more stringent
regulatory framework for unfair dismissals (e.g. Greece, Korea, Latvia, Norway, Portugal and Turkey).
Enforcement of unfair dismissal regulation
The indicators on enforcement of unfair dismissal regulation consider the maximum time to make a claim,
the burden of proof, ex-ante validation of the dismissal and pre-termination resolution mechanisms
(Figure 3.5). As explained in Section 3.2.1, this means that the coverage of enforcement issues in the
indicators remains limited overall since the indicators, for various reasons, do not take account of certain
aspects of the functioning of the judicial system, such as access to labour courts or the length of
proceedings.
The median duration for the time period during which an employee can file an unfair dismissal complaint
is two months among OECD countries. In some countries (Austria, Denmark, Hungary, Lithuania,
Slovenia, Switzerland and Turkey), the maximum time available is so short that in practice claims must be
filed before the dismissal takes effect. By contrast, it is longer than two years in Colombia, Iceland, Israel,
Japan and the United States19 (where it varies by state).
In most countries, following an unfair dismissal complaint, it falls, at least in part, on the employer to provide
evidence that the dismissal was fair. Who bears the burden of proof matters for the incentives of the firm
to dismiss a worker and of the employee to file a complaint. The only countries where the burden of proof
lies solely with the employee in cases of unfair dismissal (not based on prohibited grounds) are Australia,
Colombia, the Czech Republic, Denmark, Israel, Poland, the Slovak Republic, Switzerland and the
United States.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
184
Figure 3.5. Enforcement of unfair dismissal regulation for individual dismissals of regular workers
2019
Maximum time to make a claim Burden of proof Ex-ante validation of the dismissal Pre-termination resolution mechanism
Indicator score
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. The minimum score for burden of proof and ex-ante validation of the dismissal is 0.5.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/a914ig
Validation of the dismissal as a preventive check tends to make notification procedures more stringent, but
it has the advantage for the firm of limiting the risk that the dismissal will be judged as unfair later on. Only
in Austria and the Netherlands, all dismissals need to involve an advance validation that limits the scope
for unfair dismissal complaints; dismissals should be approved by the work council in Austria and the Public
Employment Service or the Sub-district Court in the Netherlands. A validation secures the dismissal for
the employer, but only from a given number of workers dismissed, in Belgium (in some cases), Colombia,
France, Greece, Mexico and Spain.
Resignation and some form of termination by mutual consent provide the right to unemployment benefits
in many countries (possibly with sanctions) and are thus a popular alternative to dismissals. In a number
of countries, they grant unemployment benefits under the same conditions as in the event of a dismissal.
This is the case in Austria, Chile, Colombia, France, Hungary, Japan, Korea, Lithuania, Mexico (where
there are no unemployment benefits), the Netherlands and the Slovak Republic. In other countries,
resignation and termination by mutual consent entitle workers to receive unemployment benefits with long
waiting periods (Immervoll and Knotz, 2018[71]). By contrast, resignation and termination by mutual consent
never give access to unemployment benefits (in contrast to dismissals) in Canada, Greece, Italy,
Luxembourg, Slovenia, Spain, Turkey and the United States20 (in most states and in the case of individual
termination). These countries are also among those for which the overall level of enforcement is particularly
high.
3.3.2. Aggregate assessments of dismissal regulations for regular workers
This part assesses job dismissal regulation for workers on regular contracts by aggregating the individual
elements that the previous section discussed. It does so first for individual dismissals, then for collective
dismissals and finally for a composite of individual and collective dismissals.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
185
Regulation of individual dismissals of regular workers
The OECD indicators show wide variation in the strictness of regulation of individual dismissals of regular
workers across countries (Table 3.3). Five of the ten countries with the lowest measured regulation have
a legal system with British common-law origin: the United States, Canada, Australia, the United Kingdom
and Ireland (in order). Regulation is assessed to be low as well in Switzerland, Austria, Hungary, Denmark
and Estonia. At the other end of countries with relatively strict regulation are the Czech Republic, Israel,
Portugal, the Netherlands, Turkey, Belgium, Italy, Latvia, Greece and Luxembourg. In between are the
remaining countries where policies seem to attach more equal importance to firm adaptability and job
security.
The aggregate score is determined by the four categories of regulation: procedural requirements, notice
and severance pay, the regulatory framework for unfair dismissals and enforcement of unfair dismissal
regulation. One question is whether countries with high overall regulation have strict regulations along all
four dimensions or whether certain categories of regulation are more typical of high-regulation countries.
This is investigated statistically through the correlations among the four categories of dismissal regulation
and their correlation with the aggregate score. This analysis indicates that the fourth category, enforcement
of unfair dismissal regulation, plays a different role to the other three categories, underlining the importance
of including it in the indicators to obtain a more complete picture of regulations. Procedural requirements,
notice and severance pay and the regulatory framework for unfair dismissals are all positively correlated,
suggesting that they tend to be complementary, rather than substitute policies.
The aggregate score is only weakly correlated with enforcement of unfair dismissal regulation. 21 Some
countries with low regulatory protection in aggregate have a high score on enforcement of unfair dismissal
regulation, notably Canada and the United States. Canada is one of the countries that provide no access
to unemployment benefits except after dismissals and so offer no alternative pre-termination route that
would be made attractive by granting unemployment benefits. In the United States, the time to make a
claim of unfair dismissal is long (in case restrictions to dismissals exist in the contract or “implied contract”,
in which the employer gives certain assurances for continued employment to create a contract of sorts).
The Netherlands, by contrast, a country with overall high regulatory protection, has a low score: the Public
Employment Service or the Sub-district Court provides an ex-ante validation of the dismissal, and
termination via mutual consent gives right to unemployment benefits without sanctions.
In pairwise comparisons, enforcement of unfair dismissal regulation is negatively correlated with the three
other categories.22 One reason may be that in countries, where unfair dismissals carry few rights (i.e. the
regulatory framework for unfair dismissals is less strict), a high degree of enforcement of unfair dismissal
regulation is less relevant. Another reason is design: in general, validation of a dismissal by a third party
before the dismissal occurs enters the indicator negatively under enforcement as such validation reduces
the chances for the dismissal to be qualified as unfair; and it enters the indicator positively under procedural
requirements as it makes the notification procedure more stringent. 23
The new Version 4 of the indicators changes the assessment of the stringency of job dismissal regulations
in several countries compared with the previous Version 3. Job protection against individual dismissals is,
for example, assessed to be less strict in Austria, France, Germany and the Netherlands and stricter in
Canada, Ireland and the United States. Box 3.1 mentions some of the reasons behind the new evaluation,
drawing on the analysis in Denk and Georgieff (forthcoming[62]). Conceptually, the differences in scores
are due to three reasons: i) revisions to the categories procedural requirements and the regulatory
framework for unfair dismissals; ii) the addition of the fourth category enforcement of unfair dismissal
regulation; and iii) the reduction in the weight of the three categories that have already entered the
indicator. As the examples in the box illustrate, the modifications to the indicator design ensure better
overall comparisons between countries, because they improve the way in which the indicators map the
costs for firms stemming from different regulatory aspects.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
186
Table 3.3. The OECD indicators: Strictness of regulation of individual dismissals of regular workers
White / light blue / dark blue: countries with low / middle / high regulatory protection, 2019
Procedural Notice and Regulatory Enforcement of OECD Employment
requirements severance pay framework for unfair unfair dismissal Protection
dismissals regulation Legislation indicator
United States 0.7 0.0 0.1 4.4 1.3
Switzerland 1.2 1.3 1.6 2.3 1.6
Canada 0.7 0.8 1.2 3.8 1.6
Australia 1.3 1.0 1.8 2.5 1.7
Austria 1.2 0.9 3.1 1.5 1.7
United Kingdom 1.3 1.3 1.1 3.3 1.7
Hungary 1.2 1.8 2.2 2.0 1.8
Denmark 1.2 2.1 1.9 2.3 1.8
Estonia 1.5 1.4 1.6 3.0 1.9
Ireland 1.3 1.2 1.9 3.5 2.0
Colombia 1.3 1.6 2.0 3.0 2.0
New Zealand 2.3 0.4 2.3 3.3 2.1
Japan 0.8 0.9 2.8 3.9 2.1
Iceland 1.0 1.9 1.5 4.3 2.2
Slovenia 1.3 1.5 2.4 3.5 2.2
Germany 1.7 1.3 3.1 2.9 2.2
Lithuania 2.0 3.4 1.6 2.0 2.2
Norway 1.5 1.0 3.3 3.3 2.3
Slovak Republic 2.8 1.5 2.8 2.0 2.3
Korea 2.2 1.0 3.0 3.3 2.4
Finland 2.0 1.0 2.2 4.3 2.4
Poland 2.2 2.5 2.4 2.5 2.4
Mexico 1.8 1.7 3.7 2.5 2.4
Spain 1.8 2.1 2.0 3.8 2.4
France 1.5 2.4 2.6 3.3 2.4
Chile 1.8 2.5 3.0 2.5 2.5
Sweden 2.3 1.7 2.5 3.4 2.5
Luxembourg 2.1 2.2 1.7 4.0 2.5
Greece 1.2 1.2 3.8 4.0 2.5
Latvia 2.5 1.8 3.2 3.0 2.6
Italy 1.8 2.0 3.0 4.0 2.7
Belgium 1.8 3.0 2.1 4.0 2.7
Turkey 1.3 3.4 3.1 3.5 2.8
Netherlands 4.2 2.3 2.5 2.4 2.8
Portugal 2.3 1.7 4.2 3.3 2.9
Israel 2.5 2.9 2.5 3.8 2.9
Czech Republic 3.8 2.5 3.0 2.8 3.0
Note: Range of indicator scores: 0-6. The ten countries with the lowest and highest score are classified as countries with low and high regulatory
protection. Scores are rounded to one decimal, while classification is done with the actual scores. The OECD Employment Protection Legislation
indicator is the average of the scores for the four broad categories.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
187
Box 3.1. New design, new assessment: What is different with the new OECD indicators?
The main changes to the design of the OECD indicators for job dismissal regulations of regular workers
are due to a substantive revision of the first category procedural requirements, the addition of the fourth
category enforcement of unfair dismissal regulation and the expansion of the evaluation of employment
protection regulation of collective dismissals. Each of the three aspects is important for improving the
comparisons of countries in the new Version 4 relative to the previous Version 3. This box summarises
the role of procedural requirements and enforcement of unfair dismissal regulation (Figure 3.6). Further
details, also on the role of collective dismissals, can be found in Denk and Georgieff (forthcoming[62]).
Procedural requirements is the category with the largest influence on the differences between the two
versions. Newly taking into account warning procedures and the obligation to consult the worker before
the dismissal contributes to the higher scores for Québec (Canada) and Ireland. The need to notify a
third party is now viewed as a constraint only when it involves a consultation or authorisation procedure,
which is not always, or never, the case in Austria, France, Germany and Latvia. The alignment of the
scoring scales for the time delay before notice and the length of the notice period is another explanation.
Two of the countries for which the addition of the enforcement category significantly lowers the scores
are Austria and the Netherlands. In these countries, an advance validation of the dismissal acts as a
preventive check, reducing the chances for the dismissal to be qualified as unfair, and terminations via
mutual consent grant access to unemployment benefits under the same conditions as in the case of a
dismissal. These aspects therefore enter the indicator with a minimum value for the two countries.
Figure 3.6. Changes in the assessed strictness of regulation of individual dismissals
Difference in the indicator score between the new Version 4 and the previous Version 3, 2019
Procedural requirements Notice and severance pay
Regulatory framework for unfair dismissals Enforcement of unfair dismissal regulation
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
Note: Range of indicator scores: 0-6. Countries are ordered based on the difference in the overall indicator score. A negative (positive)
number means that regulation is evaluated to be less (more) strict with the new Version 4 than the previous Version 3. For further details,
see Denk and Georgieff (forthcoming[62]), “The 2019 OECD Employment Protection Legislation indicators: New insights on job dismissal
regulation in OECD countries”, OECD Social, Employment and Migration Working Papers, OECD Publishing, Paris.
Source: OECD calculations based on OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/4l6xmo
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
188
Regulation of collective dismissals of regular workers
Periods of economic difficulty, due to for example a persistent decline in demand or required
technological change, might lead firms to restructure their workforce, involving the dismissal of a large
number of workers in relatively short time. It is common that specific regulations apply in these situations.
Until the analysis in this chapter, the database considered such specific restrictions on collective
dismissals only in selected dimensions of dismissal regulation and as a top-up to individual restrictions.
The new OECD data present, for the first time, employment protection legislation indicators for collective
dismissals that are calculated analogously to the employment protection legislation indicators for
individual dismissals. The section presents dedicated indicators for individual dismissals for economic
reasons, as collective dismissals normally occur for economic reasons and the OECD indicators for
individual dismissals described above reflect both dismissals for personal and economic reasons.
While important in its own right, a comprehensive assessment of regulation of collective dismissals has
gained further relevance given developments over the past decade. In the wake of the global financial
and economic crisis, the number of “zombie firms”, i.e. firms with difficulty to meet their financial
obligations, has increased (Andrews, Adalet McGowan and Millot, 2017 [72]), potentially related with the
regulation of collective dismissals. Moreover, the scope for firms in difficulty to adjust wa ges rather than
the workforce has shrunk in the context of downward nominal wage rigidities, low inflation and weak
nominal wage growth. Digitalisation and globalisation trends are also likely to make more firms
restructure their workforce.
As mentioned in Section 3.2.1, the indicators define a dismissal as collective if several workers are laid
off within one month, hence irrespective of whether specific regulations apply. More precisely, the
indicators score the average of the values for 10, 45 and 120 dismissals in one month. In some countries
specific regulations apply as of 10 dismissed workers or lower; in others they start only when the number
of involved workers is greater. The dismissal threshold in all countries with specific regulations for
collective dismissals is smaller than 120 workers. In this way, the indicators for collective dismissals
reflect both the stringency of regulation, when several dismissals are subject to specific regulations, and
the scope of the regulation, as captured by the dismissal threshold.
All OECD countries – with the five exceptions of Chile, Israel, Korea, Mexico and New Zealand – impose
more stringent restrictions on collective dismissals relative to individual dismissals ( Figure 3.7).24 Chile,
Israel, Korea and New Zealand regulate collective dismissals the same as individual dismissals; Mexico
is a peculiar case, as legislation does not allow individual dismissals for economic reasons, only
collective dismissals for economic reasons. Protection against collective dismissals is 10 -15% higher
than against individual dismissals in the OECD on average, mostly because of stricter procedural
requirements before notice can be given – see Section 3.3.1 and Denk and Georgieff (forthcoming[62]).
Countries’ widespread use of specific restrictions on collective dismissals is likely to reflect the greater
challenge for the economy of dealing with a collective dismissal. Nevertheless, pooling several individual
layoffs in one collective dismissal can, in some cases, reduce the administrative burden of the firm. 25
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
189
Figure 3.7. The OECD indicators: Strictness of regulation of collective dismissals (defined as
dismissals of several regular workers in one month)
2019
Collective dismissals Individual dismissals for economic reasons
Indicator score
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. Individual dismissals refer to economic reasons, because collective dismissals are always for economic
reasons. The indicators score the average of the values for 10, 45 and 120 dismissals within one month.
Source: OECD calculations based on OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/w0h42v
Where regulation of individual dismissals is strict, regulation of collective dismissals tends to be strict, as
in practice regulation of individual dismissals often serves as a minimum for that of collective dismissals.
Six of the ten countries with the lowest as well as highest regulation of individual dismissals are also among
the ten countries with the lowest, respectively highest, regulation of collective dismissals. Across countries,
additional restrictions on collective dismissals (the difference in regulation between collective and individual
dismissals) are not significantly related with the strictness of regulation of individual dismissals. Therefore,
the extent to which collective dismissals are subject to specific regulations appears to be more the choice
of governments and countries, rather than a natural consequence of the regulation of individual dismissals.
An extreme form of collective dismissals are mass dismissals which the chapter defines as layoffs of at
least 120 workers in one month. This definition ensures that the threshold for specific regulations applying
to a series of individual dismissals is passed in all OECD countries with dedicated legislation for a series
of individual dismissals. The measured degree of regulation is generally the same for mass dismissals as
for smaller-scale layoffs with a number of dismissals that is above the threshold for specific regulations.
Therefore, the higher is the threshold for specific regulations, and the more extensive are the additional
restrictions compared with individual dismissals, the greater is the difference between the indicator of
regulatory strictness for mass dismissals relative to that for collective dismissals.
In one-third of OECD countries, the dismissal threshold for collective dismissals to be subject to specific
regulations is (or is equivalent to) 10 or less workers in one month. In these countries, the regulation of
mass dismissals is identical to the regulation of collective dismissals (Figure 3.8). The dismissal threshold
is higher in some other countries, in particular in Australia, Colombia, Japan and the United States. This
explains the comparatively high additional restrictions on mass dismissals in these countries.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
190
Figure 3.8. Strictness of regulation of mass dismissals (defined as dismissals of at least
120 regular workers in one month)
2019
Mass dismissals Collective dismissals
Indicator score
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. The regulation of mass dismissals corresponds to the regulation of collective dismissals above the
country-specific threshold for a series of individual dismissals to be subject to different legislation. The definition of at least 120 regular workers
ensures that this threshold is passed in all countries with dedicated legislation for a series of individual dismissals.
Source: OECD calculations based on OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/lbgcwo
Overall regulation of dismissing regular workers
The stringency of overall regulation of individual and collective dismissals of regular workers, as
assessed with the aggregate indicators, is similar to the regulation of individual dismissals and the
regulation of collective dismissals given their high correlation (Figure 3.9). Taking the indicators at face
value suggests that dismissal regulation in the most regulated countries is close to twice as stri ct as in
the least regulated countries.
The data bring out international differences in labour market and social models quite strongly. Three of
the four countries with the least strict regulation are Anglo-Saxon: the United States, Canada and
Australia. Geographically, OECD countries in North America and Australasia have regulation that is
evaluated to be below the OECD average. 26 By contrast, regulation in the majority of the OECD countries
that are also members of the European Union is above the OECD average. Four of the five countries
with the highest regulation are EU countries: the Czech Republic, the Netherlands, Portugal and Italy.
English-speaking countries combine low costs for firms with relatively few protective measures for
workers, while dismissal costs and job security for regular workers are comparatively high in many
EU countries.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
191
Figure 3.9. The OECD indicators: Strictness of regulation of dismissing regular workers
Contributions of each component, 2019
Individual dismissals Collective dismissals
Indicator score
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. These aggregate indicators assign a weight of 5/7 to individual dismissals and 2/7 to collective dismissals.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/v8jhd9
The role of collective bargaining and case law
Employment protection regulation is not based solely on the legislation, but it also depends on collective
agreements and case law, and the indicators of employment protection legislation take this into account.
Collective agreements may influence dismissal protection for regular workers, for example by changing
notice periods (e.g. in Australia, Austria, Denmark, France, Iceland, Italy, Sweden), the trial period (e.g. in
Denmark, France, Hungary, Iceland, Italy, Sweden, Turkey), severance pay (e.g. in Australia, Denmark,
France) or the criteria for selecting which workers to dismiss (e.g. in Finland, Norway, Sweden).27 Box 3.2
examines in more detail the role of national and sectoral collective bargaining in France, Italy and Sweden,
three countries where the share of workers covered by collective agreements is high. In some countries,
firm-level collective agreements may derogate from the law or higher-level agreements, thereby reducing
the binding effect of the regulation. This is, however, beyond the scope of the indicators.
Case law also matters for assessing aspects that are little or not addressed in the legislation, for example
the freedom that judges have in their decisions or the legal value of the written statement on the reason
for the dismissal. Moreover, it can affect the interpretation of the legislation, as is the case for the transfer
requirements and reinstatement options in Italy or the maximum delay to file a complaint for unfair dismissal
in France.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
192
Box 3.2. Collective agreements and employment protection in France, Italy and Sweden
Collective agreements can affect job protection provisions, especially in countries with high collective
bargaining coverage rates like France (99%), Italy (80%) or Sweden (90%). Collective agreements
typically strengthen the standards of protection that are in the law, but sometimes they may also be
allowed to derogate, i.e. to set lower standards of protection.
The OECD indicators account for the rules in national and sectoral collective agreements. In France,
sectoral agreements for managers and professionals (cadres) usually set longer notice periods and
higher severance pay at long tenure. In Italy, sectoral agreements generally provide for shorter trial
periods and longer notice periods. In Sweden, most sectoral agreements include exemptions to the
“Last-In-First-Out” (LIFO) principle, according to which workers who were hired last are those to be
dismissed first. In addition, collective agreements for white-collar workers frequently include a 55/10
provision (Söderqvist and Lindberg, 2019[73]): the notice period for workers aged 55 or older and with
10 or more years of tenure is 12 instead of 6 months. Taking account of collective agreements increases
the job dismissal indicator for regular workers in France and Italy (Figure 3.10). In Sweden, the effect
is small as the LIFO exemptions and the 55/10 provision influence protection in opposite directions.
Figure 3.10. How taking into account collective agreements affects job protection
Strictness of regulation for individual and collective dismissals of regular workers, 2019
Indicator score
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. “No CA” refers to the hypothetical score when collective agreements are not taken into account.
Source: OECD calculations based on OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/46logn
The indicators do not capture a number of aspects of collective agreements due to limited coverage or
the small size of the regulatory deviation involved. In France, many sectoral collective agreements for
managers set longer trial periods, and derogations from several restrictions to the use of temporary
employment are allowed (since 2017). In Italy, most collective agreements extend the length of the
notification procedure. In Sweden, sectoral agreements for white-collar workers generally allow for an
extended maximum duration for successive fixed-term contracts.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
193
Small firm exemptions
All OECD indicators in this chapter focus on regulation as it applies to medium -sized and large firms,
which generally employ most employees in OECD economies. For example, in 2013, firms with
20 employees or more employed 77% of all employees in Canada, 69% in Israel, 47% in Korea, 70% in
Mexico, 73% in Turkey and 83% in the United States.28 In some countries, however, small firms are
subject to less strict regulation, which is beyond the scope of the indicators. In Australia, for example,
firms with less than 15 employees do not have to pay redundancy pay. In Austria, firms with less than
5 employees are not required to have, and therefore to inform, a work council. In Germany, firms with
10 employees or less are exempt from dismissal regulation (except in cases of discriminatory and
arbitrary dismissal). In Korea, firms with 4 employees or less are exempt from dismissal regulation,
except regarding advance notice or equivalent compensation. In Portugal, in the event of an unfair
dismissal, firms with less than 10 employees may submit a request to the court to oppose reinstatement.
In Spain, in firms with 25 employees or less, the maximum duration of the trial period is three instead of
two months (except for workers with a higher education degree) and the Wage Guarantee Fund pays
part of the redundancy pay (except if the dismissal is deemed to be unfair). In Turkey, in the event of an
unfair dismissal, firms with less than 30 employees do not have to reinstate workers or pay compensation
and back pay. A final example is the exemption from requirements for collective dismissals in firms with
less than 20 employees in Belgium, the Czech Republic, Denmark, Germany, Hungary, Iceland,
Switzerland and Turkey.
3.3.3. Regulation of hiring temporary workers
An important aspect of job protection regulations relates to the difference between regular and temporary
workers. It is more difficult for firms to lay off regular workers than to not renew temporary contracts
(OECD, 2014[60]). To counteract potential overuse of temporary contracts by firms, governments usually
impose restrictions on their use. As Section 3.2.2 described, the OECD Employment Protection
Legislation Database therefore also gathers indicators on the regulation of hiring temporary workers. It
distinguishes between regulation of fixed-term and temporary work agency contracts, attaching equal
weight to the two.
OECD countries exhibit an even wider variation in the regulatory restrictions to hiring temporary workers,
as measured by the indicators (Figure 3.11), than in the restrictions to dismissing regular workers. On
average, higher regulation of the two types of temporary contracts tend to go hand in hand, although
there are many idiosyncratic situations. The geographic pattern is similar to the one for the regu lation of
regular workers. All common-law OECD countries (Canada, the United States, the United Kingdom,
New Zealand, Australia, Ireland and Israel) are at or near the bottom of the distribution of this type of
regulation. Four of the five countries with the highest regulation are EU countries: Luxembourg, Italy,
France and Spain.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
194
Figure 3.11. The OECD indicators: Strictness of regulation of hiring temporary workers
2019
Fixed-term contracts Temporary work agency contracts
Indicator score
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
Note: Range of indicator scores: 0-6. These aggregate indicators assign the same weight to fixed-term contracts and temporary work agency
contracts.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/botx04
The correlation between the two main indicators in the database – for dismissing regular workers and
for hiring temporary workers – is highly positive (Figure 3.12). Countries with higher regulation on one
tend to have higher regulation on the other. While this is generally the case, a small number of countries
(including in particular the Czech Republic, Israel and the Netherlands) appear to step somewhat out of
line. These are assessed to have relatively low regulation of temporary contracts given their high
regulation of regular contracts.
The overall positive relationship between the regulation of regular and temporary contracts is likely to
be the result of the differences in regulation of regular contracts together with policy makers’ desire to
restrain the use of temporary contracts. Where regular contracts are not much regulated, firms have few
incentives to replace regular with temporary contracts; the need to restrict the use of temporary contracts
is therefore not there. In countries with high regulation of dismissals of regular workers, strict regulation
of temporary contracts can help avoid that these are overused. As seen in the Netherlands, Portugal
and Sweden for example (OECD, 2014[60]), relatively low regulation of temporary contracts in situations
of high regulation of regular contracts can lead to strong, unintended labour market segmentation
between highly protected regular workers and weakly protected temporary workers. Section 3.4 will shed
further light on the extent to which recent reforms in job protection have gone in the direction of
increasing or reducing the regulatory divide between regular and temporary workers.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
195
Figure 3.12. Dismissal regulation for regular workers and hiring regulation for temporary workers
are positively correlated
2019
Hiring regulation for temporary
workers
5
4.5 TUR
4
LUX ITA
3.5
FRA
EST NOR ESP
3 GRC
SVK
KOR CHL
2.5 COL PRT
SVN
AUT MEX BEL CZE
2 LTU POL
DNK
HUN DEU LVA
ISL FIN ISR
1.5 CHE IRL JPN SWE NLD
1 AUS
NZL
0.5 GBR
USA CAN
0
0.5 1 1.5 2 2.5 3 3.5
Dismissal regulation for regular workers
Note: Range of indicator scores: 0-6. The indicator for dismissals of regular workers is for individual dismissals only, as the hiring indicator for
temporary workers is also based on hiring one worker.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/2w0ibc
3.4. Recent reforms in employment protection legislation
The focus of the analysis of employment protection reforms in this section is on the period between 2013,
the year of the previous release of the OECD Employment Protection Legislation indicators, and 2019, the
latest year in the database. This period follows the more immediate aftermath of the global financial crisis,
during which a number of countries – including Greece, Portugal and several other EU countries – had
eased strict dismissal regulations for regular workers to lower dualism in the labour market.
In the 2013-19 period, 21 OECD countries undertook at least one reform, as reflected in a change in the OECD
indicator score for job dismissal regulation for regular workers or hiring regulation for temporary employment.
The section focuses exclusively on those reforms that affected the indicators. It describes their main features
and impact on the assessed level of job protection, as measured by the indicators. Country-specific OECD
reports provide further details on some of the reforms (Carcillo et al., 2019[74]; OECD, 2018[75]).
These recent reforms can be classified into three categories. A first class of reforms went into the direction
of reducing restrictions to dismissing regular workers. Second, several countries changed restrictions on
the use of temporary employment, in some instances to reduce labour market duality. A third group of
countries (Belgium and the Netherlands) aimed to make regulation fairer and simpler by standardising the
protection of regular workers against different types of dismissal. Among the reforming countries, those
that relaxed dismissal regulation were more numerous than those which strengthened it, while a similar
number of countries increased and decreased the stringency of hiring regulation for temporary workers.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
196
Three figures provide the basis for the discussion in this section. Figure 3.13 depicts the changes in the
regulatory indicator for individual and collective dismissals of regular workers and in the regulatory indicator
for hiring temporary workers from 2013 to 2019. The following two figures, Figure 3.14 and Figure 3.15,
decompose these changes in terms of the main dimensions of the regulation.
Figure 3.13. Quantifying recent reforms in employment protection legislation
A. Strictness of regulation for individual and collective dismissals of regular workers
2019 2013
Indicator score
3.5
3
2.5
2
1.5
1
0.5
0
B. Strictness of regulation of hiring temporary workers
2019 2013
Indicator score
6
5
4
3
2
1
0
Note: Range of indicator scores: 0-6. Data for Colombia and Lithuania refer to 2014 instead of 2013.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/ptrn4y
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
197
Figure 3.14. Regulatory changes in recent reforms of dismissal regulation for regular workers
Indicator score
A. Strictness of protection for individual dismissals: B. Strictness of protection for individual dismissals:
Procedural requirements Notice and severance pay
2019 2013 2019 2013
4.5 4.5
4 4
3.5 3.5
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
BEL ESP FIN FRA HUN ITA LTU NLD PRT SVN BEL ESP FIN FRA HUN ITA LTU NLD PRT SVN
C. Strictness of protection for individual dismissals: D. Strictness of protection for individual dismissals:
Regulatory framework for unfair dismissals Enforcement of unfair dismissal regulation
2019 2013 2019 2013
4.5 4.5
4 4
3.5 3.5
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
BEL ESP FIN FRA HUN ITA LTU NLD PRT SVN BEL ESP FIN FRA HUN ITA LTU NLD PRT SVN
E. Strictness of protection for individual dismissals F. Strictness of protection for collective dismissals
2019 2013 2019 2013
4.5 4.5
4 4
3.5 3.5
3 3
2.5 2.5
2 2
1.5 1.5
1 1
0.5 0.5
0 0
BEL ESP FIN FRA HUN ITA LTU NLD PRT SVN BEL ESP FIN FRA HUN ITA LTU NLD PRT SVN
Note: Range of indicator scores: 0-6. All countries with changes in the indicator of protection for individual dismissals are displayed. Data for
Lithuania refer to 2014 instead of 2013.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/2i17w9
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
198
Figure 3.15. Regulatory changes in recent reforms of hiring regulation for temporary workers
A. Fixed-term contracts
2019 2013
Indicator score
6
5
4
3
2
1
0
BEL CAN DEU DNK ESP FRA ISL ITA JPN LTU NLD NOR POL PRT SVK SVN TUR
B. Temporary work agency contracts
2019 2013
Indicator score
6
5
4
3
2
1
0
BEL CAN DEU DNK ESP FRA ISL ITA JPN LTU NLD NOR POL PRT SVK SVN TUR
C. Overall regulation for hiring temporary workers
2019 2013
Indicator score
6
5
4
3
2
1
0
BEL CAN DEU DNK ESP FRA ISL ITA JPN LTU NLD NOR POL PRT SVK SVN TUR
Note: Range of indicator scores: 0-6. All countries with changes in the indicator of regulation for hiring temporary workers are displayed. Data
for Lithuania refer to 2014 instead of 2013.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/gdmq2o
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
199
3.4.1. Reforms that reduced the stringency of employment protection against dismissals
A number of countries implemented reforms aimed at easing dismissal regulations, sometimes at the same
time as they eased restrictions on the use of temporary employment. Major reforms (or sets of reforms),
involving several aspects of regulation, took place in France, Italy, Lithuania and Slovenia. Four other
OECD countries enacted more specific relaxations of regulations concerning dismissals of regular workers,
with notable effects on the job protection indicators.
Successive reforms in France
Between June 2013 and December 2017, France enacted a number of reforms that reduced the stringency
of job dismissal regulation. Among the main measures adopted, the August 2016 Labour law clarified the
definition of fair economic reasons for dismissals. A substantial reduction (larger than a specified threshold)
in at least one of several economic indicators listed in the law, such as orders or turnover, must now be
considered as fair by the court, while economic difficulties and technological changes can still be invoked
even if the reduction does not reach the threshold (as before the reform). Subsequently, in the second half
of 2017, the Ordonnances established a schedule for compensation following an unfair dismissal and
introduced a formalised scheme of collective termination by mutual agreement (rupture conventionnelle
collective). They also clarified the definition of procedural irregularity, which is far less penalised than unfair
dismissals. In particular, an irregular notification of the reason for the dismissal in the dismissal letter is no
longer sufficient to make the dismissal unfair. In addition, the maximum period to challenge a dismissal in
court successively decreased from five years in 2013 to twelve months in 2018. With regard to collective
dismissals specifically, the June 2013 Loi relative à la sécurisation de l’emploi limited the risk of a dismissal
being classed as unfair by requiring a validation of the social plan by the administration before the dismissal
(in firms with more than 50 employees), although this came at the cost of making the notification procedure
more complex.
Overall, the reforms moderately affected employment protection for regular workers, as measured by the
aggregate indicator (see Figure 3.13, Panel A). The reduction is mostly due to the lesser legal value of the
reason stated in the dismissal letter, which considerably simplified the notification procedure in the event
of an individual dismissal. In addition, the shortening of the period during which an employee can file a
complaint to enforce unfair dismissal regulation plays a role as well (see Figure 3.14, Panels A and D).29
The previous Version 3 of the OECD indicators did not capture the legal value of the dismissal letter for
regular workers. Hence, France is a good example for how the new Version 4 better captures reforms of
dismissal regulation. Box 3.3 provides further information on these improvements.
The Jobs Act and the elimination of the mobility allowance in Italy
The March 2015 Jobs Act in Italy considerably reduced employment protection for regular workers against
individual dismissals. One of the major measures of the Jobs Act was to eliminate the possibility of
reinstatement (in firms with more than 15 employees) in case of individual dismissals for economic reasons
and most collective dismissals, as well as in some cases of dismissals for personal reasons. 30 The Act
also replaced the mandatory conciliation phase that was to take place before all individual dismissals for
economic reasons, and before a dismissal for personal reasons if the employee requested it, with an ex-
post conciliation procedure. These two changes significantly reduced the job protection indicator for regular
workers (see Figure 3.13, Panel A), by limiting the consequences of both unfair individual and collective
dismissals, and by simplifying the individual dismissal procedure (see Figure 3.14, Panels A, C and F).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
200
Box 3.3. How the new version of the indicators helps capture reforms of dismissal regulations
A comparison of the changes from 2013 to 2019 in the aggregate dismissal indicator for regular workers
between the previous Version 3 and the new Version 4 shows how the new version allows to capture
regulatory changes that were not (or very little) captured by the previous version (Figure 3.16).
This is especially the case for France, Greece and Hungary. 1 The more detailed assessment reflects
the revision of existing, and the addition of new, items in a way that takes into account the legal value
of the reason stated in the dismissal letter (for France), the authorisation requirements for collective
dismissals (Greece) and the burden of proof in an unfair dismissal case (Hungary). Also, the changes
in the indicator resulting from the reforms in Belgium, Lithuania and the Netherlands are considerably
amplified in Version 4 with respect to those due to other reforms. For Belgium, Version 4 captures the
shift of the burden of proof, which Version 3 does not. For Lithuania, the new labour code affected
collective dismissal regulation in about the same way as individual dismissal regulation and hence is
reflected in the new indicator for collective dismissal, while Version 3 only reflects collective dismissals
to the extent that they represent additional restrictions relative to individual dismissals. For the same
reason, Version 3 does not capture the extension of individual severance pay to collective dismissals
in the Netherlands. By contrast, the Italian and Slovenian reforms are relatively less marked in Version 4
than Version 3. This is because the involvement of a third party in the notification procedure (for Italy)
and transfer requirements (for Slovenia) matter less, as these are now part of a broader range of factors
accounting for procedural requirements and unfair dismissal regulation.
Figure 3.16. The new version of the indicators tends to better capture reforms of dismissal
regulations
Strictness of regulation for individual and collective dismissals of regular workers, change from 2013 to 2019
Version 3 (previous version) Version 4 (new version)
Indicator score
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
BEL ESP FIN FRA GRC GBR HUN ITA LTU NLD PRT SVN
Note: Range of indicator scores: 0-6. All countries with changes in the indicator of protection for individual and collective dismissals are
displayed. Data for Lithuania refer to 2014 instead of 2013.
Source: OECD Employment Protection Legislation Database, http://oe.cd/epl.
StatLink 2 https://stat.link/ahubcy
1. Additional exercises conducted for important job protection reforms before 2013, such as the Spanish one, also show that the new
Version 4 provides a more comprehensive assessment of reforms.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
201
The decline in the indicator for collective dismissals (see Figure 3.14, Panel F) is not only due to the Jobs
Act, but also, and more significantly, to an earlier reform of severance pay that came into force only in
January 2017. Before then, in the event of a collective dismissal, an employee with at least 12 months of
job tenure was entitled to a “mobility allowance”, replacing unemployment benefits, and the employer had
to pay a contribution to this allowance that amounted to two to seven months of salary.31 The suppression
of this allowance implies that the employer now needs to pay merely the same contribution as for individual
dismissals, i.e. usually less than one month of pay.
The new labour code in Lithuania
In Lithuania, the new labour code, which entered into force in July 2017, provides more flexibility to
employers, regarding both the dismissal of regular workers and the use of temporary forms of employment.
The reform reduced notice period and severance pay and made reinstatement subject to approval by the
employer, although a specific compensation (capped at six months of wages) needs to be awarded if there
is no reinstatement. In addition, the new labour code includes a special procedure for dismissals at the will
of the employer, in which an employee can be dismissed for any reason at very short notice (three days)
and high severance pay (six months of wages). This new procedure, together with the elimination of
reinstatement obligations, significantly reduced the employment protection indicator for regular workers
(see Figure 3.13, Panel A), by restricting both the definition and the consequences of an unfair dismissal
(see Figure 3.14, Panel C).32
The new labour code also made the use of temporary employment easier. All restrictions on valid cases
for the use of fixed-term contracts were lifted, provided that they do not account for more than 20% of all
employment contracts. There are no limits on the number of successive fixed-term contracts, although they
can only be used for a maximum of two years for a given employee in the same function and five years in
different functions. As a result of these relaxations, Lithuania is the country with the largest decline in the
indicator for temporary employment from 2013 to 2019 (see Figure 3.13, Panel B).
The new Employment Relation Act in Slovenia
In Slovenia, the new Employment Relation Act came into force in April 2013, with significant reductions in
the protection of regular workers against dismissal and more flexible rules on the use of temporary work
agency employment.
Following this reform, opposition by a trade union can no longer inhibit the dismissal procedure, notice
period and severance pay are now lower, and a dismissal can be qualified as fair even when the employer
did not attempt to retrain or transfer the worker to another position. These three aspects of the reform
affected the indicator for regular employment (see Figure 3.14, Panels A, B and C), and the overall impact
on the protection of regular workers is considerable (see Figure 3.13, Panel A).
Regarding the regulation of temporary contracts, the same Act waived the 12 months duration limit of
successive temporary work agency assignments and the requirement for temporary work agencies to
provide annual reports. At the same time, the Act established that temporary work agency employment
can no longer exceed 25% of employment at the user firm. These changes reduce the indicator for
temporary employment (see Figure 3.13, Panel B).
Reforms in other countries
Based on the indicators, four other countries softened employment protection for regular workers, without
changing restrictions on the use of temporary employment. Finland extended the trial period from four to
six months in January 2017. In Greece, collective dismissals can take place without the approval of the
administration since 2017. In Portugal, two acts in August 2013 reduced severance pay from 20 days to
12 days per year of tenure and created a “compensation fund” to help finance it. 33 In 2013, the
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
202
United Kingdom halved the minimum period between notification to the administration and a collective
dismissal from 90 to 45 days.
The regulatory changes in Finland, Greece and the United Kingdom, which were either small or limited to
collective dismissals, do not change the assessed overall level of employment protection much. The impact
of the Portuguese reforms to severance pay and its financing is of similar magnitude to that of the major
reforms in France, Italy, Lithuania and Slovenia (see Figure 3.13, Panel A).
The large majority of countries among those reforming dismissal regulations for regular workers reduced
their stringency. Hungary and Spain are the only two countries that made dismissal regulations more
stringent (see Figure 3.13, Panel A), besides Belgium and the Netherlands where the increase in
regulatory stringency came more as a by-product of standardising protection against different types of
dismissals (see Section 3.4.3). In Hungary, the new code of civil procedure that entered into force in
January 2018 shifted the burden of proof on the employer in labour law cases. 34 In Spain, employers can
since January 2019 no longer hire under a “Permanent Employment Contract to Support Entrepreneurs”.
This contract was restricted to firms with less than 50 workers and included a trial period of one year
instead of four months.35 The changes to dismissal regulations in Greece and Hungary are reflected in the
new and finer Version 4 indicators, whereas the previous Version 3 would have shown no change
(see Box 3.3).
Exceptional measures in the COVID-19 crisis
The evaluation of employment protection legislation in this chapter is particularly pertinent given the current
COVID-19 health and economic crisis which has severely increased the dismissal risk for many employees
in the private sector. Job dismissal protection, when coupled with effective short-time work schemes, has
likely preserved jobs in countries badly affected by the crisis.36 As Box 3.4 details, a few EU countries have
further strengthened job dismissal protection in the crisis. The policy priorities in the area of employment
protection over the coming months will depend on the evolution of the pandemic, restrictions to economic
activity and developments in the labour market. They will likely require a shift from the immediate need to
help preserve existing jobs and incomes to increasing support for firm-to-firm worker mobility, also in light
of the structural changes in the labour market that the crisis is bringing (for example increased demand of
workers for health care and online and delivery services).
3.4.2. Reforms of hiring regulation for temporary workers
In Lithuania and Slovenia, as discussed, one component of a broader labour market reform was to facilitate
temporary employment. Five other countries undertook specific reforms to reduce restrictions on the use
of temporary employment, with corresponding effects on the indicator. By contrast, reforms in several
countries imposed additional restrictions on temporary contracts. Some of them introduced a maximum
cumulated duration for successive contracts or assignments, while Italy and, to a lesser extent, Denmark
implemented restrictions on the valid cases for use of these forms of employment.
Reforms that focused on facilitating temporary employment
In France, contracting on temporary employment was made easier in August 2015 with an increase in the
maximum number of successive fixed-term contracts from two to three, which is reflected in the
corresponding decline in the indicator for temporary contracts (see Figure 3.13, Panel B). The possibility
for collective agreements to derogate from restrictions to the use of temporary employment, introduced by
the 2017 Ordonnances, may further facilitate the use of temporary contracts in the future, as new
agreements are signed and extended.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
203
Box 3.4. Exceptional measures in job dismissal regulations during the COVID-19 crisis
The COVID-19 health and economic crisis has drastically reduced economic activity and put many firms
in financial difficulty. It has hence severely increased the risk for many private-sector employees to be
dismissed based on economic grounds. The crisis has heightened the risk to be dismissed for personal
reasons as well, especially for workers who have difficulties maintaining high work performance in the
face of increased care responsibilities (for example because they have families with a sick household
member or young children whose schools are closed). To a certain degree, differences in labour market
developments between countries reflect the stronger protection of employees in the European Union
where, up to this point, employment has been resilient and the increase in unemployment mild (see
Chapter 1).
Several countries in the European Union have further strengthened job dismissal protection during the
COVID-19 crisis. Four EU countries (France, Greece, Italy and Spain) have taken significant,
time-limited action to discourage economic redundancies, favouring the continuation of existing
employment relationships. Two EU countries (again Italy and the Slovak Republic) have strengthened
protection of employees against personal dismissals. The remainder of the box gives an overview of
these measures.
France announced increased scrutiny of collective dismissals for economic reasons by the authorities
as part of the notification procedure in firms with more than 50 employees. Firms are allowed to dismiss
employees, if they can show that they were already in economic difficulty before the COVID-19 crisis
and if they predict to be unable to restart part of their activity in the next six months. Economic dismissals
related to the COVID-19 crisis are, in principle, only allowed when a firm ceases its activity or based on
other, rather tightly defined economic considerations. Another objective has been to relax restrictions
relating to the renewal or prolongation of fixed-term contracts.
Greece prohibited dismissals of employees in firms that have suspended their operations because of
the lockdown measures. However, employers who are significantly affected by the COVID-19 crisis
may suspend the contracts of their employees for up to one month. Upon expiry of the suspension of
the contracts, companies must maintain the same number of employees for a period equal to that of
the suspension.
Italy blocked individual and collective dismissals for economic reasons for the first two months of the
crisis. The ban applies to layoffs on grounds connected to the reduction or transformation of activities,
reorganisation of work or business closure. In addition, Italy limited the scope for dismissals for personal
reasons. The absence from the workplace of parents of a child with a disability and of parents with a
child aged between 12 and 16 cannot constitute a just cause for contract termination, provided that the
employees communicate these reasons of absence to their employer.
Spain adopted the requirement that dismissals for reasons connected to the COVID-19 pandemic need
to be reviewed by a judge and will be qualified as either null or inadmissible. If the dismissal is judged
as null, the employee will be reinstated to the position. If the dismissal is seen as inadmissible, the
employee receives a compensation of 33 days of pay per year of tenure (in addition to the statutory
severance pay), as in any other case of unfair dismissal. In addition, Spain prolonged the duration of
fixed-term contracts that expire during the health emergency.
The Slovak Republic strengthened the protection of workers against personal dismissals. The measure
considers employees with a personal obstacle to working, such as caring for a sick family member or a
young child due to school closure, as temporarily unfit for work, thereby protecting them from dismissal
for the duration of their inability to work. The same provisions apply when the employee is subject to
quarantine or isolation.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
204
Four other countries lowered the restrictions for firms to hire temporary workers. In Norway, since July
2015, firms can use fixed-term contracts without restrictions on the reasons for their use, for at most
12 months and within the limit of 15% of their workforce. Turkey, where temporary work agency
employment was prohibited, introduced this form of work in 2016 in specific industries and for specific
reasons. Belgium and Spain in 2013 extended the use of temporary work agency employment. The reforms
in Norway and Turkey significantly reduced the restrictions to temporary employment, as reflected by the
indicator, while Belgium and Spain experienced only a small decline in the assessed degree of restrictions
to temporary employment (see Figure 3.13, Panel B).
Reforms that increased restrictions to temporary employment
A number of countries introduced a legal limit for the cumulated duration of fixed-term contracts or
temporary work agency assignments: Poland with 33 months for fixed-term contracts (February 2016),
Germany with 18 months for temporary work agency assignments (April 2017) and the Slovak Republic
with 24 months for temporary work agency assignments (March 2015). In Japan, the 2013 revision of the
Labour Contract Act made it possible for workers who have had a fixed-term contract for at least five years
to have their contract converted into a permanent one. These changes resulted in small, but nevertheless
noticeable increases in the indicator (see Figure 3.13, Panel B).37
In Italy, reforms first reduced, but later increased restrictions to temporary employment. The March 2014
Poletti decree abolished the obligation to provide a rationale when using fixed-term contracts and allowed
for five successive renewals (so long as these contracts do not exceed 20% of the number of regular
contracts in firms with more than five workers). The decree also allowed the use of temporary work agency
contracts with no justification. However, the reform of July 2018 restored, and even reinforced, the
obligation to provide a rationale when using a fixed-term contract for more than 12 months. Possible
extensions (up to three) for a maximum duration of 24 months are allowed for temporary and objective
needs or to replace some workers. The use of temporary work agency assignments was restricted to the
same reasons and maximum duration. These policy changes explain why, of all countries, Italy exhibits
the largest increase in restrictions to temporary employment during 2013-19 (see Figure 3.13, Panel B).
Restrictions on the use of both fixed-term and temporary work agency contracts explain this increase (see
Figure 3.15, Panels A and B).
Denmark implemented similar restrictions in July 2013 by conditioning temporary employment on objective
reasons. However, this reform focused only on renewals of temporary work agency assignments and
therefore resulted in a small increase in the indicator (see Figure 3.13, Panel B).
3.4.3. Reforms that standardised protection against different types of dismissal
Belgium and the Netherlands undertook significant employment protection reforms to standardise
regulation, either across workers with different employment status (blue-collar and white-collar workers in
Belgium) or across different types of dismissal (via the Public Employment Service or the Labour Court in
the Netherlands).
The single status in Belgium
The January 2014 reform in Belgium introduced a single status to abolish regulatory differences between
blue-collar and white-collar workers, which the Constitutional Court had considered discriminatory and
therefore unconstitutional.
The reform harmonised the time length of the period before the dismissal takes effect (delay before notice
and notice period) and severance pay, resulting overall in an increase in protection for blue-collar workers
and a decrease for white-collar workers. The reform also standardised the definition of an individual unfair
dismissal and the course and consequences of the associated unfair dismissal procedure. In particular,
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
205
the burden of the proof is always shared between the parties (it was previously with the employee for cases
involving white-collar workers), and the compensation granted in the event of an unfair dismissal is now
aligned with the lower compensation previously granted to white-collar workers. In addition, the reason for
the dismissal now needs to be provided upon the request of the employee. The reform also abolished the
trial period and expanded the use of outplacement regimes (i.e. services provided by the employer to help
dismissed workers find a new job) following an individual dismissal, previously restricted to older workers.
As of December 2016, the newly created “Reintegration programme” has continued the trend towards
more protection, by ensuring that workers with long-term medical incapacity are transferred to suitable
jobs.
These regulatory changes considerably increased the strictness of dismissal rules, as measured by the
indicators (see Figure 3.13, Panel A). This mainly reflects the new obligation to provide a reason for
dismissal in the notification procedure, as well as the elimination of the trial period within which workers
were not protected by unfair dismissal law. The partial shift of the burden of the proof towards the employer
in unfair dismissal cases concerning white-collar workers played a role as well (see Figure 3.14, Panels A,
C and D).
The Work and Security Act in the Netherlands
The Work and Security Act in the Netherlands comprised several labour market reforms to simplify the
dismissal law. Since July 2015, employers can no longer choose the procedure for dismissal (via the Public
Employment Service or the Sub-district Court); it is now determined by the reason for the termination (the
Public Employment Service deals with dismissals for economic reasons or long-term disability). In addition,
these two procedures were made more comparable: the notice period was extended to termination via the
Sub-district Court and severance pay (the “transition allowance”) to termination via the Public Employment
Service, including for collective dismissals.
The extension of the notice period and severance pay to all dismissal procedures, and of severance pay
to collective dismissals, explains most of the notable increase in the indicator for regular workers (see
Figure 3.13, Panel A and Figure 3.14, Panels A, B, E and F).38 The Work and Security Act also increased
restrictions to temporary employment by lowering the maximum duration of successive fixed-term contracts
from three to two years, slightly increasing the indicator for temporary workers (see Figure 3.13, Panel B).
Employment protection reforms and labour market duality
Employment protection reforms can reduce labour market duality between secure and precarious jobs by
lowering the opportunities and incentives for firms to replace regular with temporary contracts. This is the
case for example for reforms that restrict the valid cases for use of temporary employment as in Italy 2018.
Policy action to reduce labour market duality goes beyond restrictions on the use of temporary
employment. In particular, it frequently involves aligning social contributions and working conditions
between temporary and regular contracts. Such changes are largely beyond the scope of the indicators.39
For instance, Slovenia’s new Employment Relation Act introduced severance pay and additional social
security contributions for fixed-term contracts. Another example is from the Netherlands, where the
Balanced Labour Market Act, which entered into force in January 2020, increased the employer’s
unemployment insurance contribution rate for fixed-term contracts.40
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
206
3.5. Concluding remarks
Employment protection legislation is a widely debated policy and, as this chapter has shown, governments
in OECD countries continuously adapt regulations in this field. During the 2013-19 period, 21 of the 37
OECD countries undertook one or several reforms in employment protection that are reflected in changes
in the OECD Employment Protection Legislation indicators. Dismissal and hiring policies involve an
inherent trade-off between job security for workers who have a job and firm adaptability to changes in
demand conditions or technology. By comparing employment protection legislation in OECD countries, the
chapter sheds light on the relative importance that different systems attach to these twin aspirations. The
descriptive evidence in the chapter and the new indicators in the OECD Employment Protection Legislation
Database can be used as tools to further analyse what design of employment protection may deliver better
outcomes than others.41
The chapter gives a detailed description of dismissal regulations for regular workers and hiring regulations
for temporary workers in OECD countries, issues that are of particular relevance in the current environment
of high dismissal risk and low hiring chances. Subsequent work will look in greater depth at the regulations
applying to dismissals of workers on fixed-term contracts and expirations of these contracts, examine the
fate of temporary workers during the COVID-19 crisis and extend the update of the OECD Employment
Protection Legislation Database to several non-OECD countries. With greater country coverage and better
comparisons of employment protection of regular workers, employment protection of temporary workers
and the differences between the two, the aspiration is to make this work the most useful for policy makers
and citizens as they decide how they would like to shape job protection provisions in the future.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
207
References
Acharya, V., R. Baghai and K. Subramanian (2014), “Wrongful discharge laws and innovation”, [81]
Review of Financial Studies, Vol. 27/1, pp. 301-346.
Andrews, D., M. Adalet McGowan and V. Millot (2017), “Confronting the zombies: Policies for [72]
productivity revival”, OECD Economic Policy Papers, No. 21, OECD Publishing, Paris,
https://doi.org/10.1787/f14fd801-en.
Andrews, D. and F. Cingano (2014), “Public policy and resource allocation: Evidence from firms [34]
in OECD countries”, Economic Policy, Vol. 29/78, pp. 253-296.
Autor, D., J. Donohue and S. Schwab (2006), “The costs of wrongful-discharge laws”, Review of [16]
Economics and Statistics, Vol. 88/2, pp. 211-231.
Autor, D., W. Kerr and A. Kugler (2007), “Does employment protection reduce productivity? [37]
Evidence from US states”, Economic Journal, Vol. 117/521, pp. F189-F217.
Bai, J., D. Fairhurst and M. Serfling (2020), “Employment protection, investment, and firm [38]
growth”, Review of Financial Studies, Vol. 33/2, pp. 644-688.
Bartelsman, E., P. Gautier and J. De Wind (2016), “Employment protection, technology choice, [41]
and worker allocation”, International Economic Review, Vol. 57/3, pp. 787-826.
Bassanini, A. and E. Caroli (2015), “Is work bad for health? The role of constraint versus choice”, [9]
Annals of Economics and Statistics 119/120, pp. 13-37.
Bassanini, A. and F. Cingano (2019), “Before it gets better: The short-term employment costs of [32]
regulatory reforms”, Industrial and Labor Relations Review, Vol. 72/1, pp. 127-157.
Bassanini, A. and A. Garnero (2013), “Dismissal protection and worker flows in OECD countries: [25]
Evidence from cross-country/cross-industry data”, Labour Economics, Vol. 21/C, pp. 25-41.
Bassanini, A., L. Nunziata and D. Venn (2009), “Job protection legislation and productivity [42]
growth in OECD countries”, Economic Policy, Vol. 24/58, pp. 349-402.
Batut, C. and E. Maurin (2019), “From ultima ratio to mutual consent: The effects of changing [64]
employment protection doctrine”, IZA Discussion Papers, No. 12440.
Bentolila, S. and G. Bertola (1990), “Firing costs and labour demand: How bad is [12]
Eurosclerosis?”, Review of Economic Studies, Vol. 57/3, pp. 381-402.
Bjuggren, C. (2018), “Employment protection and labor productivity”, Journal of Public [43]
Economics, Vol. 157(C), pp. 138-157.
Boeri, T. (2011), “Institutional reforms and dualism in European labor markets”, Handbook of [27]
Labor Economics, Card, D. and O. Ashenfelter (eds.), North Holland, Amsterdam.
Boeri, T., P. Garibaldi and E. Moen (2017), “Inside severance pay”, Journal of Public Economics, [63]
Vol. 145/C, pp. 211-225.
Boeri, T. and J. Jimeno (2005), “The effects of employment protection: Learning from variable [17]
enforcement”, European Economic Review, Vol. 49/8, pp. 2057-2077.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
208
Boeri, T. and J. van Ours (2013), The Economics of Imperfect Labor Markets, Princeton [76]
University Press, Princeton.
Bottasso, A., M. Conti and G. Sulis (2017), “Firm dynamics and employment protection: [35]
Evidence from sectoral data”, Labour Economics, Vol. 48(C), pp. 35-53.
Bravo-Biosca, A., C. Criscuolo and C. Menon (2016), “What drives the dynamics of business [36]
growth?”, Economic Policy, Vol. 31/88, pp. 703-742.
Cahuc, P., S. Carcillo and A. Zylberberg (2014), Labor Economics, MIT Press, Cambridge, MA. [8]
Cahuc, P., O. Charlot and F. Malherbet (2016), “Explaining the spread of temporary jobs and its [55]
impact on labor turnover”, International Economic Review, Vol. 57/2, pp. 533-572.
Cahuc, P., F. Malherbet and J. Prat (2019), “The detrimental effect of job protection on [29]
employment: Evidence from France”, IZA Discussion Papers, No. 12384.
Campolieti, M. and C. Riddell (2020), “Does mediation-arbitration reduce arbitration rates? [67]
Evidence from a natural experiment”, Industrial and Labor Relations Review, Vol. 73/1,
pp. 211-235.
Campos, N. and J. Nugent (2018), “The dynamics of the regulation of labour in developing and [82]
developed countries since 1960”, The Political Economy of Structural Reforms in Europe,
Campos, N., P. De Grauwe and Y. Ji (eds.), Oxford University Press, Oxford.
Carcillo, S. et al. (2019), “Assessing recent reforms and policy directions in [74]
France: Implementing the OECD Jobs Strategy”, OECD Social, Employment and Migration
Working Papers, No. 227, OECD Publishing, Paris, https://doi.org/10.1787/657a0b54-en.
Centeno, M. and Á. Novo (2012), “Excess worker turnover and fixed-term contracts: Causal [51]
evidence in a two-tier system”, Labour Economics, Vol. 19/3, pp. 320-328.
Ciminelli, G., R. Duval and D. Furceri (2018), “Employment protection deregulation and labor [46]
shares in advanced economies”, IMF Working Papers, No. 18/186.
Ciminelli, G. and D. Furceri (2020), “A new database of employment protection legislation”, [80]
Unpublished manuscript.
Cingano, F. et al. (2016), “Employment protection legislation, capital investment and access to [40]
credit: Evidence from Italy”, Economic Journal, Vol. 126/595, pp. 1798-1822.
Cingano, F. et al. (2010), “The effects of employment protection legislation and financial market [39]
imperfections on investment: Evidence from a firm-level panel of EU countries”, Economic
Policy, Vol. 25/61, pp. 117-163.
Damiani, M., F. Pompei and A. Ricci (2016), “Temporary employment protection and productivity [56]
growth in EU economies”, International Labour Review, Vol. 155/4, pp. 587-622.
Denk, O. and A. Georgieff (forthcoming), “The 2019 OECD Employment Protection Legislation [62]
indicators: New insights on job dismissal regulation in OECD countries”, OECD Social,
Employment and Migration Working Papers, forthcoming, OECD Publishing, Paris.
Dolado, J., S. Ortigueira and R. Stucchi (2016), “Does dual employment protection affect TFP? [57]
Evidence from Spanish manufacturing firms”, SERIEs: Journal of the Spanish Economic
Association, Vol. 7/4, pp. 421-459.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
209
Duval, R. et al. (2018), “A narrative database of major labor and product market reforms in [78]
advanced economies”, IMF Working Papers, No. 18/19.
Espinosa, R., C. Desrieux and M. Ferracci (2018), “Labor market and access to justice”, [65]
International Review of Law and Economics, Vol. 54, pp. 1-16.
Espinosa, R., C. Desrieux and H. Wan (2017), “Fewer courts, less justice? Evidence from the [68]
2008 French reform of labor courts”, European Journal of Law and Economics, Vol. 43/2,
pp. 195-237.
Falk, A., D. Huffman and W. Macleod (2015), “Institutions and contract enforcement”, Journal of [47]
Labor Economics, Vol. 33/3, pp. 571-590.
Fraisse, H., F. Kramarz and C. Prost (2015), “Labor disputes and job flows”, Industrial and Labor [69]
Relations Review, Vol. 68/5, pp. 1043-1077.
García Pérez, J. and V. Osuna (2014), “Dual labour markets and the tenure distribution: [30]
Reducing severance pay or introducing a single contract”, Labour Economics, Vol. 29, pp. 1-
13.
García-Pérez, J., I. Marinescu and J. Vall Castello (2018), “Can fixed-term contracts put low [58]
skilled youth on a better career path? Evidence from Spain”, Economic Journal, Vol. 129/620,
pp. 1693-1730.
Garibaldi, P. (1998), “Job flow dynamics and firing restrictions”, European Economic Review, [13]
Vol. 42/2, pp. 245-275.
Gianfreda, G. and G. Vallanti (2017), “Institutions’ and firms’ adjustments: Measuring the impact [66]
of courts’ delays on job flows and productivity”, Journal of Law and Economics, Vol. 60/1,
pp. 135-172.
Gielen, A. and K. Tatsiramos (2012), “Quit behavior and the role of job protection”, Labour [18]
Economics, Vol. 19/4, pp. 624-632.
Griffith, R. and G. Macartney (2014), “Employment protection legislation, multinational firms, and [44]
innovation”, Review of Economics and Statistics, Vol. 96/1, pp. 135-150.
Grubb, D. and W. Wells (1993), “Employment regulation and patterns of work in EC countries”, [3]
OECD Economic Studies, Vol. 21, pp. 7-58,
http://www.oecd.org/economy/growth/33942847.pdf.
Haltiwanger, J., S. Scarpetta and H. Schweiger (2014), “Cross country differences in job [19]
reallocation: The role of industry, firm size and regulations”, Labour Economics, Vol. 26/C,
pp. 11-25.
Hijzen, A., L. Mondauto and S. Scarpetta (2017), “The impact of employment protection on [52]
temporary employment: Evidence from a regression discontinuity design”, Labour Economics,
Vol. 46, pp. 64-76.
Ichino, A., M. Polo and E. Rettore (2003), “Are judges biased by labor market conditions?”, [70]
European Economic Review, Vol. 47/5, pp. 913-944.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
210
Ichino, A. and R. Riphahn (2005), “The Effect of Employment Protection on Worker Effort: [83]
Absenteeism during and after Probation”, Journal of the European Economic Association,
Vol. 3/1, pp. 120-143, http://dx.doi.org/10.1162/1542476053295296.
ILO (2015), Employment Protection Legislation: Summary Indicators in the Area of Terminating [79]
Regular Contracts (Individual Dismissals), International Labour Office, Geneva.
Immervoll, H. and C. Knotz (2018), “How demanding are activation requirements for jobseekers? [71]
New evidence on activity-related eligibility criteria for unemployment and social assistance
benefits”, OECD Social, Employment and Migration Working Papers, No. 215, OECD
Publishing, Paris, https://dx.doi.org/10.1787/2bdfecca-en.
Kahn, L. (2010), “Employment protection reforms, employment and the incidence of temporary [53]
jobs in Europe: 1996-2001”, Labour Economics, Vol. 17/1, pp. 1-15.
Leonardi, M. and G. Pica (2013), “Who pays for it? The heterogeneous wage effects of [48]
employment protection legislation”, Economic Journal, Vol. 123/573, pp. 1236-1278.
Marinescu, I. (2009), “Job security legislation and job duration: Evidence from the United [20]
Kingdom”, Journal of Labor Economics, Vol. 27/3, pp. 465-486.
Martin, J. and S. Scarpetta (2012), “Setting it right: Employment protection, labour reallocation [28]
and productivity”, De Economist, Vol. 160/2, pp. 89-116.
Martins, P. (2009), “Dismissals for cause: The difference that just eight paragraphs can make”, [49]
Journal of Labor Economics, Vol. 27/2, pp. 257-279.
Messina, J. and G. Vallanti (2007), “Job flow dynamics and firing restrictions: Evidence from [31]
Europe”, Economic Journal, Vol. 117/521, pp. F279-F301.
Micco, A. and C. Pagés (2006), “The economic effects of employment protection: Evidence from [21]
international industry-level data”, IZA Discussion Papers, No. 2433.
Millán, A. et al. (2013), “How does employment protection legislation influence hiring and firing [22]
decisions by the smallest firms?”, Economics Letters, Vol. 121/3, pp. 444-448.
Mortensen, D. and C. Pissarides (1994), “Job creation and job destruction in the theory of [14]
unemployment”, Review of Economic Studies, Vol. 61/3, pp. 397-415.
Murphy, G., I. Siedschlag and J. McQuinn (2017), “Employment protection and industry [45]
innovation”, Industrial and Corporate Change, Vol. 26/3, pp. 379-398.
Neal, D. (1995), “Industry-specific human capital: Evidence from displaced workers”, Journal of [10]
Labor Economics, Vol. 13/4, pp. 653-677.
Nickell, S. (1978), “Fixed costs, employment and labour demand over the cycle”, Economica, [15]
Vol. 45/180, pp. 329-345.
OECD (2018), Good Jobs for All in a Changing World of Work: The OECD Jobs Strategy, OECD [26]
Publishing, Paris, https://doi.org/10.1787/9789264308817-en.
OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris, [11]
https://doi.org/10.1787/empl_outlook-2018-en.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
211
OECD (2018), OECD Reviews of Labour Market and Social Policies: Lithuania, OECD [75]
Publishing, Paris, https://doi.org/10.1787/9789264189935-en.
OECD (2017), OECD Employment Outlook 2017, OECD Publishing, Paris, [54]
https://doi.org/10.1787/empl_outlook-2017-en.
OECD (2016), OECD Employment Outlook 2016, OECD Publishing, Paris, [33]
https://doi.org/10.1787/empl_outlook-2016-en.
OECD (2014), OECD Employment Outlook 2014, OECD Publishing, Paris, [60]
https://doi.org/10.1787/empl_outlook-2014-en.
OECD (2013), OECD Employment Outlook 2013, OECD Publishing, Paris, [6]
https://doi.org/10.1787/empl_outlook-2013-en.
OECD (2011), Divided We Stand: Why Inequality Keeps Rising, OECD Publishing, Paris, [59]
https://doi.org/10.1787/9789264119536-en.
OECD (2010), OECD Employment Outlook 2010, OECD Publishing, Paris, [23]
https://doi.org/10.1787/empl_outlook-2010-en.
OECD (2004), OECD Employment Outlook 2004, OECD Publishing, Paris, [5]
https://doi.org/10.1787/empl_outlook-2004-en.
OECD (1999), OECD Employment Outlook 1999, OECD Publishing, Paris, [4]
https://doi.org/10.1787/empl_outlook-1999-en.
Pissarides, C. (2010), “Why do firms offer ‘employment protection’?”, Economica, Vol. 77/308, [1]
pp. 616-636.
Ruiz-Valenzuela, J. (2020), “Intergenerational effects of employment protection reforms”, Labour [61]
Economics, Vol. 62.
Saint-Paul, G. (2014), Frictions and Institutions, Bookboon, London. [77]
Salvanes, K. (1997), “Market rigidities and labour market flexibility: An international comparison”, [24]
Scandinavian Journal of Economics, Vol. 99/2, pp. 315-333.
Scarpetta, S. (2014), “Employment protection”, IZA World of Labor, No. 12. [2]
Söderqvist, F. and M. Lindberg (2019), “A technical note on improving the granularity and scope [73]
of the OECD EPL Index”, Paper presented at the OECD Expert Workshop on employment
protection legislation, Paris, 27 June 2019.
Turrini, A. et al. (2015), “A decade of labour market reforms in the EU: Insights from the LABREF [84]
database”, IZA Journal of Labor Policy, Vol. 4/12.
van der Wiel, K. (2010), “Better protected, better paid: Evidence on how employment protection [50]
affects wages”, Labour Economics, Vol. 17/1, pp. 16-26.
Venn, D. (2009), “Legislation, collective bargaining and enforcement: Updating the OECD [7]
employment protection indicators”, OECD Social, Employment and Migration Working
Papers, No. 89, https://doi.org/10.1787/223334316804.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
212
Annex 3.A. Methodology
Annex Table 3.A.1. Structure of Version 4 of the OECD EPL indicators for dismissing regular
workers
Item Version 4 values and description Assigned score
0 1 2 3 4 5 6
Procedural requirements before notice can be given
Item 1: 0 An oral statement is enough, or a
Notification written statement without reason
procedures for the dismissal to the employee
is required.*
1 A statement of the reason for the
dismissal to the employee in
writing or to a third party is
required.*
2 A written statement sets the limits
of disputes on the reason for the
dismissal once for all (given the
information available at the time of
writing).* Multiply by (6/4.5), so that the score lies between 0 and 6.
3 A consultation of or an inspection
by a third party is required.
4 An authorisation from a third party
is required.
Add +0.5 if a warning procedure (i.e. a
series of discussions with the employee
on the issues) is required in the event of a
personal dismissal.
* Add +0.5 if an additional consultation of
the employee only is required (+0.25 for
personal reasons only, +0.25 for
economic reasons only).
Item 2: Months
Time delay Estimated time includes, where relevant,
before notice the following assumptions: six days are
can be given counted in case of a required warning
procedure, one day when dismissal can
0 ≤ 0.75 ≤ 1.25 <2 < 2.5 < 3.5 ≥ 3.5
be notified orally or the notice can be
directly handed to the employee,
two days when a letter needs to be sent
by mail and three days when this must be
a registered letter.
Notice and severance pay
Item 3: 9 months Months
0 ≤ 0.4 ≤ 0.8 ≤ 1.2 < 1.6 <2 ≥2
Length of tenure
notice period 4 years tenure Months 0 ≤ 0.75 ≤ 1.25 <2 < 2.5 < 3.5 ≥ 3.5
20 years Months
<1 ≤ 2.75 <5 <7 <9 < 11 ≥ 11
tenure
Item 4: 9 months Months of pay
0 ≤ 0.5 ≤1 ≤ 1.75 ≤ 2.5 <3 ≥3
Amount of tenure
severance pay 4 years tenure Months of pay 0 ≤ 0.5 ≤1 ≤2 ≤3 <4 ≥4
20 years Months of pay
0 ≤3 ≤6 ≤ 10 ≤ 12 ≤ 18 > 18
tenure
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
213
Item Version 4 values and description Assigned score
0 1 2 3 4 5 6
Regulatory framework for unfair dismissals
Item 5: Individual dismissals: weighted average of Items 5a, 5b, 5c and 5d, with the following weights to ensure equal weight for
Definition of dismissals for economic reasons and dismissals for personal reasons:
unfair 5a: (1/2)*(1/3)
dismissal 5b: (1/2)*(1/3)
5c: (1/2)*(1/3)
5d: 1/2.
Collective dismissals (only for economic reasons): weighted average of Items 5a, 5b and 5c, with equal weights (1/3 each).
Item 5a: 0 No justification is required, or any
Dismissal for justification is fair.
economic 1 The judges can only question
reasons: patently irrational decisions or
Degree of false reasons.
freedom of the Multiply by 2, so that the score lies between 0 and 6.
2 The judges can question the
judges operational need of the dismissal
decision.
3 Economic reasons are not a valid
justification.
Item 5b: 0 +1 for each of the following
Dismissal for 1 alternatives/obligations:
economic 2 - Transfer
reasons: 3 - Retraining
Specific 4 - Outplacement services and
alternatives to 5 training
the dismissal See previous column.
- Priority for re-hiring and/or no
and binding fixed-term contract on a similar job
obligations in - Social plan (even if it includes
the event of a some of the above obligations).
dismissal
6 Economic reasons are not a valid
justification.
Item 5c: 0 No worker selection criteria or only
Dismissal for performance criteria.
economic 1 Objective selection criteria other
reasons: Multiply by 3, so that the score lies between 0 and 6.
than performance.
Selection 2 Economic reasons are not a valid
criteria justification.
Item 5d: 0 No justification is required, or any
Dismissal for justification is fair.
personal 1 i) Insufficient performance,
reasons: ii) unsuitability for medical reasons
Fair reasons and iii) unsuitability due to
for dismissal insufficient skills/qualifications are
NA for fair reasons for dismissal.
collective 2 One reason among these three
dismissals cannot be a ground for a dismissal.
3 Two reasons among these three Multiply by (6/4), so that the score lies between 0 and 6.
cannot be grounds for a dismissal.
4 These three reasons cannot be
grounds for a dismissal.
Add +0.25 per fair reason (among the
three mentioned under value 1) for which
constraining alternatives to dismissal
(such as transfer or retraining) must be
attempted.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
214
Item Version 4 values and description Assigned score
0 1 2 3 4 5 6
Item 6: Months
Length of trial It is defined as the period within which ≥ 24 > 12 >9 >5 > 2.5 ≥ 1.5 < 1.5
period regular contracts are not fully covered by
NA for employment protection provisions and Add +1 to the score when notice period plus equivalent compensation just before
collective unfair dismissal claims usually cannot be the end of the trial period is at least two weeks.
dismissals made. Add +1 to the score when there is no trial period.
Multiply by (6/7), so that the score lies between 0 and 6.
Item 7: Compensation in months of pay
Compensation Typical compensation at 20 years of
for the tenure, including back pay and other
employee compensation, but excluding ordinary ≤3 ≤8 ≤ 12 ≤ 18 ≤ 24 ≤ 30 > 30
following an severance pay.
unfair
dismissal
Item 8: 0 No right or practice of
Possibility of reinstatement.
reinstatement 1 Reinstatement rarely or sometimes
following an made available.
unfair Multiply by 2, so that the score lies between 0 and 6.
2 Reinstatement fairly often made
dismissal available.
3 Reinstatement (almost) always
made available.
Enforcement of unfair dismissal regulation
Item 9: Duration in months
Before
Maximum time Maximum time period after the contract
dismissal
to make a termination date up to which an unfair ≤1 ≤3 ≤6 ≤9 ≤ 12 > 12
takes
claim of unfair dismissal claim can be made.
effect
dismissal
New Item 22: Does the burden of proof lie with the
Burden of employee only?
proof when the
employee files - - Yes - No - -
a complaint for
unfair
dismissal
New Item 23: Does an ex-ante validation of the
Ex-ante dismissal (e.g. by an external authority)
- - Yes - No - -
validation of limit the scope of (or prevent entirely)
the dismissal unfair dismissal complaints?
New Item 24: 0 Resignation or some form of
Pre- mutual consent gives access to
termination unemployment benefits under the
resolution same conditions as in the event
mechanisms of a dismissal.
granting 1 Resignation or some form of
unemployment mutual consent gives access to
benefits unemployment benefits, but with a
Multiply by 3, so that the score lies between 0 and 6.
longer waiting period or lower
replacement rate compared with
a dismissal.
2 Neither resignation nor any form of
mutual consent gives access to
unemployment benefits, while
dismissal gives access to
unemployment benefits.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
215
Note: Where there are differences by types of dismissal, the scored value is the average of the values for a dismissal for personal and for
economic reasons. Where there are differences by firm size, the scored value is the average of the values for a firm with 35, 150 and
350 employees. Approximately the same coding is used for individual and collective dismissals; the only differences are that, as collective
dismissals can only occur for economic reasons, Items 5d and 6 are removed, and warning procedures are not taken into account in Item 1. The
indicator for collective dismissals evaluates the average for dismissals of 10, 45 and 120 workers by a firm within one month (based on the
average of the three firm sizes for 10 dismissals, the two larger firm sizes for 45 dismissals and the largest firm size for 120 dismissals).
Annex Table 3.A.2. Weighting in the OECD EPL indicators (Version 4) for dismissing regular
workers
Weight: Weight:
Category of
Individual Collective
dismissal Lower-level elements of dismissal regulation
dismissals dismissals
regulation
(5/7) (2/7)
Procedural 1. Notification procedures 1/2 1/2
requirements (1/4) 2. Time delay before notice can be given 1/2 1/2
Notice and 3. Length of notice period 3/7 3/7
severance pay (1/4) 4. Amount of severance pay 4/7 4/7
5. Definition of unfair dismissal 1/4 1/3
Regulatory 6. Length of trial period (the initial period in which unfair dismissal claims cannot be made) 1/4 -
framework for unfair
dismissals (1/4) 7. Compensation to the worker following unfair dismissal 1/4 1/3
8. Possibility of reinstatement following unfair dismissal 1/4 1/3
9. Maximum time to make a claim of unfair dismissal 1/4 1/4
Enforcement of 22. Burden of proof when the worker files a complaint for unfair dismissal 1/4 1/4
unfair dismissal
regulation (1/4) 23. Ex-ante validation of the dismissal by an external authority 1/4 1/4
24. Pre-termination resolution mechanism granting unemployment benefits 1/4 1/4
Annex Table 3.A.3. Structure of Version 3 of the OECD EPL indicator for hiring temporary workers
Item Version 3 values and description Assigned score
0 1 2 3 4 5 6
Fixed-term contracts
Item 10: 0 There are no restrictions on the use of fixed-
Valid cases for use of term contracts.
fixed-term contracts 1 Exemptions exist on both the employer and
employee side.
2 Specific exemptions apply in situations of
employer need (e.g. starting a new activity) Multiply by 2, so that the score lies between 0 and 6.
or employee need (e.g. workers in search of
their first job).
3 Fixed-term contracts are permitted only for
“objective reasons” or “material situation”,
i.e. to perform a task which itself is of fixed
duration.
Item 11: Number
Maximum number of No
≥5 ≥4 ≥3 ≥2 ≥ 1.5 < 1.5
successive fixed-term limit
contracts
Item 12: Months
Maximum cumulated No
≥ 36 ≥ 30 ≥ 24 ≥ 18 ≥ 12 < 12
duration of successive limit
fixed-term contracts
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
216
Item Version 3 values and description Assigned score
0 1 2 3 4 5 6
Temporary work agency contracts
Item 13: 0 Temporary work agency employment is
Types of work for which generally allowed, with no or minimal
temporary work agency restrictions.
employment is legal 1 Temporary work agency employment is
generally allowed, with specified exceptions.
2 Temporary work agency employment is only Multiply by 6/4, so that the score lies between 0 and 6.
allowed for “objective reasons”.
3 Temporary work agency employment is only
allowed in specified industries.
4 Temporary work agency employment is
illegal.
Item 14: Are there restrictions on the number of renewals of
Restrictions on number temporary work agency contracts?
of renewals of - - No - Yes - -
temporary work agency
contracts
Item 15: Months
Maximum cumulated No
≥ 36 ≥ 24 ≥ 18 ≥ 12 >6 ≤6
duration of temporary limit
work agency contracts
Item 16: 0 No authorisation or reporting requirements.
Authorisation and 1 Special administrative authorisation required.
reporting obligations 2 Periodic reporting obligations required. Multiply by 2, so that the score lies between 0 and 6.
3 Both authorisation and reporting
requirements.
Item 17: 0 No requirement for equal treatment.
Equal treatment of 1 Equal treatment required regarding pay or
regular and agency working conditions; Multiply by 3, so that the score lies between 0 and 6.
workers at the user firm
2 Equal treatment required regarding pay and
working conditions.
Note: Where there are differences by firm size, the scored value is the average of the values for a firm with 35, 150 and 350 employees.
Annex Table 3.A.4. Weighting in the OECD EPL indicators (Version 3) for hiring temporary workers
Category of hiring
Lower-level elements of hiring regulation Weight
regulation
10. Valid cases for use of fixed-term contracts 1/2
Fixed-term contracts
11. Maximum number of successive fixed-term contracts 1/4
(1/2)
12. Maximum cumulated duration of successive fixed-term contracts 1/4
13. Types of work for which temporary work agency employment is legal 1/3
14. Restrictions on the number of renewals of the assignment to the user firm 1/6
Temporary work agency
15. Maximum cumulated duration of successive assignments to the user firm 1/6
contracts (1/2)
16. Authorisation and reporting obligations 1/6
17. Equal treatment of regular workers and temporary work agency workers at the user firm 1/6
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
217
Notes
1
Overviews of the main theoretical frameworks can be found in academic books such as Boeri and van
Ours (2013[76]), Cahuc, Carcillo and Zylberberg (2014[8]) and Saint-Paul (2014[77]).
2
However, a positive job tenure-protection profile tends to be efficient in the case of jobs requiring
continuous firm-specific investments by workers (Boeri, Garibaldi and Moen, 2017[63]).
3
A notable exception is Acharya, Baghai and Subramanian (2014[81]) who provide evidence for the
United States that the passage of wrongful discharge laws spurred innovation.
4
However, fixed-term contracts have been shown to induce higher effort by workers if they expect a high
probability of conversion of their contract into an open-ended one (Ichino and Riphahn, 2005[83]).
5
The collection of the information and data for the OECD Employment Protection Legislation Database is
the result of a large collaborative effort by the labour ministries of OECD countries and the OECD
Secretariat, which remains however solely responsible for the indicators. The project benefited as well
from the insightful comments and suggestions by a group of academic and policy experts.
6
One interesting avenue for future research would be to evaluate employment protection for public-sector
employees, also in comparison to employment protection for private-sector employees.
7
The only exception is the category notice and severance pay, in which severance pay carries a slightly
higher weight (4/7) than notice period (3/7). The rationale is that workers can still contribute to the firm’s
output while on notice. Consequently, the net cost to the firm is higher for a month of severance pay than
for a month of advance notice. Yet, advance notice may protect workers better than severance pay, in
particular if early support by the public employment services is in place (OECD, 2018[11]).
8
The length of the trial period enters only the employment protection indicator for individual dismissals and
not the indicator for collective dismissals, as trial periods are always associated with an individual worker.
9
This is not the case for the employment protection indicator for collective dismissals, because collective
dismissals can only occur for economic reasons.
10
As mentioned, where there are differences by firm size, the scored value is the average of the values
for a firm with 35, 150 and 350 employees. The indicator for collective dismissals is based on the average
of the three firm sizes for 10 dismissals, the two larger firm sizes for 45 dismissals and the largest firm size
for 120 dismissals.
11
As 120 workers exceeds the dismissal threshold in all countries with specific legislation for collective
dismissals (but not 10 and 45 workers), the regulatory differences between individual and mass dismissals
enter the indicator for collective dismissals with a minimum weight of around one-third.
12
An alternative approach is to focus on reforms in employment protection. Duval et al. (2018[78]) compile
a dataset of major reforms in this area with a 0-1 reform variable covering 26 advanced economies over
the period 1970-2015. The Labour Market Reforms (LABREF) Database by the European Commission
monitors qualitatively job protection reforms in the 27 EU countries since 2000 (Turrini et al., 2015[84]).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
218
13
Other researchers have developed their own databases, among them Campos and Nugent (2018[82])
and Ciminelli and Furceri (2020[80]).
14
ILO (2015[79]) shows that the OECD indicator for regular workers against individual dismissals is strongly
correlated with the EPLex (the correlation coefficient equals 0.81). Calculations by the OECD suggest a
significant, albeit weaker, correlation with a constructed CBR-LRI indicator focusing on dismissal regulation
(0.66).
15
The CBR-LRI and the Doing Business indicators go well beyond employment protection, which is only
one focus of these databases in terms of institutional coverage.
16The United States has a rising number of cases in which employees pursue wrongful termination claims
by alleging that the dismissal was based on an “implied contract” for continued employment (despite no
formal contract), because certain assurances for continued employment were given. The probability of
succeeding in claiming the existence of an “implied contract” increases with seniority and is likely to be
extremely low or zero for only a few years of job tenure.
17
In Sweden, a strict last-in-first-out rule applies except when collective agreements establish otherwise
(see Box 3.2).
18
For example, Poland has a special type of fixed-term contracts which in practice is used as a trial period.
Belgium allows the use of temporary work agency employment for “insertion”, i.e. to be hired under a
regular contract for the same job after the assignment.
19
In the United States, the prescription period in the situation of an “implied contract” depends on the state
jurisdiction, but it is typically very long (for example four years in California and six years in New York).
20
In the United States, when an employer offers voluntary separations but intends to implement involuntary
dismissals in case of an insufficient number of volunteers, employees who take the voluntary separation
package may be eligible for unemployment benefits.
21
In cross-country regressions of the aggregate indicator on the four broad categories (one at a time),
enforcement of unfair dismissal regulation is positively correlated with the aggregate indicator, but not in a
statistically significant fashion. The other three categories are positively correlated with the aggregate
indicator, statistically significant at the 1% level. The estimate is also much lower for enforcement of unfair
dismissal regulation than for the other three categories.
22
The pairwise negative correlations are not statistically significant. In line with the reasons provided,
enforcement is more negatively correlated with procedural requirements and the regulatory framework for
unfair dismissals.
23
While for individual dismissals this is only the case in the Netherlands and, to some extent, in Germany
and Sweden, for collective dismissals other countries as well require an authorisation pre-dismissal that
can serve as validation post-dismissal.
24
While additional notifications are required in Israel (to the Employment Service Bureau) and Korea (to
the Ministry of Labour), they do not impose significant additional constraints on firms and hence are not
reflected in the indicators.
25
The way the indicators evaluate regulation of collective dismissals is by the stringency that would apply
to one worker if this worker was dismissed by a collective dismissal. While this definition does not allow for
fixed (in particular procedural) costs being spread across several workers, it is not obvious how the design
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
219
of the indicators could, in a pragmatic way, take account of this possible fixed cost nature of some of the
aspects relating to dismissal regulation.
26
These are Canada and the United States (North America) and Australia and New Zealand (Australasia).
27
They can also affect restrictions on the use of temporary contracts (e.g. in France, Italy, the Netherlands,
Sweden).
28
These are calculations for the whole business economy using the OECD Structural and Demographic
Business Statistics Database.
29
By contrast, the clearer definition of fair economic reasons for dismissals on economic grounds had little
impact; it limits the freedom that judges have in their decision to classify a dismissal as unfair, but this
accounts for only one-sixth of the score for the lower-level element “Definition of unfair dismissal”.
Moreover, the new schedule for compensation following an unfair dismissal does not sufficiently reduce
the compensation to be reflected in the indicators. Finally, the introduction of the rupture conventionnelle
collective also did not change the indicators, as it only formalised the existing voluntary separation plans
which already granted eligibility for unemployment benefits.
30
Initially, the Jobs Act set a schedule of two months of wages per year of tenure for compensation
following an unfair dismissal for these types of dismissal. However, the Constitutional Court rejected this
in a November 2018 decision. The indicators do not reflect this temporary change, as it is difficult to assess
the extent to which the schedule was enforced and for how long. However, they reflect the new range of
possible values for compensation, which now lie between 6 and 36 months of wages, compared with 12
and 24 before the Jobs Act. The resulting increase in the indicators mitigates the downward effect of the
Jobs Act.
31
The contribution of the employer was higher in the case of use of Cassa Integrazione Guadagni (the
Italian short-time work scheme) before the collective dismissal.
32
The changes in notice period and severance pay induced by the different reform measures, including
the introduction of dismissals at will, exactly cancel each other out in the indicators, so that no change is
visible on the component notice and severance pay (see Figure 3.14, Panel B).
33
In May 2014, the obligations to attempt a transfer before dismissal and to follow criteria provided by the
Labour Code when selecting the worker(s) to be dismissed were reintroduced, to comply with the
September 2013 ruling of the Constitutional Court. These obligations had been lifted in August 2012. The
indicators do not reflect these temporary changes, as it is difficult to assess the extent to which these were
enforced and for how long.
34
In particular, it is the employer’s responsibility to show the contents of any collective agreement or
internal document required for adjudicating the dispute.
35
Moreover, a severance pay subsidy in the case of fair dismissals in firms with less than 25 employees
was suppressed in December 2013 (OECD, 2016[33]).
36
In the absence of effective short-time work schemes, excessively strict job protection may, however,
lead to firm bankruptcies, thereby failing to preserve jobs.
37
The indicator for temporary contracts in Portugal increases due to the end of an extraordinary regime of
renewals applying to fixed-term contracts until December 2016. An additional increase from 2.46 to 2.58
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
220
is due to the further reduction in the maximum cumulated duration of fixed-term contracts implemented in
September 2019.
38
The notice period can only start at the end of the month. Therefore, the extension of the notice period to
all dismissal procedures also increased the delay before notice can be given.
39
Still, the indicators include one item on equal treatment of regular and temporary work agency workers,
which makes it possible to reflect part of these changes. For example, in Québec (Canada), since June
2018, employers can no longer set different wages for workers solely based on their employment status,
and temporary work agencies are prevented from paying wages that are lower than those of regular
employees performing the same tasks. Similarly, in Iceland, since 2013, temporary work agency workers
should be paid at least the same wages and benefit from the same facilities as regular employees. In the
Netherlands, the possibility to deviate from the basic wage during the first 26 weeks of a temporary work
agency assignment has been removed from the Collective Labour Agreement for Temporary Agency
Workers. These changes explain the increase in the indicators for temporary work agency employment for
these countries (see Figure 3.15, Panel B).
40
The Balanced Labour Market Act also made dismissals easier (employers can now combine personal
reasons for dismissal that are by themselves insufficient to justify a dismissal), aligned severance pay for
permanent and temporary workers, modified their calculation rules and extended the maximum cumulated
duration for fixed-term contracts. The reform did not change the indicator of protection for regular workers,
because the indicator does not capture the possible combination of insufficient grounds and the opposite
changes in severance pay at different tenures cancel each other out. However, it reduced the indicator for
temporary workers from 1.48 to 1.23.
41
The chapter mostly focuses on developments since 2013. Researchers and other users interested in
the complete time series of the data and information available since 1985 can find these on the dedicated
website (http://oe.cd/epl). Using these data, Denk and Georgieff (forthcoming[62]) depict long time series
for employment protection.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
221
4 What is happening to middle-skill
workers?
Driven by mega trends such as automation, ageing and globalisation, the
share of middle-skilled jobs has been declining in the majority of OECD
labour markets (a process also referred to as job polarisation). Middle-skill
jobs are defined as occupations in the middle of the occupation-wage
distribution. One little explored question is what is happening to the workers
who have traditionally occupied these jobs? This chapter starts by
examining whether the fall in the share of middle-skill employment is
explained primarily by attrition or transitions. Attrition accounts for fewer
younger workers entering these jobs compared to older workers retiring.
Transitions explain changes in career patterns after a person has started
working. The chapter then studies the characteristics of what would have
been a “typical” middle-skill worker and uses this profile to examine how the
jobs they hold have changed over time.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
222
In Brief
Key findings
The share of middle-skill jobs – defined as occupations whose average wages place them in the middle
of the wage distribution – declined in OECD countries over the past two decades. From the mid-1990s
(1994-1996) to the latest available period (2016-2018), the share of total employment accounted for by
middle-skill occupations – e.g. truck drivers and machine operators for men, cashiers and secretaries
for women – declined over 11 percentage points. This contrasts with growth of 9 percentage points in
high-skill occupations and of 3 percentage points in low skill occupations. The employment share of
middle-skill occupations declined because the number of jobs in low- and particularly high-skill
occupations grew strongly, while the number of middle-skill jobs held broadly steady.
The key question addressed in this chapter is how this decline in employment shares of middle-skill
occupations has taken place. Is it due to attrition or transitions? This difference amounts to whether the
decline is borne by new entrants to the labour market, or mid-career workers. Attrition is primarily driven
by new cohorts of workers entering the labour force in middle-skill occupations at lower rates than
previous cohorts. Transitions are changes in the career patterns of different occupational groups, in
which workers reallocate to other skill groups (including possible spells of non-employment) during their
careers at different rates than in the past. Transitions could include earlier cohorts who started their
working lives in middle-skill jobs being laid off mid-career, for example, while attrition could take the form
of young workers entering the labour market in different occupations. Attrition is about labour market
entry, while transitions account for mid-career changes.
The main findings of the chapter are as follows:
In contrast to popular perceptions (and anecdotal evidence concerning mass dismissals), this
chapter shows, using data primarily from European OECD countries, that the share of
middle-skill employment has declined more because of attrition than transitions. Changing
patterns of labour market entry appear to be key. For cohorts born before 1970, some 32.8% of
workers were employed in middle-skill occupations when aged 25-29. For cohorts born after
1970, this share decreased to 26.5%. The share of those in high-skill employment exhibited the
reverse pattern.
Meanwhile, except during the financial crisis, workers separating from middle-skill jobs have
tended to transition to other middle-skill jobs or to non-employment at similar rates as in the past.
The changing demographic composition of younger cohorts entering the labour force is a
contributing factor to the decline in middle-skill employment. New entrants to the labour force
are more likely to be women and have a tertiary education compared with 20 years ago. Workers
without a tertiary education were the most likely to hold middle-skill jobs in the past. There are
therefore fewer workers entering the labour force who would typically hold middle-skill jobs,
which would be expected to cause a mechanical decline in the share of middle-skill employment.
However, the analysis confirms that this explains only part of the contraction in middle-skill
employment shares.
New cohorts of workers without tertiary education are less likely to start their careers in
middle-skill jobs. In fact, even those who would have been regarded as “typical” middle-skill
workers in the past are now much less likely to start working in middle-skill jobs, and more likely
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
223
to be in low-skill employment. A formal decomposition of changes in employment shares
confirms that attrition accounts for most of the decline in middle-skill employment.
Employment shares of women without a tertiary degree in low-skill occupations have increased
dramatically. The largest percentage point increases for those holding low-skill jobs have been
among men and women with a middle level of education (an upper-secondary qualification rather
than a tertiary degree).
In most countries, middle-educated workers are now also less likely to hold high-skill jobs than
in the 1990s. A few countries appear to be exceptions, however. Over the past two decades,
Sweden, Germany, Norway and Denmark have seen a significant rise in the propensity of
middle-educated workers to be employed in high-skill occupations, for both men and women.
Although not addressed explicitly in this chapter, these countries place an emphasis on
vocational education and training, as well as maintaining a tradition of cooperative social
dialogue (see also Chapter 5).
Introduction
At its peak, manufacturing employed millions of workers at wages solidly in the middle of the pay
distribution, which helped support a strong middle class across industrialised nations (Helper, Krueger
and Wial, 2012[1]; OECD, 2019[2]). However, since at least the 1970s, employment in OECD economies
has been shifting from manufacturing to service industries. At the same time, the share of employment
in middle-skill occupations within industries has steadily decreased. While these jobs declined,
employment grew in high-skill occupations such as human resources administrators and information
technology support. Low-skill service jobs such as janitors, home care workers and retail sales assistants
flourished as well. Employment moved from assembling cars on the shop floor to stocking shelves on
the sales floor.
Economists and policy makers termed this trend job polarisation. Defining the skill level of jobs by the
average wage in an occupation (similarly using task content, or education), economists found that the
employment shares of both higher- and lower-skill occupations have increased in many but not all
countries, while shares of employment in middle-skill occupations have declined (Autor, Levy and
Murnane, 2003[3]; Goos and Manning, 2007 [4]; Goos, Manning and Salomons, 2009 [5]). Subsequent
research identified automation as the main cause of this job polarisation (OECD, 2017[6]; Autor and Dorn,
2013[7]).1 Increasing penetration of information technology and robotics have eroded jobs consisting of
routine tasks. These routine jobs were traditionally situated in the middle of the skill distribution.
A common concern is that the decline of middle-skill employment may have resulted in distress, job
insecurity and displacement for workers who held those jobs – see e.g. Autor (2010[8]), Cortes (2016[9]),
and OECD (2017[10]). Yet this need not be the case. If job polarisation results in workers transitioning to
higher-paid jobs, its effects may be more benign than previously thought. Moreover, the theory of j ob
polarisation and the history of job destruction is ambiguous with respect to overall employment and
wages in the long run (Autor, 2015[11]; Autor, 2015[12]; Acemoglu and Restrepo, 2018 [13]). The important
question is then how are middle-skill workers who need to transition to new jobs affected by job
polarisation?
This chapter reviews where workers holding middle-skill jobs are going in the face of the shrinking share
of such jobs. The first question it seeks to answer is how this adjustment has taken place. Are firms
increasingly dismissing middle-skill workers, forcing mid-career workers to find new employment in other
skill groups (“transitions”), or are older middle-skill workers gradually retiring and younger workers
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
224
entering other, growing occupations (“attrition”)? In addition to the adjustment mechanism, this chapter
asks what types of jobs workers who fit the traditional profile of a middle-skill worker are taking.
Understanding both the nature of adjustment and the final destinations of middle -skill workers will help
inform policy makers on the types of policies that can aid this restructuring of OECD labour markets. For
example, if job polarisation is driven mainly by workers losing middle-skill jobs, then labour market policy
needs to focus on helping these workers make the transition to other occupations where employment
opportunities are emerging (OECD, 2018[14]). By contrast, if patterns of labour market entry are the key
factor shaping this process, then policies need to accompany young workers in starting their career and
ensuring its sustainability over time (see e.g. Chapter 5).
The analysis begins by briefly documenting the near universal trend of job polarisation, and disentangling
the dynamics of the shifting share of jobs across occupation groups (Section 4.1 and Section 4.2). This
includes evidence for whether middle-skill employment shares have shrunk due to transitions of mid-
career workers, or through attrition and differences in the labour market entry patterns of younger
cohorts. In Section 4.3 the chapter turns to building the profile of the “typical” middle-skill worker of the
past and identifies the characteristics associated with middle-skill work two decades ago. Section 4.4
uses the profile of middle-skill workers to shed light on where they are going. Specifically, the analysis
shows the types of occupations that workers of different demographic groups are employed in compared
with two decades ago.
4.1. How are middle-skill jobs changing?
In order to better assess how middle-skilled jobs are changing, this section provides descriptive evidence
on how the share of middle-skill jobs has changed over the preceding decade. The analysis in this
section confirms that the share of employment in middle-skill jobs – defined by the average wage in an
occupation (Box 4.1) – declined across OECD countries over the past decade.
Confirming earlier analyses, the share of employment in middle-skill occupations in OECD countries
declined from the mid-1990s (1994-1996) to the mid-2010s (2016-2018). The analysis in this chapter
relies on cross-sectional and panel survey data from across Europe and the United States (see Annex
4.A). Figure 4.1 depicts the share of employment in middle-skill jobs at both time periods for a broad
range of OECD countries. 2 Across countries, the share of middle-skill employment declined by a little
less than 11 percentage points. This compares to increasing shares of high-skill employment
(9 percentage points) and low-skill employment (3 percentage points).
The fall in the share of middle-skill employment across countries accompanied an increase in high- and
low-skill employment shares. Two decades earlier, middle-skill employment comprised slightly more
than 42% of employment in OECD countries compared to about 35% and 24% for high - and low-skill
employment, respectively. In twenty years the share of middle-skill employment fell to be closer to
low-skill employment than high-skill, with the shares averaging 32%, 43% and 27% for middle-, high- and
low-skill employment respectively.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
225
Figure 4.1. Employment shares of middle-skill occupations declined sharply
Employment in middle-skill occupations as a share of total employment, average of years 1994-1996 and 2016-2018
1994-1996 2016-2018
60%
50%
40%
30%
20%
10%
0%
NOR NLD SWE DNK USA LVA FIN GBR CHE EST DEU POL OECD IRL FRA BEL SVK ESP CZE HUN LUX AUT PRT ITA SVN GRC
Notes: For countries with no data in 1994, "1994-1996" is the three earliest years of data. The earliest years are: 1995 (Austria), 1996
(Netherlands, Norway, Slovenia), 1997 (Estonia, Finland, Hungary, Sweden), 1998 (Czech Republic, Latvia, Slovak Republic), 2002 (Poland).
OECD is an unweighted average of the countries shown.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/orp1zm
Box 4.1. Defining occupations in terms of “skill levels”
The term middle-skill seems simple, but it typically means different things depending on the data
available, the research question, or the country involved. The original research from Autor, Levy and
Murnane (2003[3]), who documented the declining share of middle-skill jobs, defined “skill” by the
underlying tasks performed in an occupation (“routine” or “non-routine”) for the United States. Since this
original research, further work on job polarisation employs an ever expanding conception of “skill”,
including task, wage, and education while sometimes finding contradictory evidence for job polarisation
(Hofer, Titelbach and Vogtenhuber, 2017[15]; Tåhlin, 2019[16]; Oesch and Piccitto, 2019[17]).
This chapter will use “skill” or “occupation” as short-hand for a wage-based ranking of occupations. For
continuity, and in order to build on previous OECD work, this chapter will use the classification of
occupations employed in OECD (2017[6]). The occupation groups follow the classification defined in
Goos, Manning and Salomons, (2014[18]), who define International Standard Classification of
Occupations (ISCO) occupations by their average wage. The authors use data from the European
Community Household Panel (ECHP, the predecessor of EU-Statistics on Income and Living
Conditions EU-SILC), which has wage information, to define the occupations by their average wage at
the same occupation level available in the EU-LFS. Workers whose occupation had an average wage
in the middle of the occupation-wage distribution would be classified as middle-skilled regardless of
their formal education, training, or labour market experience. This chapter uses the same EU-LFS data,
and therefore employs the same country-invariant classification:
High-skill, or high-occupation. ISCO-88 one-digit occupations 1-3.
Middle-skill, or middle-occupation. ISCO-88 one-digit occupations 4, 7, 8.
Low-skill, low-occupation. ISCO-88 one-digit occupations 5, 9.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
226
The chapter uses a similar occupation-based grouping for the United States. OECD (2017[10]) mapped
the ISCO-88 classification to Standard Occupational Classification (SOC) 2000 codes. This chapter
then merged this mapping to Dorn (2009[19]) to provide harmonised Census codes from the 1990s to
the present. The U.S. classification scheme shares many characteristics with classifications which
group occupations based on tasks or education (Acemoglu and Autor, 2011[20]).
At its simplest, this chapter is concerned with how educational attainment maps into wage-based
rankings of occupational outcomes, and how this relationship has changed. When referring to measures
of education the following terms will be employed:
High-education. This refers to a person who has at least a tertiary degree corresponding to
International Standard Classification of Education (ISCED) level 5 and above.
Middle-education. For persons with an upper-secondary degree or a post-secondary
non-tertiary degree corresponding to ISCED levels 3-4.
Low-education. This refers to all persons without an upper-secondary degree: ISCED level 2
and below.
4.2. What drives the fall in the share of middle-skill jobs?
To answer where middle-skill workers are going, it is important to understand how the share of
middle-skill jobs declined. The share of middle-skill employment can decline for several reasons,
which may evolve over time. A central question addressed in this chapter is whether the share of
middle-skill employment adjusted due to gradual adjustment through attrition, and/or more abrupt
adjustment through transitions.
The two different paths both lead to a diminished middle-skill employment share, but they point to
different policy responses. With attrition, workers enter into different occupation groups early in their
careers. This is primarily driven by new cohorts of workers entering the labour force b y starting in low-
and high-skill occupations at higher rates than previous cohorts. In the case of adjustment via
transitions, workers reallocate to other skill groups (with possibly spells of non -employment) due to
increased separations in middle-skill employment.
4.2.1. With the exception of the global financial crisis, middle-skill separation rates
were stable
The level of middle-skill employment held steady in most OECD countries until the 2008 -09 financial
crisis, however employment growth was more robust in low- and high-skill employment. Table 4.1
shows average rates of hires and separations across European OECD countries for four time periods. 3
The time periods roughly align to the 1990s, 2000s pre-crisis, the crisis and immediate aftermath, and
post-crisis. Before the financial crisis, on average across countries, middle-skill hiring and separation
rates were about equal, while hiring rates clearly exceeded separation rates for low- and high-skill
occupations. This implies that employment in absolute numbers was growing in low - and high-skill
occupations before the crisis, while remaining more or less constant for middle -skill employment. The
higher employment growth rate in low- and in particular high-skill employment led to a decreasing
share of middle-skill employment.
During the financial crisis, the level of middle-skill employment declined, while high and low-skilled
occupations continued to add jobs. For all occupation groups across countries, hiring rates have
gradually declined over the past twenty years. 4 For low- and high-skill occupations, separation rates
remained remarkably consistent, and below hiring rates. However, for middle -skill jobs, the crisis
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
227
resulted in a sharp increase in separation rates. During the recovery from the crisis, separation rates
for all groups returned to their (lower) pre-crisis levels. For the dynamics of employment adjustment
across occupation groups, the crisis accelerated the declining share of middle-skill employment by
actually destroying jobs, rather than employment simply growing more slowly than other occupation
groups as happened in the 15 years preceding the crisis.
Table 4.1. Middle-skill employment held steady until the global financial crisis
Hiring and separation rates by skill grouping in four time periods
Low-skill Middle-skill High-skill
Hires (%) Separations (%) Hires (%) Separations (%) Hires (%) Separations (%)
1995-2000 23.6 21.5 16.4 15.9 13.8 10.8
2001-2007 22.8 20.7 15.8 15.8 12.4 10.6
2008-2012 21.5 20.4 14.4 18.2 11.4 10.4
2016-2018 22.3 21.0 16.4 15.2 13.6 11.0
Notes: Countries included and time periods: Austria, Belgium, the Czech Republic, Estonia, Denmark, Finland, France, Germany, Greece,
Hungary, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and the United Kingdom are included in
all time periods averaging available years. Latvia, Poland and the Slovak Republic are included starting in 2001.
Source: European Labour Force Survey (EU-LFS).
StatLink 2 https://stat.link/fvkiwa
Looking more closely at individual countries reveals the large variation in separation rates between
and within countries. Countries that experienced large decreases in middle -skill employment saw large
increases in middle-skill separation rates during the crisis driving the overall increase in separations
(Figure 4.2). These countries included Greece, Luxembourg and Spain. However, almost all countries
in the sample have returned to, or experienced lower separation r ates than their pre-crisis rates.
Employment adjustment across skill groupings is a tale of two time periods. Before 2007, labour markets
experienced a declining share of middle-skill jobs caused by higher rates of growth in low- and high-skill
occupations. The quantity of middle-skill employment remained mostly constant. During the crisis,
low- and high-skill employment continued to grow, but at slower rates than pre-crisis. The quantity and
share of middle-skill employment declined sharply because of a hike in separation rates. During the
recovery, separation and hiring rates largely returned to their pre-crisis averages. In short, with the
exception of the global financial crisis, separations have declined gradually, and middle -skill employment
has held steady or declined slightly over the past twenty years.
With the exception of the financial crisis, labour market destinations for workers
separating from middle-skill jobs also remained stable
Although middle-skill separations have remained mostly stable (except for the financial crisis period),
one question that naturally arises is whether the destinations for workers separating from middle -skill
jobs changed in any meaningful way. This has two important implications. First, a changing composition
of the destinations (occupation groups for new jobs, or non-employment) of workers separating from
middle-skill jobs is a signal that the change in middle-skill employment shares is happening through
transitions (rather than attrition) of employed middle-skill workers. Conversely, a mostly stable
distribution of labour market outcomes for workers separating from middle-skill jobs combined with
mostly stable separation rates (as above) points to a diminished role of the transition channel.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
228
Figure 4.2. Changes in separation rates were large during the global financial crisis, but have
mostly returned to pre-crisis levels
Hiring and separation rates for middle-skill occupations, 2001-2018
2001-2007 2008-2012 2016-2018
A. Hiring rates
30%
25%
20%
15%
10%
5%
0%
GRC ESP LVA LUX FRA SVK GBR PRT IRE POL DEU DNK SWE SVN HUN NLD EST CHE BEL FIN ITA CZE AUT NOR
B. Separation rates
30%
25%
20%
15%
10%
5%
0%
GRC ESP LVA LUX FRA SVK GBR PRT IRE POL DEU DNK SWE SVN HUN NLD EST CHE BEL FIN ITA CZE AUT NOR
Notes: Countries are sorted by their change in employment growth for middle-skill occupations from 2001-2007 to 2008-2012.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany.
StatLink 2 https://stat.link/7f51rq
Second, the destinations of workers previously in middle-skill jobs provide evidence for the normative
implications of the decline in middle-skill employment. If these workers are increasingly moving into
high-skill jobs, policy makers may worry less about the implications of shrinking middle-skill employment.
By contrast, increasing transitions into low-skill employment or non-employment may signal to policy
makers greater distress among workers who previously held middle-skill employment.
Across European OECD countries, the destinations for workers separating from middle -skill jobs have
remained stable on average. Figure 4.3 shows the four mutually exclusive and collectively exhaustive
destinations for workers who separated from middle-skill jobs one year earlier: low-, middle-, high-skilled
employment and non-employment. The propensities are calculated one year after separating, which
allows for short non-employment durations. The analysis compares the average of years 2005-2007,
2008-2012, and 2015-2017. Although separations grew during the financial crisis, this chapter is most
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
229
concerned with the long-term decline in the share of middle-skill employment, and omitting the crisis
period attenuates movements due to cyclical variation.
Non-employment and transitions to other middle-skill jobs are the most likely destinations, and have
remained so post-crisis. The countries are ordered by the percentage point change in the share of
employment in middle-skill jobs one year later pre- and post-crisis – i.e. the difference between dark
blue diamonds and bars in Panel B. Non-employment was the most likely destination pre-crisis with
51.4% of workers separating from middle-skill jobs to non-employment one year later followed by a
different middle-skill job with 35.1%. Post-crisis the likelihood of ending up in non-employment fell to
47.6%, while the probability of working in another middle-skill job increased slightly to 35.6%.
Figure 4.3. Outcomes for workers separating from middle-skill jobs remain stable
Skill-group of employment and non-employment outcomes for workers who separated from a middle-skill job one
year prior
2005-2007 2008-2012 2015-2017
A. Non-employment B. Middle-skill
90% 90%
80% 80%
70% 70%
60% 60%
50% 50%
40% 40%
30% 30%
20% 20%
10% 10%
0% 0%
C. Low-skill D. High-skill
25% 25%
20% 20%
15% 15%
10% 10%
5% 5%
0% 0%
Notes: Each data point depicts the frequency of each skill-group or non-employment for middle-skill workers who held a middle-skill job and
separated from their employer or firm one year prior. Countries are ordered by the percentage point change in the share of employment in
middle-skill jobs one year later between 2005-2007 and 2015-2017 – i.e. the difference between dark blue diamonds and bars in Panel B. Data
are yearly averages over the indicated years. OECD is an unweighted average of the countries shown.
Source: EU-Statistics on Income and Living Conditions (EU-SILC), and the German Socio-Economic Panel (SOEP) for Germany.
StatLink 2 https://stat.link/disrl2
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
230
Once again, the largest changes in the patterns of workers separating from middle-skill jobs occurred
during the crisis. The share of workers separating from middle-skill jobs into non-employment grew to
59.4% during 2008-2012. The shares separating into the other occupation groups all declined. In
particular, the share separating into another middle-skill job one year later declined by over 7 percentage
points to 27.8%.
The propensity to find low- or high-skill jobs within one year after separation increased uniformly, but
modestly from pre- to post-crisis periods. On average, the low-skill propensity increased from 7.1% to 8.4%
and the high-skill post-separation employment propensity increased from 7.4% to 8.2%. All but six
countries in the sample saw an increase in the low-skill propensity with Sweden, Poland and the
United Kingdom experiencing the largest percentage point increases. Transitions into high-skill
employment were similarly broad-based. Austria, Sweden and the United Kingdom saw some of the largest
percentage point increases. Of course, the composition of the labour force likely changed over this time
period, a dimension which will be explored further in Section 4.4.
Younger workers were disproportionately affected by the financial crisis and concurrent
spike in separation rates for middle-skill jobs
In addition to how employment is adjusting across skill groupings, the question of who is involved in the
adjustment is just as important. The adjustment out of middle-skill work was mostly born by the young.
Table 4.2 presents the same hiring and separation rates as Table 4.1 but limited to workers in middle-skill
occupations and further divided into workers who are less than 30 years old and those that are 30 years
old and older. The rate of hires and separations is much higher for workers younger than 30 than those
that are older. This is expected as younger workers have more volatile employment histories due to
increased job hopping, higher rates of temporary contracts, and generally trying to find their way in the
labour market.
Table 4.2. Middle-skill employment adjustment led by younger workers
Hiring and separation rates for middle-skill occupations by age in four time periods
Younger workers Older workers
Hires (%) Separations (%) Hires (%) Separations (%)
1995-2000 31.5 32.3 10.2 9.7
2001-2007 31.5 32.2 10.4 10.5
2008-2012 29.1 36.2 9.8 12.6
2016-2018 34.4 33.7 11.3 10.2
Notes: Younger workers are those aged 16-29, older workers are aged 30-64. Countries included and time periods: Austria, Belgium, the
Czech Republic, Estonia, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Luxembourg, the Netherlands, Norway,
Portugal, Spain, Sweden, Switzerland and the United Kingdom are included in all time periods averaging available years. Latvia, Poland and
the Slovak Republic are included starting in 2001.
Source: European Labour Force Survey (EU-LFS).
StatLink 2 https://stat.link/skj3bc
The crisis led to decreased middle-skill employment in all age groups including prime age and older
workers. The impact was much stronger for younger workers. During the crisis hires declined and
separations increased for both older and younger workers in middle-skill occupations. The magnitudes
differed greatly with younger workers experiencing a net employment loss of 7 percentage points annually
compared to 3 percentage points for older workers. In sum, both older and younger age group workers
experienced serious employment declines in middle-skilled occupations during the crisis, but younger
workers incurred a disproportionate burden.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
231
The results in this section compare well with the previous literature, which finds that polarisation is a result
of differential hiring rates, as well as layoffs. Earlier research from the United States found that the decline
in middle-skill employment was concentrated during the past two recessions (Jaimovich and Siu, 2014[21]).
The work in this chapter suggests that the global financial crisis acted as an accelerant for job polarisation
on average across European OECD countries as well. In these countries, adjustment took place gradually
until the crisis. Separation rates then spiked for workers in middle-skill jobs during the crisis before mostly
returning to pre-crisis levels. The results further generalize the finding – again, previously found in the
United States – that job polarisation is the result of higher flows into low- and high-skill occupations, and
lower flows into middle-skill occupations, with the pattern most pronounced among the young (Smith,
2013[22]).
4.2.2. Differential labour market entry of younger workers explains most of the change in
middle-skill employment shares
The remainder of this section will examine workers by birth cohort to see how their employment shares
across different occupation groups change over their working-ages. By examining workers by birth cohort,
the analysis decomposes employment shares in different occupational groups by labour market entry and
labour market history. Figure 4.4 shows employment shares in low-, middle-, and high-skilled occupations
as well as non-employment by birth cohorts.5
The analysis uses eight different European birth cohorts of six contiguous years with the number and size
of the cohorts determined by the 24 years of available data.6 Each panel of the figure displays the share
for a given skill group and non-employment. For ease of interpretation, the figure averages the birth
cohorts’ shares into cohorts born before and after 1970. 7 Each line represents one of the two average birth
cohorts. The figures present the shares such that they allow comparison for cohorts at the same age.
The entry shares for the two average cohorts shows the greatest divergence. Workers aged 25-29 worked
in middle-skill occupations at a rate of 32.8% before 1970. For cohorts born after 1970, the share
decreased to 26.5%. However, the employment trajectories of cohorts as they aged diverged only slightly.
Birth cohorts before 1970 saw a slight decreasing propensity to work in middle-skill employment over the
life cycle. For the post 1970 birth cohorts, the trajectory was mostly flat.
The shares in high-skill employment exhibited the reverse pattern. Workers aged 25-29 before 1970
entered high-skill employment at a rate of 22.9%. For birth cohorts after 1970 the rate was 30.7%. Both
average birth cohorts saw a slight upward trajectory for the shares employed in high-skill jobs over the life
cycle, but birth cohorts born after 1970 were and are employed in high-skill jobs at a higher rate at every
age group.
Employment shares in low-skill jobs and non-employment did not diverge appreciably across average birth
cohorts. Workers in the pre-1970 cohort entered low-skill employment at a rate of 16.3% as 25-29 year
olds, while workers born post-1970 entered low-skill employment at a rate of 18.1%. Entry shares for
non-employment also diverged only slightly. The pre-1970 birth cohorts were slightly more likely to be
non-employed at age 25-29. For the pre-1970 birth cohorts the rate was 28%, while for post-1970 birth
cohorts averaged 24.8%. After the youth entry share, the trajectories and propensities for the pre- and
post-1970 average birth cohorts were remarkably similar. However, the post-1970 birth cohorts were
always more likely to be both employed, and working in low-skill employment, although only slightly.
Although each successive cohort entered skill groups at different rates, their career trajectories followed
markedly similar paths. Put differently, when examining Figure 4.4, the lines showing different employment
shares at different ages for each cohort are largely parallel. After entering the labour market in their youth,
each cohort has progressed in the labour market similarly.
The exception is the life-cycle patterns for employment shares in middle-skill and high-skill occupations.
The employment shares do not seem to be decreasing for the post-1970 cohorts as they did for the
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
232
pre-1970 cohorts. Both have become flatter suggesting perhaps that young workers entering middle-skill
occupations are less likely now to upgrade towards high-skill occupations. This seems consistent with
findings (Chapter 5) on vocational graduates, which appear to have a flatter career pattern now with
respect to the past. It also consistent with recent findings on the decline in internal labour markets (Maurin
and Signorelli, 2019[23]).
Figure 4.4. Propensities to work in different skill groups have been driven by differential entry of
younger cohorts
Shares of different occupation-skill groups at different ages by birth cohort
Pre-1970 Post-1970
A. Middle-skill B. Low-skill
50% 50%
40% 40%
30% 30%
20% 20%
10% 10%
0% 0%
25-29 30-34 35-39 40-44 45-49 50-54 25-29 30-34 35-39 40-44 45-49 50-54
C. High-skill D. Non-employment
50% 50%
40% 40%
30% 30%
20% 20%
10% 10%
0% 0%
25-29 30-34 35-39 40-44 45-49 50-54 25-29 30-34 35-39 40-44 45-49 50-54
Note: Pooled data represents unweighted country averages. Shares are computed as the ratio of the number of workers in a given skill group
(or non-employment) to the population of the relevant age group. Data encompass years 1994-2018. Pre-1970 encompasses birth years
1946-1970. Post 1970 consists of birth years 1970-1993. Countries included: Austria, Belgium, the Czech Republic, Estonia, Denmark, Finland,
France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Luxembourg, the Netherlands, Norway, Poland, Portugal, Spain, the Slovak Republic,
Sweden, Switzerland and the United Kingdom.
Source: European Labour Force Survey (EU-LFS).
StatLink 2 https://stat.link/juclpr
The analysis in this section shows that, with the exception of the years after the Global Financial crisis, the
rate of separations, and the destinations of workers separating from middle-skill jobs have stayed constant.
Younger workers are more likely to enter low- and high-skill employment and are less likely to take
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
233
middle-skill jobs highlighting the greater importance of attrition for the decline in the share of middle-skill
employment. Annex 4.B provides a formal decomposition of the forces of attrition, transition and cohort
size on the change in middle-skill employment shares, which confirms that attrition is the dominant channel
accounting for the decline in middle-skill occupations.
4.3. Who were middle-skill workers?
Looking at only the transitions of workers currently in middle-skill jobs misses the labour market decisions
of new cohorts entering the labour market. As the previous section showed, this is likely the most salient
mechanism for the shifting shares of skill groups. As the share of middle-skill occupations declines, the
probability of any given worker finding a middle-skill job also declines and the probability she finds a job in
either high- or low-skill occupations increases. To account for the full picture of “where middle-skill workers
are going”, one must therefore also consider what those workers who in the past would have been a
middle-skill worker are doing today. In short, one needs to account for the possibly shifting demographics8
of middle-skill workers.
To build a clearer picture of a “middle-skill worker”, it is essential to define the profile of the “typical”
middle-skill worker. Using variables which best identify a middle-skill worker in the period when job
polarisation began, this chapter sketches the profile of a middle-skill worker from twenty years prior.
Section 4.4 will examine workers with this profile to compare how their labour market outcomes compare
to workers with the same profile twenty years prior. This approach helps remedy the problem of
non-observability of counterfactual outcomes for cohorts who enter the labour market in times of more or
fewer middle-skill jobs, and who may therefore look vastly different to someone holding that job in the past.
The rest of this section is concerned with identifying the characteristics that best describe a middle-skill
worker from times past. The following analysis paints a picture of a typical middle-skill worker which
remains relatively constant over time and does not change with shifts in the labour market.
4.3.1. Middle-skill workers are predominantly workers without a tertiary degree
To determine where middle-skill workers are going, it is necessary to find a set of characteristics that best
predict work in middle-skill occupations. These characteristics should ideally be independent of labour
market conditions and outcomes.9
Education was the strongest indicator of middle-skill work
Education was the single best predictor of being a middle-skill worker twenty years prior. Figure 4.5 shows
the share of middle-skill workers divided into four categories: men and women separately who had at least
an upper-secondary degree but no tertiary degree, and men and women with less than an upper-secondary
degree. The shares are averages for each country of the years 1994-1996 (years vary depending on
availability, see figure notes). Across the OECD countries for which data are available, slightly more than
90% of middle-skill workers lacked a tertiary degree with the share increasing to over 95% in ten of the
countries. The Czech Republic and Austria had the highest share while Estonia and Belgium had the
lowest shares with 83% and 86.8%, respectively.
Among those without a tertiary degree, a majority of middle-skill workers did have at least an
upper-secondary non-tertiary degree. On average in the OECD, 56.8% of middle-skill workers possessed
an upper-secondary education. The highest shares were in the Slovak Republic, Poland and the
Czech Republic, while the lowest shares of middle-skill workers holding at least an upper-secondary
diploma were found in Portugal, Spain and Greece.10 Over a third of middle-skill workers possessed less
than an upper-secondary diploma implying that middle-skill jobs were accessible to even those with little
education.11
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
234
Figure 4.5. Middle-skill workers were workers without a tertiary degree
Share of middle-skill workers by gender and education, 1994-1996 (average)
Female, low-education Female, middle-education Male, low-education Male, middle-education
CZE
AUT
SVK
PRT
SVN
ITA
HUN
POL
GRC
LVA
FRA
NLD
DNK
SWE
LUX
NOR
USA
GBR
FIN
DEU
ESP
IRL
BEL
EST
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Notes: Mid-2000s refer to 2004, 2005 and 2006. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland).
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/28oxqj
Middle-skill workers were more likely to be male
Middle-skill employment was also dominated by men. Among middle-skill workers without a tertiary degree,
a little less than two thirds were men on average across the OECD. Among countries for which data are
available, the highest shares were found in Spain and Luxembourg, which had male shares of middle-skill
workers without a tertiary degree of over 70%. The lowest shares were found in the United States and
Slovenia. Despite having the lowest shares of men, men still exceeded 50% of middle-skill workers in each
of these countries.
Further enforcing the gender disparities in middle-skill work, the share of men in middle-skill work without
an upper-secondary degree exceeded the share of women with one. Men without an upper-secondary
degree made up 26% of middle-skill workers on average. Women who held at least an upper-secondary
degree, but less than a tertiary degree, comprised only 18.6% of middle-skill workers.
4.3.2. Both male and female middle-skill workers were most likely to work in
manufacturing
Although substantial gender differences existed among middle-skill workers, both men and women were
most likely to work in manufacturing. Figure 4.6 shows the share of middle-skill workers for men and
women, respectively, in the three industries they were most likely to work in. For both men and women in
middle-skill work, manufacturing represented the modal industry, employing 37.7% of men, and 35.3% of
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
235
women. Slovenia and Italy employed the highest share of male middle-skill workers in manufacturing. For
women, Slovenia and Portugal employed the highest share of middle-skill women in manufacturing, with
58.3% and 55.7%, respectively.
Figure 4.6. Male and female middle-skill workers were concentrated in manufacturing
Share of middle-skill workers by industry for men (Panel A) and for women (Panel B), 1994-1996 (average)
A. Men
Transport, Storage and Communication Construction Manufacturing
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
EST USA LVA CHE GRC NLD FRA LUX GBR SVK IRL ESP PRT HUN POL BEL DNK ITA NOR SWE CZE AUT DEU FIN SVN
B. Women
Wholesale & Retail Public Administration Manufacturing
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
LUX NLD FIN CHE NOR USA GBR DEU SWE IRL DNK FRA AUT BEL EST LVA ESP POL GRC CZE HUN SVK ITA SVN PRT
Notes: Industry share of middle-skill workers in mid-1990s, 1994-1996. For countries with no data in 1994, "mid-1990s" is the three earliest
years of data. The earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary,
Sweden), 1998 (Czech Republic, Latvia, Slovak Republic), 2002 (Poland).
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/0nym5b
The gender differences for middle-skill workers are most apparent in industries other than manufacturing.
For women, the next most probable industry was wholesale and retail trade which averaged 11.7% of
women in middle-skill employment followed by public administration with 9.4%. The United States and the
Netherlands employed the largest share of middle-skill women in wholesale and retail trade with 21.6%
and 19.1%, respectively. For public administration, Belgium with 17.8%, and Greece with 16.9%, employed
the highest shares.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
236
Middle-skilled men were more likely to be employed in construction, as well as in transportation and
storage. After manufacturing, construction employed the largest share of middle-skill male workers across
the OECD followed by transportation and storage with 16.8% and 13.3%, respectively. The highest share
of workers in construction were found in Luxembourg and Austria, with 24.3% and 21% respectively. For
transportation and storage, the highest shares of middle-skilled male workers were found in Finland and
Latvia with a little over 16% in both countries.
Occupations differed greatly for middle-skill workers by gender
For men, the industry distribution of middle-skill workers was reflected in their most common occupations.
Table 4.3 shows the three most likely detailed occupations across OECD countries for middle-skill workers
from two decades prior by gender.12 The most likely occupations were drivers, building finishers, and
machinery mechanics and repairers. The three most likely occupations are indicative of the three most
likely industries: transportation & storage, construction, and manufacturing.
Table 4.3. Middle-skill occupations varied greatly by gender
Most common middle-skill occupations by gender, 1994-1996 (average)
Women Men
Secretaries Truck, Delivery Drivers
Cashiers Machine Operators
Bookkeepers and Accounting Clerks Building Finishers (floors, roofing, insulation)
Notes: First row is the most prevalent occupation for each gender, the last row is the third most prevalent. For countries with no data in 1994,
"mid-1990s" is the three earliest years of data. The earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997
(Estonia, Finland, Hungary, Sweden), 1998 (Czech Republic, Latvia, Slovak Republic), 2002 (Poland).
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
The occupations most likely to be held by middle-skill women did not follow as clearly from the most likely
industries to employ middle-skill women. Middle-skill women were most likely to be employed as
secretaries, cashiers and book keepers or auditing clerks. The latter two reflect two of the modal industries
most likely to employ female middle-skill workers: wholesale and retail trade, and public administration.
The most likely occupation, secretaries, are employed across industries and therefore make up the modal
occupation without any explicit tie to the modal industry, manufacturing. 13
4.4. Where are middle-skill workers going?
This section ties together the results of the previous two sections. First, it shows that changes in the
composition of the employed population are not the main cause of the declining share of middle-skill
employment shares, an issue complicating the results of Section 4.2. It also establishes that groups likely
to have been middle-skill workers in the past (Section 4.3) experienced a decreased tendency to work in
middle-skill jobs. The rest of the section shows that groups who were previously likely to work in middle-skill
jobs, especially workers with an upper-secondary degree but no tertiary degree, are now more likely to
work in low-skill employment.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
237
4.4.1. The decline in middle-skill employment is not primarily due to changing
demographics
The working-age population is more highly educated today than twenty years prior. This shift alone may
account for the decline in middle-skill employment. This would complicate the results in Section 4.2
because the types of workers holding middle-skill jobs today have a different demographic profile
compared to middle-skill workers two decades prior. To answer the question of whether or not shifts in the
demographic composition of the working-age population are causing the shares of middle-skill employment
to shrink, the analysis turns to a shift-share analysis.14 The shift-share analysis decomposes the change
in the share of middle-skill employment into shifts induced by changes in the composition of the workforce
and changes in propensity to be employed in middle-skill employment within groups. In other words, it
shows what the share of middle-skill employment would have been if the skill composition of the work-force
had not changed in each country over 20 years (composition effect), as well as if the propensity to work in
middle-skill jobs among individuals of each skill groups had not changed over the same period of time
(propensity).15
The analysis in this section is most similar to Cortes, Jaimovich and Siu (2017[24]) who perform a similar
analysis for the United States. Researchers have performed similar analyses for Germany (Bachmann,
Cim and Green, 2018[25]), Finland (Maczulskij and Kauhanen, 2017[26]), and the United Kingdom (Salvatori,
2015[27]).
A simple example helps to explain the shift-share and argue for its importance. From Section 4.3 it is
apparent that workers with a tertiary degree are less likely to work in middle-skill jobs compared to workers
with less education. If the share of the labour force with a tertiary degree increases over time, the share of
employment in middle-skill work will likely decline. In this case, middle-skill workers are not “moving.”
Workers across the education distribution are possibly employed in different skill groups at the same rates
as before, but the shift in composition makes it look like middle-skill employment has declined.
A key part of the shift-share is the division of workers into distinct groups. The analysis divides workers
into mutually exclusive and collectively exhaustive groups based on education and gender. The analysis
in the previous section found that they are the best predictors of middle-skill employment. The association
between these predictors and middle-skill employment is the deciding factor for their inclusion. They are
similar to the demographic characteristics used in Cortes, Jaimovich and Siu (2017[24]) who follow the same
methodology for the United States.16
The results of the shift-share show that both changes in composition and changes in the propensity to work
in middle-skill jobs contributed to the decline in the share of middle-skill employment. Across OECD
countries in the sample, the share of middle-skill employment declined by 4.2 percentage points.17 That
decline can be decomposed into the part due to composition changes, 2.1 percentage points, and
decreased propensity to work in middle-skill jobs within groups, 2.3 percentage points. The decrease in
middle-skill employment due to compositional changes is not surprising given the increased share of the
population with a tertiary degree.18
Decreases in the propensity to work in middle-skill employment exceed compositional effects in the
majority of countries in the sample (Figure 4.7). Luxembourg, Slovenia, and Norway saw the largest
declines in middle-skill employment due to decreased propensity in absolute numbers. The countries with
the largest declines in propensity as a share of the total decline in middle-skill shares were Estonia, Spain,
Ireland and Belgium.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
238
Figure 4.7. Shrinking share of middle-skill employment due (slightly more) to diminished
propensity to work more than changing composition
Percentage point change in share of middle-skill employment due to composition and propensity, mid-1990s to
mid-2010s
Composition Propensity Total
0.06
0.04
0.02
0
-0.02
-0.04
-0.06
-0.08
-0.1
-0.12
-0.14
LUX SVN NOR GRC CHE DNK AUT GBR USA SWE FRA BEL FIN OECD PRT IRL ITA NLD ESP EST LVA SVK CZE DEU POL HUN
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s refer to 2016-2018. The share of middle-skill employment includes the
entire working-age population and includes non-employment as a possible category. OECD is an unweighted average of the countries shown.
Source: European labour force survey (EU-LFS) and the Current Population Survey (CPS) for the United States.
StatLink 2 https://stat.link/hp16qt
4.4.2. Workers without a tertiary degree are more likely to be employed in low-skill
occupations
The preceding shift-share analysis confirmed that propensity to work in middle-skill jobs contributed to the
decline in the share of middle-skill employment. Changes in the composition of the labour force – greater
educational achievement and women’s increased participation – are a contributing factor. The analysis did
not show where workers are increasingly likely to work. To see where they are working, this analysis digs
deeper into the changing propensities of where workers likely to be middle-skill are employed.
Middle-educated men have seen only a modest drop in middle-skill employment shares
Across OECD countries for which data are available, workers without a tertiary degree have become less
likely to work in middle-skill occupations. This is not entirely surprising given the decline in middle-skill
employment overall. However, it was not a given that all groups would experience a drop in their propensity
to work in middle-skill occupations. Middle-educated men have been the least affected with the share of
the working-age population in middle-skill occupations dropping a little over 2 percentage points
(Figure 4.8). Low-educated men saw their share decrease by 7 percentage points. Women with low and
middle-education experienced a decline of a little over 4 percentage points each.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
239
Figure 4.8. The propensity to work in middle-skill occupations fell for all workers without a tertiary
degree
Percentage point change in the shares in middle-skill jobs by gender-education groups, mid-1990s to mid-2010s
Low-education Middle-education
A. Women
0.2
0.1
0
-0.1
-0.2
-0.3
AUT BEL CHE CZE DEU DNK EST ESP FIN FRA GRC HUN IRL ITA LUX LVA NLD NOR POL PRT SWE SVN SVK GBR USA
B. Men
0.2
0.1
0
-0.1
-0.2
-0.3
AUT BEL CHE CZE DEU DNK EST ESP FIN FRA GRC HUN IRL ITA LUX LVA NLD NOR POL PRT SWE SVN SVK GBR USA
Notes: For each group, the bars report the change in the ratio of the number of workers of that group holding middle-skill jobs to the working-age
population of that group. Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of
data. The earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden),
1998 (Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s refer to 2016-2018.
Source: European labour force survey (EU-LFS), and the Current Population Survey (CPS) for the United States.
StatLink 2 https://stat.link/lnpk16
Middle-educated workers are much more likely to be in low-skill employment
As the share of middle-occupation employment has declined, workers without a tertiary degree are
increasingly likely to work in low-skill occupations. Figure 4.9 shows the percentage point increase in the
propensity to work in low-skill occupations for women without a tertiary degree. The figure further divides
employed women into those without an upper-secondary degree, and those with at least an
upper-secondary degree, but no tertiary degree. The shift is most pronounced for middle-educated women.
The propensity of employed middle-educated women to work in low-skill occupations increased from
18.8% to 27% twenty years later. For low-educated women the increase was more muted. Low-educated
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
240
women saw their propensity to work in low-skill occupations grow from 17.8% in the mid-1990s to 21.2%
in the mid-2010s.19
Figure 4.9. Women without a tertiary degree more likely to work in low-skill occupations
Women’s shares in low-skill occupations by education, mid-1990s and mid-2010s
Mid-1990s Mid-2010s
A. Low-education
50%
40%
30%
20%
10%
0%
POL GRC SVN LVA HUN SVK EST IRL ITA BEL ESP USA CZE DEU FIN NLD FRA AUT LUX PRT SWE GBR CHE DNK NOR
B. Middle-education
50%
40%
30%
20%
10%
0%
POL GRC SVN LVA HUN SVK EST IRL ITA BEL ESP USA CZE DEU FIN NLD FRA AUT LUX PRT SWE GBR CHE DNK NOR
Notes: For each group, the bars report the ratio of the number of women of that group holding low-skill jobs to the working-age population of
that group. Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The
earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s refer to 2016-2018.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/br0if4
Middle-educated women in Finland, Portugal and Spain experienced the largest increase in the propensity
to work in low-skill occupations. The percentage point increase in all three countries exceeded 17%.
Low-educated women experienced the largest increase in the share working in low-skill occupations in
Spain, Portugal and Estonia.
The propensity for employed men without a tertiary degree to work in low-skill occupations also increased
across OECD countries. Figure 4.10 shows the change in the share in low-skill occupations for male,
middle- and low-educated workers. Though still large, men saw a more muted shift towards low-skill
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
241
occupations compared to women. Low-educated men’s propensity to work in low-skill occupations
increased from 12.1% to 14% over the previous twenty years across OECD countries. For middle-educated
men, the increase was larger, rising from 11.2% to 15.1%.
Figure 4.10. Men without a tertiary degree are also increasingly in low-skill occupations
Men’s shares in low-skill occupations by education, mid-1990s and mid-2010s
Mid-1990s Mid-2010s
A. Low-education
30%
20%
10%
0%
POL SVN SWE SVK FIN FRA EST HUN BEL CZE GRC DEU CHE LVA NOR NLD AUT LUX PRT GRB IRL ESP ITA USA DNK
B. Middle-education
30%
20%
10%
0%
POL SVN SWE SVK FIN FRA EST HUN BEL CZE GRC DEU CHE LVA NOR NLD AUT LUX PRT GBR IRL ESP ITA USA DNK
Notes: For each group, the bars report the ratio of the number of men of that group holding low-skill jobs to the working-age population of that
group. Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s refer to 2016-2018.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/9p1fwn
When examining individual countries, it is clear middle-educated men experienced a larger increase in the
share of low-skill employment compared to low-educated men. Low-educated men in the Netherlands,
Sweden and Hungary experienced the largest increase in the propensity to work in low-skill occupations.
Their propensities increased by 8.6, 7.8 and 6.3 percentage points, respectively. For middle-educated
men, the greatest increase occurred in Portugal, the United Kingdom and Spain, where increases topped
8 percentage points in each country.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
242
Middle-educated workers of both genders are less likely to work in high-skill occupations,
with some notable exceptions
There is not a corresponding increase in the propensity to work in high-skill occupations. Across OECD
countries, the propensity to work in high-skill occupations declined by 0.5 percentage points for
middle-educated men, and 0.4 percentage points for middle-educated women (Figure 4.11). For
low-educated men, the change was a decline of 0.7 percentage. Low-educated women in OECD countries
saw no change in the propensity to work in high-skill jobs.20
Figure 4.11. Middle-educated men and women are only slightly more likely to work in high-skill
occupations
Shares of middle-educated workers in high-skill occupations by gender, mid-1990s and mid-2010s
Mid-1990s Mid-2010s
A. Men
40%
30%
20%
10%
0%
POL HUN LVA SVK ESP EST DEU SVN IRL GRC FIN CZE BEL FRA DNK AUT SWE NOR GBR ITA PRT LUX USA CHE NLD
B. Women
40%
30%
20%
10%
0%
POL HUN LVA SVK ESP EST DEU SVN IRL GRC FIN CZE BEL FRA DNK AUT SWE NOR GBR ITA PRT LUX USA CHE NLD
Notes: For each group, the bars report the ratio of the number of workers of that group holding high-skill jobs to the working-age population of
that group. Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The
earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s refer to 2016-2018.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/zgo1n4
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
243
Although propensities to work in high-skill occupations declined, there were some countries that saw
substantial increases in the share of workers in high-skill occupations. In Denmark, Germany, Norway and
Sweden the share of middle-educated women in high-skill occupations increased by more than
5 percentage points. Denmark, Germany, Norway and Sweden (together with Estonia) also had the largest
increases in the share of middle-educated men moving into high-skill occupations over the preceding
twenty years. Interestingly, many of these countries have a vocational education and training system (VET)
at the intermediate education level that seem to work particularly well (see Chapter 5).
Changing entry propensities for middle- and low-skill employment are almost entirely due to
workers without a tertiary education
To bring the analysis to a close, the chapter returns to the cohort analysis introduced at the end of
Section 4.2. Figure 4.12 shows the life-cycle labour market employment shares for pre- and post-1970
birth cohorts again, but this time further broken out by educational attainment. Breaking out the life-cycle
employment shares by education further targets workers most likely to have been middle-skilled in the past
(Section 4.3).
Figure 4.12. Propensities to work in different skill groups mostly driven by differential entry of
younger cohorts without a tertiary degree
Shares in different occupation-skill groups at different ages by birth cohort and education
Pre 1970, no tertiary degree Post 1970, no tertiary degree Pre 1970, tertiary degree Post 1970, tertiary degree
A. Middle-skill B. Low-skill
50% 50%
40% 40%
30% 30%
20% 20%
10% 10%
0% 0%
25-29 30-34 35-39 40-44 45-49 50-54 25-29 30-34 35-39 40-44 45-49 50-54
C. High-skill D. Non-employment
100% 50%
90%
80% 40%
70%
60% 30%
50%
40% 20%
30%
20% 10%
10%
0% 0%
25-29 30-34 35-39 40-44 45-49 50-54 25-29 30-34 35-39 40-44 45-49 50-54
Notes: Pooled data represents unweighted country averages. Shares are computed as the ratio of the number of workers in a given skill group
(or non-employment) to the population of the relevant age-education group. Data encompass years 1994-2018. Pre 1970 encompasses birth
years 1946-1970. Post 1970 consists of birth years 1970-1993.
Source: European labour force survey (EU-LFS).
StatLink 2 https://stat.link/6gw358
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
244
Changing entry propensities for middle- and low-skill employment are almost entirely due to workers
without a tertiary education. This finding reinforces the main results from the beginning of this section.
There is little change in either entry propensities or life-cycle trajectories for high-skill employment and
non-employment by educational attainment. Similarly, workers with a tertiary education show no
discernible differences pre- and post-1970 with regard to employment in middle- and low-skill employment.
The major change is that workers without a tertiary degree are less likely to work in middle-skill employment
and more likely to work in low-skill employment. The trajectories to the eye are quite similar with entry
patterns showing the greatest difference between pre- and post-1970 birth cohorts without a tertiary
education.
4.5. Concluding remarks
The share of middle-skill jobs in OECD labour markets has declined over the past three decades.
Middle-skill jobs once made up a large share of overall employment, but automation and offshoring have
reduced the share of middle-skill employment relative to low-skill and high-skill occupations, a trend that
has been termed job polarisation. What is happening to workers who could have previously expected to
be employed in middle-skill occupations is an enduring question for governments in OECD countries.
In the past, employment in middle-skill occupations provided many workers with a good standard of living.
The decline in employment opportunities in these occupations has meant that workers who previously
would have held these jobs are increasingly employed in low-skill occupations.
The shift in the share of middle-skill jobs towards low- and high-skill employment has mostly taken place
through attrition, at least in European countries. That is to say through successive cohorts of younger
workers being less likely to enter the labour force in middle-skill jobs, and more likely to start in low-skilled,
and to a lesser extent, high-skilled jobs. Their subsequent job trajectories over the life cycle are a
contributing but secondary factor. Workers without a tertiary degree are sliding down the job ladder.
Compared with twenty years ago, workers without a tertiary degree are less likely to work in middle-skill
occupations. This has been matched almost exactly by an increase in low-skill employment for this group.
The analysis in this chapter, however, relies on cross-sectional and panel survey data from across Europe
and the United States. Further analysis would be required to fully generalise these findings to the whole
OECD.
Although the answer to the question of what is happening to middle-skill workers appears bleaker than
expected, some countries are performing well in mitigating the adverse effects of job polarisation. Over the
past two decades in Sweden, Germany, Norway and Denmark, the rise in the employment shares of
middle-educated workers in high-skill occupations was almost as sizeable as the rise in their employment
in low-skill occupations. The relative success of these countries shows that good jobs for formerly
middle-skill workers is not necessarily due to insurmountable structural forces. Automation and
globalisation have reduced the number of middle-skill employment opportunities for workers without a
tertiary degree. However, countries that have improved employment opportunities for middle-skill workers
share a set of common policies. They have strong institutions and practices around social dialogue, as
well as an emphasis on vocational education and training (Chapter 5). The explicit application of these
policies to the dynamics studied in this chapter will be left for future research.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
245
References
Acemoglu, D. and D. Autor (2011), “Skills, Tasks and Technologies: Implications for Employment [20]
and Earnings”, in Ashenfelter, O. and D. Card (eds.), Handbook of Labor Economics,
http://dx.doi.org/10.1016/S0169-7218(11)02410-5.
Acemoglu, D. and P. Restrepo (2018), “Artificial Intelligence, Automation and Work”, NBER [13]
Working Paper Series, No. 24196, NBER, Cambridge, MA.
Autor, D. (2015), Polanyi’s Paradox and the Shape of Employment Growth. [11]
Autor, D. (2015), “Why are there still so many jobs? The history and future of Workplace [12]
automation”, Journal of Economic Perspectives, Vol. 29/3, pp. 3-30.
Autor, D. (2010), The Polarization of Job Opportunities in the U.S. Labor Market Implications for [8]
Employment and Earnings, The Hamilton Project, Washington, D.C.,
https://www.hamiltonproject.org/papers/the_polarization_of_job_opportunities_in_the_u.s._la
bor_market_implica (accessed on 13 January 2020).
Autor, D. and D. Dorn (2013), “The Growth of Low-Skill Service Jobs and the Polarization of the [7]
US Labor Market”, American Economic Review, Vol. 103/5, pp. 1553-1597.
Autor, D. and D. Dorn (2009), “This Job is “Getting Old”: Measuring Changes in Job [30]
Opportunities using Occupational Age Structure”, American Economic Review Papers and
Proceedings, Vol. 99/2, pp. 45-51.
Autor, D., F. Levy and R. Murnane (2003), “The Skill Content of Recent Technological Change: [3]
An Empirical Exploration”, The Quarterly Journal of Economics, Vol. 118/4, pp. 1279-1334,
https://economics.mit.edu/files/11574 (accessed on 8 December 2017).
Bachmann, R., M. Cim and C. Green (2018), “Long-run Patterns of Labour Market Polarisation: [25]
Evidence from German Micro Data”, Ruhr Economic Papers, No. 748, RWI.
Cazes, S. and M. Tonin (2010), “Employment protection legislation and job stability: A European [32]
cross-country analysis”, International Labour Review, Vol. 149/3, pp. 261-285,
http://dx.doi.org/10.1111/j.1564-913X.2010.00087.x.
Cortes, G. (2016), “Where Have the Middle-Wage Workers Gone? A Study of Polarization Using [9]
Panel Data”, Journal of Labor Economics, Vol. 34/1, pp. 63-105,
http://dx.doi.org/10.1086/682289.
Cortes, M., N. Jaimovich and H. Siu (2017), “Disappearing Routine Jobs: Who, How, and Why?”, [24]
Journal of Monetary Economics, Vol. 91, pp. 69-87.
Dorn, D. (2009), “Data Appendix”, in Essays on Inequality, Spatial Interaction, and the Demand [19]
for Skills, Dissertation University of St. Gallen no. 3613, September.,
https://www.ddorn.net/data/Dorn_Thesis_Appendix.pdf (accessed on 17 April 2019).
Goos, M. and A. Manning (2007), “Lousy and Lovely Jobs: The Rising Polarization of Work in [4]
Britain”, The Review of Economics and Statistics, Vol. 89/1, pp. 118-133,
http://www.mitpressjournals.org/doi/pdf/10.1162/rest.89.1.118 (accessed on 4 August 2017).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
246
Goos, M., A. Manning and A. Salomons (2014), “Explaining Job Polarization: Routine-Biased [18]
Technological Change and Offshoring”, American Economic Review, Vol. 104/8, pp. 2509-
2526, http://dx.doi.org/10.1257/aer.104.8.2509.
Goos, M., A. Manning and A. Salomons (2009), “Job Polarization in Europe”, American [5]
Economic Review, Vol. 99/2, pp. 58-63, http://dx.doi.org/10.1257/aer.99.2.58.
Helper, S., T. Krueger and H. Wial (2012), Why Does Manufacturing Matter? Which [1]
Manufacturing Matters? A Policy Framework, Metropolitan Policy Program at the Brookings
Institution.
Hofer, H., G. Titelbach and S. Vogtenhuber (2017), “Polarisierung am österreichischen [15]
Arbeitsmarkt?”, Wirtschaft und Gesellschaft, Vol. 43/3, pp. 379-404.
Hyatt, H. and J. Spletzer (2013), “The recent decline in employment dynamics”, IZA Journal of [29]
Labor Economics, Vol. 2/1, p. 5, http://dx.doi.org/10.1186/2193-8997-2-5.
Jaimovich, N. and H. Siu (2014), “The Trend is the Cycle: Job Polarization and Jobless [21]
Recoveries”, NBER Working Paper Series, No. 18334, NBER,
http://www.nber.org/papers/w18334.
Maczulskij, T. and M. Kauhanen (2017), “Where do workers from declining routine jobs go and [26]
does migration matter?”, Työpapereita Working Papers, No. 314, Labour Institute for
Economic Research.
Maurin, E. and S. Signorelli (2019), The Decline in Internal Labor Markets and Technological [23]
Change, Working Paper.
OECD (2019), Under Pressure: The Squeezed Middle Class, OECD Publishing, Paris, [2]
https://dx.doi.org/10.1787/689afed1-en.
OECD (2018), OECD Employment Outlook 2018, OECD Publishing, Paris, [14]
https://dx.doi.org/10.1787/empl_outlook-2018-en.
OECD (2017), “How technology and globalisation are transforming the labour market”, in [6]
Annexes of Chapter 3 of the OECD Employment Outlook 2017, OECD Publishing, Paris,
https://www.oecd.org/els/emp/Employment-Outlook-2017-Annexes-Chapter-3.pdf.
OECD (2017), OECD Employment Outlook 2017, OECD Publishing, Paris, [10]
https://dx.doi.org/10.1787/empl_outlook-2017-en.
OECD (2009), OECD Employment Outlook 2009: Tackling the Jobs Crisis, OECD Publishing, [31]
Paris, https://dx.doi.org/10.1787/empl_outlook-2009-en.
Oesch, D. and G. Piccitto (2019), “The Polarization Myth: Occupational Upgrading in Germany, [17]
Spain, Sweden, and the UK, 1992-2015”, Work and Occupations, Vol. 46/4, pp. 441-469,
http://dx.doi.org/10.1177/0730888419860880.
Salvatori, A. (2015), “The anatomy of job polarisation in the UK”, IZA Discussion Paper, [27]
No. 9193, IZA, https://www.iza.org/publications/dp/9193.
Schmitt, J., H. Shierholz and L. Mishel (2013), “Don’t Blame the Robots”, EPI-CEPR Working [28]
Paper, Economic Policy Institute, Washington, DC,
https://www.epi.org/publication/technology-inequality-dont-blame-the-robots/.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
247
Smith, C. (2013), “The Dynamics of Labor Market Polarization”, Finance and Economics [22]
Discussion Series, No. 2013-57, Federal Reserve Board,
https://www.federalreserve.gov/pubs/feds/2013/201357/201357abs.html.
Tåhlin, M. (2019), “Polariseringsmyten. Försvinner verkligen de medelkvalificerade jobben?”. [16]
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
248
Annex 4.A. Data Sources
Unless otherwise noted, all samples use the working-age population defined as individuals ages 16-64.
Some samples further restrict the age range to younger workers (ages 16-24), prime-age (25-54) and older
workers (55-64).
European Labour Force Survey (EU-LFS). The EU-LFS is the largest European household
sample survey covering labour force participation of people aged 15 years and older as well as
people outside the labour force. The national statistical institutes design and execute their own
respective labour force surveys. Eurostat harmonises and distributes the microdata in consultation
with the respective national statistical institutes. All OECD members who are members of the
European Union or are EFTA countries are included in this study with the exception of Iceland.
German Socio-Economic Panel (GSOEP). The GSOEP is a longitudinal survey of private
households in Germany from 1984 to 2016. The database is produced by the Deutsches Institut
für Wirtschaftsforschung (DIW). The survey covers labour force topics, as well as household
composition, health, and satisfaction.
Current Population Survey (CPS) including supplements. The CPS is a monthly household
survey conducted by the U.S. Census Bureau on behalf of the Bureau of Labor Statistics. The
survey provides labour force information on household members age 15 years and older. In
addition to the main questionnaire, each month one fourth of respondents rotate out of the survey
and provide information on earnings.
European Union Statistics on Income and Living Conditions (EU-SILC). The EU-SILC is a
European household sample survey covering income, poverty, social exclusion, living conditions
and labour market outcomes. The EU-SILC began in 2004 and it now covers all EU countries,
Iceland, Norway and Switzerland. The survey contains both a cross-sectional survey and a
longitudinal element which observes individuals over-time. The longitudinal portion is employed in
this chapter to track labour market transitions.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
249
Annex 4.B. Attrition Transition Decomposition
The graphical analysis found in Figure 4.4 of the difference between transitions and attrition for the causes
of the decline in middle-skill employment is useful, but incomplete. Annex Figure 4.B.1 shows the results
of a more rigorous analytical decomposition of the contribution of attrition, transitions and cohort sizes to
the decline in middle-skill employment shares. Attrition is meant as the contribution of younger cohorts
showing different propensities to enter different occupation groups and non-employment when they are
young (25-29 or 30-34, for example) while older cohorts exit the labour market. Transitions are meant to
capture how propensities change as cohorts age, and whether these trajectories have meaningfully
changed for subsequent birth cohorts. Finally, cohort size captures the fact that the analysis looks at the
change in middle-skill employment over time, and birth cohorts appear at different ages at different points
in time. Thus, the share of middle-skill employment can change – holding attrition and transitions constant
– if birth cohort sizes vary.
Annex Figure 4.B.1. Attrition accounts for the majority of the decline in the share of middle-skill
employment
Decomposition of the change in the share of middle-skill employment by attrition, transitions and cohort size,
1995-2018
Attrition Cohort size Transition Total change
0.04
0.02
0
-0.02
-0.04
-0.06
-0.08
-0.1
-0.12
-0.14
-0.16
LVA LUX SVK FIN SWE GRC EST CHE NOR CZE AUT OECD DNK PRT FRA GBR ESP BEL IRE NLD ITA HUN DEU
Notes: Units are percentage point changes. Time period for change in the share of middle-skill employment is the average of 1995-2001 and
2013-2018. OECD is an unweighted average of the countries shown.
Source: European Labour Force Survey (EU-LFS).
StatLink 2 https://stat.link/pu9f27
The more formal analysis shows that attrition is the primary mechanism for declining middle-skill
employment shares. The figure shows the share of the change in middle-skill employment shares due to
attrition, transitions and cohort size. Reported total changes differ from those shown in Figure 4.1 due to
different time-periods. The first main finding is that the attrition component is sharply negative for almost
all countries, and in all countries except five – Estonia, Switzerland, Norway, Hungary and Sweden – the
contribution from attrition is greater than that from transitions. This indicates that the lower propensity to
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
250
enter middle-skill employment for younger workers is driving the change. Second, the transition component
is heterogeneous across countries. Few countries show a sharp negative contribution from transitions –
Estonia, Switzerland, and Norway – while others show even a positive contribution. In countries where the
transition component is positive, this implies that if entry rates were equal across age cohorts, the share
of middle-skill employment would actually increase over-time. On average, attrition accounts for a little
under two thirds of the change in middle-skill employment with the rest split almost evenly among
transitions and cohort size.
Derivation of attrition and transition decomposition
What follows is the derivation of the decomposition into the two components (attrition and transition) as
well as the age-cohort population weights. The derivation covers the general case of when the time period
of the decomposition perfectly aligns with the minimum and maximum of the age groups. However, the
decomposition generalizes to arbitrary time spans and the cases when cohorts are not perfectly
observable. The analysis focuses on the cohort decomposition for middle-skill jobs (but the same
reasoning/calculations hold for other types of jobs and non-employment).
There are three dimensions: age class j, cohort c and time t. Note that there is an unambiguous
correspondence between any couple of these indexes and the third one. For example, class j at time t
completely identifies cohort c. But at the same time cohort c and class j perfectly identifies time t, etc. To
fix ideas, a unit of time in the derivation of the same span as the span of age classes. In Figure 4.4 age
classes cover 5 years.
𝑗
The derivation first proceeds with definitions. Let 𝑀𝑖,𝑡 be the number of people of age class j in middle-skill
𝑗
jobs at time t (for cohort i) and 𝑃𝑖,𝑡 the number of people of age class j at time t. We have, assuming that
cohort c is in age class 1 at time t:
𝑗
𝑀𝑐−𝑗+1,𝑡 𝑗
𝑗
∑𝑘𝑗=1 ( 𝑗 ) 𝑃𝑐−𝑗+1,𝑡 𝑘 𝑘
∑𝑘𝑗=1 𝑀𝑐−𝑗+1,𝑡 𝑃𝑐−𝑗+1,𝑡 𝑗 𝑗 𝑗 𝑗 𝑗
𝑆𝑡 = 𝑘 𝑗
= 𝑗
= ∑(𝑀𝑐−𝑗+1,𝑡 /𝑃𝑐−𝑗+1,𝑡 )𝑤𝑡 = ∑ 𝑠𝑐−𝑗+1,𝑡 𝑤𝑡
∑𝑗=1 𝑃𝑐−𝑗+1,𝑡 ∑𝑘𝑗=1 𝑃𝑐−𝑗+1,𝑡 𝑗=1 𝑗=1
𝑗
𝑗 𝑀𝑖,𝑡
where 𝑆𝑡 stands for shares at time 𝑡, 𝑠𝑖,𝑡 = 𝑗 are age-class-specific share, w are population weights (the
𝑃𝑖,𝑡
share of population with age class j in total population within the range defined by the selected set of age
classes, e.g. 25-54) and k is the number of age classes.
The decomposition begins with a shift-share analysis between t and t+k-1 (with age-classes as the only
grouping variable). This yields:
𝑘 𝑘
𝑗 𝑗 𝑗 𝑗 𝑗 𝑗
𝑆𝑡+𝑘−1 − 𝑆𝑡 = ∑ 𝑤𝑡 (𝑠𝑐+𝑘−𝑗,𝑡+𝑘−1 − 𝑠𝑐−𝑗+1,𝑡 ) + ∑ 𝑠𝑐+𝑘−𝑗,𝑡+𝑘−1 (𝑤𝑡+𝑘−1 − 𝑤𝑡 )
⏟
𝑗=1 ⏟
𝑗=1
𝑆ℎ𝑖𝑓𝑡[𝑡;𝑡+𝑘−1] 𝑆ℎ𝑎𝑟𝑒[𝑡;𝑡+𝑘−1]
The Share component simply gives the share of the change due to cohort size. The Shift component will
yield the attrition and transition components. The idea is to compare each age-by-time specific share with
the corresponding age-by-time specific share in the benchmark cohort.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
251
We have:
𝑘 𝑘
𝑗 𝑗 𝑗 𝑗
𝑆ℎ𝑖𝑓𝑡[𝑡; 𝑡 + 𝑘 − 1] = ∑ 𝑤𝑡 𝑠𝑐+𝑘−𝑗,𝑡+𝑘−1 − ∑ 𝑤𝑡 𝑠𝑐−𝑗+1,𝑡
𝑗=1 𝑗=1
𝑘 𝑘 𝑘 𝑘
𝑗 𝑗 𝑗 𝑗 𝑗 𝑗 𝑗 𝑗
= ∑ 𝑤𝑡 𝑠𝑐+𝑘−𝑗,𝑡+𝑘−1 − ∑ 𝑤𝑡 𝑠𝑐,𝑡+𝑗−1 + ∑ 𝑤𝑡 𝑠𝑐,𝑡+𝑗−1 − ∑ 𝑤𝑡 𝑠𝑐−𝑗+1,𝑡
𝑗=1 𝑗=1 𝑗=1 𝑗=1
𝑘 𝑘
𝑗 𝑗 𝑗 𝑗 𝑗 𝑗
= ∑ 𝑤𝑡 (𝑠𝑐+𝑘−𝑗,𝑡+𝑘−1 − 𝑠𝑐,𝑡+𝑗−1 ) − ∑ 𝑤𝑡 (𝑠𝑐,𝑡+𝑗−1 − 𝑠𝑐−𝑗+1,𝑡 )
𝑗=1 𝑗=1
Before proceeding, define ∏𝑗𝑠=1 𝑝 𝑠
as the transition path for the benchmark cohort, which gives the
𝑗 1 𝑗
trajectory of the benchmark cohort from its initial entry share. This implies that 𝑠𝑐,𝑡+𝑗−1 = 𝑠𝑐,𝑡 (∏𝑠=1 𝑝 𝑠 ).
Similarly, define ∏𝑘𝑠=𝑗 𝑞 𝑠 as the inverse transition path of the benchmark cohort, which maps the benchmark
𝑘 𝑗
cohort back from its terminal share at age group k implying 𝑠𝑐,𝑡+𝑘−1 (∏𝑘𝑠=𝑗 𝑞 𝑠 ) = 𝑠𝑐,𝑡+𝑗−1 . For clarity, in this
example the benchmark cohort is cohort c, and enters the labour market (j=1) at time t.
The derivation into the transition and attrition components proceeds with the following two steps. First,
substitute the two identities defined in the previous paragraph into the second and third terms of the most
𝑗 𝑗 𝑗 𝑘 𝑗
recent equation and then add and subtract ∑𝑘𝑗=1 𝑠 (∏𝑠=1 𝑝 𝑠 )𝑤𝑡 and ∑𝑘𝑗=1 𝑠 (∏𝑘𝑠=𝑗 𝑞 𝑠 )𝑤𝑡 ,
𝑐+𝑘−𝑗,𝑡+𝑘−1 𝑐−𝑗+1,𝑡+𝑘−𝑗
respectively. This yields
𝑆ℎ𝑖𝑓𝑡[𝑡; 𝑡 + 𝑘 − 1] = 𝑇𝑊𝑇𝑟[𝑡; 𝑡 + 𝑘 − 1] + 𝑇𝑊𝐴𝑡[𝑡; 𝑡 + 𝑘 − 1]
Where TWTr represents the (weighted) transition component across dates and TWAt represents the
(weighted) attrition component across dates. Operationally we have:
𝑇𝑊𝑇𝑟[𝑡; 𝑡 + 𝑘 − 1] =
𝑘 𝑗 𝑘 𝑘
𝑗 1 𝑗 𝑘 𝑗
= ∑ (𝑠𝑐+𝑘−𝑗,𝑡+𝑘−1 − 𝑠𝑐+𝑘−𝑗,𝑡+𝑘−𝑗 (∏ 𝑝 )) + ∑ 𝑤𝑡 (𝑠𝑐−𝑗+1,𝑡+𝑘−𝑗 (∏ 𝑞 𝑠 ) − 𝑠𝑐−𝑗+1,𝑡 )
𝑠
𝑗=1 𝑠=1 𝑗=1 𝑠=𝑗
The first term fixes the benchmark trajectory to each post-benchmark cohort and compares their
hypothetical share (using the benchmark trajectory) to their actual share in the t+1 time period. The second
term does the same with the pre-benchmark cohorts using the inverse transition path and comparing the
hypothetical and actual share in period t. The result is the contribution of transitions. For attrition we have:
𝑇𝑊𝐴𝑡[𝑡; 𝑡 + 𝑘 − 1] =
𝑘 𝑗 𝑘 𝑘
𝑗 1 𝑗 𝑘 𝑘
= ∑ 𝑤𝑡 (∏ 𝑝 ) (𝑠𝑐+𝑘−𝑗,𝑡+𝑘−𝑗 − 𝑠𝑐,𝑡 ) + ∑ 𝑤𝑡 (∏ 𝑞 𝑠 ) (𝑠𝑐,𝑡+𝑘−1
𝑠 1
− 𝑠𝑐−𝑗+1,𝑡+𝑘−𝑗 )
𝑗=1 𝑠=1 𝑗=1 𝑠=𝑗
The intuition is similar. The first term holds the transition component constant allowing the comparison of
the entry share between the benchmark cohort, and each subsequent cohort. The second term does the
same, except allowing the comparison of the benchmark cohort in its terminal age group (j=k) with the
share in the terminal age group for all earlier cohorts. The result is the attrition component.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
252
Annex 4.C. Additional Figures
Annex Figure 4.C.1. Education shares of the working-age population
Mid-1990s Mid-2010s
A. Low-education
80%
70%
60%
50%
40%
30%
20%
10%
0%
AUT BEL CZE DEU DNK EST ESP FIN FRA GRC HUN IRL ITA LUX LVA NLD NOR POL PRT SWE SVN SVK GBR USA
B. Middle-education
80%
70%
60%
50%
40%
30%
20%
10%
0%
AUT BEL CZE DEU DNK EST ESP FIN FRA GRC HUN IRL ITA LUX LVA NLD NOR POL PRT SWE SVN SVK GBR USA
C. High-education
80%
70%
60%
50%
40%
30%
20%
10%
0%
AUT BEL CZE DEU DNK EST ESP FIN FRA GRC HUN IRL ITA LUX LVA NLD NOR POL PRT SWE SVN SVK GBR USA
Notes: Education share are of population ages 15-64. Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, “mid-1990s”
is the three earliest years of data. The earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia,
Finland, Hungary, Sweden), 1998 (Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s refer to 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/iv3e8b
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
253
Annex Figure 4.C.2. Alternative shift-share analysis
Composition Propensity Total
A. Population
0.05
0
-0.05
-0.1
-0.15
-0.2
-0.25
-0.3
LUX SVN DNK GRC GBR CHE NOR AUT PRT FRA SWE USA DEU IRL BEL ITA FIN NLD ESP LVA EST SVK CZE HUN POL
B. Employed
0.05
0
-0.05
-0.1
-0.15
-0.2
-0.25
-0.3
LUX SVN GRC FRA IRL GBR AUT DNK ESP PRT DEU ITA SWE CHE BEL NOR FIN LVA NLD POL HUN USA EST SVK CZE
Notes: Units are percentage point changes. Both shift-share analyses include age (young, prime, old) in addition to sex and education. Panel A
decomposes changes in the share of middle-skill workers for the working-age population. Panel B does the same for the working-age employed
only. Shift-share is from mid-1990s to mid-2010s. Mid-1990s refer to 1994, 1995, and 1996. For countries with no data in 1994, "mid-1990s" is
the three earliest years of data. The earliest years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia,
Finland, Hungary, Sweden), 1998 (Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/x5ua7r
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
254
Annex Figure 4.C.3. Shares of low-education, prime-age workers in middle-skill employment
Mid-1990s Mid-2010s
A. Women
70%
60%
50%
40%
30%
20%
10%
0%
ESP IRL CHE NLD BEL ITA LUX SVK DEU USA NOR DNK HUN FRA GBR GRC POL FIN LVA AUT SWE CZE PRT EST SVN
B. Men
70%
60%
50%
40%
30%
20%
10%
0%
SVK HUN LVA GBR CHE USA CZE IRL POL FIN DNK EST ESP AUT NLD BEL ITA DEU FRA NOR PRT SWE SVN GRC LUX
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the CPS for the United States.
StatLink 2 https://stat.link/3yvgta
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
255
Annex Figure 4.C.4. Share of middle-educated, prime-age workers in middle-skill employment
Mid-1990s Mid-2010s
A. Women
70%
60%
50%
40%
30%
20%
10%
0%
NLD ESP LUX LVA SVK GRC FIN SWE DEU NOR EST ITA POL CZE HUN BEL AUT IRL USA DNK FRA GBR PRT SVN
B. Men
70%
60%
50%
40%
30%
20%
10%
0%
PRT NLD ITA ESP USA IRL FIN SWE GBR LUX LVA BEL GRC POL FRA DNK NOR AUT HUN DEU EST SVK SVN CZE
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the CPS for the United States.
StatLink 2 https://stat.link/w9qt7n
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
256
Annex Figure 4.C.5. Share of low-educated, prime-age workers in low-skill employment
Mid-1990s Mid-2010s
A. Women
70%
60%
50%
40%
30%
20%
10%
0%
GRC USA IRL ITA ESP POL SVN BEL NLD HUN CHE EST DEU LUX GBR PRT LVA FRA AUT SVK FIN CZE SWE DNK NOR
B. Men
70%
60%
50%
40%
30%
20%
10%
0%
CHE SWE SVN POL FIN NOR NLD FRA EST GRC DEU GBR BEL PRT USA SVK LUX HUN LVA ESP IRL AUT DNK ITA CZE
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/z57o3n
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
257
Annex Figure 4.C.6. Share of middle-educated, prime-age workers in low-skill employment
Mid-1990s Mid-2010s
A. Women
70%
60%
50%
40%
30%
20%
10%
0%
ITA PRT LUX USA GRC IRL ESP DEU NLD HUN POL BEL GBR FRA FIN AUT SVK SVN CZE LVA DNK EST SWE NOR
B. Men
70%
60%
50%
40%
30%
20%
10%
0%
EST POL NOR CZE DEU SVN SWE FIN BEL LUX NLD LVA AUT FRA SVK ITA USA HUN GBR PRT DNK IRL GRC ESP
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
for the United States.
StatLink 2 https://stat.link/cxbr2h
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
258
Annex Figure 4.C.7. Share of low-educated, prime-age workers in non-employment
Mid-1990s Mid-2010s
A. Women
70%
60%
50%
40%
30%
20%
10%
0%
SWE FIN PRT NOR SVN CZE AUT GBR DNK EST FRA LVA DEU CHE SVK USA HUN LUX NLD POL BEL GRC ITA ESP IRL
B. Men
70%
60%
50%
40%
30%
20%
10%
0%
LUX GRC PRT AUT ITA NLD FRA SWE NOR CHE BEL GBR DNK DEU ESP SVN USA IRL CZE FIN EST LVA HUN POL SVK
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/h6sfw5
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
259
Annex Figure 4.C.8. Share of middle-educated, prime-age workers in non-employment employment
Mid-1990s Mid-2010s
A. Women
70%
60%
50%
40%
30%
20%
10%
0%
SVN SWE NOR CZE DNK PRT USA GBR EST SVK AUT FIN FRA LVA NLD HUN BEL DEU ITA POL LUX IRL ESP GRC
B. Men
70%
60%
50%
40%
30%
20%
10%
0%
LUX CZE NLD NOR AUT BEL GRC FRA USA IRL PRT DNK SVN DEU GBR ITA SVK SWE ESP EST HUN LVA FIN POL
Notes: Mid-1990s refer to 1994, 1995 and 1996. For countries with no data in 1994, "mid-1990s" is the three earliest years of data. The earliest
years are: 1995 (Austria), 1996 (Netherlands, Norway, Slovenia, Switzerland), 1997 (Estonia, Finland, Hungary, Sweden), 1998
(Czech Republic, Latvia, Slovak Republic), 2002 (Poland). Mid-2010s are 2014-2016.
Source: European labour force survey (EU-LFS), The German Socio-Economic Panel (SOEP) for Germany, and the Current Population Survey
(CPS) for the United States.
StatLink 2 https://stat.link/zwk1xn
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
260
Notes
1
The causes of job polarisation are not uniformly agreed upon. Researchers noted that the U-shaped
pattern underpinning job polarisation does not hold across time periods. In addition, changes in occupation
shares do not explain skill patterns as the theory intended (Schmitt, Shierholz and Mishel, 2013[28]). This
chapter notes the existing research and makes no original claims about why the shares of middle-skill
occupations are declining. Rather, it focuses on the dynamics of how the share has declined, and where
workers who would have held those jobs in the past are doing now.
2
European employment data beyond 2010 was mapped from ISCO-08 to ISCO-88 using a many-to-many
mapping technique. This mapping technique is described in Annex 3.A4 (OECD, 2017[6]). For the
United States, uniform Census codes come from Dorn (2009[19]).
3
Employment flows for each group 𝑖, are derived following OECD (2009[31]): ∆𝐸𝑚𝑝𝑖 = 𝐻𝑖𝑟𝑒𝑠𝑖 − 𝑆𝑒𝑝𝑖 ,
Where ∆𝐸𝑚𝑝𝑖 is the net employment change in group 𝑖 between years 𝑡 and 𝑡 − 1. 𝐻𝑖𝑟𝑒𝑠𝑖 is the gross hires
for group 𝑖 at time 𝑡, which is determined by the number of workers with tenure less than one year. 𝑆𝑒𝑝𝑖 is
the number of gross separations and is pinned down by the identity. In practice gross hires and separations
are presented as a rate by dividing by average employment in years 𝑡 and 𝑡 − 1.
4
This is a longer term trend, which is well documented in the United States (Hyatt and Spletzer, 2013[29]),
with varying trends in other OECD countries (Cazes and Tonin, 2010[32]).
5
Unlike the previous section, this analysis uses labour force surveys, and does not follow the same worker
over time.
6
The size of each birth cohort is limited by sample size in the survey data. Ideally one would use each
year as the size of a birth cohort, but that would leave too small a sample with labour force surveys. The
number of age ranges is similarly limited by sample size, but also the maximum span of the surveys to
20 years of data.
7
This allows for roughly equal years before and after 1970 with birth years between 1946 and 1993. It also
roughly aligns with “baby boom” and “post baby boom” generations.
8
This includes education.
9
For example, age, and for the most part gender are fixed and pre-determined at birth, and do not change
with labour market conditions. Education and region of residence are more problematic and partly reflect
local labour market conditions. Both also involve high switching costs, and may reflect exogenous factors
such as historical family ties (place of residence) and labour market conditions many years prior to
observation (education). Given their importance in predicting middle-skill workers, both are included in the
set of possible predictors. Variables conditional on employment, such as industry and occupation, are not
included as predictors. Skill groups are determined by occupation, and its inclusion will make any
inferences about occupation tautological. Due to the strong correlation between occupations and industry,
industry is not included for similar reasons to occupation. Industry and occupation will be explored to
complete the picture of the typical middle-skill worker from 20 years prior, but they will be excluded from
defining a middle-skill worker.
10
This cross-country variation partially reflects the relative abundance of low- and middle-educational
attainment in the population.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
261
11
As noted, education is included as one of the main predictors, but its inclusion is potentially problematic.
Education is an endogenous choice for young workers or young people deciding whether to enter the
labour force or pursue further schooling. If a given young person is, for example, confronted with
satisfactory job opportunities once she obtains an upper-secondary degree, she may decide to forgo
further education. That same person entering a labour market at a different, hypothetical, point in time may
instead pursue further schooling in the face of a poor job market. This choice for people at the margin of
pursuing further education complicates the interpretation of the outcomes of education cohorts at different
points in time. Fixing this problem is not trivial, and all results should be interpreted with this in mind.
12
These are ISCO-88 3-digit occupations and the uniform Census codes from Dorn (2009[19]).
13
While the decline in manufacturing shifted the industry distribution of middle-skilled jobs, perhaps
surprisingly, the modal occupations remained relatively consistent compared to 20 years prior.
14
See Annex 4.C for a depiction of the changing education shares of the working-age population.
15
The shift share decomposes the change in the share of middle-skill employed according to:
𝑗 𝑗 𝑗 𝑗 𝑗
𝜋̅1 − 𝜋̅0 = ∑𝑔 ∆𝑤𝑔1 𝜋𝑔0 + ∑𝑔 𝑤𝑔0 ∆𝜋𝑔1 . The term 𝜋̅𝑡 is the share of the population in skill group 𝑗 at time 𝑡.
𝑗
The term 𝑤𝑔𝑡 is the share of the population in demographic group 𝑔 at time 𝑡, and 𝜋𝑔𝑡 is the share of group
𝑔 in skill group 𝑗 at time 𝑡. The left-hand side of the equation is the change in the share of the population
in skill group 𝑗. The two terms on the right-hand side of the equation are (from left to right) the composition
and propensity effects, respectively.
16
Other analyses include age as an important factor in middle-skill employment (Autor and Dorn, 2009[30]).
The analysis also undertook the shift-share using age as a factor. All results are qualitatively similar. The
change in propensities are stronger for younger workers, however. Results for the shift-share including
age, and propensities for prime-age education by sex groups to be employed in different skill groups is
available in Annex 4.C.
17
This is significantly more modest than the percentage point decrease presented at the beginning of
Section 4.1. In the first section, the shares are defined as shares of the employed as originally constructed
in the literature (Autor, Levy and Murnane, 2003[3]). The formulation here constructs the shares as a share
of the working-age population, which allows the shift-share to account for shifts out of employment.
18
For some perspective, across OECD countries in this chapter, the share of the population aged 16-64
without an upper-secondary degree decreased from 38.5% to 24.5%, while the population with at least a
tertiary degree increased from 15.9% to 28.9%. The share with at least an upper-secondary degree but
without a tertiary degree increased from 44.7% to 45.8%.
19
The increase in propensity for middle-educated women to work in low-skill employment was partly the
result of increased rates of employment. Low-educated women saw no meaningful increase in their
employment to population ratio.
20
Not in the chart.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
262
5 Smooth transitions but in a
changing market: The prospects of
vocational education and training
graduates
This chapter looks at current labour market outcomes of young graduates
from mid-level vocational education and training (VET), as well as how they
have changed in the past 10 to 15 years and what can be expected in the
medium-term. It looks at indicators of job quality and quantity, and zooms in
on the types of occupations that employ VET graduates. The outcomes of
VET graduates are compared to those of general education graduates (at
the same qualification level), tertiary education graduates and graduates
without an upper-secondary education degree. Differences in outcomes
based on the features of each country’s VET system are discussed. Finally,
based on these findings, the chapter discusses key policy directions to
improve VET graduates’ access to high-quality secure jobs.
This chapter was produced with the financial assistance of the European Union. The views expressed herein can in
no way be taken to reflect the official opinion of the European Union.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
263
In Brief
Key findings
Vocational Education and Training (VET) embraces education, training and skills development in a wide
range of occupational fields. VET programmes can include work-based components or be completely
school based, but they generally lead to qualifications that are relevant in the labour market. They
therefore enhance student engagement in education, through lower school dropout, and facilitate
school-to-work transitions. However, structural changes in the labour market, as documented in
Chapter 4, have raised concerns around the labour market outcomes of VET graduates. Many VET
programmes are organised at medium levels of education and therefore prepare students for
employment in middle-skill jobs that are declining. This is a particular concern in countries with weaker
VET systems. This chapter looks at the current labour market outcomes of young graduates from
mid-level VET, how these have changed in the past 10 to 15 years and what can be expected in the
medium term. It compares the experiences of young middle-educated VET graduates and people
leaving general education with the same level of qualifications. Middle-educated VET graduates are
those who have obtained at most an upper-secondary education qualification (i.e. ISCED level 3) or a
post-secondary non-tertiary education qualification (i.e. ISCED level 4). They are compared with
individuals completing general education with non-vocational qualifications at the same level. These two
groups are referred to as “VET graduates” and “general education graduates” for simplicity. For
completeness, the outcomes of these two groups are also compared to those of individuals with higher
and lower education levels. The key findings are as follows:
VET plays a prominent role in education systems in OECD countries. On average, almost a third
of 15-34 year-olds hold a mid-level VET qualification as their highest qualification. However, the
way in which VET is organised and delivered has an impact on its quality, which in turn influences
the attractiveness of VET and the labour market outcomes of graduates. Large differences exist
between countries in the importance of VET in the education system, with mid-level VET playing
a very prominent role in countries like Austria, the Czech Republic, Germany and the
Slovak Republic, and only a limited role in Canada, Israel, Japan and Mexico. Countries also
differ in the way they organise VET, with some countries, for example, having an important
workplace learning component in their VET curricula (e.g. Austria, Germany, Norway and
Switzerland) and others taking a predominantly school-based approach (e.g. Belgium, Finland
and Slovenia).
Young VET graduates in OECD countries have higher employment rates and lower
unemployment rates than general education graduates on average, and these differences have
remained stable over the past 15 years. This is particularly the case in Austria, Germany,
Norway and the United States where the labour market performance of VET graduates is
substantially better than that of general education graduates. At the same time, on average
across countries, the difference disappears for older age groups, suggesting that VET does
particularly well in facilitating rapid and successful school-to-work transitions, but potentially
loses its comparative advantage over the working life
At the start of their careers, VET graduates enjoy better job quality than general education
graduates. Median wages of young VET graduates are slightly higher than those of general
education graduates, but this difference cannot be observed for older age groups on average
across countries. On the other hand, VET graduates earn less than tertiary education graduates
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
264
at all ages, even when employed in similar jobs. VET graduates are also more likely to have a
permanent contract and to have supervisory responsibilities in their job than general education
graduates at the very start of their careers.
Just over a quarter of young VET graduates are employed as service and sales workers, and
an additional 22% as craft and related trades workers, compared to 30% and 9%, respectively,
of general education graduates. Furthermore, 20% of VET graduates are employed in high-skill
occupations (i.e. managers, professionals and associate professionals and technicians),
compared with 26% of general education graduates. However, this average hides significant
differences across VET systems. In Germany, Switzerland and the United States, more than
one-third of young VET graduates work in high-skill occupations.
The occupational composition of VET graduates’ employment has changed in recent decades,
with service and sales occupations and elementary occupations representing a growing share
of VET graduates’ employment. High-skill occupations and clerical occupations have become
relatively less prevalent for VET graduates. Interestingly, the share of young VET graduates
employed in craft and related trades occupations has remained stable over the past 15 years,
in spite of an overall decline in the importance of those occupations in the labour market. VET
graduates’ comparative advantage in these occupations therefore seems to have given them
some protection against structural changes in the labour market.
The short-term labour market prospects for VET graduates are on average better than for
general education graduates, although they are less bright than for tertiary education graduates.
Occupations that have a larger share of VET graduates among their young workers are more
likely to have a surplus of labour (i.e. the supply of labour with the relevant skills exceeds the
demand) than those that do not employ many VET graduates. Nonetheless, there are some
exceptions, and the occupation that has the largest share of VET graduates among their
workers, i.e. electrical and electronics trades, is facing a large shortage on average across
countries. Labour surpluses are even more common in occupations that employ many general
education graduates. The opposite is true for young tertiary education graduates, as occupations
that employ a large share of these graduates are more likely to face labour shortage.
On average, 21% of jobs held by young VET graduates are highly automatable, which is about
the same as for general education graduates (22%), but much higher than for jobs held by
tertiary education graduates (9%). Graduates without an upper-secondary qualification face the
highest risk of job automation, with 28% of them working in jobs at high risk. A simulation of the
impact of a surge in the adoption of technology to automate tasks on the occupational structure
of employment suggests that VET and tertiary education graduates would mostly see relative
employment gains in high-skill jobs, whereas the relative gains for general education graduates
and those without an upper-secondary education degree would mainly be in low-skill
occupations. Moreover, while automation and other structural factors imply that few new jobs
might be created in the coming years in some occupations that are important for VET graduates,
like crafts and related trades occupations, a significant number of job openings can still be
expected to replace workers who leave these occupations (e.g. due to retirement).
In a changing world of work, certain aspects of VET systems in certain countries might need to be
re-engineered to further strengthen the positive impact VET can have on education and labour
market outcomes. VET graduates are not facing the same challenges in all countries, and the need
for intervention depends on the specific quality of each VET system and its ability to adapt to
changes. Focusing on countries where VET graduates perform well provides some insights into
the specific features of VET systems that might foster positive labour market outcomes. For
instance, in Austria, Denmark, Germany and Switzerland – where labour market outcomes of VET
graduates are good – the ties between VET institutions and social partners are very strong. This
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
265
suggests that co-operation between VET systems and the world of work is essential to ensure that
graduates enter the labour market with skills that correspond to labour market needs. Such co-
ordination will involve opening up the VET system to non-traditional fields of study linked to growing
occupations and sectors, to help VET graduates access available job opportunities.
As graduates need to be able to adapt to change, it is crucial that they have strong foundation
skills. However, literacy, numeracy and digital problem-solving skills of young VET graduates
are generally lower than those of graduates from general education. Exceptions are Canada,
Ireland, New Zealand and the United States, where VET is predominantly organised at the
post-secondary non-tertiary level and VET graduates are therefore exposed to more years of
general education, but also Japan. Additionally, for VET graduates to remain resilient in light of
changing skill needs, they need access to upskilling and reskilling opportunities. Only 43% of
young VET graduates participate in formal or non-formal training. This is slightly higher than
among general education graduates, but much lower than among tertiary education graduates.
Introduction
As a result of global megatrends, such as technological progress and globalisation, the demand for skills
has undergone substantial changes in recent decades. At the same time, educational attainment has risen
drastically in OECD countries, significantly altering the supply of skills. As documented in Chapter 4, labour
markets have polarised, with middle-skill jobs becoming less important relative to high- and low-skill jobs.
These changes have meant that middle-educated workers increasingly end up in low-skill jobs. This raises
the question of the extent to which graduates from vocational education and training (VET), whose training
generally prepares them for middle-skill jobs, are impacted by these changes. Are graduates from VET
more strongly affected because many of the typical VET jobs are the ones most exposed to automation?
Or are they better at withstanding the negative consequences of structural changes because VET systems
are able to adapt and prepare students for the jobs that are in demand in the labour market?
VET is a comprehensive term commonly used to refer to education, training and skills development for a wide
range of occupational fields. Many VET programmes have work-based components (e.g. apprenticeships,
traineeships, dual-system education programmes), but VET programmes can also be entirely school based.
Successful completion of VET programmes leads to market-relevant, vocational qualifications recognised
as occupationally oriented by the relevant national authorities and in the labour market (OECD, 2018[1]).
This chapter compares the labour market outcomes of young middle-educated VET graduates and general
education graduates with the same level of qualifications. Middle-educated VET graduates comprise those
who have obtained a vocationally oriented upper-secondary education qualification (ISCED 3) or a
post-secondary non-tertiary education qualification (ISCED 4) at most (see Annex 5.B for an overview of
education programmes at these levels). These are compared with general education graduates with
non-vocational qualifications at the same level (ISCED 3 or 4). Hereafter, these two groups are referred to
as “VET graduates” and “general education graduates”. For comparison, the analysis also includes tertiary
education graduates (i.e. all graduates with qualifications higher than ISCED 4, whether vocational or
general)1 and graduates who left education without an upper-secondary education. Individuals who are
still in education are excluded from the sample.2 It is important to note that differences in labour market
outcomes between the education groups not only reflect the differences in quality, relevance and duration
of education, but also other factors such as selection effects. Students entering the vocational track in
secondary education might have very different characteristics than the ones opting for the general track.
For example, PISA data show that 15-year old students in pre-vocational or vocational tracks have, on
average across countries, lower skill levels than students in general tracks, even when comparing students
with similar socio-economic characteristics (OECD, 2016[2]; 2016[3]).3,4
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
266
The chapter concentrates on 15 to 34-year-olds who are no longer in education (referred to as
“graduates”), as these people all left the education system relatively recently, thus ensuring that
comparisons are being made between individuals who enrolled in similar education and training
programmes. Nonetheless, some of the analyses presented below compare young graduates’ outcomes
with those of older age groups. When interpreting these comparisons, it should be kept in mind that i n
these cases the different age groups did not necessarily go through similar education and training
programmes.
The first section of this chapter looks at the importance of mid-level VET in the overall education system
in OECD countries, highlighting the large differences between countries. In Section 5.2, job quality and
quantity outcomes of VET graduates are compared with those of other types of grad uates, including an
analysis of the occupational composition of employment. Section 5.3 looks at the short- and
medium-term labour market outlook for VET graduates. This includes a discussion of short-term
employment prospects linked to current labour market imbalances and an analysis of the medium -term
outlook related to the automation of tasks. Section 5.4 discusses possible avenues for making VET
systems more resilient in a changing world of work.
5.1. The importance of medium-level VET in the education system
When looking at the types of qualifications held by individuals aged 15 to 34 years old, it is clear that
VET plays a prominent role in many OECD countries (see Figure 5.1). On average, the highest
qualification obtained by almost one out of three individuals aged 15 to 34 is mid-level VET
(i.e. upper-secondary (ISCED 3) or post-secondary non-tertiary level (ISCED 4) with vocational
orientation). Seen from a different angle, VET graduates account for 64% of indi viduals whose highest
qualification is at ISCED Level 3 or 4. The reason for a relatively low share of general education
graduates among those with mid-level qualifications in most countries, is that the majority of general
education graduates continue into higher education. In this chapter, general education graduates only
include those individuals who left education after obtaining an upper-secondary or post-secondary
non-tertiary degree (with a general orientation) or who continued to higher education bu t did not finish it
(excluding those who are still enrolled).
In countries such as Austria, the Czech Republic, Germany or the Slovak Republic, almost everyone
whose highest qualification is at ISCED Level 3 or 4 has a vocationally oriented degree (more than 90%
among 15-34 year-olds). In other countries, such as Korea, Norway, Spain or the United Kingdom, VET
graduates are not very common overall (less than 25% of all 15-34 year-olds), but still constitute at least
half of those who completed at most a mid-level degree. Yet, there are also countries where VET is
relatively rare, both in the full 15-34 year-old population, and among those with at most mid-level degrees
(e.g. Canada, Israel, Japan, Mexico, Turkey and the United States where less than 15% of all
15-34 year-olds and less than half of those with at most a medium-level qualification have a vocationally
oriented degree). OECD countries not only differ in the importance of VET in the education system, but
also in the way VET is organised and delivered and the two are likely to be related. Box 5.1 describes
some key differences with respect to workplace learning, employer engagement and the education level
at which VET is organised.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
267
Figure 5.1. VET is an important part of many education systems
Educational attainment of young adults (percentage of adults aged 15 to 34 not in formal education)
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Upper-secondary & post-secondary non-tertiary - Unknown
Tertiary
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Note: VET: Vocational Education and Training. The chart includes individuals aged 15/16 to 34 who are not enrolled in formal education. Data
refer to 2018 for all countries except Australia, Canada (2019), Korea, Japan (2011/12), Chile, New Zealand, Israel, Turkey (2014/15) and the
United States (2011/12-2014-17). OECD is an unweighted average of the countries shown. Countries are ordered by ascending order of
upper-secondary and post-secondary non-tertiary VET.
Source: European Union Labour Force Survey (2018), Turkish Labour Force Survey (2015), Australian Survey of Education and Work (2019),
Canadian Labour Force Survey (2019), OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/y9j2v8
Box 5.1. Understanding international differences in VET systems
There is wide variation across countries in how VET programmes are organised and delivered, as well
as the stages of education at which individuals pursue VET. One of the key distinctions between VET
systems concerns the use of workplace learning. Generally, VET programmes are classified as
school-based, work-based or combined school- and work-based, and often countries provide different
types of programmes in parallel.
In school-based programmes, at least 75% of the curriculum is presented in the school
environment (this may include distance education). This includes special training centres run by
public or private authorities, or enterprise-based special training centres if they qualify as
educational institutions. In countries like Belgium, Finland, Japan and Slovenia,
upper-secondary or post-secondary non-tertiary VET programmes are predominantly
school-based.
In combined school- and work-based programmes, at least 10% but less than 75% of the
curriculum is presented in the school environment or through distance education, with the
remainder organised as work-based learning. These programmes can be organised in
conjunction with education authorities or institutions. They include apprenticeship programmes
that involve concurrent school-based and work-based training (e.g. in Denmark and Norway),
and programmes that involve alternating periods of attendance at educational institutions and
participation in work-based training (like in the dual system in Germany and Switzerland).
In work-based programmes, the school-based component makes up less than 10% of the time.
Such programmes are usually non-formal education programmes leading to a qualification that
is recognised by national education authorities (or equivalent).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
268
The quality of work placements in work-based and mixed systems is crucial. How work placements are
regulated and organised determines whether students will systematically participate in quality-assured
placements, which allow them to develop useful skills and connect with employers or whether work
placements remain an optional add-on and of limited value. A lack of comparative data on the design
features of workplace learning limits the international comparability of VET programmes and makes it
difficult to categorise different VET systems.
Another important aspect of VET is the involvement of employers beyond their role in providing
workplace learning opportunities. Employers can be involved in designing curricula, qualification
standards and student evaluation guidelines. Moreover, employers can share information about student
outcomes and skill needs to feed into the re-design of curricula, and can be involved in determining the
optimal timing for curriculum re-design. In countries like Austria, Switzerland, Denmark and Germany,
employer involvement in these aspects is strong (KOF Swiss Economic Institute, 2016[4]).
It is important to note that not all OECD countries offer VET programmes at both the upper-secondary
and post-secondary non-tertiary level. In the United States, for example, there is no distinctive
vocational path at upper-secondary level, but vocational courses are offered (optionally) within the
general track. VET is provided mainly at post-secondary level, with community colleges being the main
provider. Similarly, in New Zealand students in upper-secondary education can integrate vocational
courses in general programmes, but fully fledged vocational programmes mostly exist at the
post-secondary level. In Australia, Canada and Ireland, VET programmes are typically delivered at the
post-secondary level (with Quebec being the exception in Canada). Also within the group of countries
that have a substantial VET offer at the secondary education level, there are differences in terms of the
age at which students are tracked into the different streams (i.e. VET versus general education).
Finally, countries also differ in terms of the pathways available for VET graduates to enter higher
education. Many OECD countries allow for direct access of VET graduates to higher education or
access through a bridging programme. In EU countries, two thirds of VET students at the
upper-secondary level are enrolled in programmes that give direct access to tertiary education (2017
data – Cedefop (2020[5])). However, internationally comparable data on progression from mid-level VET
to higher education is limited, and hence little is known about the use of the different pathways.
Source: Kís (forthcoming[6]), Improving evidence on VET: Data and Indicators; Cedefop (2017[7]), Education and labour market outcomes
for vocational education and training graduates in different types of VET systems in Europe.
In the majority of countries, the most common fields of study for young VET graduates are “engineering,
manufacturing and construction”, “social sciences, business and law”, and “services”. In some countries,
like the Netherlands and the United States, “health and welfare” is also a common field of study for VET
graduates. In most countries, women are significantly less likely to have a VET degree, and gender
differences in field of study choice are large, with very few female VET graduates specialised in
“engineering manufacturing and construction” but more in “social sciences, business and law”, “health and
welfare” and “services”. These gender differences in VET specialisation could result in gender differences
in labour market outcomes.
The share of young graduates who have at most a mid-level education qualification with vocational
orientation decreased across OECD countries with available data in the period 2004-18, from 38% to 32%
(Figure 5.2). The share of young graduates with qualifications at the same level but with a general
orientation remained stable at 16%. The share of young graduates from tertiary education increased
strongly (from 23% in 2004 to 36% in 2018), while the share of young graduates from the lowest
educational levels (i.e. below upper-secondary) decreased from 23% in 2004 to 16% in 2018. The average
trends mask substantial differences across countries. In Australia, Finland, Italy, Spain and Turkey the
share of young graduates with at most a medium-level VET degree increased slightly over the period
2004-17, whereas this share declined substantially in countries like Denmark, Hungary, Poland, the
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
269
Slovak Republic and Switzerland. The decline could be the result of a decrease in attractiveness of VET,
but could also mean that more VET graduates continue to (and complete) tertiary education.
Figure 5.2. Tertiary education is on the rise
Trend in educational attainment of young adults (percentage of adults aged 15 to 34 not in formal education)
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
40%
35%
30%
25%
20%
15%
10%
5%
0%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Note: VET: Vocational Education and Training. The chart includes individuals aged 15 to 34 not in education or training. Unweighted average
based on a balanced panel of countries, including Australia, Canada, Chile, Czech Republic, Estonia, Finland. Germany, Greece, Hungary,
Latvia, Lithuania, Netherlands, Poland, Slovak Republic, Spain, Sweden, Switzerland and Turkey.
Source: European Union Labour Force Survey, Encuesta de Caracterización Socioeconómica Nacional, Household, Income and Labour
Dynamics in Australia Survey, Turkish Labour Force Survey, Canadian Labour Force Survey.
StatLink 2 https://stat.link/42i1sw
As the educational composition of the population is changing, this could imply that the characteristics of
VET graduates relative to other graduates have changed as well. As shown in Annex 4.A, the share of
women decreased among young VET graduates relative to graduates from general or tertiary education.
The extent to which the education level of parents influence the choice between vocational and general
programmes has also changed significantly over time: young adults with tertiary educated parents are less
likely to obtain a VET degree than a general or tertiary degree, and this difference has increased over time.
Changes in the educational composition of graduates could also have altered the relative skill levels of
graduates, and Box 5.2 looks more closely into this using data from the OECD Survey of Adult Skills
(PIAAC).The findings from Box 5.2 suggest that literacy and numeracy skills of VET graduates leaving the
education system in the past 15 years have remained roughly stable relative to general and tertiary
education graduates’, while the skills of graduates who left education without an upper-secondary
education degree worsened compared to VET graduates’ skills.5
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
270
Box 5.2. Changing skill levels of young graduates
As educational attainment rates rise, one could expect the composition of graduates from different
education groups to change, including their skill levels. Measuring these changes is not straightforward,
as comparable information on the skill levels of graduates over time is not readily available. Using the
OECD Survey of Adults Skills (PIAAC), one can compare graduates who obtained their degree in different
time periods. The downside of such an analysis is that graduates who left the education system several
years prior to the skill assessment carried out in the survey might have experienced skills depreciation or,
conversely, their skills might have further developed because of the activities they have engaged in after
leaving education.
The figure below compares the skill level of different graduates at different times since graduation. This
allows comparing the skill levels of those who graduated recently (i.e. less than 5 years ago) to those who
obtained their degree 5 to 9 years ago and those who graduated 10 to 15 years ago. To eliminate as much
as possible the differences related to skills depreciation and/or skills development after education, the skill
levels reported in the figure are corrected for a range of aspects. These include: accumulated work
experience, current employment status, current occupation, current industry, self-reported willingness to
learn, participation in formal or non-formal training in the last 12 months, gender, migrant status and
number of children. Adults who graduated longer than 15 years ago are not included to avoid comparing
adults who went through substantially different education systems.
Figure 5.3. The skill composition of graduates from the last 15 years has remained stable
Skill levels of different cohorts
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
Ln (numeracy Numeracy Literacy
Ln (literacy
proficiency) proficiency)
5.75 5.75
5.7 5.7
5.65 5.65
5.6 5.6
5.55 5.55
5.5 5.5
5.45 5.45
5.4 5.4
5.35 5.35
10-14 years ago 5-9 years ago 0-4 years ago 10-14 years ago 5-9 years ago 0-4 years ago
Time of graduation Time of graduation
Note: VET: Vocational Education and Training. Based on an OLS regression of skills (ln numeracy and literacy, respectively) on education
group, years since graduation (+ an interaction of those two), labour market status, occupation (1-digit), industry (1-digit), years of work
experience, willingness to learn, participation in formal or non-formal training, gender, migrant status, number of children, country (all included
as dummy variables). The sample includes all individuals who are not enrolled in formal education and obtained their highest education
qualification at most 15 years prior to the interview in all OECD countries except Colombia, Iceland, Latvia, Luxembourg and Portugal.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/4pdz2k
The results show that the numeracy and literacy proficiency of recent graduates is slightly lower than that
of adults who graduated between 5 and 15 years ago (with the exception of literacy skills of VET
graduates). The decline is steeper for graduates who left education without an upper-secondary education
degree. Recent graduates from that group have lower literacy and numeracy skills than those who left
education at least five years ago. The decline is also slightly stronger for general education graduates than
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
271
for VET graduates, although differences are small. These results suggest that all graduates from a specific
education level/orientation who left the education system at most 15 years ago entered the labour market
with a broadly similar skillset, at least concerning general skills, with the exception of those without an
upper-secondary degree who now enter the labour market with lower skills than they did 5 to 15 years ago.
Therefore, these results cautiously suggest that the relative skills of graduates leaving the education
system in the past 15 years have remained roughly stable over time for VET, general and tertiary education
graduates, and improved relative to those without an upper-secondary degree.
5.2. Job quality and quantity
5.2.1. VET facilitates labour market entry of graduates
One of the most cited benefits of VET is that it helps graduates with their transition from school to
work. Brunello and Rocco (2017[8]), for example, find that upper-secondary and post-secondary
non-tertiary graduates from a vocational field have slightly lower wages, but better employment
outcomes than graduates from general fields, both through higher probabilities of employment and
larger shares of working life spent in paid employment.
Figure 5.4 (Panel A) shows that employment rates are indeed higher among young VET graduates
than among graduates from general programmes at similar levels (except in Estonia and the
United Kingdom, where VET graduates’ employment rate is marginally lower than that of general
education graduates) and for those without an upper-secondary degree. In several countries, VET
graduates’ employment rates are almost the same as those of tertiary education graduates. Similarly,
young VET graduates are less likely to be unemployed than graduates from general education (except
in Estonia, France, Greece, Japan, Portugal and the United Kingdom; Panel B). However, this
difference is small in many countries. VET graduates’ unemployment rates are higher than among
graduates from tertiary programmes (except in Denmark, Korea, Mexico and Turkey), and significantly
lower than for graduates without an upper-secondary degree. These findings suggest that it might be
easier for young VET graduates to find work after leaving education than for general education
graduates. They also support the view that VET programmes provide a valuable education pathway
to retain youth at risk of dropping out of school without a qualification – i.e. without an upper-secondary
degree – through more applied, often work-based, learning. Box 5.3 looks at school-to-work transitions
of different types of graduates, and confirms that VET graduates have an advantage compared to
general education graduates at the start of their career.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
272
Figure 5.4. Young VET graduates have relatively strong employment outcomes
Employment (Panel A) and unemployment (Panel B) rates of graduates (aged 15 to 34 not in formal education)
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
A. Employment rate
100%
90%
80%
70%
60%
50%
40%
30%
B. Unemployment rate
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
Note: VET: Vocational Education and Training. Panel A includes all individuals aged 15/16 to 34 who are not enrolled in formal education.
Panel B includes employed and unemployed individuals aged 15/16 to 34 not in education or training. Data refer to 2018 for all countries except
Australia, Canada (2019), Korea, Japan (2011/12), New Zealand, Israel, Turkey (2014/15), and the United States (2011/12-2014-17). OECD is
an unweighted average of the countries shown.
Source: European Union Labour Force Survey (2018), Turkish Labour Force Survey (2015), Australian Survey of Education and Work (2019),
Canadian Labour Force Survey (2019), OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18), Encuesta de Caracterización Socioeconómica
Nacional (2017).
StatLink 2 https://stat.link/irtg5k
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
273
Box 5.3. School-to-work transitions in Europe
As a proxy of the smoothness of school-to-work transitions, Figure 5.5 shows the average time between
obtaining one’s highest education degree and the start of the first significant work experience. These
data are available only for European countries. On average across countries, VET graduates have a
gap of 7.5 months between ending their studies and starting their first significant job. This is shorter than
for general education graduates (8.9 months). However, in a few countries, like Romania, France and
Finland, VET graduates have slower transitions than general education graduates.
On average, the school-to-work transition of VET and general education graduates is slower than for
tertiary education graduates, who spend 6.3 months between finishing their studies and starting their
first job. However, in some countries the gap is of similar duration for VET and tertiary education
graduates (e.g. Belgium, Denmark, Lithuania, Spain and the United Kingdom). In a few countries this
gap is even shorter for VET than tertiary education graduates (Estonia, Ireland and Latvia). In all
countries with available data, except Romania, VET graduates have shorter spells between graduating
and their first significant job than those who left education without an upper-secondary education degree,
and in many countries the differences are substantial. This is not the case for general graduates, whose
performance is close to (or even worse than) that of those without an upper-secondary education degree
in a number of countries.j
Figure 5.5. VET graduates spend less time than general education graduates between the end of
their studies and their first job
Number of months between end of education and start of first significant job
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General Upper-secondary & post-secondary non-tertiary - VET Tertiary
Months
20
18
16
14
12
10
8
6
4
2
0
Note: VET: Vocational Education and Training. Data were collected in 2009 and refer to individuals who obtained their qualification in the
period 2004-09 (aged between 15 and 34 at the time of data collection). Significant work experience is identified as a job for pay or profit (as
employee, self-employed or family worker) with a minimum duration of 3 months. Apprenticeships or unpaid traineeships, compulsory military
or community service, and summer jobs are excluded. Germany and Switzerland excluded because of different data collection regarding the
starting date of the first significant job. The average is calculated as the unweighted average of EU countries with available data for all four
education groups.
Source: European Union Labour Force Survey 2009 ad hoc module.
StatLink 2 https://stat.link/c64vmf
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
274
While school-to-work transitions might be smoother for VET than for general education graduates,
evidence suggests that these positive employment effects disappear for older age groups. Brunello and
Rocco (2017[8]), Forster, Bol and van de Werfhorst (2016[9]), Hanushek et al. (2017[10]), and Rozer and Bol
(2019[11]) indeed show that individuals with a VET qualification have higher employment rates than those
with a general qualification at the start of their career, but this pattern disappears later in life. This
age-employment profile is more pronounced in countries that have a larger work-based learning
component in their VET programmes, as the initial gains are relatively large (Hanushek et al., 2017[10]).
Rozer and Bol (2019[11]) find that this life-cycle pattern did not change over time in the Netherlands (in the
period 1996-2012). A declining labour market advantage for VET graduates is not found in all countries.
Silliman and Virtanen (2019[12]), for example, show that admission to the vocational track in Finland
significantly increases annual income compared to the general education track, and that these benefits do
not diminish with time. However, their analysis only follows individuals for 15 years after entry into VET.
As discussed by Rozer and Bol (2019[11]) less steep long-run returns to VET could be caused by several
mechanisms: i) VET preparing students for employment in manual and craft jobs that have limited potential
growth opportunities; ii) VET graduates mostly having job-specific skills rather than transferrable skills; and
iii) VET graduates participating less in on-the-job training, making them less flexible in light of structural or
technological changes. Hanushek et al. (2017[10]) indeed find that VET graduates participate less in
job-related training, and argue that this might lead to skills obsolescence which could be one of the reasons
for poorer employment outcomes later in life. In addition, they link the decreasing employment advantage
for VET graduates to poorer basic skills and hence lower adaptability. Brunello and Rocco (2017[8]) confirm
that VET graduates have lower basic skills than graduates from general fields at the same education level.
Moreover, Verhaest et al. (2018[13]) show that VET graduates at the start of their career are less likely to
be mismatched by qualification and have a lower degree of over-skilling compared to general education
graduates. VET programmes which combine a specific focus with workplace learning are found to be most
effective in avoiding most types of educational and skill mismatches during the first part of the career of
medium-skilled workers. However, the authors also find that this advantage of VET graduates declines
with time elapsed since graduation and therefore conclude that VET graduates are more employable when
they leave initial education because of the labour market focus of their qualification, but that their skills
gradually become obsolete because of structural and organisational changes in the labour market.
Figure 5.6 confirms that the employment (and unemployment) advantage for VET graduates with respect
to general education graduates is smaller for older age groups. 6 The gap in unemployment rates of
individuals with VET and general education qualifications gradually declines and disappears by age 35,
while the gap in employment rates disappears by age 45. Declining employment gaps with age are also
confirmed when comparing individuals with similar skill levels and other personal characteristics. 7
Individuals with VET degrees maintain their advantage relative to those without upper-secondary education
in all age groups. Individuals with tertiary education degrees have higher employment rates than all other
education groups at all ages, and lower unemployment rates. 8
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
275
Figure 5.6. The job quantity advantage for VET versus general education is smaller for older age
groups
Percentage point difference in (un)employment rates between individuals with qualifications from below
upper-secondary, general and tertiary education relative to individuals with VET qualifications
No controls Controls
A. Employment rate B. Unemployment rate
15 *** ****** 8
*** *** ***
10 *** *** 6 ***
*** ***
5 4 *** ***
*** ** *
0 2 *
-5 *** 0
***
-10 ****** *** -2
*** **
-15 ***
***
-4 ******
*** *** ** ***
*** *** *** *** ******
-20 -6
General Tertiary General Tertiary General Tertiary General Tertiary General Tertiary General Tertiary General Tertiary General Tertiary
Below upper-secondary Below upper-secondary Below upper-secondary Below upper-secondary Below upper-secondary Below upper-secondary Below upper-secondary Below upper-secondary
16-34 35-44 45-54 55-65 16-34 35-44 45-54 55-65
Age groups Age groups
Note: Marginal effects after probit regressions. All regressions include country fixed effects. Controls include gender, migrant status, number of
children (no, one, two or more), age (5-year categories), numeracy proficiency and literacy proficiency. The sample only includes individuals
who are not enrolled in formal education in all OECD countries except Colombia, Iceland, Latvia, Luxembourg and Portugal. VET and general
refers to ISCED levels 3 and 4. ***Significant at the 1% level, ** 5% level, *10% level.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/57oxia
Repeating this exercise by gender, shows that the advantage in terms of higher employment rates for
young VET graduates relative to general education graduates is the same for men and women on average
across countries (controlling for skill levels and other personal characteristics as in Figure 5.6). For the
unemployment rate, the gap between young VET and general education graduates is only found for men.
Both for men and women the advantage disappears for older age groups. For men, the gap in employment
rates disappears by age 45, while for women it already disappears by age 35. The gap in unemployment
rates for men disappears by age 35.
The evolution of employment rates in the period 2004-18 is comparable between young VET graduates
and those with a general degree at a similar level (see Figure 5.7), but the former experienced a somewhat
larger decline during the global financial crisis (2008-10). The crisis also had a stronger impact on
employment rates of VET graduates than tertiary education graduates. The decline in the period 2008-10
was the strongest for those without an upper-secondary education degree. Likewise, the trend in
unemployment rates is very similar for young VET and general education graduates, but the gap in
unemployment rates temporarily closed in 2010. The increase in unemployment rates in 2008/2010 was
also less pronounced for VET graduates than for those without an upper-secondary education degree, but
stronger than for young graduates with a tertiary education qualification. Overall, these results suggest that
the job quantity advantage of young VET graduates relative to general education graduates has remained
stable in recent years, although VET graduates have been somewhat more exposed to the global financial
crisis.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
276
Figure 5.7. The job quantity advantage of young VET graduates relative to general education
graduates has remained stable in recent years
Employment and unemployment rates of young graduates (aged 15 to 34 not in formal education)
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
A. Employment rate
100%
90%
80%
70%
60%
50%
40%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
B. Unemployment rate
35%
30%
25%
20%
15%
10%
5%
0%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Note: VET: Vocational Education and Training. The sample includes individuals aged 15 to 34 who are not enrolled in formal education.
Unweighted average based on a balanced panel of countries, including Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark,
Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Latvia, Lithuania, Netherlands, Poland, Slovak Republic, Spain, Sweden,
Switzerland and Turkey.
Source: European Union Labour Force Survey, Encuesta de Caracterización Socioeconómica Nacional, Household, Income and Labour
Dynamics in Australia Survey, Turkish Labour Force Survey, Canadian Labour Force Survey.
StatLink 2 https://stat.link/yhk7qf
5.2.2. Young VET graduates mainly work in crafts, sales and services jobs
The occupational composition of graduate employment (see Figure 5.8) shows that most young VET graduates
are employed in middle-skill occupations (mostly crafts and related trades jobs: 22% of employed VET
graduates) and low-skill occupations (typically services and sales jobs: 26%). Only 20% of young VET
graduates are employed in high-skill occupations. However, there are substantial country differences in the
occupational composition of VET employment. For example, the share of young VET graduates in high-skill
occupations amounts to more than one third in Germany, Switzerland and the United States, where VET
graduates often work as technicians and associate professionals. Gender differences in the occupational
composition of VET graduates’ employment are substantial. While crafts and related trades occupations employ
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
277
around one third of male VET graduates on average, these occupations only account for 4% of female VET
graduates’ employment. Sales and service jobs are the most important occupations for female VET graduates,
accounting for 44% of employment, compared to only 15% of male VET graduates’ employment. Employment
in high-skill occupations is slightly more common for female than for male VET graduates (22% versus 19%).
The occupational structure of graduate employment differs strongly between education groups, see
Figure 5.9. Young graduates from general education are mostly employed as service and sales
workers, as are VET graduates. However, in contrast to VET graduates, only a small share of gen eral
education graduates work in craft and related trades jobs. General education graduates are more
likely to work in high-skill jobs than VET graduates (6 percentage point gap in 2018). Young tertiary
education graduates are predominantly employed as professionals or technicians, while those without
an upper-secondary degree work mostly in elementary occupations, in service and sales jobs and in
craft and related trades occupations (see Annex Figure 5.A.2).
Figure 5.9 also shows the trends in the occupational composition for VET and general education graduates.9
It shows that young general and vocational education graduates are increasingly employed as service and
sales workers (ISCO 5), while a declining share of these graduates end up in clerical jobs (ISCO 4).
Elementary occupations are growing in importance for both education groups, albeit at a faster pace for
general education graduates. Interestingly, while the share of general education graduates employed in crafts
and related trades occupations (ISCO 7) has been on the decline – consistent with the overall trend in the
labour market – the share of young VET graduates employed in this occupation group remained relatively
stable. As shown in Annex Figure 5.A.2, young adults without an upper-secondary education degree mainly
saw an increase in employment shares in elementary occupations and service and sales occupations, and a
fall in agricultural and crafts and related trades occupations. For young tertiary education graduates,
professional and technician occupations grew in relative importance, while relative employment in
management and crafts and related trades occupations was on the decline.
Figure 5.8. A small share of VET graduates work in high-skill occupations
Percentage of employed VET graduates (aged 15 to 34)
Low-skill Middle-skill High-skill
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 15/16 to 34 who are not enrolled in formal education.
Countries are sorted by the share of VET graduates who work in high-skill occupations (ISCO-08 1-3). Middle-skill occupations are defined as
ISCO-08 4 plus ISCO-08 6-8; low-skill occupations as ISCO-08 5 plus ISCO-08 9. See Chapter 4 for details on the skill groups. Australian data are
mapped from ANZSCO to ISCO, Canadian data from NOC to ISCO. Data refer to 2018 for all countries, except Australia, Canada (2019), Korea,
Japan (2011/12), New Zealand, Israel, Turkey (2015), United States (2011/12-2014-17). OECD is an unweighted average of the countries shown.
Source: European Union Labour Force Survey (2018), Turkish Labour Force Survey (2015), Canadian Labour Force Survey (2019), Australian Survey
of Education and Work (2019), OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18), Encuesta de Caracterización Socioeconómica Nacional (2017).
StatLink 2 https://stat.link/bq7kcx
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
278
Figure 5.9. The occupational composition of young graduate employment is changing
Percentage of employed graduates (aged 15 to 34), 2004-18
Managers Professionals Technicians and associate professionals
Clerical support workers Service and sales workers Skilled agricultural, forestry and fishery workers
Craft and related trades workers Plant and machine operators, and assemblers Elementary occupations
A. Upper secondary & post-secondary non-tertiary - VET
30%
25%
20%
15%
10%
5%
0%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Year
B. Upper secondary & post-secondary non-tertiary - General
35%
30%
25%
20%
15%
10%
5%
0%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Year
Note: VET: Vocational Education and Training. The sample includes employed graduates aged 15 to 34 who are not enrolled in formal education.
Unweighted average based on a balanced panel of countries including, Australia, Austria, Canada, Chile, Denmark, Estonia, Finland, France,
Germany, Greece, Hungary, Ireland, Latvia, Lithuania, the Netherlands, Poland, Slovak Republic, Spain, Sweden, Switzerland, Turkey. Data
refer to 1-digit ISCO-08 occupations (except in Australia where ISCO-88 occupations are used). For European countries and Turkey, the
occupational classification used in the data changes in 2011 or 2012 (from ISCO-88 to ISCO-08), and the pre-2011/12 data are recoded to
minimise the series break. In the Chilean and Australian data, occupations are coded in ISCO-88 in all years. In the Chilean data, the ISCO-88
occupations are recoded into ISCO-08 using a one-to-one crosswalk, while in Australia occupations are not recoded and the ISCO-88
classification is used (because of data limitation). Canadian data are mapped from NOC to ISCO-08.
Source: European Union Labour Force Survey, Encuesta de Caracterización Socioeconómica Nacional, Household, Income and Labour
Dynamics in Australia Survey, Turkish Labour Force Survey, Canadian Labour Force Survey.
StatLink 2 https://stat.link/3ygq0k
Changes in the occupational composition of young graduates’ employment are the result of changes
in the overall occupational composition of the labour market, as well as changes of the educational
composition of young graduates in the entire labour market and within occupations. To disentangle
these components, Figure 5.10 decomposes the change in the share of young VET and general
education graduates by occupation (as shown in Figure 5.9) into the contribution of: i) the change in
overall graduate employment by occupation, i.e. the relative size of that occupation in the labour
market for 15-34 year-olds, ii) the change in the share of young VET and general education graduates
among all 15-34 year-old workers within each occupation, relative to the change in the share of VET
and general education graduates in the overall labour market for 15-34 year-olds.10
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
279
Figure 5.10. Craft jobs remain important for young VET graduates despite an overall declining
importance of those jobs in the labour market for young graduates
Percentage point change in employment shares of VET/general graduates (aged 15 to 34) between 2004 and 2018
Contribution of changes in the relative size of the occupation in the labour market for 15-to-34 year-olds Total change
Contribution of changes in the share of the education group among 15-to-34 year-old workers within the occupation relative to changes in the size of the education
group among employed 15-to-34 year olds
Percentage points A. VET
6
4
2
0
-2
-4
-6
Managers Professionals Technicians and Clerical support Service and sales Skilled agricultural, Craft and related Plant and machine Elementary
associate workers workers forestry and fishery trades workers operators, and occupations
professionals workers assemblers
Percentage points B. General
6
4
2
0
-2
-4
-6
Managers Professionals Technicians and Clerical support Service and sales Skilled agricultural, Craft and related Plant and machine Elementary
associate workers workers forestry and fishery trades workers operators, and occupations
professionals workers assemblers
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 15 to 34 who are not enrolled in formal
education. Average of changes in Australia, Austria, Canada, Chile, Denmark, Estonia, France, Finland, Greece, Hungary, Ireland, Latvia,
Lithuania, Netherlands, Poland, Slovak Republic, Spain, Sweden, Switzerland and Turkey. Decomposition based on 1-digit ISCO-08
occupations (except in Australia where ISCO-88 occupations are used). For European countries and Turkey, the occupational classification
used in the data changes in 2011 or 2012 (from ISCO-88 to ISCO-08), and the pre-2011/12 data are recoded to minimise the series break. In
the Chilean and Australian data, occupations are coded in ISCO-88 in all years. In the Chilean data, the ISCO-88 occupations are recoded into
ISCO-08 using a one-to-one crosswalk, while in Australia occupations are not recoded and the ISCO-88 classification is used (because of data
limitation). Canadian data are mapped from NOC to ISCO-08.The bars represent the relative contribution of the three different components to
the overall change, but cannot be interpreted as the size (in percentage points) of these components.
Source: European Union Labour Force Survey, Encuesta de Caracterización Socioeconómica Nacional, Household, Income and Labour
Dynamics in Australia Survey, Turkish Labour Force Survey, Canadian Labour Force Survey.
StatLink 2 https://stat.link/m27q08
The chart shows that the importance of craft and related trades jobs remained stable for young VET
graduates (i.e. “Total change”, represented by the diamond shape) because of two counterva iling
effects: i) an overall decline in the importance of this occupation in total employment of
15-34 year-olds (light blue), and ii) an increase in the share of VET graduates among the
15-34 year-olds working in this occupation that is larger than the overall decline of VET graduates in
total employment of young graduates (dark blue). Another way to interpret the results for these
occupations is that although total graduate employment in these occupations is decreasing ( light blue),
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
280
the share of young VET graduates who find employment in this type of occupation is stable over time
(diamond); therefore, the share of VET graduates within these occupations has increased despite an
overall decline of VET in total employment. Services and sales jobs, on the oth er hand, gained in
importance for young VET graduates, both because it is an occupation that is growing overall for
15-34 year-olds (light blue bar) and because the decline in the share of VET graduates within the
occupation is substantially smaller than the overall decline in the importance of VET in the labour
market (dark blue). Professionals are also growing occupations for young graduates, but because the
share of VET graduates within these occupations is on the decline (and this decline is stronger th an
the overall decline of VET in the labour market), these occupations have become less important for
young VET graduates in the period 2004-18.
Things look similar for general education graduates in most occupations. Unlike for VET graduates,
the importance of crafts and related trades jobs declined for general education graduates, as the
overall decline in the size of this occupation group was not made up for by a sufficiently large rise in
the share of general education graduates employed within this occ upation group. The importance of
professional occupations only declined modestly for general education graduates compared to VET
graduates, because the decline in the share of general education graduates within the occupation was
relatively small.
Overall, these changes in the occupational structure show that both young VET and general education
graduates have been impacted by structural changes in the labour market that reduce the relative
importance of middle-skilled jobs (that is: clerical support workers; craft and related trades; and plant and
machine operators – see Chapter 4 for a discussion of the partition of occupations into high-, middle- and
low-skill jobs). However, the impact of these structural changes has been smaller for VET than for general
education graduates, partially because young VET graduates managed to secure the remaining crafts and
related trades jobs. For both groups of graduates, employment growth was strongest in low-skill
occupations (sales, service and elementary jobs). The importance of high-skill occupations (managers,
professionals, and technicians and associate professionals) remained almost the same for young general
education graduates, but these jobs became less important for VET graduates. Looking at the occupation
mobility among young graduates, Box 5.4 shows that VET graduates who change occupations are less
likely than general education graduates to move into jobs with at a higher skill level or with a lower
probability of automation.
Box 5.4. Moving between occupations
On average across European countries included in the EU-SILC dataset, 17.5% of young VET
graduates change occupations from one year to the other.1 This is slightly lower than for general
education graduates (19.6%) and graduates without an upper-secondary education degree (19.3%),
and roughly the same as for tertiary graduates (17.8%). In the vast majority of countries, differences
between the education groups are small. Notable exceptions are Finland, the Netherlands and Norway,
where general education graduates have considerably higher probabilities of changing occupations.
Only in Portugal are VET graduates substantially more likely to change occupations than other young
graduates.
However, there are substantial differences in the types of changes that are made by the different
graduates. Table 5.1 compares young graduates from different education groups who changed
occupations from one year to the next, to see if they have different probabilities of moving into jobs with
higher skill levels (measured as average numeracy proficiency of workers in the occupation), higher
wages (measured as median wage in the occupation) or lower risk of automation (see below for a
discussion on the risk of automation).2 The results show that young general education graduates are
significantly more likely than VET graduates employed in similar occupations to move into occupations
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
281
that have a higher average numeracy skill level and/or lower risk of automation. Also compared to
tertiary education graduates employed in similar occupations, VET graduates have a lower probability
of making these positive moves (including moving to occupations with higher median wage levels), and
the differences are much larger than between VET and general education graduates. Graduates who
do not have an upper-secondary education degree are even less likely than VET graduates to make
these positive moves.
Table 5.1. Among occupation changers, VET graduates are less likely than general education
graduates to make positive occupation changes
Probit (marginal effects) and OLS regression estimates
Skill level Median wage level Risk of automation
Probability of % Probability of % Probability of % difference in
moving to difference moving to difference moving to average risk of
occupation with in average occupation in median occupation with automation
higher average skill level with higher wage level lower risk of
skill median wage automation
General (vs. VET) 0.050** 1.046*** 0.027 3.617*** 0.037* -2.914***
(0.020) (0.373) (0.018) (1.251) (0.019) (0.937)
Tertiary (vs. VET) 0.258*** 5.079*** 0.187*** 16.519*** 0.206*** -12.648***
(0.016) (0.301) (0.015) (0.952) (0.016) (0.727)
Below upper- -0.054*** -1.731*** -0.069*** -4.584*** -0.056*** 2.051***
secondary (vs. VET) (0.018) (0.335) (0.017) (0.890) (0.018) (0.726)
Number of obs. 21 514 21 514 21 107 21 107 21 419 21 419
Note: VET: Vocational Education and Training. The sample includes graduates aged 15 to 34 who are not enrolled in formal education (at
year t and t-1) and changed occupations from one year to the next. Individuals who have a different education level in year t compared to
year t-1 are excluded. The estimated specifications include controls for age (15-24, 25-29, 30-34), gender, previous and current employment
status (full-time employees, part-time employee, full-time self-employed, part-time self-employed, unemployed, inactive), previous
occupation (at year t-1) and country. For the probit regressions, the independent variable is equal to 0 when the move is negative and 1
when the move is positive, and the estimated coefficient needs to be multiplied by 100 to obtain the percentage point difference in
probabilities between the education groups. The sample includes all EU countries except Germany. Robust standard errors in parenthesis.
***Significant at the 1% level, ** 5% level, *10% level.
Source: European Union Statistics on Income and Living Conditions (2014-17).
1. Occupations are defined at the 2-digit ISCO level. This means that all occupation changes that happen within a two-digit occupation will
not be recoded as changes (e.g. from cook (ISCO 512) to waiter or bartender (ISCO 513)).
2. Skill, wage and automation levels are measured by occupation and country. For EU countries that did not participate in the OECD Survey
of Adult Skills, the average across participating EU countries is used for skills and automation.
5.2.3. Young VET and general education graduates have similar job quality
Earnings quality
Individual returns to education and skills, measured as the increasing earnings associated with
additional years of schooling and/or higher skills, are a well-researched topic (Willis, 1986 [14];
Heckman, Lochner and Todd, 2006 [15]; Peracchi, 2006 [16]; Pritchett, 2006 [17]; Deere and Vesovic,
2006[18]). Wages are an important component of job quality and are therefore a key incentive for
individuals to invest in education (Becker, 1993 [19]). Figure 5.11 shows median hourly wages of each
education group relative to tertiary education graduates. In all countries, VET and general education
graduates have lower hourly wages than tertiary education graduates. The wag es of VET graduates
are on average higher than those of general education graduates but there is considerable
heterogeneity across countries. In particular, in Canada 11, Denmark, Iceland, the Netherlands and
Norway, VET graduates have considerably higher wages than general education graduates, while in
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
282
Estonia, Luxembourg and Portugal, the opposite holds. In all countries for which data are available,
VET graduates earn more than those without an upper-secondary education degree, while this is not
systematically the case for general education graduates.
The wage difference between education groups persists when controlling for additional personal
characteristics, including skill levels, and workplace characteristics ( Figure 5.12).11,12 This exercise is
repeated for older age groups, to see if the wage differences remain the same over time or become
less (or more) pronounced. As with other age group analyses, it has to be noted that differences
between these groups do not only reflect how differences between graduates evolve over time, but
also how VET systems, the socio-demographic composition of graduates and educational attainment
levels have changed. The (small) wage advantage of VET graduates over general education
graduates among youth (aged 16 to 34) disappears entirely when looking at older age groups. Among
middle-age workers (aged 35 to 44), general education graduates actually earn more than VET
graduates, but this difference disappears when the occupation and industry are taken into account.
This implies that the wage difference between middle-aged VET and general education graduates can
to a large extent be explained by the fact that general education graduates tend to work in occupations
and industries with relatively higher wages.
Figure 5.11. Young VET graduates have slightly higher wages than general education graduates
Index: 1 = median hourly wage of tertiary graduates
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General Upper-secondary & post-secondary non-tertiary - VET
1
0.9
0.8
0.7
0.6
0.5
0.4
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 15/16 to 34 not in education or training. Hourly wages
are derived from annual, monthly or weekly earnings, by accounting for months spent in employment in the income reference year and information on
average hours worked. Data refer to 2014-18 for all countries, except Korea, Japan, Canada, Germany, Ireland (2011/12), New Zealand, Israel, Turkey
(2014/15), Australia, Chile (2017), United States (2011/12-2014-17). OECD is an unweighted average of the countries shown.
Source: European Union Statistics on Income and Living Conditions (2014-18), Turkish Labour Force Survey (2015), Encuesta de
Caracterización Socioeconómica Nacional (2017), Household, Income and Labour Dynamics in Australia Survey (2017), OECD Survey of Adult
Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/yd2usz
Repeating this analysis by gender shows that the wage advantage for young VET graduates relative
to general education graduates is the same for men and women. This advantage disappears by age 35
for both men and women. When comparing individuals employed in similar occupations and industries,
only young women with a VET degree are found to have significantly higher wages than their general
education counterparts. These results show that both men and women with a VET degree have an
advantage at the start of their career in terms of wage levels relative to graduates from general
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
283
education. While this effect for young men is entirely due to VET graduates working in higher paying
industries and occupations than general education graduates, for women it is due to a combination of
selection into higher paying industries and occupations and higher pay than graduates from general
education within the same industries and occupations.
Figure 5.12. The wage advantage for VET vs. general education disappears for older age groups
Percentage difference in hourly wages between individuals with qualifications from below upper-secondary, general
and tertiary education relative to individuals with VET qualifications
Basic controls only Within occupations/industries
35%
30% ***
*** ***
25% ***
20% *** ***
***
15%
***
10%
**
5%
0%
-5%
*** *** *** *** ***
-10% *** *** ***
*** ***
-15%
Below General Tertiary Below General Tertiary Below General Tertiary Below General Tertiary
upper-secondary upper-secondary upper-secondary upper-secondary
16-34 35-44 45-54 55-65
Age groups
Note: Basic controls are gender, literacy proficiency (5 categories), age (5-year categories), number of children, migrant status, job tenure with
current employer (4 categories), firm size (5 categories), part-time working hours, contract type and country. Occupation controls are 1-digit
ISCO, industry controls are 1-digit ISIC. The within occupation/industry regression also include the basic controls. Hourly wages include bonuses
and are trimmed at the bottom and top 1% per country. The sample only includes individuals who are not enrolled in formal education. It includes
all OECD countries except Colombia, Iceland, Latvia, Luxembourg and Portugal. *** Statistically significant at the 1% level, ** 5% level, * 10%
level.
Source: OECD Survey of Adult Skills data (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/6mcy42
Job security
As discussed above, young VET graduates have relatively low unemployment rates, contributing to high
job security. In addition, when they are employed, young VET graduates are less likely to have a temporary
contract than general education graduates or those without an upper-secondary degree (17% versus 22%
and 26%, respectively, see Figure 5.13) but equally likely as tertiary graduates. This contributes further to
their job security, as workers on temporary contracts enjoy lower job protection (see Chapter 3) and are
forced to change jobs more frequently when their contract is not renewed (OECD, 2014[20]). The only
exception is Portugal, where VET graduates are the most likely group to have a temporary contract. In
contrast, in several countries the prevalence of temporary contracts is lower among VET graduates than
among graduates from all other educational levels/types. These cross-country differences might be
explained by differences in the way VET systems are organised, as well as differences in labour market
institutions (which might affect graduates from different education groups differently, depending on their
occupational or sectoral employment composition).
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
284
Figure 5.13. Young VET graduates are less likely to be employed on a temporary contract than
general education graduates
Percentage of employed graduates with a temporary contract (aged 15 to 34)
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General Upper-secondary & post-secondary non-tertiary - VET Tertiary
60%
50%
40%
30%
20%
10%
0%
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 15/16 to 34 not enrolled in formal education.
Data refer to 2018 for all countries, except Australia (2017), Canada (2019), Korea, Japan (2011/12), New Zealand, Israel, Turkey (2014/15)
and the United States (2011/12-2014-17). Temporary employment in Australia includes casual workers. OECD is an unweighted average of the
countries shown.
Source: European Union Labour Force Survey (2018), Turkish Labour Force Survey (2015), Canadian Labour Force Survey (2019 Encuesta
de Caracterización Socioeconómica Nacional (2017), Household, Income and Labour Dynamics in Australia Survey (2017), OECD Survey of
Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/9swz0f
In general, the most common reason for having a temporary contract is that the person could not find a
permanent job (i.e. involuntary temporary work). Although VET graduates are less likely to have a
temporary contract, in countries with available data 13, VET graduates who do have a temporary contract
are more likely to be involuntary temporary workers than graduates from general and tertiary education
(63%, compared to 52% and 60%, respectively). Only in Italy, Portugal and Turkey are VET graduates
who are employed on a temporary contract less likely than general education graduates to be an
involuntary temporary worker.
Temporary contracts have become increasingly common among young graduates from all education
groups in the period 2004-18, see Figure 5.14. The increase was relatively small for VET graduates and
graduates from tertiary education (+1.5 percentage points), but was more widespread for general
education graduates (+2.9 percentage points) and especially for those who left education without an
upper-secondary education degree (+5.7 percentage points). As such, the advantage of VET relative to
general education graduates in access to permanent employment has increased over time.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
285
Figure 5.14. The incidence of temporary employment is on the rise for all education groups
Percentage of employed graduates with a temporary contract (aged 15 to 34)
Below upper-secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
35%
30%
25%
20%
15%
10%
5%
0%
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 15 to 34 not enrolled in formal education.
Unweighted average based on a balanced panel, including Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Estonia,
Finland, France, Germany, Greece, Hungary, Ireland, Latvia, Lithuania, Netherlands, Poland, Slovak Republic, Spain, Sweden, Switzerland and
Turkey. Temporary employment in Australia includes casual workers.
Source: European Union Labour Force Survey, Encuesta de Caracterización Socioeconómica Nacional, Household, Income and Labour
Dynamics in Australia Survey, Turkish Labour Force Survey, Canadian Labour Force Survey.
StatLink 2 https://stat.link/i5lp0h
One may expect that the probability of having a permanent contract increases as graduates accumulate
more work experience and firms have completed their screening of recent hires – see e.g. Booth,
Francesconi and Frank (2002[21]) and Faccini (2013[22]). For all types of graduates, and controlling for
personal characteristics (including literacy and numeracy skills), workplace characteristics and occupation
and industry, the probability of having a temporary contract is indeed lower among those who graduated
longer ago (see Figure 5.15).14 This declining age-probability profile is particularly steep for general
education graduates. While VET graduates have a significantly lower probability of being employed on a
temporary contract at the beginning of their working life compared to general education graduates, this
gap disappears among those with at least five years of work experience. Moreover, while VET and tertiary
education graduates are equally likely to have a temporary contract at the start of their career, tertiary
education graduates are less likely to have this type of contract later in their career. Individuals without an
upper-secondary degree are more likely to have a temporary contract than VET graduates, irrespective of
the number of years since they left education.
Looking at this separately for men and women shows some interesting differences. Male VET graduates
only have a significantly lower probability of being employed on a temporary contract relative to general
education graduates in the first five years after graduation. In later years these probabilities are the same
for male VET and general education graduates. Irrespective of the number of years since graduation, male
VET graduates are more likely to be employed on a temporary contract than male tertiary education
graduates. By contrast, female VET graduates are less likely than general education graduates to have a
temporary contract in the first ten years after graduation. Moreover, female VET graduates are also less
likely than tertiary education graduates to have a temporary contract in the first five years after graduation
(but this difference is much smaller than between VET and general). For female VET graduates who
graduated between 10 and 15 years ago, there is no statistically significant difference in the probability of
temporary employment relative to general education graduates, but also relative to tertiary education
graduates.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
286
Figure 5.15. Young VET graduates are less likely to have a temporary contract than general
education graduates only at the start of their career
Employed graduates’ (aged 16 to 34) probability to have a temporary contract
Below upper secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
50%
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
0-4 5-9 10-14
Years since graduation
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 16 to 34 who are not enrolled in formal
education. The figure shows marginal effects with 95% confidence intervals after a probit regression analysis, with a temporary/permanent
dummy as the dependent variable. Independent variables are type of education, years since graduation (and an interaction of those two), literacy
and numeracy proficiency, gender, migrant status, number of children, firm size, country fixed effects, occupation fixed effects (1-digit ISCO)
and industry fixed effect (1-digit ISIC). The sample includes all OECD countries except Colombia, Iceland, Latvia, Luxembourg and Portugal.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18)
StatLink 2 https://stat.link/tkpm6y
Quality of the work environment
The nature and content of the work performed, working-time arrangements and workplace relationships,
are equally important dimensions of job quality. Working more than 50 hours per week is an important
indicator of job strain (OECD, 2014[23]). On average, around 8% of individuals aged 15 to 34 indicate that
their usual workweek exceeds 50 hours. This percentage is similar across education groups. In the past
15 years, the share of young graduates whose usual workweek exceeds 50 hours has decreased, and the
decrease happened at a similar pace in all education groups. Another aspect of job strain is the physical
burden of the job, and as Box 5.5 describes, this is higher among VET graduates than among general
education graduates.
Having career progression opportunities in your job, such as the option of being promoted to a job with
more supervisory responsibilities, is a key driver of job motivation, and therefore job quality. In most
countries, young VET graduates are equally likely as graduates from general education to have
supervisory responsibilities in their job. Only in Australia, New Zealand and Norway are VET graduates
substantially more likely to have supervisory responsibilities, while the opposite holds in Korea. On
average, 19% of VET and 18% of general education graduates have supervisory responsibilities,
compared to 27% of tertiary graduates and 12% of those without an upper-secondary degree. The share
of VET graduates with supervisory responsibilities has remained relatively stable over time (2004-17), in
line with the trend observed among general education graduates.
One could expect that the probability of having supervisory responsibilities increases with age and
experience (i.e. years since graduation) as people progress in their career. Figure 5.16 shows that for or
all types of graduates, and controlling for personal characteristics (including literacy and numeracy skills),
workplace characteristics and occupation and industry, the probability of having supervisory
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
287
responsibilities is indeed higher among those who graduated longer ago. 15 However, for VET graduates
the probability to supervise others increases at a much slower rate than for graduates from general
education. Recent VET graduates (less than five years after graduation) have a higher probability of having
supervisory responsibilities in their job than recent graduates from general education, but this advantage
rapidly disappears with time: VET graduates who obtained their degree at least five years ago, have the
same probability of carrying out supervisory tasks as general education graduates. This pattern suggests
that VET graduates enter the labour market with an advantage over general graduates (potentially because
they have stronger job-specific skills and/or acquired work experience during education), but have fewer
opportunities for upward mobility over time. The difference in the probability of having supervisory
responsibilities between VET graduates and tertiary education graduates and those without an
upper-secondary degree remains substantial and statistically significant over time. Repeating this analysis
by gender shows that the advantage in terms of supervisory responsibilities for VET graduates relative to
general education graduates at the start of their career only exists for men.
Figure 5.16. Young VET graduates are more likely than general education graduates to have
supervisory responsibilities at the start of their career, but this advantage disappears later on
Employed graduates’ (aged 16 to 34) probability of having supervisory responsibilities on the job
Below upper secondary Upper-secondary & post-secondary non-tertiary - General
Upper-secondary & post-secondary non-tertiary - VET Tertiary
45%
40%
35%
30%
25%
20%
15%
10%
5%
0%
0-4 5-9 10-14
Years since graduation
Note: VET: Vocational Education and Training. The sample includes employed individuals aged 16 to 34 who are not enrolled in formal
education. The figure shows marginal effects with 95% confidence intervals after a weighted probit regression, with an indicator variable with
value 1 if the respondent supervises others in their current job as the dependent variable. Independent variables include type of education, years
since graduation (and an interaction of those two), literacy and numeracy proficiency, gender, migrant status, number of children, firm size,
country fixed effects, occupation fixed effects (1-digit ISCO) and industry fixed effect (1-digit ISIC). The sample includes all OECD countries
except Colombia, Iceland, Latvia, Luxembourg and Portugal.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/acg3sf
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
288
Box 5.5. Physically demanding jobs
One aspect of job strain is the extent to which workers are exposed to physical health risk factors, such
as tiring or painful positions and carrying or moving heavy loads (OECD, 2014[23]). 59.6% of young VET
graduates report working physically for long periods every day, which is higher than for general
graduates (52.8%) and especially tertiary graduates (22.1%) – see Figure 5.17. However, it is
substantially lower than among graduates without an upper-secondary degree (66.6%). Over 70% of
VET graduates work physically for long periods in Lithuania, Poland, the Slovak Republic and Turkey,
while only around 45% of VET graduates do so in Italy, Japan, Korea and Mexico.
When controlling for skill levels and personal characteristics (gender, age, migrant status, number of
children), as well as for firm size, industry and occupation 1, the differences in the probability of working
physically for long periods every day between education groups remain statistically significant. The
difference between VET and general education graduates remains roughly the same when controlling
for these characteristics, whereas the difference between VET and tertiary education graduates, as well
as those without an upper-secondary education degree, increases slightly.
Figure 5.17. Young VET graduates work physically for long periods more often than general
education graduates
Percentage of employed graduates (aged 16 to 34)
Never Less than once a month At least once a month but less than once a week At least once a week but not every day Every day
100%
90% 22.1
80%
52.8 8.9
70% 59.6
66.6 5.6
60%
12.9
50%
40% 11.6
5.3 11.4
30% 10.4 6.7 4.4 50.5
20% 4.2 6.5
5.0
10% 23.6
13.8 18.1
0%
Below upper-secondary Upper-secondary & post-secondary Upper-secondary & post-secondary Tertiary
non-tertiary - General non-tertiary - VET
Note: VET: Vocational Education and Training. Unweighted average of OECD countries (except Colombia, Iceland, Luxembourg, Latvia and
Portugal). The sample includes employed individuals aged 16 to 34 who are not enrolled in formal education.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/wzjknd
1. This is estimated using a probit regression with as dependent variable a dummy that equals one if the worker says to work physically for
long every day. Explanatory variables include educational attainment (4 categories), gender (dummy), age (5-year categories), migrant status
(dummy), number of children (3 categories), literacy and numeracy proficiency, firm size (5 categories), industry (1-digit ISIC), occupation
(1-digit ISIC) and country fixed effects.
Exposure to high performance work practices (HPWP), which include both aspects of work organisation –
team work, autonomy, task discretion, mentoring, job rotation, applying new learning – and management
practices – employee participation, incentive pay, training practices and flexibility in working hours – is
another aspect of job quality. Higher exposure to HPWP has been linked to higher wages, higher job
satisfaction, lower job-related stress, and higher labour productivity (OECD, 2016[24]). On average, young
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
289
VET graduates are slightly more likely than general education graduates to be employed in jobs with high
HPWP16. However, their exposure to HPWP is lower than among tertiary education graduates (with the
exception of graduates in Australia and Denmark), and higher than among graduates without an
upper-secondary degree (except in Belgium, Czech Republic, Greece, Ireland and Poland). Tertiary
education graduates are especially more likely to organise their own time, plan their own activities, teach
other people, participate in training, and have flexible working hours. By contrast, VET and general
education graduates more frequently cooperate with others in their job than tertiary education graduates.
Differences between general and vocational education graduates are relatively small for all aspects of
HPWP, with the exception of performance pay which is more common among VET graduates.
When comparing graduates employed in similar occupations and industries and with similar personal, job
and firm characteristics, the differences in exposure to HPWP between the education groups persist
although they become smaller (especially the gap between tertiary education and VET graduates).17
However, repeating this analysis by years since graduation shows that the slightly higher exposure to
HPWP among VET graduates than among general education graduates is mainly driven by differences at
the start of their career, as no differences are found for those who graduated between five and 15 years
ago. Moreover, the gap between VET and tertiary education graduates increases with years since
graduation. HPWP has been linked to better skill use, and Box 5.6 describes differences in skill use
between graduates. Consistent with the findings on exposure to HPWP, skill use is similar for general and
VET graduates, but it is substantially lower than among tertiary education graduates (even when having
similar skill levels and employed in similar jobs).
Box 5.6. The use of literacy, numeracy and problem-solving skills at work
Graduates might not only differ in the level of their skills, but also in the intensity at which they use those
skills. OECD (2016[24]) shows that countries rank differently in terms of skills proficiency and use,
suggesting that these are indeed two different, albeit related, concepts. Higher skills use has been found
to contribute to higher individual earnings and job satisfaction, as well as, at the aggregate level,
productivity growth (OECD, 2016[24]). The use of HPWP is important to foster better skill use.
Young VET graduates use their literacy, numeracy and digital problem-solving skills at work to the same
extent as general education graduates with the equivalent numeracy, literacy and problem-solving
proficiency and similar personal characteristics (Figure 5.18).1 Only numeracy skill use is slightly higher
among young general education graduates than among VET graduates. Skill use is also similar for young
VET and general education graduates employed in the same occupations and industries (except for
literacy use which is slightly lower for general education graduates). Compared to tertiary education
graduates with similar skill levels, young VET graduates use their skills significantly less intensively, even
when employed in the same occupation and industry. By contrast, VET graduates use their skills more
intensively than graduates without an upper-secondary education degree who have similar skill levels (also
when working in similar occupations and industries).
Results look similar for older age groups, with individuals with VET and general education degrees working
in similar occupations and industries having the same skill use. The only exception is numeracy use among
35-54 year-olds, which is higher for individuals with general education than for those with VET
qualifications. Nonetheless, general education graduates aged 35 and more work more frequently in
industries and occupations where literacy, numeracy and digital problem-solving skills are used more
intensively (which explains why differences are significant when not controlling for industry and
occupation). Differences in skill use between individuals with VET and tertiary education degrees employed
in similar occupations and industries remain substantial in all age groups, as do differences between
individuals with VET degrees and those without an upper-secondary education qualification.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
290
Figure 5.18. Young VET graduates use their literacy, numeracy and problem-solving skills less
intensively at work than tertiary education graduates with similar skill levels
Percentage difference in skill use intensity between individuals with qualifications from below upper-secondary,
general and tertiary education relative to individuals with VET qualifications
Basic controls only Within occupations/industries
A. Literacy B. Numeracy C. Digital problem-solving
40 *** 40 40
*** *** ***
30 30 30 *** *** ***
*** *** *** *** ***
20 *** *** 20 *** ****** *** *** 20 *** ***
*** ** **** *** *** *** *** *** *** **
10 10 * * 10
0 0 0
-10 * -10 -10 *
**
-20 *** ****** ****** ****** ***
****** ****** **
-20 *** *** *** *** -20 ***
-30 *** *** *** *** -30 -30
General General General General
Below upper-secondary Below upper-secondary
Tertiary Tertiary Tertiary Tertiary
Below upper-secondary Below upper-secondary Below upper-secondary Below upper-secondary
General General
Tertiary Tertiary
Below upper-secondary Below upper-secondary
General General
Tertiary Tertiary
Below upper-secondary Below upper-secondary
General General
Tertiary Tertiary
Below upper-secondary Below upper-secondary
General General
Tertiary Tertiary
16-34 35-44 45-54 55-65 16-34 35-44 45-54 55-65 16-34 35-44 45-54 55-65
Age groups Age groups Age groups
Note: The sample includes all employed individuals who are not in formal education. All regressions include country fixed effects and controls
for proficiency (literacy, numeracy and problem-solving, respectively), age, gender, migrant status, number of children and firm size. Occupations
are included at the 2-digit level and industries at the 1-digit level (in addition to the basic controls). The skill use variables are a combination of
more detailed tasks (which are measured on a 1-5 scale), see OECD (2016[24]). The sample includes all OECD countries except Colombia,
Iceland, Luxembourg, Latvia and Portugal.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18) and OECD (2016[24]), OECD Employment Outlook 2016,
https://dx.doi.org/10.1787/empl_outlook-2016-en.
StatLink 2 https://stat.link/ovk7m1
1. The results in Figure 5.18 are based on an OLS regression of log skill use on educational attainment (4 categories), skill proficiency, age
(5-year categories), gender (dummy), migrant status (dummy), number of children (3 categories) and country dummies.
5.3. The labour market outlook for VET graduates
5.3.1. VET and general education jobs are likely to face excess supply
The ease of finding a job that matches one’s skills depends on the demand for and supply of those skills.
When the demand for a certain skill is higher than its supply (i.e. a situation of skills shortages), firms have
difficulties finding the right workers for their vacancies. Individuals with those skills will find it easy to get a
job matching their skills. In the opposite case, when the demand for a skill is lower than the supply (i.e. skills
surplus), there is an abundance of workers with these skills, which makes it more difficult for individuals
with these skills to find jobs that match their skill profile. The type of skills in shortage or surplus, and the
intensity of these imbalances, will therefore be an important factor to understand the short-term
employment outlook for adults with different skill profiles.
Using information from the OECD Skills for Jobs database18, Figure 5.19 compares the shortage or surplus
intensity of an occupation to the share of workers (aged 15 to 34) in that occupation with a certain education
level. Current shortages and surpluses in occupations reflect relative changes in employment, hours
worked, hourly wages and under-qualification, as well as relative unemployment rates. The negative
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
291
relationship in Panels A, B and C suggest that occupations that employ mostly low-educated (Panel A)
and/or middle-educated (Panels B and C) workers are more likely to experience (relatively large)
surpluses. This negative relationship is stronger for young graduates from general education than for VET
graduates. Moreover, the two occupations that have more than 60% of workers with VET degrees among
their workers aged 15 to 34 (i.e. metal and machinery workers and electrical and electronics trades
workers), do not align with the overall pattern, as they experience no imbalance or a substantial shortage,
respectively, on average across OECD countries. The picture looks very different for those with tertiary
education degrees (Panel D), as occupations that mostly employ this type of graduates are experiencing
substantial shortages. Hence, while the occupations that predominantly employ tertiary education
graduates are facing excess demand, occupations that mostly employ lower-educated graduates face
excess supply, albeit to a lesser extent for VET graduates.
Figure 5.19. Occupations employing mostly below upper-secondary, general education or VET
graduates are more likely to face excess supply
Shortage(+)/surplus(-) intensity and graduate (15 to 34 years old) employment by occupation
A. Below upper secondary B. Upper sec. & post-secondary non-tertiary - General
Imbalance Imbalance
0.5 0.5
y = -1.2268x + 0.1525 y = -2.0403x + 0.2715
0.4 R² = 0.5764 0.4 R² = 0.4567
0.3 0.3
0.2 0.2
0.1 0.1
0 0
-0.1 -0.1
-0.2 -0.2
-0.3 -0.3
-0.4 -0.4
0% 5% 10% 15% 20% 25% 30% 35% 40% 0% 5% 10% 15% 20% 25% 30%
Share below upper-secondary by occupation Share general by occupation
C. Upper sec. & post-secondary non-tertiary - VET D. Tertiary
Imbalance Imbalance
0.5 0.5
y = -0.5965x + 0.1859 y = 0.4867x - 0.2035
0.4 R² = 0.277 0.4 R² = 0.5696
0.3 0.3
0.2 0.2
0.1 0.1
0 0
-0.1 -0.1
-0.2 -0.2
-0.3 -0.3
-0.4 -0.4
0% 10% 20% 30% 40% 50% 60% 70% 0% 20% 40% 60% 80% 100%
Share VET by occupation Share tertiary by occupation
Note: VET: Vocational Education and Training. Unweighted average of OECD countries with available data (Colombia, Iceland, Israel, Japan,
and Luxembourg not covered). Each dot represents a 2-digit ISCO-08 occupation. The occupational shortage index (on the y-axis) aggregates
information on employment growth, wage growth, hours worked growth, change in under-qualification and unemployment rates, and ranges
between -2.5 and 2.5 (with negative values referring to surpluses and positive values to shortages). Values refer to the latest available year. For
each occupation and education group, the employment share is the ratio of employment in that occupation from that education group to total
employment of that occupation. Employment shares refer to employed individuals aged 15/16 to 34 who are not enrolled in formal education.
Source: European Union Labour Force Survey (2017), Encuesta de Caracterización Socioeconómica Nacional (2017), Turkish Labour Force
Survey (2015) and OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18) data and the OECD Skills for Jobs database.
StatLink 2 https://stat.link/ewvhrl
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
292
Medium-term projection exercises, such as the European one described in Box 5.7, but also the
occupational projections from Canada and the United States19, suggest that employment growth in
some of the common occupations for VET graduates will be modest or even negative in the coming
decade(s). This is especially the case for craft and related trades occupations, which have already
seen declining employment relative to other occupations in recent years. While VET graduates seem
to have managed to secure the remaining craft jobs (as discussed above), these projections suggest
that fewer and fewer jobs will be available in those occupations. Nonetheless, negative or modest
employment growth does not mean that no job opportunities will be available in those occupations.
Job openings will continue to be created, mostly due to substantial replacement demand that exceed
the number of new or lost jobs. Hence, while many graduates specialised in those declining
occupations will still be able to find a suitable job, there is a risk of over-supply if VET systems do not
adapt. Employment levels are projected to continue to grow in high -skill occupations (professionals
and technicians), but also in sales and service occupations. If VET systems adapt to prepare students
for high-skill jobs, such as technicians and associate professionals, VET graduates can benefit from
the growing job opportunities in those types of jobs. Higher VET and smooth pathways for VET
graduates to tertiary education are crucial in this respect (see Section 5.4).
Box 5.7. Medium-term employment projections in Europe
According to projections from Cedefop and Eurofound (2018[25]), the job polarisation trend observed in
recent decades in EU countries is likely to continue in the next decades (2016-30). These forecasts
show that across EU countries, total employment will grow the most for professionals, as well as
technicians and associate professionals (new/lost jobs in the below chart). The sectors that are projected
to have the strongest employment growth are R&D, consulting services, computer programming and
advanced manufacturing. Managers, service and sales workers, and elementary occupations will also
experience growth, but at a slower pace. Employment levels will go down for craft and related trades
workers, clerks and skilled agricultural workers.
However, the demand for graduates does not only come from the creation of new jobs, but also from
replacement demand, i.e. job openings arising from a worker leaving a job, temporarily or permanently.
A very large part of the demand for skills in Europe is replacement demand (Cedefop and Eurofound,
2018[25]). Shrinking employment does not mean that an occupation disappears altogether, and openings
will still be created to replace workers who leave those occupations. Replacement demand reflects to a
large extent the size of the occupation, and as occupations shrink their replacement demand declines.
The patterns look largely the same across countries (for broad occupation groups), although there are
some notable exceptions. In Ireland, for example, a substantial number of new jobs are projected in
crafts and related trades occupations.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
293
Figure 5.20. Job creation in middle-skill occupations is projected to be negative in the EU in the
period 2016-30
Projected number of job openings, by occupation and type (thousands)
New/lost jobs Replacement demand Job openings
35 000
30 000
25 000
20 000
15 000
10 000
5 000
0
-5 000
Legislators, Professionals Technicians and Clerks Service workers Skilled Craft and related Plant and Elementary
senior officials associate and shop and agricultural and trades workers machine occupations
and managers professionals market sales fishery workers operators and
workers assemblers
Source Cedefop and Eurofound (2018[25]), “Skills forecast: trends and challenges to 2030”, Cedefop reference series, No. 108, Cedefop,
http://dx.doi.org/10.2801/4492.
StatLink 2 https://stat.link/ca0bfd
5.3.2. Many VET jobs could be affected by automation
One of the key drivers of changes in the occupational structure of the labour market is technological
progress. Technology has led to automation of certain tasks in the workplace, and this process is expected
to continue to contribute to further changes in the occupational composition of the labour market but also
in the task composition within occupations – see e.g. OECD (2019[26]). As discussed in Chapter 4, the
occupational composition of employment has indeed changed in recent decades. Moreover, Box 5.8
shows that the task content of jobs has also changed in recent years, with communication- and ICT-related
skills gaining in importance in the period 2012-17 (in the United States). According to estimates from
Nedelkoska and Quintini (2018[27]), some of the occupations that employ many VET graduates, like craft
and related trades occupations, have a relatively high probability of significant changes because of
automation. On the other hand, some other typical VET occupations, like certain sales and personal care
and service jobs, have a much lower risk of change due to automation.
Across OECD countries, 21.3% of jobs held by young VET graduates are highly automatable, meaning
that a very large share of the tasks in those jobs could potentially be automated. This is slightly lower than
for general education graduates (22.4%), but much higher than for jobs held by tertiary education
graduates (9%) – see Figure 5.22. Graduates without an upper-secondary degree face the highest risk of
automation, with 28% of them working in jobs at high risk. In the majority of countries, this risk of automation
is similar for general education and VET graduates. In Denmark, Ireland and New Zealand, VET graduates
have a substantially lower risk than general education graduates, while the opposite holds in France, Israel,
Lithuania, the Slovak Republic and Sweden. In all countries, the share of workers at high risk of automation
is higher among VET graduates than among tertiary education graduates. Differences between VET and
tertiary education graduates are smallest in Denmark, Korea, Mexico and the United States, and largest in
Belgium (Flanders), Lithuania and the Slovak Republic. In all countries except Japan and Lithuania, VET
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
294
graduates are less likely to be at high risk than graduates without an upper-secondary degree, although in
some countries the difference is only very small (e.g. Belgium, Canada and Turkey). Gender differences
in the risk of automation are small for all education groups.
Differences in the risk of automation between VET and other graduates can be explained by the fact that
these graduates work in different occupations (which have a different risk of automation), but also by the fact
that they carry out more or less automatable tasks in the same occupations. Using a standard shift-share
decomposition, the importance of these two components in explaining the difference in the average risk of
automation between VET and other graduates can be disentangled.20 On average across OECD countries,
the risk of automation of general education graduates is only marginally different from that of VET graduates,
and can fully be explained by occupational composition differences. Young VET graduates have a higher
average risk of automation than those with a tertiary education degree, and 68% of this difference can be
explained by the fact that VET graduates work in more automatable occupations (i.e. the “between
occupation” component). The remaining 32% is due to VET graduates carrying out more automatable tasks
even when they are working in the same occupation as tertiary education graduates (i.e. “within occupation”
component). The importance of the within component is larger for the difference between VET graduates and
those without an upper-secondary education degree: 46% of the lower average risk of automation of VET
graduates relative to those without an upper-secondary education degree is due to them carrying out less
automatable tasks in the same occupations.
Box 5.8. The changing task content of jobs in the United States
The task-approach to labour markets argues that changes in the allocation of workplace “tasks” between
capital and labour, and between domestic and foreign workers, has altered the structure of labour
demand, contributing to labour market polarisation (Autor, 2013[28]) (see Chapter 4). Moreover, the
research on the risk of automation takes tasks as a starting point to estimate automation probabilities,
based on experts’ views on their automation probability (Nedelkoska and Quintini, 2018[27]).
Using the United States’ data from the OECD Survey of Adult Skills (PIAAC), it is possible to analyse
how the task content of jobs changed in the period 2012-17. PIAAC collects information on the frequency
of carrying out a range of tasks at work, including literacy, numeracy and ICT-related tasks, but also
several more specific tasks, such as selling, problem-solving and communicating. The new cycle of
PIAAC data, which will be collected in 2021-22, will allow repeating this analysis for other OECD
countries and over a longer time period.
The figure below shows the results of regressing the probability of carrying out certain tasks at least once a
week on a time dummy (0 if 2012/2014, 1 if 2017) (with and without occupation controls).1 The reported
coefficients shows the change in the probability of frequently carrying out the respective tasks in 2017
relative to 2012/2014. The results show that for some tasks this probability changed significantly. This is
particularly the case for ICT-related tasks, which increased considerably (with the exception of the use of
email, word processing and programming). Other tasks that are more frequently carried out are
communication and – related to this- writing and reading letters, memos and emails. On the other hand,
workers carry out tasks like reading articles in newspapers and professional journals, and calculating
fractions less frequently. Moreover, workers are less often involved in tasks that require the use of hands or
fingers, as well as negotiation-related tasks, which are both considered difficult-to-automate tasks
(i.e. bottlenecks to automation). In spite of these changes in the frequency of carrying out certain
difficult-to-automate tasks, the average risk of automation did not change significantly in the analysed period.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
295
Figure 5.21. The use of certain ICT tasks increased significantly in the period 2012-17 in the
United States
Percentage point changes in the probability of frequently carrying out a certain task at work (marginal effects)
Occupation controls No controls
A. Bottleneck
2
-3
-8
Dexterity PS simple PS complex Teaching Planning Influencing Negociating Selling Advising Communicating
B. ICT
7
2
-3
-8
Mail Internet for Conduct Spreadsheet Word Programming Real time
work-related info transactions language discussions
C. Literacy
6
4
2
0
-2
-4
-6
-8
Read Read letters, Read Read prof. Read books Read Read financial Read Write Write Write Fill in
directions memos and newspapers Journals manuals or statements diagrams, letters, articles reports forms
or instructions mails or magazines or reference maps or memos
publications materials schematics and mails
D. Numeracy
0
-2
-4
-6
-8
Calculating costs or budgets Use or calculate fractions or Use a calculator Prepare charts, graphs or Use simple algebra or Use advanced mathematics
percentages tables formulas or statistics
Note: Marginal effects after a probit regression of an indicator variable of frequency of carrying out a certain tasks (1 if carried out at least
once a week) on a time dummy (1 if year is 2017) and occupation dummies (ISCO-08 2-digits). Lighter blue bars indicate statistical
significance at (at least) the 10% level (others not statistically significant). Regressions include around 7 100 observations. The 2012/2014
sample is a combination of the initial sample analysed in 2012 and the top-up sample added in 2014.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/759b0u
1. This is estimated using a probit regression with a dummy variable equal to one of the worker carries out a certain tasks at least once per
week as dependent variable. The explanatory variables are a time dummy (equal to 1 for 2017 observations) and country fixed effects.
Occupation controls are added at the 2-digit ISCO level.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
296
Figure 5.22. Young graduates with tertiary education degrees face the lowest risk of automation of
their jobs
Percentage of employed graduates (aged 16 to 34) with jobs at high risk of automation
Below upper-secondary Upper secondary & post-secondary non-tertiary - General Upper secondary & post-secondary non-tertiary - VET Tertiary
60%
50%
40%
30%
20%
10%
0%
Note: VET: Vocational Education and Training. High risk of automation is defined as having a probability of automation of at least 0.7. Belgium
refers to Flanders only, the United Kingdom to England and Northern Ireland. The sample includes employed individuals aged 16 to 34 who are
not enrolled in formal education. OECD is an unweighted average of the countries shown.
Source: OECD Survey of Adult Skills (2011/12, 2014/15, 2017/18).
StatLink 2 https://stat.link/jsl406
Using a simulation modelling approach, this chapter further explores how the types of jobs available are
altered when bottlenecks to automation are overcome. To analyse the potential impact of the ensuring “burst”
of automation, the model uses the Cedefop (2018[29]) sectoral employment forecasts as a baseline (see
Box 5.7)21, and incorporates the insight from Nedelkoska and Quintini (2018[27]) that automation affects jobs
through its impact on individual tasks. Using estimates of task automation provided by Brandes and
Wettenhofer (2016[30]), the model probabilistically automates certain tasks during a burst of automation.
O*NET occupational data (National Center for O*NET Development, 2020[31]) are used to provide detailed
breakdowns of task importance and frequency within occupations, as well as the knowledge requirement for
workers to fulfil those tasks. After a burst of automation, firms no longer need workers for the automated
tasks within an occupation, and thus the frequency of the remaining tasks (i.e. those not automated)
increases. Annex 5.C describes the simulation model in detail.
According to the simulation results22, the task content of jobs changes as a result of automation, and therefore
the skillset that employers are looking for when they hire workers also changes. This means that employers
might hire workers with a different education and work experience background than they did in the past.
Figure 5.23 shows that an automation episode is expected to change the occupational composition of VET
graduates’ employment, with middle-skill jobs becoming less important for them, and low- and high-skill jobs
gaining in importance. For all education groups middle-skill jobs become less important, as these are the jobs
most exposed to automation (see Annex Figure 5.A.2), and the impact is smallest for VET and tertiary education
graduates. Moreover, for VET and tertiary education graduates the employment structure shifts mostly towards
high-skill jobs whereas general education graduates and those without an upper-secondary education degree
mostly see relative employment gains in low-skill occupations. Despite the fact that the employment structure
of general education graduates shifts more strongly to low-skill occupations than to high-skill occupations, their
high-skill employment share change is still larger than the change in the high-skill employment share for VET
graduates. For VET graduates, the larger employment share in high-skill occupations is fully driven by
occupations that demand a management-type skillset, with VET graduates being more often employed as
specialised managers (e.g. construction managers, wholesale and retail managers) after an automation
episode.23,24 For general education graduates, the employment shift towards high-skill occupations is mostly
due to increased relative employment in professional and – to a lesser extent – management occupations.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
297
These results suggest that automation can potentially further reduce the importance of middle-skill
occupations in the labour markets of OECD countries. While this could affect all education groups, the
impact on VET graduates’ employment structure can be expected to be less strong than for general
education graduates and for those without an upper-secondary education degree. VET graduates have a
comparative advantage in middle-skilled jobs relative to other graduates, as their education specifically
prepared them for those jobs. Employers might therefore prefer VET graduates over general education
graduates (and graduates without an upper-secondary education degree) for the remaining middle-skill
jobs. At the same time, the skills demanded in high-skill occupations (i.e. managers, professionals,
technicians) are unlikely to be automated and these professions remain in demand (see Annex
Figure 5.A.2). The ability of VET graduates to get access to these jobs crucially depends on the skills they
acquired in education, as well as the willingness of employers to hire workers who do not fully fit the profile
they are looking for and fill skill gaps through training.
Figure 5.23. The impact of a burst of automation on the skill composition of VET graduates’
employment is relatively limited
Percentage point change in the employment share of each occupation group (by education level) that is attributable
to a burst of automation
Low-skill occupations Middle-skill occupations High-skill occupations
6
4
2
0
-2
-4
-6
Below upper secondary Upper secondary & post- Upper secondary & post Tertiary Total
secondary non-tertiary - General secondary non-tertiary - VET
Note: VET: Vocational Education and Training. Figure presented the simulated change in employment within each educational group that is
directly attributable to automation. Changed presented as a share of total employment within each educational group over the period. The total
of the data points for each educational group sums to zero. High-skill occupations are defined as ISCO 1-3; middle-skill occupations as ISCO 4
plus ISCO 6-8; low-skill occupations as ISCO 5 plus ISCO 9. See Chapter 4 for details on the skill groups.
Source: OECD simulations, see Annex 5.C for details.
StatLink 2 https://stat.link/ke7yz8
5.4. Improving the future-readiness of VET systems
VET systems around the world have a crucial role to play in the education system. As documented above,
they facilitate school-to-work transitions, resulting in better labour market outcomes for VET graduates
compared to general education graduates at the start of their career. Earlier research has also shown that
VET helps reduce high school dropout, especially for high-risk students (Kulik, 1998[32]; Henriques et al.,
2018[33]). In this respect, VET is crucial for engaging students in education and therefore improving their
labour market prospects. Nonetheless, in a changing world of work, certain aspects of VET systems might
need to be re-engineered to further strengthen the positive impact VET can have on education and labour
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
298
market outcomes. As the characteristics of VET systems and the labour market outcomes of VET
graduates differ widely between OECD countries, the extent to which a re-engineering of the system is
important also varies.
5.4.1. Creating responsive VET systems
With many of the jobs commonly targeted by VET undergoing substantial changes, VET programmes need
to be responsive, so that they remain relevant for students and employers. In responsive VET systems,
existing VET programmes are updated in a timely way to reflect changing needs in the labour market, and
new programmes are created when there is sustained demand for them. Strong coordination between the
VET system and the world of work allows for a better understanding of how jobs and skill needs are changing
and how VET systems should react to these changes. Strong ties between VET providers and social partners
also facilitates the implementation of work-based learning. Social partners can be involved in different aspects
of the VET system. According to KOF Swiss Economic Institute (KOF Swiss Economic Institute, 2016[4]),
employer engagement can take place in the curriculum design, application and feedback phase. In the
curriculum design phase, employers can be involved in setting qualification standards, as well as in the
development of student evaluation guidelines. Employer involvement in the application phase mainly
happens through the provision of work-based learning, but employers can also be involved in other areas,
such as quality assurance of work-based learning, cost-sharing agreements, the provision of equipment and
teachers, and the inclusion of a workplace component in student evaluations. Finally, in the feedback phase,
employers can share information about student outcomes and skill needs to feed into the re-design of
curricula, and they can be involved in determining the optimal timing for curriculum re-design. Among a range
of countries that are deemed to have well performing VET systems, Austria, Switzerland, Denmark and
Germany are found to score highest on employer engagement across the different dimensions (KOF Swiss
Economic Institute, 2016[4]). In Germany, for example, employers have an important role in providing
apprenticeship places, but are also key players in determining the content and organisation of VET
programmes. Social partners can make the case for updating existing training regulations or developing new
ones, and nominate experts who are involved in the development of these regulations (OECD, 2019[34]).
As discussed in the previous section, employment in high-skill occupations is expected to continue to
increase at a faster pace than in medium-skill occupations. These changes imply that there is an increased
need for higher-level vocationally oriented qualifications (at ISCED levels 5 and above) and for easy
pathways between medium-level VET and these higher-level qualifications. Many countries have opened
up higher education to individuals with vocational qualifications and/or with work experience, but actual
use of these non-traditional access routes is still relatively low (Cedefop, 2019[35]). In addition to helping
meet the demand for high-level skills, effective learning pathways can help increase the attractiveness of
VET, support lifelong learning, reduce inequalities and promote social inclusion and mobility (Field and
Guez, 2018[36]). In practice, many barriers hinder smooth pathways between mid-level VET and higher
education, including fragmented education systems with limited transparency, limited development of
general skills in mid-level VET to be successful in higher education, and a lack of flexibility in higher
education programmes. In Austria, graduates from the dual system and 3-4 year VET schools can enter
universities and Fachhochschulen, by completing special exams (Berufsreifeprüfung). Students can
participate in preparatory courses provided by several institutions (OECD, 2014[37]). Since 2008,
apprentices have the option of pursuing a double degree (Lehre mit Matura), combining the occupational
qualification and the special higher education entrance degree. In 2018, only around 6% of apprentices
opted for this combined degree (Dornmayr and Nowak, 2018[38]). In Norway, graduates from the vocational
track at the upper-secondary level have the option to continue to higher education after a one-year bridging
course (Norwegian Directorate for Education and Training, 2013[39]). This bridging course covers six key
academic subjects: Norwegian, English, Mathematics, Natural Sciences, Social Sciences, and History. For
certain higher education programmes, mainly in the engineering field, entry is allowed for vocational
qualification holders without going through the bridging programme.
OECD EMPLOYMENT OUTLOOK 2020 © OECD 2020
Full text shows the first 300 of 370 pages; the complete document is the PDF above.File and source
- File
- oecd-employment-outlook-2020-worker-security-and-the-covid-19-crisis.pdf
- Size
- 8,221,876 bytes
- SHA-256
- 2b5bd5fb3c68cf069431b8206de76ea3d9733377ad40b622ec17611c1f2d5465
- Original
- www.oecd.org