How Effective Were Job-Retention Schemes During COVID-19?
Summary
An International Monetary Fund working paper, WP/23/3, from the Fiscal Affairs Department, dated January 2023 and prepared by W. Raphael Lam and Alexandra Solovyeva. It uses the EUROMOD microsimulation model and household data to assess how well job-retention schemes stabilized household income across European countries during the COVID-19 pandemic. The abstract reports that these schemes, with other measures, absorbed nearly 80 percent of market income shocks and helped mitigate the rise in the unemployment rate by about 3 percentage points. The paper contrasts EU labor markets, where hours per worker fell by almost 12 percent in the second quarter of 2020, with the United States, and mentions the Paycheck Protection Program as a limited US job-retention scheme. It closes with a list of references.
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
How Effective were Job-Retention
Schemes during the COVID-19
Pandemic? A Microsimulation
Approach for European Countries
W. Raphael Lam and Alexandra Solovyeva
IMF Working Papers describe research in
progress by the author(s) and are published to
elicit comments and to encourage debate.
The views expressed in IMF Working Papers are
those of the author(s) and do not necessarily
represent the views of the IMF, its Executive Board,
or IMF management.
2023
JAN
© 2023 International Monetary Fund WP/23/3
IMF Working Paper
Fiscal Affairs Department
How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Prepared by W. Raphael Lam and Alexandra Solovyeva 1
Authorized for distribution by Paulo Medas
January 2023
IMF Working Papers describe research in progress by the author(s) and are published to elicit
comments and to encourage debate. The views expressed in IMF Working Papers are those of the
author(s) and do not necessarily represent the views of the IMF, its Executive Board, or IMF management.
ABSTRACT: The COVID-19 pandemic had posed a dramatic impact on labor markets across Europe. Forceful fiscal
responses have prevented an otherwise sharper contraction. Many countries introduced or expanded job-retention
schemes to preserve jobs and support households. This paper uses a microsimulation approach (EUROMOD) and
household data to assess the effectiveness of those schemes in stabilizing household income during the pandemic
across European countries. Empirical evidence shows that job-retention schemes were effective in stabilizing income
and, along with other measures, absorbed nearly 80 percent of market income shocks—almost doubling the extent of
the automatic stabilization of the pre-p andemic tax and benefit systems. The large effects are related to the
widespread use and scaling up of those schemes and a deep but short-lived disruption to labor markets during the
pandemic. Along with other fiscal support measures, job-retention schemes helped mitigate the rise in the
unemployment rate, by about 3 percentage points, and income inequality during the pandemic. Our results show that
job-retention schemes were largely targeted, in which households more vulnerable to income losses, such as lower-
income families, youth, and low-skilled workers, are able to stabilize their income.
RECOMMENDED CITATION: Lam, W. Raphael and Solovyeva, Alexandra. 2023. How Effective were Job-Retention
Schemes during the COVID-19 Pandemic? A Microsimulation Approach for European Countries. IMF Working Paper
23/3, International Monetary Fund, Washington DC.
JEL Classification Numbers: D1, D12, D31, D61; E2, E21, E24, E64, E65; H3, H31; J3, J6
Job-retention schemes; COVID-19 pandemic; short-time work;
Keywords:
inequality; income stabilization; Okun’s Law
Author’s E-Mail Address: WLam@imf.org; ASolovyeva@imf.org
1
The authors would like to thank Vitor Gaspar, Paolo Mauro, Paulo Medas, Ricardo Reis, Camille Landais, Alberto Tumino, and
participants in IMF-FAD seminar and July 2022 Fiscal Monitor Workshop for their constructive comments and suggestions. Andrew
Womer and Andre Vasquez provided excellent research assistance and editorial support. The paper is based on data from Eurostat,
2019 European Statistics on Income and Living Conditions (2018 reference year). The responsibility for all conclusions drawn from
the data lies entirely with the authors. The results presented in the paper are based on EUROMOD version I4.0+. Originally
maintained, developed and managed by the Institute for Social and Economic Research (ISER), since 2021 EUROMOD is
maintained, developed and managed by the Joint Research Centre (JRC) of the European Commission, in collaboration with
EUROSTAT and national teams from the EU countries. We are indebted to the many people who have contributed to the
development of EUROMOD. The report is based on data from Eurostat 2018-20.
WORKING PAPERS
How Effective were Job-Retention
Schemes During the COVID-19
Pandemic? A Microsimulation
Approach for European Countries
Prepared by W. Raphael Lam and Alexandra Solovyeva
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Contents
I. INTRODUCTION_________________________________________________________________________3
II. IMPACT OF THE PANDEMIC ON EU LABOR MARKETS ___________________________4
III. DATA AND METHODOLOGY __________________________________________________________8
IV. SIMULATION RESULTS _________________________________________________________________9
A. Income stabilization before the pandemic_____________________________________________9
B. Income stabilization during the pandemic___________________________________________ 11
C. Regression analysis ____________________________________________________________________ 16
V. POLICY IMPLICATIONS AND CONCL USIONS_____________________________________ 17
ANNEX I. SIMULATIONS OF SHOCKS IN EUROMOD ______________________________________________ 19
REFERENCES _____________________________________________________________________________________________ 20
FIGURES
1. Labor Market Developments in the European Union During the COVID-19 _______________________________ 5
2. Relationship between Output Fluctuations and Labor Market Dynamics __________________________________ 7
3. EU-27: Changes in Unemployment ________________________________________________________________________ 7
4 Income Stabilization in EU before the COVID-19 Pandemic, 2019________________________________________ 10
5. Simulated Changes in Market and Disposable Incomes During the Pandemic___________________________ 11
6. Income Stabilization after the COVID Shock, by Country and Income Level _____________________________ 12
7. Correlation of Simulated Income ________________________________________________________________________ 13
8. Income Stabilization During the Pandemic across Households __________________________________________ 13
9. Redistribution Effects of Fiscal Measures During the Pandemic__________________________________________ 14
10. Simulated Income Stabilization Coefficients by Worker Groups and Sectors ___________________________ 15
TABLES
1. Features of Job-Retention Schemes in Selected European Countries During the COVID-19 Pandemic ____ 6
2. Regression Results on Differences of Income Stabilization across EU Households_______________________ 17
INTERNATIONAL MONETARY FUND 2
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
I. Introduction
The COVID-19 pandemic generated widespread economic disruptions and consequently led to a sharp
deterioration in labor markets across Europe. Despite a dramatic economic contraction, the impact on
employment appeared to be muted, with employment rate remaining at 0.7 percentage points below the pre-
crisis levels, far above projections based on the pre-pandemic relationship between output growth and
unemployment (Okun’s relationship). Most of the adjustment in the labor markets was through a sharp
reduction in working hours per worker, by 12 percent in the second quarter of 2020. These observations were
in contrast with the experience of the United States, where the unemployment rate surged by 11 percentage
points in the first two months of the pandemic, while working hours per worker fell moderately.
Swift and forceful fiscal support has cushioned the adverse impact of the pandemic. Yet the diverse experience
in EU and US labor markets was likely driven by specific fiscal measures. Many EU countries introduced new
or expanded the existing job-retention schemes which prevented a surge in unemployment, while the United
States expanded the federal unemployment income support despite that many states had some form of job-
retention schemes in place. 2 This is usually seen as a key contributing factor for different developments in the
labor markets (Ando and others 2022; Giupponi, Landais, and Lapeyre 2022). Within the European Union, the
design and coverage of job-retention schemes varied significantly across countries. Given that job-retention
schemes can become a prominent tool for future shocks, it is therefore important to assess whether job-
retention schemes are effective in terms of stabilizing household income and to what extent those schemes
target well to workers vulnerable to job losses.
The paper uses a microsimulation approach (EUROMOD) and household data to assess the effectiveness of
those schemes in stabilizing household income during the pandemic across EU countries. Our paper is related
to Christl and others (2022) that analyzes the aggregate stabilization of tax and benefit systems during the
pandemic. Our paper extends their analysis and quantifies not only the aggregate effects of pandemic-related
fiscal support, including job-retention schemes, in stabilizing household income, but also focuses on the impact
on different socio-economic groups, including age, gender, occupations, and level of educational attainment.
The paper is also related to other strands in the literature. First, it contributes to the literature on the size of
automatic stabilizers—the built-in components in the budget that adjust automatically to cyclical changes in the
economy. The paper uses a micro-simulation approach that relies on household data and detailed policies in
the tax and benefit systems (Auerbach and Feenberg 2000; Dolls, Fuest, and Peichl 2012). It allows an
analysis of the direct effects of specific tax or expenditure policies on household income during an adverse
shock by household characteristics. Although the microsimulation approach does not account for feedback
effects, other approaches using cyclical budget balances or semi-elasticities are likely less applicable in an
environment of sharp adjustments and high uncertainty such as during the pandemic. The relationship between
fiscal variables and output inherent in those approaches could change dramatically (for example, due to
lockdown restrictions). 3
Second, our microsimulation results provide an estimate on the degree of income stabilization of tax and
benefit systems during the pandemic. This updates the previous estimates in the literature that focus on the
2
The widespread use of job-retention schemes was in part mobilizing the EU funds under the temporary Support to mitigate
Unemployment Risks in Emergency (SURE) instrument.
3
Other approaches include the conventional statistical method that uses the cyclical component of the government budget balance
or makes use of the semi-elasticities of revenue and expenditures (IMF 2015) or general equilibrium models. The general
equilibrium modeling approach can estimate both direct and indirect effects of behavioral responses and their feedback (Krusell,
Mukoyama, and Sahin 2010; McKay and Reis 2016).
INTERNATIONAL MONETARY FUND 3
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
pre-pandemic period, including (Coady and others 2023; Dolls, Fuest, and Peichl 2012; European Commission
2017; Mohl Mourre, and Stovicek 2019).
Third, the paper contributes to the research that examines the role of pandemic-related fiscal support in
European labor markets. For example, Ando and others (2022) find that job-retention schemes across the euro
area were crucial in mitigating the adverse impact of the pandemic, otherwise, unemployment rates could have
risen further by another 2½ percentage points in 2020. Aiyar and Dao (2021) uses data on state-level
Kurzarbeit, short-time work program in Germany, and finds that the unemployment rate would have increased
by 3 percentage points without the job-retention schemes and consumption would have contracted even
further. Giupponi, Landais, and Lapeyre (2022) compares the experience between EU and US policy
responses and concludes that cyclical job-retention schemes can be an efficient and expedient way to
complement unemployment insurance. Our paper supports these findings and provides further analyses across
different household groups. The paper does not examine the effects of prolonged use of job-retention schemes
on labor market allocations partly because those schemes were quickly phased out when labor markets
recovered during the pandmeic, although some studies suggest the potential drawback on disincentives to
work from such schemes if not withdrawn timely (Basso and others 2020).
The remainder of the paper is organized as follows. Section II discusses the impact of the pandemic on the EU
labor markets and provides an overview of policies to protect workers against job and income losses. Section
III describes the data and methodology of the microsimulations of the tax and benefit systems and the design of
various scenarios. Section IV presents simulation results for the pre-pandemic automatic stabilizers and
analyzes the extent of income stabilization in EU countries from job-retention schemes and other tax and
benefit components during the pandemic. Section V concludes with some takeaways and policy implications.
II. Impact of the Pandemic on EU Labor Markets
During the COVID-19 pandemic, labor markets in the European Union adjusted mostly through reduction of
hours worked per employee (‘intensive margin’) rather than employment levels (Figure 1). Average hours
worked per worker dropped by almost 12 percent year-on-year in the second quarter of 2020, while
employment fell by just under 3 percent over the same period. There was some variation across EU countries,
depending on the differences in the severity of the pandemic and the sectoral structure.
The developments in EU labor markets are very different from those in the United States, where employment
plunged by 12 percent at the onset of the pandemic but working hours per workers remained steady (Figure 1).
The adverse impact of large decline in working hours but muted employment loss was in contrast to the global
financial crisis. During the onset of the global financial crisis, the decline in employment was broadly similar to
the drop in working hours per worker, and the decline was more protracted over several years.
The widespread use of job-retention schemes in EU countries has contributed to the muted loss of employment
during the pandemic. Job-retention schemes encompass policies that subsidize workers’ wages in firms that
have reduced working hours but preserved workers’ jobs, which broadly include short-time work schemes and
wage subsidies. Before the pandemic, many EU countries had already some forms of job-retention schemes in
place. Some schemes have always been active 4, such as Kurzarbeit in Germany and Activité Partielle in
France, while other governments took action to introduce or expand those schemes to protect workers. During
the pandemic, countries introduced new or expanded existing job-retention schemes by simplifying access,
4
A firm can always apply but needs to prove that all the eligibility criteria are met.
INTERNATIONAL MONETARY FUND 4
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
relaxing eligibility criteria, and raising the benefit levels (Table 1). Some schemes gave more generous support
for workers in contact-intensive sectors that were affected the most, such as in Austria and Luxembourg.
At the beginning of the pandemic, an average of 14 percent of working-age population were under some job-
retention schemes in the four largest EU economies (compared to 2 percent during the global financial crisis).
More than half of EU countries had the take-up rate higher than 12 percent of working-age population. This
dwarfed the increase of people receiving unemployment income support, which increased only modestly, by
about 2 percentage points, given the muted impact on employment (Figure 1). The rise in unemployment rates
in the EU during the pandemic was lower than predicted by the estimates based on the Okun’s Law, partly
reflecting the widespread use of job-retention schemes, as well as a drop in the labor force participation when
workers did not actively search for a new job at the onset of the pandemic (Ando and others 2022). The
average fiscal cost was about 2 percent of GDP in advanced EU economies (1.4 percent in emerging market
economies in EU), about one-third of total fiscal support during the pandemic (Ando and others 2022).
On the other hand, the US has also introduced a limited job-retention scheme to firms—the Paycheck
Protection Program (PPP)—that provided small firms with loans to cover labor costs, forgivable if the payroll
level was maintained. Some estimates suggest that PPP saved about 3.6 million jobs, equivalent to 2.2 percent
of total employment (Autor and others 2022). In many cases, however, workers were temporarily laid off and
received unemployment income support at the onset of the pandemic. The US scaled up its federal
unemployment support by about 3 percent of GDP to provide weekly supplements to standard unemployment
insurance, expand the eligibility to include independent workers, and extend the duration of the federal benefits.
Figure 1. Labor Market Developments in the European Union During the COVID-19 Pandemic
1. Employment Growth 2. Growth in Average Hours per Worker
(Percent, year-on-year) (Percent, year-on-year)
10 15
10
5
5
0
0
-5 United States -5
United States
EU-27
-10 -10 EU-27
-15
-15
2007Q2 2008Q2 2009Q2 2010Q2 2011Q2 2012Q2 2013Q2 2014Q2 2015Q2 2016Q2 2017Q2 2018Q2 2019Q2 2020Q2 2021Q2
2007Q2 2008Q2 2009Q2 2010Q2 2011Q2 2012Q2 2013Q2 2014Q2 2015Q2 2016Q2 2017Q2 2018Q2 2019Q2 2020Q2 2021Q2
Sources: Eurostat, B ureau of Labor Statistics, and authors’ calculations.
Note: Shaded areas refer to CEPR-based recessions for the EU countries and NBER recessions for the United States.
3. Changes in Employment and Working Hours, 4. Take-up of Job-Retention Schemes and
2019-20 Unemployment Income Support
(Percentage change) (Percent of working-age population)
2
USA ROU 14 Global financial crisis, 2008-09 14 COVID-19, 2020-21
0 HRV
FIN 12 12
POL
SWE
DNK 10 10
-2
Hours per worker
BGR NLD 8 8
LVA HUN
-4 EST LTU 6 6
CZE DEU SVN 4 4
-6
EU-27 LUX
CYP 2 2
ESP SVKAUT
IRL FRA 0 0
-8 PRT MLT
BEL Dec-07 Mar-08 Jun-08 Sep-08 Dec-08 Mar-09 Jun-09 Sep-09 Dec-09 Jan-20 Mar-20 May-20 Jul-20 Sep-20 Nov-20 Jan-21 Mar-21 May-21
ITA
-10 GRC
-8 -6 -4 -2 0 2 4
Employment
Sources: Bureau of Labor Statistics, Eurostat, and authors’ Sources: Giupponi, Landais, and Lapeyre 2022; and
calculations. The figure uses International Organization for authors’ calculations.
Standardization (ISO) country codes. Note: Data for EU-4 is a weighted average of Germany,
France, Spain, and Italy.
INTERNATIONAL MONETARY FUND 5
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Table 1. Features of Job-Retention Schemes in Selected European Countries During the COVID-19 Pandemic
Pre-existing job- Increased access and Increased New job-retention Take-up rate
Remarks
retention scheme coverage generosity scheme Maximum Average
Austria x x x 17.7 7.9 Longer duration, more flexible rules for extension of duration and administrative simplification. Up to 100
precent working time reduction in the hospitality sector.
Belgium x x x 16.9 7.6
Bulgaria x
Croatia x
Czech Republic x x x 8.6 4.5
Cyprus x
Denmark x x x 7.3 3.5 Introduced temporararily with no membership in unemployment scheme required.
Estonia x 14.4 3.3
Finland x x x 4.8 2.6
France x x x 20.6 8.8 No condition on type of contract, part or full time, seniority. The maximum duration is extended from 6 to 12
months. The subsidy is 70 percent of gross wage. Most employers do not bear any cost of hours no worked.
Germany x x x 11.2 6.4 Firms can apply if 10 percent of their workforce is subject to reduction of hours (30 percent before).
Replacement rate for lost earnings is raised to 70 and 80 percent (from 4th month and 7th month, respectively).
Greece x 11.0 4.8 Available for employers with at least a 20 percent revenue loss. Only full-time dependent employees are
eligible. Replacement rate for lost earnings is 60 percent of net wages.
Hungary x 3.2 2.3 Job-retention scheme with 30-50 percent working time reduction.
Ireland x x 14.8 11.8 The existing short-time work was replaced by a temporary wage subsidy.
Italy x x x 14.7 7.3 Firms of any size and from all sectors can apply. Evidence of economic need is no longer required. Employers
do not bear any cost for hours not worked.
Latvia x 5.0 3.2 Short-time work with full- and part-time reduction.
Lithuania x 9.4 3.8 Short-time work with full- and part-time reduction.
Luxembourg x x x 22.2 6.4 Up to 100 percent working time reduction; temporary workers and apprentices eligible.
Wage Supplement scheme provided eligible employees with a basic wage (March 2020-May 2022). Funds were
Malta x forwarded via the employer.
Netherlands x x 23.8 14.2 The existing short-time work program was replaced by a temporary wage subsidy; employees received 100
percent of their wage.
Poland x 6.3 2.4
Portugal x x x 12.0 4.5
Romania x Short-time work program with up to 80 percent of reduced work time.
Slovakia x x x 12.6 8.0
Slovenia x 13.4 5.6
Spain x x x 11.5 4.9
Sweden x x x 7.6 4.3 Larger working time reduction of 80 percent between May and July 2020.
United Kingdom x 21.2 12.8 Replacement rate for unworked hours is 80 percent of gross salary. The cost of unworked hours faced by
firms set at zero. All workers who were on payroll on March 19 are eligible.
Sources: Ando and others (2022), Drahokoupil and Müller (2021), Giupponi, Landais, and Lapeyre (2022), and OECD (2020).
Note: The take-up rates of job-retention schemes refer to the maximum and the average of share of workers enrolled in job-retention schemes during March to December
2020. The indicators are expressed in percent of working-age population in the country.
INTERNATIONAL MONETARY FUND 6
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
The COVID-19 pandemic had a disproportionate impact on certain groups of workers. For example, young
workers experienced the largest decline in employment and the largest rise in the unemployment rate between
2019 and 2020 (Figure 3). Workers with low-level of education also saw a large decline in employment rate, 5
percentage points on average at the EU level. In contrast, elderly and high-skilled workers were less affected,
with their employment rates rising slightly by 1.7 and 2.4 percent, respectively. The findings that young and
low-skilled workers were more sensitive to economic fluctuations coincided with empirical estimates before the
pandemic. For example, pre-pandemic Okun’s Law estimates suggest that unemployment rates of these
groups are more responsive to output fluctuations (Figure 2), consistent with Ando and others (2022).
Figure 2. Relationship between Output Fluctuations and Labor Market Dynamics across Worker
Groups
1. Estimated Okun’s Law Coefficients, 1995-2019 2. Actual Unemployment Rates Rose Less than
Predicted Levels based on Okun’s Law
Estimates, 2020
(Coefficients, average across EU countries) (Percentage points, average)
0 2
2020Q1 2020Q2 2020Q3 2020 average
-0.05
0
-0.1
-0.15
-2
-0.2
-0.25
-4
-0.3
-0.35 -6
-0.4
-0.45 -8
15-24 25-54 55-64 Men Women Low Medium High 15-24 25-54 55-64 Men Women Low Medium High
Total Age Gender Education Total Age Gender Education
Sources: Eurostat and authors’ calculations.
Notes: Okun’s Law coefficient measures the impact of GDP growth on changes in the unemployment rate. The estimation
sample includes 15 EU countries and covers the period during 1995Q1-2019Q4 (or earliest available). Coefficients for
individual countries not statistically significantly at the 5 percent level are set to zero. Panel 2 shows the average prediction
errors (the difference between the actual and predicted unemployment rates based on Okun’s Law coefficients). Low
education corresponds to less than primary and lower secondary. High education corresponds to tertiary education.
While job-retention schemes have contributed to Figure 3. EU-27: Changes in Unemployment
preventing a surge in unemployment in EU countries, Rates and Employment Growth, 2019-20
questions arise whether those schemes are effective to (Percentage points for changes in unemployment
stabilize household incomes and target well to workers rate; percent for employment growth)
3 3
that are more vulnerable to job losses. The following 2 2
section analyzes these by conducting micro- 1 1
simulations. 0 0
-1 -1
-2 -2
-3 -3
-4 -4
-5 -5
-6 -6
-7 -7
Medium Medium
Men High Men High
Women Women
Low Low
15-24 25-54 55-64 15-24 25-54 55-64
Age Gender Education Age Gender Education
Unemployment rate change Employment growth
Sources: Eurostat and authors’ calculations.
INTERNATIONAL MONETARY FUND 7
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
III. Data and Methodology
The analysis uses a microsimulation approach that quantifies how well existing tax and benefit systems or new
policy measures buffer shocks to household market income (income before taxes and transfers). This approach
allows a detailed analysis based on household characteristics, although it does not account for the feedback
effects on aggregate income when policies change.
The analysis uses a static microsimulation model EUROMOD (version I4.0+) and microdata from the 2019
European Statistics on Income and Living Conditions (EU-SILC) for 26 EU countries. 5 The EUROMOD is a
tax-benefit microsimulation model developed by the Joint Research Center (JRC) of the European Commission
in collaboration with Eurostat and national teams from the EU countries. The model simulates country-specific
direct tax liabilities and in-cash benefit entitlements for samples of representative households in EU countries
(Sutherland and Figari 2013). The model allows us to calculate, in a comparable manner, the effects of tax and
benefit policies on the income of individual households (see Annex I).
The paper conducts two sets of simulations to analyze how a change in taxes and benefits would stabilize
income during an adverse shock.
1. The first set of simulations aims to assess to what extent income was stabilized in an adverse shock before
the pandemic. The adverse shock is assumed to be a uniform 5-percent decline in market incomes of all
households under the 2019 tax and benefit system (see Annex I). This help assess the size of automatic
stabilizers—the built-in components in the budget that adjust automatically to cyclical changes in the
economy—prior to the pandemic. The analyses simulate and compare two scenarios—the baseline and
the income shock scenario. The baseline scenario is based on the tax-benefit system of individual EU
countries in 2019 and the 2019 EU-SILC household-level microdata (assuming no major discretionary
fiscal measures in 2019), while the income shock scenario is based on the 2019 tax-benefit system with a
5-percent decline in market incomes of all individuals in the 2019 EU-SILC microdata.
2. The second set of simulations assesses to what extent the announced fiscal support during the pandemic
has stabilized household incomes. The analyses simulate and compare two scenarios—a “COVID-19”
scenario and a counterfactual “no COVID-19” scenario. The COVID-19 scenario is based on the 2020 tax-
benefit system and the 2019 EU-SILC microdata, adjusted to match the actual labor market conditions
observed in 2020. The adjustment follows Christl and others (2022) and employs the Labor Market
Adjustment (LMA) Add-on that simulates transitions between employment, unemployment and job-
retention schemes based on the data from the European Labor Force Survey and other detailed
administrative sources. 6 This scenario essentially captures how household incomes would change under a
COVID-19 shock, taking into account the announced fiscal support measures. In the “no COVID-19”
scenario, the analysis assumes that there were no pandemic-related labor market transitions (i.e., no rise
in unemployment or decline in working hours as observed during the pandemic). The simulation uses the
2020 tax-benefit system and the latest available 2019 EU-SILC microdata.
In cases where the reference year of the microdata is different from that of the tax and benefit systems, the
EUROMOD adjusts monetary variables in the microdata. The adjustment through uprating factors follows the
5
The EUROMOD version I4.0+ is developed by the Institute for Social and Economic Research, University of Essex; Joint Research
Centre, European Commission, 2022. The EUROMOD input files are based on 2019 EU-SILC microdata, which are made
available by Eurostat for all EU countries except Germany.
6
See the EUROMOD LMA Add-on documentation for further information.
INTERNATIONAL MONETARY FUND 8
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
EUROMOD modeling conventions. For example, an uprated adjustment is made to align the differences in
reference years for household income (2018) and the tax and benefit system (2019).
The paper quantifies the extent of income stabilization by the tax and benefit system during an adverse income
shock using an income stabilization coefficient, in line with the literature (Dolls and others 2012; Mohl, Mourre,
and Stovicek 2019). The coefficient measures the average share of the market income shock that is absorbed
by the tax and benefit system. It is defined as follows:
∑ 𝑁𝑁
ℎ=1 ∆𝑌𝑌ℎ ∑ 𝑁𝑁
ℎ=1 ∆𝑇𝑇ℎ ∑ 𝑁𝑁
ℎ=1 ∆𝑆𝑆ℎ ∑ 𝑁𝑁
ℎ=1 ∆𝐵𝐵ℎ
𝐼𝐼𝐼𝐼𝐼𝐼 = �1 − 𝑁𝑁
� × 100 = � 𝑁𝑁
+ 𝑁𝑁
− 𝑁𝑁
� × 100 (1)
∑ ℎ=1 ∆𝑀𝑀ℎ ∑ ℎ=1 ∆𝑀𝑀ℎ ∑ ℎ=1 ∆𝑀𝑀ℎ ∑ ℎ=1 ∆𝑀𝑀ℎ
where ∆𝑀𝑀ℎ (∆𝑌𝑌ℎ ) is the change in market (disposable) income of household h following the shock. Variables
∆𝑇𝑇ℎ , ∆𝑆𝑆ℎ , and ∆𝐵𝐵ℎ refer to the changes in personal income taxes, social insurance contributions, and social
benefits, respectively. Social benefits include unemployment benefits, social assistance and housing benefits,
family and education benefits, health and disability benefits. In the simulations of the COVID-19 shock, social
benefits also include the monetary compensation received by workers on short-time work programs, wage
subsidies, as well as similar schemes for self-employed, which the paper refers to broadly as job-retention
schemes. 7 The income stabilization coefficient is equal to 100 percent when the market income shock is fully
absorbed by the tax and benefit system. A zero coefficient means that the tax and benefit system does not
compensate for income losses such that the change in disposable income (after taxes and transfers) is the
same as the change in market income (before taxes and transfers).
IV. Simulation Results
A. Income stabilization before the pandemic
The first set of simulations suggests that the tax and benefit system in the EU countries, on average, can
absorb 37 percent of an adverse income shock (5 percent decline in market income), reflecting the size of
automatic stabilizers in the tax and benefit systems before the pandemic (Figure 4, panel 1). The estimates of
income stabilization are in line with other estimates such as Mohl, Mourre, and Stovicek (2019) and Coady and
others (2023), in which the latter finds that the size of automatic stabilizers had been stable during 2011-19 for
the EU countries on average. 8
Within the tax and benefit system, personal income tax has been the largest contributor to the income
stabilization during an adverse income shock. It absorbs 24 percent of the adverse income shock on average,
accounting for more than half of the total income stabilization in most countries. However, there is considerable
variation across countries in the EU, with coefficients ranging from 14 percent in Bulgaria to 57 percent in
Belgium. The large variation reflects differences in the progressivity of the tax systems. Countries with more
7
The complete list of monetary compensation schemes and other pandemic-related policies in the LMA Add-On is available in
Christl and others (2022).
8
Using the tax-benefit microsimulation model EUROMOD for 19 countries, Dolls and others (2012) assess the effectiveness of tax-
benefit systems to provide income insurance through automatic stabilizers during the global financial crisis and find that automatic
stabilizers absorbed 38 percent of a proportional market income shock in the EU, compared to 32 percent in the United States.
European Commission (2017) analyzes the direct and total effects of automatic stabilizers on income in the 28 Member States for
2014 using the microsimulation model EUROMOD and the macrosimulation model QUEST, respectively. They estimate that the
direct automatic income stabilization is about 33 percent on average, slightly higher than the total macro-based stabilization (29
percent) due to behavioral responses and macroeconomic feedback effects. Coady and others (2023) quantify the extent of
automatic income and consumption stabilization in the EU prior to the COVID-19 pandemic, and find that, in 2019, tax-benefit
policies absorbed 41 percent of the proportional market income shock on average, while only 15 percent of the market income
shock was transmitted to household consumption.
INTERNATIONAL MONETARY FUND 9
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
progressive personal income taxes, such as Belgium, Denmark, and Ireland, tend to have higher stabilization
from taxes, with personal income taxes absorbing at least 40 percent of the adverse income shock. This means
that the applicable personal income tax would fall when households face an adverse hit in their income. Social
insurance contributions and benefits are another key component of the automatic stabilizes in protecting
households against income losses. Together they absorb a total of 12 percent of the income shock, amid large
country variations. 9 In countries with limited progressivity of income taxes, such as Bulgaria and Romania,
social protection systems contribute over 70 percent of income stabilization during adverse shocks.
Figure 4 Income Stabilization in EU before the COVID-19 Pandemic, 2019
(Percent, share of the shock absorbed by the tax and benefit system)
1. By country 2. By household income quintile
70 70 70 70
Personal income taxes Personal income taxes
60 Social insurance contributions 60 60 60
Social insurance contributions
Social benefits Social benefits
50 50 50 50
40 40
40 40
30 30
30 30
20 20
20 20
10 10
10 10
0 0
First quintile Second Third quintile Fourth quintile Fifth quintile
0 0
BGR (lowest quintile (highest
EU
EST
CYP
MLT
income) income)
POL
LVA
CZE
GRC
ESP
SVK
HUN
FRA
PRT
LTU
ITA
SWE
ROU
SVN
NLD
LUX
FIN
IRL
DNK
AUT
BEL
Source: Authors’ calculations.
Note: Estimates are based on the EUROMOD I.40+ and microdata from the 2019 EU-SILC (excluding Germany) under the
under the 2019 tax-benefit system and an illustrative 5-percent negative shock to market incomes for all households. For
EU and each income quintile, the chart reports the average stabilization coefficients across countries. Income quintiles are
calculated at the country level based on the household’s market income in the baseline scenario. The difference between
stabilization coefficients for top income quintile and the first income quintile is positive and statistically significant for 18 out
of 26 EU countries. Social benefits include unemployment benefits, social assistance and housing benefits, family and
education benefits, health and disability benefits. Pensions are excluded. Data labels in the figure use International
Organization for Standardization (ISO) country codes. EU=European Union.
The size of income stabilization varies across household income distribution within a country, reflecting the
progressivity of the tax system and the strength of the social protection system. Across the EU countries,
household income tends to be stabilized more for higher-income households, with income stabilization
coefficients ranging from 32 percent for the poorest income quintile to 39 percent at the top income quintile
(Figure 4, panel 2). Nonetheless, social insurance and benefits play a more important role for lower-income
households, contributing about two-thirds to the overall stabilization (20 percent out of 32 percent) relative to
about one-quarter (10 out of 39 percent) for the households in the top income quintile. As for the social
protection system, the social insurance contributions stabilize income by about 11 percent, broadly at the same
degree across the household income distribution. In countries with stronger social safety nets, such as Ireland,
Luxembourg, and the Netherlands, social benefits contribute more to the income stabilization at the lower end
9
As the scenario considers a uniform 5-percent decline in household income without a change in unemployment, the income
stabilization from unemployment insurance and assistance is muted. We have conducted alternative scenarios in which
households face higher likelihood of unemployment in the event of an adverse income shock. The overall stabilization coefficient is
similar to the baseline scenario, but unemployment income support would play a more important role, particularly in countries with
strong unemployment income support systems. While the income stabilization coefficients are not entirely linear in the size of
income shocks, alternative scenarios show that the income stabilization coefficients in individual countries are broadly unchanged
if income shocks are between 0 and 10 percent.
INTERNATIONAL MONETARY FUND 10
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
of the income distribution. In contrast, income taxes absorb a larger share of adverse income shocks for the
high-income groups.
B. Income stabilization during the pandemic
Households suffered from a large decline in incomes during the pandemic, although the magnitudes vary
across countries reflecting differences in labor market dynamics, severity of the pandemic, and policy
responses. Simulations show that market incomes for households fell by 5.3 percent on average, with the
largest drop of more than 10 percent in Malta, Ireland, Italy, and Slovakia (Figure 5). In contrast, some
economies faced a smaller decline in market income, less than 2 percent (Denmark, Luxembourg, the
Netherlands). Fiscal support measures mitigated part of the income shock, resulting in a milder drop of the
disposable income (after taxes and transfers), by 1.6 percent in 2020. In a few countries, disposable income
actually rose slightly because of large fiscal support measures at the onset of the pandemic (Croatia, Denmark,
and Luxembourg).
Across the household income distribution, lower-income households tend to experience a larger decline in
market incomes before accounting for fiscal support. For example, market incomes fell by a median of 4.6
percent for households in the lowest income quintile, while only by 3.7 percent for households in the top income
quintile (Figure 5). Once accounting for the stabilization role of tax and benefit systems, the disproportionate
impact on lower-income households was more than offset by pre-existing automatic stabilizers and fiscal
support measures implemented during the pandemic. The disposable income after taxes and transfers
remained broadly unchanged for lower-income households (implying fiscal measures were able to absorb the
income shock), while it declined by about 2 percent for households at the top income quintile. These suggest
that fiscal policy was impactful and progressive, mitigating largely the income shocks across the board,
particularly for low-income households.
Figure 5. Simulated Changes in Market and Disposable Incomes During the Pandemic
(Percent)
1. By country 2. By household income quintile
4 4 4
2 2 Median Average 2
0 0 0
-2 -2 -2
-4 -4 -4
-6 -6 -6
-8 -8
-8
-10 -10
-10
Disposable income -12 -12
-12
Market income -14 -14
-14
-16 -16
-16 MLT Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5
EU
IRL
SVK
ITA
GRC
AUT
FRA
CYP
BEL
ESP
LTU
SVN
HRV
Market income Disposable income
PRT
EST
ROU
CZE
HUN
POL
LVA
FIN
BGR
SWE
DNK
NLD
LUX
Source: Authors’ calculations.
Note: Estimates are based on the EUROMOD I.40+ and microdata from the 2019 EU-SILC (excluding Germany). Labor
market shock is simulated to replicate the 2020 labor market conditions using Labor Market Adjustment (LMA) Add-On. For
EU and each income quintile, the average percentage change in market and disposable incomes across countries are
reported. The box-whisker shows the variation across EU countries, with the interquartile range (box), and 5 th and 95 th
percentiles (whiskers). Income quintiles are calculated at the country level based on the household’s market income in the
baseline “no COVID-19” scenario. Data labels in the figure use International Organization for Standardization (ISO) country
codes. EU=European Union.
INTERNATIONAL MONETARY FUND 11
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Fiscal support was impactful in stabilizing household incomes during the pandemic in the EU. Together with
pre-existing automatic stabilizers, fiscal support had absorbed 78 percent of the decline in market incomes
across countries (Figure 6), almost double the stabilization observed before the pandemic. 10 The income
stabilization varied across countries, ranging from 55 percent in Malta and Poland to almost 100 percent in
Belgium and Denmark. These explain large differences between changes in market and disposable incomes
during the pandemic, as well as a relatively muted impact on the disposable income even in countries that
faced a severe pandemic shock.
Among various components in the tax and benefit systems, job-retention schemes (including short-time work,
wage subsidies, and other similar schemes for self-employed) had contributed significantly to stabilizing
incomes. Such schemes preserved jobs and compensated workers for the reduction of working hours. They
stabilized 47 percent of the income shock on average (or about 60 percent the overall income stabilization, i.e.,
47 out of 78 percent), far exceeding other components such as personal income taxes or social insurance
contributions (Figure 6). Job-retention schemes are more impactful in countries where workers receive a higher
compensation rate for hours not worked, such as in Czech Republic, Denmark, and Slovak Republic
(OECD 2021). As the EU countries only experienced a modest rise in unemployment rates during the
pandemic, income stabilization from unemployment income support compensated for a mere 9 percent of the
adverse shocks.
Figure 6. Income Stabilization after the COVID Shock, by Country and Income Level
(Percent, share of the shock absorbed by the tax-benefit system)
1. By country 2. By household income quintile
110 100 Personal income taxes Social insurance contributions
Personal income taxes Social insurance contributions
Job retention schemes Unemployment benefits
100 Job retention schemes Unemployment benefits 90
Other benefits
90 Other benefits
80
80
70
70
60
60
50
50
40
40
30 30
20 20
10 10
0 0
-10 First quintile Second quintile Third quintile Fourth quintile Fifth quintile
(lowest income) (highest income)
MLT
EU
POL
FIN
ITA
GRC
CYP
ESP
SWE
IRL
PRT
CZE
LVA
EST
HUN
FRA
AUT
ROU
SVK
BGR
SVN
HRV
LUX
NLD
LTU
BEL
DNK
Source: IMF staff calculations.
Note: Estimates are based on the EUROMOD I.40+ and microdata from the 2019 EU-SILC (excluding Germany). Labor
market shock is simulated to replicate the 2020 labor market conditions using Labor Market Adjustment (LMA) Add-On. The
chart reports the average stabilization coefficients for EU countries and each household income quintile. Income quintiles
are calculated at the country level based on the household’s market income in the “no COVID-19” scenario. Job-retention
schemes include compensation received by employees on short-time work schemes, wage subsidies, as well as similar
schemes for self-employed. Other benefits include social assistance and housing benefits, family and education benefits,
health and disability benefits. Pensions are excluded.
10
Christl and others (2022) find that the European tax-benefit systems absorbed about 75 percent of the market income shock at
the EU level during the pandemic in 2020, of which the monetary compensation (job-retention) schemes played a major role. They
also estimated that consumption was largely stable (at 90 percent of the shocks) based on the marginal propensity to consume
proxied by the likelihood of liquidity constraints before the pandemic. However, during the pandemic, consumption was restricted
not just because of liquidity constraints or income deterioration but also because of lockdown restrictions and social distancing
requirements. This suggests the stabilization of consumption cannot be easily estimated based on the pre-pandemic parameters.
INTERNATIONAL MONETARY FUND 12
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Our estimates of the income stabilization coefficients
Figure 7. Correlation of Simulated Income
have a strong correlation with the actual data.
Stabilization Coefficients and Actual Change in
Countries with a strong income stabilization during the Per-Capita Consumption, 2019-20
pandemic tend to experience a smaller decline in per- (Percentage points)
capita real disposable income and real per-capita 0
private consumption expenditure at the aggregate -2
DNK
SVK
SWE LTU
Real per capita consumption % change
level. It suggests some evidence of correlation, not EST
-4 FIN
necessarily a causation (Figure 7).
y = 0.1163x - 11.605
-6 CYP
POL PRT LVA R² = 0.1687
Simulations also allow an assessment on how job- ROU NLD
FRA
GRC SVN
-8
retention schemes can stabilize income across income BEL LUX
AUT HUN
groups. For each country, we calculate the income -10
ITA CZE
IRL
stabilization coefficients for five income groups -12 ESP
(quintiles) according to the household market income in MLT
-14
the “no-COVID19” scenario. Results show that the tax 0 10 20 30 40 50 60 70 80
Change in income stabilization coefficient
and benefit systems, together with the pandemic-
Source: Eurostat and authors’ estimates.
related support measures, stabilized household Note: Consumption growth corresponds to the percentage
incomes more strongly for lower-income households. change in real final consumption expenditure of households
and non-profit institutions serving households in per capita
They absorbed 88 percent of the income shock during terms between 2019 and 2020. The change in income
the pandemic for lower-income households, more than stabilization coefficient is the difference between coefficients
simulated during the pandemic and pre-pandemic levels.
for households at the top quintile (72 percent of the
income shock) (Figure 6, panel 2). 11 Job-retention schemes account for the bulk of the overall stabilization
Figure 8. Income Stabilization During the Pandemic across Households
(Percent, share of the shock absorbed by the tax-benefit system)
1. Income stabilization coefficient 2. Contribution of job-retention schemes
120
120
100
100
80
80
60
60
40
40
20
20
0
0
First quintile Second quintile Third quintile Fourth quintile Fifth quintile
First quintile Second quintile Third quintile Fourth quintile Fifth quintile
(lowest income) (highest income)
(lowest income) (highest income)
Source: Authors’ calculations.
Note: Estimates are based on the EUROMOD I.40+ and microdata from the 2019 EU-SILC (excluding Germany). Labor
market shock is simulated to replicate the 2020 labor market conditions using Labor Market Adjustment (LMA) Add-On. The
box-whisker shows the variation across EU countries, with the median level (marker), interquartile range (box), and 5 th and
95 th percentiles (whiskers). Income quintiles are calculated at the country level based on the household’s market income in
the “no COVID-19” scenario. Job-retention schemes include compensation received by employees on short-time work
schemes, wage subsidies, as well as similar schemes for self-employed.
11
Similarly, the temporary expansion of unemployment income support in the United States was more progressive with most
benefits accruing to low-income workers (Ganong and others 2022).
INTERNATIONAL MONETARY FUND 13
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
across the board for all income groups, particularly for lower-income households compensating 64 percent of
their income shocks. They contributed to the overall income stabilization much more than other components in
the tax and benefit system. These results are robust across countries and household income groups,
suggesting that job-retention schemes were impactful to protect household income and well targeted as they
stabilized income more strongly for lower-income households (Figure 8). 12
Fiscal measures during the pandemic helped protect households against large income losses. They also
played a redistributive role by mitigating the rise in income inequality and protecting people’s livelihoods.
Simulations suggest that the pandemic could have led to a rise in the market income inequality—measured by
Gini coefficients—by 0.65 percentage points as lower-income households were more disproportionately
affected during the pandemic. Without job-retention schemes, the disposable income inequality could have
risen by 0.38 percentage points compared to the “no COVID-19” scenario (Figure 9). 13 But once accounting for
the pandemic-related fiscal support, the tax and benefit systems were able to protect household incomes,
especially of lower-income groups. Inequality in disposable income during the pandemic had decreased by
0.24 percentage points on average. The impact of job-retention schemes on disposable income inequality
varies across countries, reflecting differences in the design of job-retention schemes, the differences of the
pandemic impact, and the heterogeneity of labor markets.
Figure 9. Redistribution Effects of Fiscal Measures During the Pandemic
(Change in Gini coefficients, percentage points)
1. Impact of the COVID Shock on Inequality 2. Impact of Job-Retention Schemes on
Disposable Income Inequality
3.0 2.5
Disposable income inequality
Market income inequality
2
2.0
1.5
1.0
1
0.0
0.5
-1.0 0 NLD
EU
HUN BGR
EU
HRV
NLD
POL
ROU
CYP
DNK FIN
HUN
POL
SWE
LVA
DNK
IRL
SWE
BGR
BEL
CZE
PRT
FRA
FIN ITA
CYP
EST
CZE
PRT
SVN
LVA
SVN
EST
GRC
LTU
ESP FRA
ROU
ESP
HRV
LTU
GRC
BEL
LUX
ITA
SVK
MLT
IRL
AUT MLT
AUT
SVK
LUX
Source: Authors’ calculations.
Note: Estimates are based on the EUROMOD I.40+ and microdata from the 2019 EU-SILC (excluding Germany). Labor
market shock is simulated to replicate the 2020 labor market conditions using Labor Market Adjustment (LMA) Add-On.
Panel 1 shows changes in Gini coefficients of market (disposable) income following the COVID shock (between “no
COVID-19” scenario and the pandemic scenario). Panel 2 shows the change in Gini coefficients of disposable income
under the pandemic scenario with and without job-retention schemes.
Certain workers were more vulnerable to adverse shocks. For example, workers in the contact-intensive
sectors were exposed to greater job losses during the pandemic. A strong and effective job-retention scheme
should have stabilized income for workers who are more vulnerable. Our simulation results compare the impact
of job-retention schemes across different worker groups and assess if the stabilization effects are greater for
12
In a few countries, the job-retention schemes stabilized more than 100 percent of the income shocks for the lowest income
quintile because some schemes provided a lump sum support and may have more than compensated for the income losses for
households earning very little income.
13
The rise in disposable income inequality would also be affected by progressive income taxes and other means-tested benefits,
which partly compensate for the decline in income even without the job-retention schemes.
INTERNATIONAL MONETARY FUND 14
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
vulnerable workers, such as young workers and those with low skills. We classify workers according to their
age (18-24 years old, 24-55 years old, and 55-64 years old), gender, and education attainment (low-skilled and
high skilled). Income stabilization coefficients are calculated for all individuals of the working age (aged 15-64)
from the corresponding group j:
𝑗𝑗 𝑗𝑗
𝐼𝐼𝐼𝐼𝐶𝐶 𝑗𝑗 = �1 − ∑ 𝑁𝑁
𝑖𝑖=1 ∆𝑌𝑌𝑖𝑖
𝑖𝑖=1 ∆𝑀𝑀𝑖𝑖 × 100 (2)
Figure 10. Simulated Income Stabilization Coefficients by Worker Groups and Sectors
(Percent, share of the shock absorbed by the tax-benefit system)
1. Average, by age group 2. Differences across countries
90
Personal income taxes Social insurance contributions 20
80 Unemployment benefits Job retention schemes
Other benefits 15
70
10
60
50 5
40 0
30
-5
20
-10
10
-15
0
Women relative to 15-24 relative to 55-64 relative to Low relative to
Women Men 15-24 25-54 55-64 Low High Men 25-54 25-54 High
Gender Age Education Gender Age Education
3. Average, by sector 4. Country variation, by sector
90 120
Personal income taxes Social insurance contributions
Unemployment benefits Job retention schemes
80
Other benefits
100
70
60 80
50
60
40
30 40
20
20
10
0 0
Trade,Transp. Industry Prof. serv Inf. & Comm. Public serv. Real estate Construction Finance Agriculture Trade,Transp. Industry Prof. serv Inf. & Comm. Public serv. Real estate Construction Finance Agriculture
Food & Arts & Ent. Food & Accom. Arts & Ent.
Accom.
Source: Authors’ calculations.
Note: Estimates are based on the EUROMOD I.40+ and microdata from the 2019 EU-SILC (excluding Germany). Labor
market shock is simulated to replicate the 2020 labor market conditions using EUROMOD Labor Market Adjustment (LMA)
Add-on. Panel 1 shows average stabilization coefficients across countries for each group. Job-retention schemes include
compensation received by employees on short-time work schemes, wage subsidies, as well as similar schemes for self-
employed. Other benefits include social assistance and housing benefits, family and education benefits, health and disability
benefits. Pensions are excluded. Low level of education corresponds to upper secondary or below. High level of education
corresponds to post-secondary and tertiary education. Panel 2 shows the variation across countries for each worker group.
Panel 3 shows the average coefficients for each sector across countries. The boxes in panels 2 and 4 correspond to the
interquartile range, the marker to the median, and whiskers to 5 th and 95 th percentiles.
Simulation results show that the income stabilization coefficients are on average stronger for workers aged
18-24 years old and those with lower educational attainment (Figure 10, panel 1). The tax and benefit systems,
including the job-retention schemes, absorbed about 75 percent of the income losses for the young workers
during the pandemic, compared to 70 percent of those aged 25 and older. Income stabilization was also
stronger for workers with lower education (absorbing over 72 percent of the income shock) and females (72
percent). Job-retention schemes played an important role across all worker groups, absorbing almost half of
the income shock for young workers and about 40 percent of the shock in other groups. Unemployment income
INTERNATIONAL MONETARY FUND 15
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
support also helped stabilize income during the pandemic, particularly for low-skilled workers. There is no
major difference in the stabilization effects of social insurance contributions across worker groups.
Differences across worker groups were observed in many countries (Figure 10, panel 2). Income stabilization
for females is stronger in over 80 percent of EU countries (by as much as 6 percentage points on average).
Similarly, majority of countries have their tax and benefit systems stabilizing income more for young workers
and those with lower levels of education.
The job-retention schemes also stabilized income more for workers in contact-intensive sectors that were
affected the most by lockdown restrictions and social distancing (Figure 10, panel 3). Income stabilization was
the strongest in trade, transport, food and accommodation sectors (absorbing 75 percent of income losses),
while the stabilization is lower for less contact-intensive sectors such as finance and agriculture. Country
variations remain large (Figure 10, panel 4).
C. Regression analysis
While the simulation results point to some differences across household income groups and worker
characteristics in terms of the effectiveness of the job-retention schemes, a regression analysis would help
assess whether those differences are statistically significant or not, controlling for other factors such as the size
of pandemic-related support, and pre-existing social protection systems. The regression specification follows:
𝐼𝐼𝐼𝐼𝐼𝐼𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
𝑖𝑖,𝑐𝑐 = α + 𝑋𝑋𝑖𝑖,𝑐𝑐β + 𝑌𝑌𝑐𝑐γ + 𝐷𝐷𝑐𝑐δ + 𝜀𝜀𝑖𝑖,𝑐𝑐 (3)
where 𝐼𝐼𝐼𝐼𝐼𝐼𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶
𝑖𝑖,𝑐𝑐 is the income stabilization coefficient for an individual i from country c folllwing the pandemic
shock. The income stabilization coefficient for an individual i is calculated as 𝐼𝐼𝐼𝐼𝐶𝐶𝑖𝑖 = (1 − ∆𝑌𝑌𝑖𝑖⁄∆𝑀𝑀𝑖𝑖 ) × 100. 𝑋𝑋𝑖𝑖,𝑐𝑐 is
a vector of individuals’ characteristics, such as age, gender, the level of education, contact intensity of the
sector that individual works, and the market income quintile the individual belongs to; 𝑌𝑌𝑐𝑐 is a vector of county-
specific macro-fiscal variables, including allowance of the job-retention scheme (i.e. percent of lost income that
a worker receives for hours not worked), the change in cyclically adjusted primary deficit between 2019 and
2020 in percent of potential GDP, a net replacement rate in unemployment, and a percentage change in the
average number of hours worked per worker between 2019 and 2020. The specification includes a vector of
country dummy variables 𝐷𝐷𝑐𝑐.
Estimation results provide evidence that the tax and benefit systems, alongside with the pandemic-related
support, have stabilized income more strongly for female workers, workers employed in contact-intensive
industries and those with lower level of education. The corresponding coefficients are positive and statistically
significant (Table 2). The coefficients for those households in the lower-income quintiles tend to be higher,
suggesting the tax and benefit systems are able to stabilize their income more in face of an adverse shock.
At the same time, empirical results show that countries with higher allowance rates of the job-retention
schemes or more generous unemployment benefits tend to exhibit a stronger income stabilization. But the
income stabilization effects are weaker in countries that experienced a larger decline in working hours for
workers. Countries that have stronger counter-cyclical fiscal responses, as measured by the change in
cyclically adjusted primary balance, have greater income stabilization, suggesting discretionary fiscal support
can help stabilize income for individual households.
INTERNATIONAL MONETARY FUND 16
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Table 2. Regression Results on Differences of Income Stabilization across EU Households
Dependent variable: Income Stabilization Coefficient
(1) (2) (3) (4) (5)
Age between 15-24 1.953 1.602 1.594 -1.653 -0.772
Age between 55-64 -2.307*** -2.164** -2.164** -1.651** -1.769**
Female 4.447*** 4.027*** 4.087*** 1.740*** 1.918***
Education level, low 4.937*** 4.441*** 5.064*** 0.906 1.308
Contact-intensive 3.876*** 5.593*** 2.896*** 2.607***
Contact-intensive × Education level, low -2.383** -0.597 -1.462
Market income, the lowest quintile 3.713 3.939
Market income, 20th to 40th percentile 3.813** 3.867*
Market income, 60th to 80th percentile -3.605** -3.253*
Market income, the top quintile -11.40*** -11.13***
Job retention scheme allowance, percent of lost income 0.495***
Change in cyclically adjusted primary deficit 2020 2.378***
Net replacement rate in unemployment, percent of previous income 0.407***
Hours per worker, percentage change in 2020 0.372***
Numer of country dummies 25 25 25 25 16
Constant 80.70*** 80.17*** 79.76*** 85.26*** 14.67***
Observations 48,945 48,945 48,945 48,945 41,365
Number of countries 26 26 26 26 21
Source: Authors’ estimates.
The table reports results of the pooled ordinary least squares estimation. The dependent variable is the individual’s income
stabilization coefficient estimated based on simulations in the EUROMOD I.40+ and microdata from the 2019 EU-SILC
(excluding Germany). Labor market shock is simulated to replicate the 2020 labor market conditions using EUROMOD
Labor Market Adjustment (LMA) Add-on. Income quintiles used to construct dummy variables are calculated at the country
level based on the individual’s market income in the “no COVID-19” scenario. Low level of education corresponds to upper
secondary or below. Contact intensive sectors include trade, transport, food and accommodation, professional services, arts
and entertainment. Standard errors are clustered at the country level. ***p<0.01, **p<0.05. *p<0.1
V. Policy Implications and Conclusions
Diverse and forceful fiscal responses during the pandemic opened new grounds to support households against
large income or job losses. The preceding analyses on job-retention schemes provide some takeaways that
can inform the policy design.
The use of job-retention schemes across the EU helped prevent widespread job losses and stabilized
household incomes when people suffered from a severe shock. Such schemes have proved to be timely,
effective, and well-targeted—providing significant income stabilization in general and particularly to those
workers that are vulnerable to job and income losses. Our microsimulation approach points to the evidence that
job-retention schemes during the pandemic absorbed nearly 80 percent of market income shocks—almost
doubling the extent of the automatic stabilization of the pre-pandemic tax and benefit systems. In the absence
of those schemes, unemployment rates in the EU could have risen by additional 3 percentage points and
income inequality could have deteriorated further. Empirical results also show strong evidence that job-
retention schemes were well-targeted, with stronger income stabilization of vulnerable households, such as
lower-income families, youth, and low-skilled workers, after controlling for other factors.
Job-retention schemes are complementary to the unemployment income support because they work on
different margins (working hours versus unemployment) and tend to insure different types of workers
(Giupponi, Landais, and Lapeyre 2022). Both policies can provide a timely buffer and cushion the loss of labor
INTERNATIONAL MONETARY FUND 17
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
income. In case of a severe shock, governments are wary of the risks of massive layoffs, which could
undermine the valuable employer-employee relationships, especially in countries with more rigid labor markets
that are less capable to absorb unemployed workers quickly or in countries where social safety nets are
inadequate. Job-retention schemes could help limit productivity losses from unemployment (IMF 2022).
Job-retention schemes can become a more prominent part of the resilience toolkit in response to adverse
shocks. Once the job-retention schemes are in place, they can be expanded or broadened depending on the
severity of the shocks, particularly to those who are not qualified or fall outside the regular unemployment
income support, such as workers who have not worked long enough for the unemployment assistance.
Experience during the pandemic shows that many governments can unwind the support through those
schemes once economic conditions improve, as the take-up rates of those schemes have quickly returned to
the pre-pandemic low levels. Nonetheless, it is crucial to link the generosity of those schemes to economic
activity and incentivize workers and firms to return to normal working hours (Ando and others 2022). Job-
retention schemes are best used in situations when the adverse shocks are deep but pose short-lived
disruptions to labor markets. If the adverse shock turns out to be more persistent, preserving jobs through job-
retention schemes would hinder necessary reallocation, and policies should gradually transition from protecting
jobs to supporting workers and facilitating job-to-job transitions.
INTERNATIONAL MONETARY FUND 18
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Annex I. Simulations of Shocks in EUROMOD
The annex describes how the simulations are conducted based on the household level data and the tax-benefit
systems for EU countries. The simulations are done in EUROMOD version I4.0+ for all EU countries except
Germany and United Kingdom.
The EUROMOD is a tax and benefit microsimulation model that simulates individual and household tax
liabilities and benefit entitlements according to the policy rules in place in each country. It calculates, in a
comparable manner and based on representative micro-data on individuals and households drawn from
national household income surveys, the static effects of the tax and benefit system on household incomes for
each country. EUROMOD aims to simulate as much as possible of the tax and benefit components of
household disposable income, includng income taxes, social contributions, family benefits, housing benefits,
social assistance and other income-related benefits. Some instruments, such as contributory benefits and
pensions, are not simulated due to lack of information on previous employment and contribution history in the
cross-sectional surveys and are taken directly from the data. Please see Sutherland and Figari (2013) for the
detailed description of the EUROMOD model.
Two scenarios are simulated to assess the size of income stabilization prior to the pandemic. The baseline
scenario uses the 2019 EU-SILC household-level microdata and the 2019 tax and benefit policies. In the
scenario of a uniform market income shock, the simulation uses a negative 5-percent reduction of market
income across all households with nonnegative income. We then compute changes in market and disposable
incomes between the baseline and the uniform market income shock scenario for each household and
calculate the income stabilization coefficients for each country. To estimate income stabilization across
households’ income distribution, income stabilization coefficients are calculated for every income quintile of
each country.
Annex Table I.1. Uniform Market Income Shock
Input microdata
Input microdata 2019 SILC & 5%
2019 tax-benefit 2019 tax-benefit
2019 SILC negative market
policy rules policy rules
income shock
Baseline Income shock
Note: M stands for market income; T(M) includes direct income taxes and social insurance contributions payable by
households, and B(M) is the benefits accrued to households. Y is the disposable income, which is market income net of
taxes and transfers (Y=M-(T(M)-B(M)). Variables with an apostrophe denote the shock scenario.
We simulate two hypothetical scenarios to gauge the size of income stabilization during the pandemic. The “no
COVID-19” scenario uses the 2019 EU-SILC household-level microdata and the 2020 tax-benefit policies. The
“COVID-19” scenario is based on the 2020 tax-benefit policies and the 2019 EU-SILC microdata adjusted to
match the 2020 labor market conditions. The Labor Market Adjustment (LMA) Add-on is used to adjust and
simulate transitions between employment, unemployment and job-retention schemes based on the data from
the European Labor Force Survey and other detailed administrative data. We then compute changes in market
and disposable incomes between the “no COVID-19” and “COVID-19” scenarios and calculate the income
stabilization coefficients for each country. To estimate income stabilization across households’ income
distribution and worker groups, income stabilization coefficients are calculated for every income quintile and
worker group of each country, respectively.
INTERNATIONAL MONETARY FUND 19
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Annex Table I.2. COVID-19 Shock
2020 tax-benefit Input microdata 2020 tax-benefit Input microdata
policy rules 2019 SILC policy rules 2019 SILC
LMA add-on
No COVID-19 Adjust microdata to
account for 2020 labor
market conditions
COVID-19
Note: M stands for market income; T(M) includes direct income taxes and social insurance contributions payable by
households, and B(M) is the benefits accrued to households. Y is the disposable income, which is market income net of
taxes and transfers (Y=M-(T(M)-B(M)). Variables with an apostrophe denote the shock scenario.
References
Aiyar, S., & Dao, M.C. (2021). The Effectiveness of Job-Retention Schemes: COVID-19 Evidence from the
German States. IMF Working Paper No. 21/242, International Monetary Fund, Washington, DC.
Ando, S., Balakrishnan, R., Gruss, B., Hallaert, J.J., Jirasavetakul, L.F., Kirabaeva, K., Klein, N., Lariau, A., Liu,
L.Q., Malacrino, D., Qu, H., & Solovyeva, A. (2022). European Labor Markets and the COVID-19
Pandemic: Fallout and the Path Ahead. IMF Departmental Paper 2022/004, International Monetary Fund,
Washington, DC.
Auerbach, A. J., & Feenberg, D. (2000). The Significance of Federal Taxes as Automatic Stabilizers. Journal of
Economic Perspectives 14 (3): 37–56.
Autor, D., Cho, D., Crane L.D., Goldar, M., Lutz, B., Montes, J., Peterman, W.B., Ratner, D., Villar, D., &
Yildirmaz, A. (2022). An Evaluation of the Paycheck Protection Program using Administrative Payroll
Microdata. Journal of Public Economics 211.
Basso, G., Boeri, T., Caiumi, A., & Paccagnella, M. (2020). The New Hazardous Jobs and Worker Reallocation.
CEPR Discussion Paper 15100, Center for Economic and Policy Research, London.
Blanchard, O., Philippon, T., & Pisani-Ferry, J. (2020). A New Policy Toolkit is Needed as Country Exit
COVID-19 Lockdowns. Policy Brief 20-8, Peterson Institute for International Economics, Washington, DC.
Carroll, C. D., Slacalek, J. & Tokuoka, K. (2015). The Distributions of Wealth and the MPC: Implications of New
European Data. American Economic Review 104 (5): 107–111.
Christl, M., De Poli, S., Hufkens, T., Ivaškaitė-Tamošiūnė, V., Leventi, C., Papini, A., & Tumino, A. (2021).
Simulating labor market transitions in EUROMOD: EUROMOD LMA Add-on and COVID-related policies.
Joint Research Centre, European Commission.
Christl, M., De Poli, S., Figari, F., Hufkens, T., Leventi, C., Papini, A., & Tumino, A. (2022). Monetary
Compensation Schemes during the COVID-19 Pandemic: Implications for Household Incomes, Liquidity
Constraints and Consumption across the EU. Working Paper 613, ECINEQ, Society for the Study of
Economic Inequality.
Coady, D., De Poli, S., Hernández, A., Papini, A., & Tumino, A. (2023). The Extent and Composition of
Automatic Stabilization in EU Countries. IMF Working Paper.
INTERNATIONAL MONETARY FUND 20
IMF WORKING PAPERS How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation
Approach for European Countries
Dolls, M., Fuest, C., & Peichl, A. (2012). Automatic Stabilizers and Economic Crisis: US vs. Europe. Journal of
Public Economics 96 (3-4): 279–294.
Drahokoupil, J., & Müller, T. (2021). Job retentions schemes in Europe. A lifeline during the COVID-19
Pandemic. European Trade Union Institute Working Paper, 2021.07.
European Commission (2017). Impact of Fiscal Policies on Income Distribution. Report on Public
Finances in EMU, 71-131.
Ganong, P., Greig, F., Noel, P., Sullivan, D., & Vavra, J. (2022). Unemployment Insurance. Recession
Remedies: Lessons from the U.S. Economic Policy Response to COVID-19, edited by Edelberg, W.,
Sheiner, L., & Wessel, D. Hutchins Center on Fiscal and Monetary Policy and the Hamilton Project,
Brookings Institution, Washington, DC.
Giupponi, G., Landais, C., and Lapeyre, A. (2022). Should We Insure Workers or Jobs during
Recessions? Journal of Economic Perspectives 36 (2): 29–54.
Institute for Social and Economic Research, University of Essex; Joint Research Centre, European
Commission, EUROMOD: Version I4.0+ (software), 2022.
International Monetary Fund (IMF). 2015. Fiscal Monitor. Now Is the Time: Fiscal Policies for Sustainable
Growth. Washington, DC, April.
International Monetary Fund (IMF). 2022. Fiscal Monitor. Helping People Bounce Back. Washington, DC,
October.
Krusell, P., Mukoyama, T., and Şahin, A. (2010). Labour-Market Matching with Precautionary Savings and
Aggregate Fluctuations. Review of Economic Studies 77(4): 1477-1507.
Mohl, P., Mourre, G., and Stovicek, K. (2019). Automatic Fiscal Stabilizers in the EU: Size and Effectiveness.
Economic Brief No. 045, European Commission, May 2019.
McKay, A., & Reis, R. (2016). The Role of Automatic Stabilizers in the US Business Cycle. Econometrica 84
(1): 141-194.
Organisation for Economic Co-operation and Development (OECD). 2020. Job Retention Schemes during the
COVID-19 Lockdown and Beyond. Paris: OECD Publishing.
Organisation for Economic Co-operation and Development (OECD). 2021. OECD Employment Outlook 2021:
Navigating the COVID-19 Crisis and Recovery. Paris: OECD Publishing.
Sutherland, H., & Figari, F. (2013). EUROMOD: the European Union tax-benefit microsimulation model.
International Journal of Microsimulation 6(1), 4-26.
INTERNATIONAL MONETARY FUND 21
How Effective were Job-Retention Schemes during the COVID-19 Pandemic? A Microsimulation Approach for
European Countries
Working Paper No. WP/2023/3
File and source
- File
- how-effective-were-job-retention-schemes-during-the-covid-19-pandemic.pdf
- Size
- 1,338,158 bytes
- SHA-256
- 66a5f87f517a3b63b917ee44d3d7751a5c7db4730915d3da6c11dcc4ae58b8cf
- Original
- www.imf.org