Une étude de la Commission européenne compare la générosité des systèmes de chômage

EUROPEAN
ECONOMY
Economic Papers 454 | May 2012

Benchmarking Unemployment
Beneﬁt Systems

Klara Stovicek and Alessandro Turrini

Economic and
Financial Aﬀairs

Economic Papers are written by the Staff of the Directorate-General for Economic and Financial
Affairs, or by experts working in association with them. The Papers are intended to increase awareness
of the technical work being done by staff and to seek comments and suggestions for further analysis.
The views expressed are the author’s alone and do not necessarily correspond to those of the European
Commission. Comments and enquiries should be addressed to:

European Commission
Directorate-General for Economic and Financial Affairs
Publications
B-1049 Brussels
Belgium
E-mail: Ecfin-Info@ec.europa.eu

This paper exists in English only and can be downloaded from the website
ec.europa.eu/economy_finance/publications

A great deal of additional information is available on the Internet. It can be accessed through the
Europa server (ec.europa.eu)

KC-AI-12-454-EN-N
ISBN 978-92-79-22975-6
doi: 10.2765/26084

© European Union, 2012

Benchmarking unemployment benefit systems

Klara Stovicek
Alessandro Turrini

Directorate General for Economic and Financial Affairs
European Commission

Abstract

This paper proposes a methodology for benchmarking unemployment benefits systems, with a
view to assess reform needs and priorities. The methodology permits to assess different
dimensions of unemployment benefit systems and to consider alternative relevant
benchmarks. Looking at all relevant dimensions allows to better gauge how unemployment
benefit systems perform in terms of their multi-faceted objectives (such as income support
and stabilisation, incentives to take up work) and to have a more thorough assessment of each
objective. Comparisons with alternative benchmarks offer the possibility of assessing against
more meaningful country comparators, which take into account similarities in terms of
economic fundamentals, institutions and policy settings. The methodology is applied to EU
countries and results are discussed.

JEL classification: J65, J68, H20, H53

Keywords: Unemployment insurance, unemployment assistance, tax and benefit policies,
benchmarking, flexicurity.

The views expressed in this paper are those of the author and do not reflect necessarily those of the European
Commission. We thank Anne Bucher, Alfonso Arpaia, Santiago Loranca for helpful comments. Adam Kowalski
provided excellent statistical assistance.

3

1. Introduction

Unemployment benefits are a key instrument to deal with labour market risks. They permit to
insure individual incomes during temporary unemployment and provide assistance during
longer unemployment periods. From a macroeconomic perspective, unemployment benefit
systems perform a role of automatic stabilisation, thereby contributing to smooth aggregate
shocks. The main efficiency cost of unemployment benefits are reduced incentives to work as
a result of moral hazard which induces sub-optimal job search intensity (e.g. Mortensen,
1997; Shavell and Weiss, 1979), unemployment and inactivity traps linked to the tax and
benefit system (Carone at al., 2003), benefit dependence associated with long benefit duration
and falling employability over the unemployment spell.1

With unemployment rates becoming persistently high in most European countries after the
crisis, and with a lengthening of unemployment duration, reforms are becoming necessary to
ease the trade-off between income smoothing and incentives to take up jobs. A number of
elements in the design of unemployment benefit systems help easing such trade-off, including
for what concerns the level and time profile of unemployment benefit replacement rates or the
implementation of adequate activation policies aimed at strengthening job search effort. As
unemployment benefit systems differ considerably across countries, and since they operate
within heterogenous labour market and institutional contexts, desirable and feasible reform
paths may differ considerably from one country to the other. Moreover, notably in Europe and
in the current context, reforms need to carefully balance social protection, incentives for the
unemployed and fiscal costs.

Benchmarking of existing policies and institutions presents a necessary step to assess reform
needs and define desirable reform paths in individual countries. This paper proposes a
benchmarking methodology that permits to take into account a number of relevant dimensions
of unemployment benefit systems and that goes beyond the simple comparison with cross-
country summary statistics by taking into account similarities across countries in terms of
economic fundamentals and policy settings. The methodology is applied to EU countries

1
Findings from a vast empirical literature corroborate the expectation that more generous UI benefits are
associated with higher incidence of unemployment and longer periods in unemployment (e.g. Layard et al, 1991;
Nickell, 1998; Krueger and Meyer, 2002). In addition, a reduction in the maximum benefit duration is related to
shorter unemployment spells (e.g., Krueger and Meyer, 2002 Van Ours and Vodopivec, 2005; Lalive, 2008;
Caliendo et al., 2009).

4

using information from the European Commission – OECD Tax and Benefits project and
covers to the period over 2001-2010.

The remainder of the paper is organised as follows. Section 2 describes a series of dimensions
of unemployment benefit systems relevant for a cross-country comparison. Section 3
describes the methodology based on different benchmarks and proposes a set of indicators
used for a cross-country comparison. Section 4 presents results, while section 5 concludes.

2. Comparing unemployment benefit systems: the relevant dimensions

Unemployment benefits systems consist of two main instruments: unemployment insurance
and unemployment assistance. Unemployment insurance aim at insuring individual incomes
during the unemployment spell and are typically based on an insurance principle.
Unemployment insurance benefits are payable to job losers that, within a certain reference
period, have completed a minimum period of employment or paid contributions.
Unemployment assistance aims at preventing unemployment-related poverty. It is means-
tested and based on welfare principle. It is usually paid either to the long term unemployed
with insufficient means who have exhausted their unemployment insurance benefits or to
those who failed to qualify for unemployment insurance benefits. To qualify for
unemployment assistance, the unemployed often do not need to have any
employment/contribution period or the required period is shorter than in case of
unemployment insurance benefits. It is relevant to distinguish between unemployment
assistance and unemployment insurance as these two instruments generally have different
implications on incentives to take up work.2

A series of dimensions need to be taken into account when assessing unemployment benefit
systems in a comparative perspective. The assessment of various dimensions allows
evaluating unemployment benefit systems in terms of effectiveness in ensuring income
support and effects on incentives to take up jobs.

2
Unemployment assistance is means-tested and available only to low-income unemployed. The unemployed
who exhaust their unemployment insurance benefits and are not entitled to unemployment assistance loose their
income support entirely. This strongly increases their financial incentives to find a job. In addition, in some
countries, job-search requirements differ depending on the scheme.

5

2.1. Effectiveness in ensuring income support

Entitlement conditions regulate access to unemployment benefits and are in this way essential
in ensuring an effective income support to the pool of the unemployed. The aim of
unemployment benefits is to provide consumption smoothing and income support to the
unemployed during their transition period to employment. The income smoothing property of
unemployment benefit systems, however, depends not only on their generosity but also on
entitlement conditions, i.e. the conditions on the employment and/or contribution period the
unemployed have to fulfil to qualify for benefits as well as sanctions in case of voluntary
unemployment. If effectively designed, entitlement conditions ensure that the benefit
coverage, i.e., the share of jobless people who are receiving unemployment benefits, is
sufficiently large to provide income protection for those in need, provide effective income
smoothing thereby playing a stabilisation role for the economy, while ensuring the
compatibility with fiscal constraints.

2.2. Effects on incentives to take up jobs

Depending on a design, unemployment benefits can raise a risk of low incentives to take up
work and benefit dependence. Unemployment benefit systems involve an inevitable trade-off
between income smoothing and economic efficiency. By providing income replacement,
unemployment benefits reduce incentives of unemployed to search and take up jobs. This is
relevant above all in the context of generous and long-lasting transfers which can create
unemployment traps and benefit dependence, thus entrenching long-term unemployment.

• Unemployment traps. Unemployment traps can be defined as cases of low net income
gain from taking up work from unemployment. The traps stem from an inappropriately
designed tax and benefits system and are usually associated with high net replacement
rates. Large unemployment traps may entrench unemployment above all of low wage
persons who tend to have higher net replacement rates and low-wage job prospects. As
low-wage persons are generally at higher risk of unemployment than high wage
persons and concern a larger share of unemployed the negative impact on incentives is
likely to be more pervasive. Therefore, the negative repercussions on effective labour
supply and potential output could be large.

• Inactivity traps. Inactivity traps can be defined as cases of low net income gain from
taking up work from inactivity. In this case, means-tested social assistance provides

6

income replacement; typically to the unemployed with an insufficient contribution
period, long term unemployed and hardly employable persons. Large inactivity traps
raise a risk of reducing attachment of persons from the labour market and boost
structural unemployment.

• Benefit dependence. A risk of benefit dependence increases with the benefit duration
and is stronger when replacement rates do not fall over the unemployment spell and
when job search and availability requirements (valid reasons for refusing job offers,
availability to participate in active labour market policy programmes, availability to
occupational and geographical mobility…) are absent or insufficient.

2.3. Generosity of unemployment benefit system

The total income support provided to the unemployed depends on the generosity of benefits
throughout the whole unemployment spell. The overall generosity of the unemployment
benefit system depends on the level of replacement rates of both unemployment insurance and
unemployment assistance, their duration, and their structure throughout the unemployment
spell. Section 3.2 describes the construction of a synthetic indicator that captures total income
support available to an eligible unemployed and that will be used in the following analysis.

3. Benchmarking unemployment benefits systems across the EU: methodology

3.1. Income support, impact on incentives

A series of dimensions of unemployment benefit systems in individual countries are
considered. The dimensions of unemployment benefit systems presented in the Section 2 can
be assessed by various indicators such as entitlement conditions, unemployment and inactivity
traps, benefit duration, time profile of net replacement rates and job search and availability
requirements. The source of these indicators is the European Commission-OECD Tax and
Benefit project based on the OECD Tax and Benefits models (Carone, 2003). A description of
legislation used for the computation of most parameters is available in the Tax and benefits
country chapters published on the OECD website.3 Depending on the dimension analysed, the
indicators refer to characteristics of the most appropriate income support instrument

3
http://www.oecd.org/document/3/0,3343,en_2649_34637_39617987_1_1_1_1,00.html.

7

(unemployment insurance to measure the risk of unemployment traps at the beginning of the
unemployment spell, social assistance to track the risk of inactivity traps, both unemployment
insurance and unemployment assistance to measure the extent to which benefits fall over time
and a related risk of benefit dependence).

• Entitlement conditions describe the conditions that determine initial entitlement to
unemployment insurance. The indicator is an index and provides a score between one
(least strict) and five (most strict). The aim of this indicator is to provide some
measure of unemployment benefit coverage.

• Unemployment traps are captured by an indicator that measures financial incentives to
take up work for individuals benefiting form unemployment insurance and at the
beginning of the unemployment spell. The indicator measures the marginal effective
tax rate (METR) on labour, namely the extent to which the gains from moving from
unemployment to work are reduced as a result of the presence of the tax and benefit
system. It usually ranges between 0 and 100; high values correspond to low incentives
to take up work.4

• Inactivity traps are captured by a similar indicator as that for unemployment traps, but
this indicator measures incentives to take up work from inactivity (after exhaustion of
unemployment benefits or with no entitlement), while benefiting from social
assistance if available.

• Duration of unemployment insurance benefits. The indicator is constructed with
reference to an unemployed who is 40 years old and has 22 years of contributions.

• Time profile of benefits is measured as the ratio of the average net replacement rates of
either unemployment insurance or unemployment assistance (whatever available)
between the second and fifth year in unemployment and that in the first year.

• Job search and work availability. The indicator describes the extent to which an
individual entitled to unemployment insurance must be available and actively search
for work to fully qualify as a benefit recipient, and accept suitable job offers. The
indicator ranges from 1 (least strict) to 5 (most strict).

4
The METR is computed as
METR = 1 − Δy net / Δygross , where Δy net and Δygross are, respectively the

change in net and gross income resulting from moving from unemployment to work. It is visible that
unemployment traps may exceed 100 in the extreme case where taking up work reduces net disposable income.

8

The following benchmarks are chosen for a comparison of the various dimensions of EU
unemployment benefit systems:

• EU average. It is a standard summary statistics used for benchmarking of EU
countries. It has been widely used in the context of the Lisbon Assessment Framework
(LAF) and the Joint Assessment Framework (JAF). The idea is to compare various
dimensions of the unemployment benefit system in individual countries on the basis of
the EU average.

• Average within relevant country groupings. An overall design of unemployment
benefit systems varies considerably across EU countries, depending inter-alia on
country specific labour market and welfare institutions. The unemployment benefit
system can comprise different passive (unemployment insurance, unemployment
assistance) and activation and active labour market policy instruments and could be
oriented towards either the insurance or the welfare principle.5 It is customary to
classify EU countries into groups characterised by relatively similar combinations of
welfare and labour market policies and institutions. The average for these homogenous
groups lends itself as an alternative benchmark for the assessment of unemployment
benefits. This paper will use the taxonomy proposed by European Commission (2007)
where countries are classified into five groups on the basis of their flexicurity model.6
The taxonomy permits to identify relatively homogenous country groups in terms of
the characteristics of their unemployment benefit systems. Table 1 reports relevant
summary statistics by country groups.

Nordic countries (Denmark, Finland, the Netherlands, Sweden). These
countries are characterised by a highly generous unemployment benefit system
both in terms of entitlement conditions and income support per unemployed.
Long benefit duration is often coupled with high net replacement rates, in

5
The insurance principle aims at consumption smoothing, contributions are the main source of financing, and
benefits are strongly related to previous earnings. Conversely, the welfare principle aims at redistributing
income, financing is by means of flat-rate contributions or a general tax financing, and benefits are not strongly
linked to past income (i.e. benefits are capped at relatively low levels).
6
The taxonomy proposed by the European Commission (2007) largely confirms the findings of the seminal
work on the taxonomy of socio-economic systems carried out by Esping and Andersen (1990). It covers 22 EU
countries which are classified into five groups on the basis of principal component analysis. The five missing EU
countries were for the purpose of this paper allocated on the basis of unemployment benefit systems
characteristics as follows: Malta and Cyprus were allocated to the Anglo-Saxon group of countries, Luxembourg
to the Continental group, Romania and Latvia to the Central, Eastern group of countries.

9

particular within the first two years of unemployment. The benefit system is
highly redistributive, and caps on maximum benefits reduce generosity of
benefits for higher wage persons. To keep work incentives high, activation and
active labour market policies have a prominent role and the conditions on job
search and work availability are very strict.

Continental countries (Austria, Belgium, Germany, France, Luxembourg).
Continental countries generally have a generous unemployment insurance
system, though benefit duration is generally shorter and net replacement rates
lower than in Nordic countries. Unemployment assistance often complements
income support. The benefit coverage is large and ensures that a large pool of
the unemployed receive benefits. The risk of unemployment and inactivity
traps and benefit dependency is addressed by activation and active labour
market policies.

Anglo-Saxon countries (Ireland, the UK, Malta, Cyprus). In Anglo-Saxon
countries, unemployment insurance benefits are relatively modest, while
unemployment assistance is a relevant means-tested instrument to provide
income protection to the unemployed. Expenditure on means tested full time
unemployment benefits often exceeds non-means tested unemployment
benefits. While unemployment traps are not a major issue, benefit dependency
is a risk on account of nearly flat-rate and long-lasting unemployment
assistance. To counter benefit dependency, monitoring of job-search activity is
strict whilst active labour market policies play a less important role.

Southern countries (Spain, Italy, Portugal). The generosity of unemployment
insurance generosity varies widely depending on age and contribution period
of the unemployed. Eligibility rules for unemployment insurance benefits are
relatively strict. Unemployment assistance is of little relevance, as well as the
risk of inactivity traps. Activation policies play a relatively minor role, while
making the continuation of provision of unemployment insurance conditional
to the participation in active labour market policies is more widespread.

Central and Eastern countries (Greece, Bulgaria, Czech Republic, Estonia,
Hungary, Lithuania, Poland, Slovakia, Slovenia, Romania, Latvia). This group
of countries tends to exhibit a tight unemployment benefit system both in terms
of benefit support per unemployed and benefit coverage. Although

10

replacement rates at the beginning of the unemployment spell can be high in
some cases, benefits drop sharply after the first year of unemployment. While
unemployment assistance plays a minor role, the unemployed can draw on
substantial means-tested income support provided by housing and social
benefits. Relatively strict conditions on job search and active participation of
labour market policies often apply.

3.2. Generosity of unemployment benefit systems

A synthetic indicator is built to capture the overall available income support per unemployed.
The overall generosity of the unemployment benefit system depends not only on the level of
replacement rates of both unemployment insurance and unemployment assistance, but also on
their pattern and evolution throughout the unemployment spell. With a view of capturing
these features, a synthetic indicator, new in the existing literature, is computed as follows.

k
UBgenerosity = ∑ nrrUI , i * durationUI , i + nrrUA * durationUA
i =1 , (1)

where nrr stands for net replacement rate, UI and UA at the pedix of variables denote,
respectively, unemployment insurance and unemployment assistance, the index i refers to the
different replacement levels for unemployment insurance over the unemployment spell.

The indicator corresponds to the sum of all benefits received during the unemployment spell
in terms of previous labour earnings. The indicator measures ex ante the maximum potential
income support available to those unemployed that fulfil all eligibility criteria, e.g. fulfil the
criteria for job search. It can be interpreted as the number of months of previous job income to
which the unemployed are entitled. All necessary parameters for its computation, notably net
replacement rates over the unemployment spell and duration of both unemployment insurance
benefits and unemployment assistance are taken from the joint European Commission-OECD
Tax and Benefit project. Note that the actual, ex post generosity depends on the effective
expenditure per unemployed, which is determined also by policy-driven factors, notably job
search requirements, as discussed in the previous section of the paper. There is no obvious
way to take into account cross-country differences in generosity linked to job search
conditions into a synthetic indicator, but this is an element that deserves to be considered in
the overall assessment.

11

The overall generosity of unemployment benefits is assessed also on the basis of a benchmark
which takes into account country-specific labour market institutions and policies. The overall
generosity of unemployment benefits depends on a series of factors, including the need to
provide sufficient income support, the risk of depressing incentives to take up jobs, fiscal
constraints. Hence, in addition to simply comparing against the EU average or the average
within a homogenous country groups, a more sophisticated benchmark will be used, which
permits to take into account these differences. The benchmark is the prediction obtained from
cross-country regressions on the following variables (see the Appendix for details on the
econometric estimations and results): 7

• Real GDP per capita. Demands for income support during unemployment tend to be
stronger in countries with higher income per capita.

• Unemployment rate. A larger pool of unemployed corresponds to a larger
unemployment risk and lower probability of finding a job and raises demands for more
generous protection.8

• Long term unemployment rate. The larger the number of the unemployed that are
already in unemployment for more than one year, the higher the risk that generous
benefits create a problem of benefit dependence, thus depressing incentives.

• Government budget balance. A surplus in the government budget allows for more
room to increase a state share of financing of unemployment benefits.

• Expenditures on activation policies. Well-designed activation policies and strictly
enforced job search conditionality can effectively offset disincentives stemming from
more generous benefit support.

7
The explanatory variables do not include some relevant labour market institutions such as employment
protection legislation due to short time span (the data on employment protection legislation are available until
2008) and incomplete country coverage (the employment protection legislation is not available for the EU
countries which are not OECD members).
8
This result suggests the counter-cyclical adjustment of unemployment benefits, i.e. longer benefit duration and
higher net replacement rates during downturns when unemployment rate increases.

12

4. Benchmarking unemployment benefits systems across the EU: results

4.1. Income support and incentives

Results are reported in Table 2. The information reported is for the latest available years. It is
important to note that the data do not permit to capture the effects of reforms that become
operational after 2010. A synthesis for year 2010 by means of radar graphs is provided in
Graphs 5 to 10. A few words of caution are warranted when interpreting these graphs. First,
all variables are re-scaled to values between zero and 100.9 Second, all variables are presented
in such a way that a higher value (or a larger radar perimeter) correspond to more generous
unemployment benefit systems. To this purpose, the entitlement and the job search indexes
were adjusted in such a way that a higher value implies less strict conditions.10

A glance to radar graphs (Graphs 5 to 10) allows for a quick comparison of the various
characteristics of unemployment benefit systems to EU and country group averages. This
simple comparison helps assessing unemployment benefits systems according to the main
relevant dimensions, thereby permitting a comprehensive view. For instance, for the case of
Denmark, it stands out quite clearly that this country is characterised by a relatively generous
system regardless of the benchmark, but that it also has in place instruments to fight a high
risk of benefit dependence.

A few countries stand out as having particularly generous systems along all dimensions, while
some countries appear to be characterised by strict systems in all respects. A cursory
inspection of graphs quickly reveals that Austria, Belgium, Denmark, Finland, are more
generous that the EU average in most dimensions. The opposite holds for Estonia, Italy,
Poland, Romania and Slovakia. When compared with respective country groups, Finland
however appears more in line with the Nordic average.

Concerning the benchmarking of each unemployment benefits system characteristic taken in
isolation, results can be summarised as follows:

9
This is done in such a way that the minimum value of each variable is set to zero and the maximum value to
100. The values between the minimum and the maximum value are re-scaled as follows: Rescaled Value =
(Original value - Minimum) / (Maximum - Minimum).
10
The usual interpretation of the entitlement conditions index and the job search and work availability index is
such that a higher value of the index means more strict conditions and less generous benefit system. By taking an
inverse of both indices (1/index) the interpretation could be reversed. Overall, low values of these indices, high
unemployment and inactivity traps, long benefit duration and a small decline of net replacement rate over the
unemployment spell suggest that the benefit system is generous and the respective radar graph large.

13

• Entitlement conditions. In Austria, Denmark, Cyprus and the Czech Republic income
support is provided to a large pool of unemployed. Conversely, in Italy and Portugal
entitlement conditions are particularly strict.11 Entitlement conditions are also tight in
most Central and Eastern European countries, with the notable exceptions of the
Czech Republic and Latvia. Belgium has entitlements broadly in line with the EU
average, which are however relatively strict as compared with the average for
Continental countries.

• Unemployment traps. In Denmark, Luxembourg, and Latvia marginal effective tax
rates are relatively high, exceeding 80%. However, it also appears that marginal
effective tax rates in Denmark and Luxembourg are not deviating substantially from
the averages of Nordic and Continental group respectively. An opposite consideration
applies to countries like Ireland and Sweden: the risk of unemployment traps for these
countries appears to deviate more from their own country groups than with respect to
the EU average. Unemployment traps appear instead particularly low in Estonia,
Malta, Slovakia, the UK.

• Inactivity traps. The risk of inactivity traps appears high in Denmark, Finland,
Germany, the Netherlands, Slovenia when compared to the EU average. However,
while Finland, Germany, and the Netherlands are broadly in line with the average of
their respective country groups, Slovenia appears to deviate also in this respect,
together with few other Eastern European countries (Czech Republic, Latvia,
Lithuania). Conversely, relatively low risks of inactivity traps emerge in countries
without generous social assistance, notably Estonia, Italy and Greece, Spain, Portugal,
Romania, Bulgaria and Slovakia. Within their country groups, however, Spain and
Estonia appear broadly in line, and Portugal relatively generous.

• Duration. Duration of unemployment insurance benefits appears particularly long in
Belgium, Denmark and Portugal. Denmark, however, does not have duration much
higher than what is observed on average in Nordic countries.12 In contrast, benefit
duration is relatively short in Italy and in most countries belonging to the Anglo-Saxon

11
To qualify for benefits the unemployed should have in Italy two years of insurance and 52 weekly
contributions during the last two years while in Portugal 450 days of employment record in the last 2 years.
12
The assessment depends largely on the fact that Denmark in 2010 halved the duration of unemployment
insurance from 48 to 24 months.

14

group and the Central and Eastern group.13 The short duration of unemployment
insurance benefits in Anglo-Saxon countries is explained by the availability of means-
tested unemployment assistance after the exhaustion of unemployment insurance.
Austria also has a relatively short unemployment insurance benefit duration, but it
provides earnings-related, means tested unemployment assistance of indefinite
duration. Duration of unemployment insurance benefits in Austria, Sweden,
Luxemburg and Germany is closer to EU average than to their own group averages.14

• Time profile of replacement rates. In Ireland, Malta and the UK unemployment
insurance benefits are fully flat, and followed by a flat rate unemployment assistance
of indefinite duration. Net replacement rates decline only marginally also in Belgium
and Austria. Conversely, Cyprus (an exception within the Anglo-Saxon group), Italy,
Luxembourg and Latvia do not operate unemployment assistance systems and have a
strong drop in net replacement rates due to short unemployment insurance duration.

• Job search conditions. Belgium, Bulgaria, Greece have relatively lax job search and
availability conditions. In contrast, Germany and Denmark appear to have the most
stringent job search conditions in the EU. In both countries, strict conditions apply to
job refusals and generally require participants in active labour market policy measures
to be available for work and actively search for a job.15 In this respect, Germany
scores well above the average for the group of Continental countries.

Overall, benchmarking with respect to EU or homogenous country group averages matter, and
results are not always obvious. Benchmarking of individual dimensions of benefit systems to
group averages as opposed to EU average generally tends to reduce the dispersion in the value
of variables with respect to the benchmark. However, there are also cases in which
benchmarking with respect to country groups increases variation (see Table 4). This often
happens when variable values for single countries and country groups deviate from the EU
average in opposite directions.

13
Italy increased unemployment insurance duration in the last decade, yet the duration remains well below both
the EU average and the average of the Southern group.
14
The data do not permit to capture yet the recent reform in Hungary reducing the duration of unemployment
insurance to 3 months.
15
In both countries, unemployment benefit recipients are required to accept any job that they are capable of
doing, regardless of their previous occupation. After three months of unemployment long commuting times for

15

4.2. Unemployment benefit generosity

The design of unemployment benefits over the unemployment spell differs widely across EU
countries. Graph 1 compares the pattern of unemployment benefit replacement rates for
unemployment insurance and unemployment assistance over the unemployment spell across
EU countries. Net replacement rates drop to some extent in almost all EU countries, either due
to a reduction in unemployment insurance benefits or to the replacement of unemployment
insurance benefits with less generous unemployment assistance.16 Net replacement rates can
in some countries be quite high at the beginning of the unemployment spell and drop
substantially after a short period of time (e.g. Latvia, Luxembourg). In other countries, net
replacement rates can persist at a relatively high level for an extended period of time (e.g.
Belgium, Denmark, Portugal). Graph 1 also shows that unemployment insurance benefits
provide the most important income support in Belgium, Denmark, Portugal and Spain while
means-tested unemployment assistance is of main relevance for income protection in Austria,
Malta, Ireland and the UK.

The overall generosity of unemployment benefit systems varies widely across the EU and is
relatively stable over time, except in countries having carried out major reforms in the past.
Graph 2 displays the synthetic benefit generosity indicator (see section 3.2) for each country
since 2001. The graph clearly shows that the total available income support per unemployed
varies very widely across the EU, with countries with higher income per capita generally
providing more generous systems.17 Over past decade, benefit support per unemployed
appears relatively stable in most countries. However, some countries saw a substantial drop in
total benefit generosity (e.g. Germany, Denmark, Sweden, France, Slovenia) while some
increases were observed in Belgium, Ireland and the Netherlands. These visible changes are
associated with reforms carried out throughout the past ten years.18

job reasons are required in Denmark while in Germany the required salary for a suitable job falls with the
duration of unemployment.
16
It is not always the case that unemployment insurance has more generous net replacement rates than
unemployment assistance. Net unemployment assistance benefits exceed that of unemployment insurance in
Malta because unemployment insurance is taxed while this is not the case for unemployment assistance.
17
For computation purposes, indefinite unemployment insurance duration in Belgium and indefinite
unemployment assistance duration in Austria, Germany, Finland, Ireland, Malta and the UK is capped to 60
months. Therefore, for these countries, the total available benefit generosity is understated and should be in
principle unlimited as long as the unemployed fulfil the eligibility criteria.
18
In Germany a large drop in benefit support in 2005 followed the Hartz IV reform which merged the
unemployment assistance with social assistance and introduced a flat-rate instead of an earnings-related income

16

The simple comparison of unemployment benefit generosity with the EU average does not
permit to account for the fact that differences are to some extent related to structural
differences in economic fundamentals, institutions, policy settings. Appropriate benchmarking
helps taking into account such differences.

• As shown in Table 3, Belgium, Malta, Austria, Denmark, Ireland, Finland and
Portugal stand out as having more generous benefit systems than the EU average,
while the opposite holds for the Czech Republic, Lithuania, Slovakia, Italy, Poland,
Hungary, Latvia and Estonia.

• When generosity is compared to the average of more homogenous country groups,
deviations intuitively tend to narrow. In particular, all Central and Eastern European
countries that have much less generous benefit systems than the EU average do not
deviate much from their own group average. However, there are also cases in which
deviations from benchmark become more evident when referred to country groups.
For instance, Luxembourg, Sweden and the UK tend to have less generous benefits
when compared to their own group averages than when compared to the EU average.

• When the comparison is performed with respect to benchmarks predicted from
multivariate regressions (see section 3.2. and Appendix), it appears that once
differences in fundamentals are taken into account, new insights emerge (Table 3,
Graphs 3 and 4). While unemployment benefit systems in Malta, Belgium and
Portugal appear even more clearly highly generous, the evidence becomes weaker for
Denmark and Finland. The reason is that high income per capita, strong activation
policies, and sufficient fiscal space in these countries, justify more generous benefit
systems. These benchmarks strengthen the case that Luxemburg and Sweden are
characterised by relatively tight benefit systems, while for Slovakia, Greece, Poland
and Hungary results become more nuanced: benefit systems are still relatively tight,
but, in light of a combination of low income per capita, fiscal deficits, and high long-

support. Denmark, Sweden and France reduced its benefit support by cutting benefit duration in 2010, 2008 and
2003 respectively. In addition, Sweden introduced a declining profile of net replacement rates in 2007. A decline
of benefit support in Slovenia follows scrapping off the unemployment assistance in 2007. On the other hand,
Belgium saw a gradual increase in benefit support over time, largely due to a flattening of the benefit profile
over the unemployment spell. In Ireland, benefit support was steadily increased due to a solid rise in weekly
benefits. The Netherlands reduced substantially the maximum benefit duration from 60 to 38 months and
introduced a declining profile of benefit support in 2007. However, the benefit duration for a person with 22
years of contribution period increased from 18 to 22 months.

17

term unemployment, not as much as coming out from a simple comparison with EU
average.

5. Conclusions

The paper evaluates unemployment benefit systems in the EU using various benchmarks. The
data are taken from the European Commission – OECD Tax and Benefit Project and are
available over the period 2001-2010. The analysis considers a multiplicity of dimensions of
unemployment benefit systems - such as entitlement conditions, unemployment and inactivity
traps, benefit duration, time profile of replacement rates and job search and work availability.
Cross-country comparisons are performed not only on the basis of the EU average but also on
the basis of the average of a homogenous group of countries with similar labour market
institutions.

A synthetic measure of the total generosity of unemployment benefit systems is constructed
taking into account net replacement rates over the whole unemployment spell of both
unemployment insurance and unemployment assistance. This synthetic indicator is compared
with a benchmark estimated from multivariate cross-country regressions and takes into
account differences across countries in GDP per capita, labour market performance, fiscal
constraints and expenditures on activation policies.

A number of relevant results emerge as follows from the comparison of the different
dimensions of unemployment benefit systems:

• Belgium, Denmark, Portugal, Spain, Finland and the Netherlands appear to be
relatively generous in terms of unemployment insurance replacement rates and
duration compared with the EU average, while in the UK, Malta, Slovakia, Estonia,
Poland and Romania benefit conditions are relatively tight.

• Regarding entitlements conditions, they are particularly loose in Denmark, Austria,
Cyprus, while in Portugal, Italy, Slovenia, they are rather tight. As for job search
conditionality, it is loose in Belgium, Greece, Bulgaria, while tight conditions apply in
Germany, Denmark, Slovakia. While a decline in the net replacement rate is nearly
absent in Ireland, Malta, the UK, Austria and Belgium, the decline is rather abrupt in
Italy, Latvia, Luxembourg, Greece, Estonia, Poland, Slovenia and Lithuania.

18

• Within the EU, there are groups of countries with relatively homogenous benefit
systems. Nordic and Continental countries are characterised by relatively generous
unemployment benefit systems both in terms of entitlement conditions and income
support per unemployed. In both groups, activation and active labour market policies
have a prominent role, with job search conditionality being strong especially in Nordic
countries. In Anglo-Saxon countries, unemployment insurance benefits are relatively
modest, while unemployment assistance plays a major role. Monitoring of job-search
activity is strict whilst active labour market policies play a less important role. In
Southern countries, access to unemployment insurance is strict and benefit generosity
varies widely depending on age and contribution period. Activation policies play a
relatively minor role, while participation in active labour market policies is
widespread. Finally, Central and Eastern countries tend to exhibit a tight
unemployment benefit system both in terms of benefit support per unemployed and
benefit coverage. Although replacement rates at the beginning of the unemployment
spell can be high in some cases, benefits drop sharply over the unemployment spell.
Strict conditions on job search and availability often apply.

• Some countries that stand out as particularly generous in terms of replacement rates
and duration as compared with the overall EU average turn out instead broadly in line
when compared with the average for their respective groups. This is the case of
Finland, the Netherlands, Spain. Symmetrically, Malta, the UK, Poland, Romania and
Estonia which have tight benefits as compared to the whole EU appear broadly in line
with the other countries of their own groups.

The benchmarking of the synthetic measures of overall generosity of unemployment benefit
systems reveals a number of insights as follows:

• The overall generosity of unemployment benefit systems exhibits a high degree of
variation across EU countries, with Belgium, Malta, Austria, Denmark, Ireland,
Finland and Portugal having much more generous benefit systems than EU average
while opposite is the case in the Czech Republic, Lithuania, Slovakia, Italy, Poland,
Hungary, Latvia and Estonia.

• A somehow different picture emerges when the overall generosity of unemployment
benefit systems is benchmarked against the prediction from multivariate regressions
that take into account differences in fundamental factor affecting unemployment

19

benefit systems. While unemployment benefit systems in Malta, Belgium and Portugal
appear even more clearly highly generous, the evidence becomes weaker for Denmark
and Finland. The reason is that high income per capita, strong activation policies, and
sufficient fiscal space in these countries, justify more generous benefit systems. These
benchmarks strengthen the case that Luxemburg and Sweden are characterised by
relatively tight benefit systems, while for Slovakia, Greece, Poland and Hungary
results become more nuanced: benefit systems are still relatively tight, but, in light of
a combination of low income per capita, fiscal deficits, and high long-term
unemployment, not as much as coming out from a simple comparison with EU
average.

Overall, results confirm the usefulness of considering a wide range of elements when
comparing unemployment benefit systems and of referring to alternative benchmarks.
Looking at all relevant dimensions permits to better gauge how unemployment benefit
systems perform in terms of their multi-faceted objectives (income support and stabilisation,
incentives to take up work,…). Using alternative benchmarks offers the possibility of
assessing against more meaningful country comparators, which take into account similarities
in terms of economic fundamentals, institutions and policy settings.

20

References

Caliendo, M., K. Tatsiramos and A. Uhlendorff (2009), 'Benefit Duration, Unemployment
Duration and Job Match Quality: A Regression-Discontinuity Approach', IZA Discussion
Paper 4670, Institute for the Study of Labor (IZA).
Carone, G., A. Salomaki, H. Immervoll and D.Paturot (2003), "Indicators of Unemployment
and Low-Wage Traps”, OECD Social, Employment and Migration Working Paper No. 18,
OECD, Paris.
Esping-Andersen, G. (1990), The three worlds of welfare capitalism, Princeton, Princeton
University Press.
European Commission (2007), "Working time, work organisation and internal flexibility -
flexicurity models in the EU", Employment in Europe.
Katz, l. and Bruce D. Meyer (1990): 'The Impact of the Potential Duration of Unemployment
Benefits on the Duration of Unemployment,' NBER Working Papers 2741, National
Bureau of Economic Research, Inc.
Krueger, A., B., and B. D. Meyer, 2002. "Labor Supply Effects of Social Insurance," in A. J.
Auerbach & M. Feldstein (ed.), 2002. Handbook of Public Economics, Elsevier, edition 1,
volume 4, number 4, April.
Lalive, R. (2008), 'How do extended benefits affect unemployment duration? A regression
discontinuity approach', Journal of Econometrics 142, pp. 785–806.
Layard, R, S Nickell and R Jackman (1991), Unemployment: Macroeconomic Performance
and the Labour Market, Oxford University Press.
Mortensen, D. (1977), 'Unemployment Insurance and Job Search Decisions', Industrial and
Labor Relations Review 30, 505-517.
Nickell, S. (1998), "Unemployment: Questions and some answers", Economic Journal, 108:
802-16.
Shavell, S and L. Weiss (1979), "The Optimal Payment of Unemployment Insurance Benefits
over Time", Journal of Political Economy 87, 1347-1362.
Van Ours, J. C. and M. Vodopivec (2005), 'How changes in benefits entitlement affect the
duration of unemployment', Discussion Paper 30, Tilburg University, Center for
Economic Research.

21

Table 1. Unemployment benefits systems, activation and active labour market policies,
by groups of countries, averages over 2001-2010
Policy area* Variable* Nordic Continental Anglo- Southern Central
countries countries Saxon countries and
countries Eastern
countries
Unemployment benefit UI generosity_100% of AW 16.4 13.9 2.5 13.1 4.3
generosity (in months of UB generosity_100% of AW 19.7 22.8 16.4 14.5 4.6
previous net income) UI generosity_67% of AW 21.1 16.2 3.2 13.5 5.1
UB generosity_67% of AW 25.6 25.3 22.1 15.4 5.6

Passive labour market Full time UB 1.40 1.09 0.53 0.83 0.28
policies - expenditures (% Full time UB_non means tested 1.11 0.89 0.20 0.65 0.26
of GDP) Full time UB_means tested 0.28 0.20 0.33 0.19 0.02

Passive labour market UB 67 68 51 40 18
policies - coverage

Activation, active labour Labour market services (cat.1 - 0.22 0.18 0.18 0.09 0.07
market policies - activation)
expenditures (% of GDP) Total LMP measures (categories 2- 1.09 0.66 0.30 0.53 0.18
7 - ALMP)

Activation, active labour Labour market services (cat.1 - 23 28 7 0.08 23
market policies - coverage activation)
Total LMP measures (categories 2- 47 48 20 45 11
7 - ALMP)
Activation of registered 25 30 10 25 16
unemployed

Unemployment benefit Minimum UI duration 21.4 16.4 7.3 5.8 5.2
duration (in months) Maximum UI duration 36.4 30.8 8.9 23.9 11.7
UI duration if 22 years of a 27.9 23.4 8.9 18.7 9.9
contribution period
UA duration 36.9 37.5 60.0 15.4 7.5

Net replacement rates NRR_UB_year 1_average 0.66 0.70 0.46 0.64 0.45
over the unemployment NRR_UB_year 2_average 0.63 0.50 0.38 0.48 0.10
spell NRR_UB_year 5_average 0.14 0.40 0.38 0.05 0.08
NRR_UB+HB+SA_year 0.75 0.73 0.67 0.64 0.61
1_average
NRR_UB+HB+SA_year 0.74 0.66 0.66 0.48 0.46
2_average
NRR_UB+HB+SA_year 0.66 0.61 0.66 0.25 0.42
5_average

Unemployment and In_trap_single_67% of AW 0.80 0.66 0.60 0.34 0.46
inactivity traps In_trap_1earner couple_2 0.93 0.79 0.72 0.40 0.62
children_67% of AW
In_trap_2earners couple_2 0.46 0.36 0.32 0.27 0.35
children_67% of AWs
Un_trap_single_67% of AW 0.83 0.79 0.60 0.77 0.74
Un_trap_1earner couple_2 0.92 0.84 0.70 0.73 0.76
children_67% of AW
Un_trap_2earners couple_2 0.82 0.81 0.53 0.80 0.77
children_67% of AW

Incentives structure Entitlement conditions 2.8 3.0 2.8 3.9 2.9
(index) Job-search and availability 3.4 2.6 3.0 2.9 3.4
requirements
Monitoring of job-search activity 2.5 2.8 4.5 3.0 3.4
Sanctions 3.4 3.2 3.1 4.6 4.4
Legend: UB = unemployment benefits; UI = unemployment insurance; UA = unemployment assistance; NRR = net
replacement rate; PLMP = passive labour market policies; ALMP = active labour market policies; HB = housing benefits; SA
= social assistance; AW = average wage; In_trap = inactivity trap; Un_trap = unemployment trap
UB generosity is computed for a single person, 40 years old and with 22 years of a contribution period. Coverage is measured
as the number of participants per 100 persons wanting to work. Decline in NRR is computed as the ratio of the average net
replacement rates in the years 2 to 5 to the net replacement rate in the year 1. UI in Belgium and UA in some countries are of
indefinite duration which is for computation purposes capped at 60 months.
Source: Eurostat, ESSPROS, Joint European Commission-OECD project, using OECD Tax-Benefits models.

Table 2. Selected dimensions of unemployment benefit systems
Job search
Decline in net
Entitlement UI duration if 22 years and
Country Unemployment trap Inactivity trap replacement
conditions of a contribution period availability
rate
conditions
2011 2004 2010 2009 2008 2007 2010 2009 2008 2007 2010 2009 2008 2007 2010 2009 2011 2004
AT 2.0 2.0 70 69 71 70 56 55 57 57 9.0 9.0 9.0 9.0 95 95 2.5 2.5
BE 3.0 3.0 79 79 74 74 59 59 59 59 60.0 60.0 60.0 60.0 91 91 2.0 2.0
BG 3.5 - 71 72 67 - 36 37 38 - 11.0 11.0 11.0 - - - 2.3 -
CY 2.0 - - - - 62 - - - 48 - - - 6.0 6 - 2.8 -
CZ 2.0 2.0 74 72 69 71 54 53 54 56 5.0 5.0 6.0 6.0 30 30 4.0 4.5
DE 3.0 3.0 75 77 76 77 62 64 63 64 12.0 12.0 12.0 10.0 64 64 4.5 4.0
DK 2.0 2.0 87 86 86 87 83 83 84 84 24.0 48.0 48.0 48.0 79 79 4.1 4.1
EE 4.0 4.0 55 54 53 55 39 37 33 35 11.8 11.8 11.8 11.8 14 14 3.5 2.8
ES 4.0 4.0 72 71 70 71 37 35 34 37 23.7 23.7 23.7 23.7 46 46 3.5 3.5
FI 2.5 2.5 74 74 74 75 61 61 62 62 23.0 23.0 23.0 23.0 65 65 2.5 2.8
FR 3.0 - 69 69 71 71 54 54 53 52 24.0 24.0 23.0 23.0 58 58 2.9 -
GR 3.5 - 72 71 71 66 15 16 15 15 12.0 12.0 12.0 12.0 11 11 2.0 -
HU 3.0 - 75 77 77 78 42 52 52 53 8.9 8.9 8.9 8.9 28 28 3.5 -
IE 3.0 - 70 69 67 65 60 63 59 59 12.0 12.0 15.0 15.0 100 100 3.3 -
IT 4.5 4.0 72 72 72 66 25 26 24 23 8.0 8.0 8.0 7.0 3 3 4.0 2.3
LT 2.5 2.5 75 82 75 72 58 55 45 43 6.0 6.0 6.0 6.0 23 11 2.5 2.5
LU 3.5 - 83 82 83 82 60 59 57 56 12.0 12.0 12.0 12.0 11 11 3.5
LV - 2.0 80 75 74 78 57 50 48 52 9.0 9.0 9.0 9.0 5 5 2.8 2.8
MT 3.5 3.5 56 57 57 59 50 50 51 52 6.0 6.0 6.0 6.0 100 99 4.0 3.5
NL 3.5 3.5 76 76 78 78 68 68 69 66 22.0 22.0 22.0 22.0 27 27 3.6 3.6
PL 3.5 - 68 63 63 65 47 47 48 50 12.0 12.0 18.0 18.0 17 17 3.8 -
PT 4.5 3.5 71 72 73 73 40 40 40 41 27.6 27.6 27.6 27.6 53 56 3.3 2.8
RO 4.0 - 64 66 61 - 37 36 35 - 12.0 15.0 12.0 - - - 4.0 -
SE 3.0 3.0 71 73 74 77 55 56 57 59 13.8 13.8 13.8 27.7 59 59 3.8 3.4
SI 4.0 4.0 78 79 79 77 62 64 65 65 9.0 9.0 9.0 9.0 22 22 3.5 3.0
SK 3.0 3.0 59 58 69 69 38 37 46 47 6.0 6.0 6.0 6.0 27 24 4.0 4.0
UK 2.5 2.5 50 50 50 50 49 49 49 49 6.0 6.0 6.0 6.0 99 99 2.9 2.8
EU27 3.2 3.0 71 71 71 71 50 50 50 51 14.8 15.9 16.1 16.5 45 46 3.3 3.1
Nordic countries 2.8 2.8 77 77 78 79 67 67 68 68 20.7 26.7 26.7 30.2 57 57 3.5 3.5
Continental
2.9 2.7 75 75 75 75 58 58 58 58 23.4 23.4 23.2 22.8 64 64 3.1 2.8
countries
Anglo-Saxon
2.8 3.0 59 59 58 59 53 54 53 52 8.0 8.0 9.0 8.2 76 99 3.2 3.1
countries
Southern countries 4.3 3.8 72 72 72 70 34 33 33 34 19.8 19.8 19.8 19.4 34 35 3.6 2.8
Central and
3.3 2.9 70 70 69 70 44 44 44 46 9.3 9.6 10.0 9.6 20 18 3.3 3.3
Eastern countries
Notes:
Data are presented for the most recent available years.
Unemployment and inactivity trap are averages over six different stylised family types (single, one-earner couple, two-earner
couple, with and without children) and two earnings levels (67% and 100% of average full-time wages).
Decline in net replacement rate is computed as the ratio of the average net replacement rates in the years 2 to 5 to the net
replacement rate in the year 1. UI in Belgium and UA in some countries are of indefinite duration which is for computation
purposes capped at 60 months.
Source: Joint European Commission-OECD project, using OECD Tax-Benefits model.

Graph 1. Net replacement rates over the unemployment spell, 2010

AT BE BG CZ DE DK
0 .2 .4 .6 .8

EE ES FI FR GR HU
0 .2 .4 .6 .8
Net replacement rate

IE IT LT LU LV MT
0 .2 .4 .6 .8

NL PL PT RO SE SI
0 .2 .4 .6 .8

0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90

SK UK
0 .2 .4 .6 .8

0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90

Benefit duration (in months)
Net replacement rate_unemployment insurance Net replacement rate_unemployment assistance

Notes: Net replacement rates and benefit duration are computed for a single, low-wage person, 40 years old and
with 22 years of a contribution period.

Graph 2. The evolution of unemployment benefit generosity over time

0 10 20 30 40 AT BE BG CY CZ DE
Unemployment benefit generosity

DK EE ES FI FR GR
0 10 20 30 40

HU IE IT LT LU LV
0 10 20 30 40

MT NL PL PT RO SE
0 10 20 30 40

2000 2005 2010 2000 2005 2010 2000 2005 2010

SI SK UK
0 10 20 30 40

2000 2005 2010 2000 2005 2010 2000 2005 2010

year

Notes: Unemployment benefit generosity is computed for a single, low-wage person, 40 years old and with 22 years of a contribution period.
See Section 3.2. for the definition and interpretation of the unemployment benefit generosity. The drop of benefit generosity in Germany is to some extent attributed to the
separation of a living allowance and a rent allowance payment after 2004, whereby only a living allowance is considered as unemployment assistance.

Graph 3. Benchmarks (predictions) and actual unemployment benefit generosity

AT BE CZ DE DK
0 10 20 30 40

EE ES FI FR GR
0 10 20 30 40

HU IE IT LT LU
0 10 20 30 40

LV MT NL PL PT
0 10 20 30 40

2000 2005 2010

SE SI SK UK
0 10 20 30 40

2000 2005 2010 2000 2005 2010 2000 2005 2010 2000 2005 2010

year
Unemployment benefit generosity_actual values Unemployment benefit generosity_predictions

Notes: Unemployment benefit generosity is computed for a single, low-wage person, 40 years old and with 22 years of a contribution period.
See Section 3.2. for the definition and interpretation of the unemployment benefit generosity. Predictions are obtained from the multivariate regression (1) specified in the
column 2 of Table A1 in the Appendix.

Graph 4. Benchmarks (predictions) and actual unemployment insurance generosity

AT BE CZ DE DK
0 10 20 30 40

EE ES FI FR GR
0 10 20 30 40

HU IE IT LT LU
0 10 20 30 40

LV MT NL PL PT
0 10 20 30 40

2000 2005 2010

SE SI SK UK
0 10 20 30 40

2000 2005 2010 2000 2005 2010 2000 2005 2010 2000 2005 2010

year
Unemployment insurance generosity_actual values Unemployment insurance generosity_predictions

Notes: Unemployment insurance generosity is computed for a single, low-wage person, 40 years old and with 22 years of a contribution period.
See Section 3.2. for the definition and interpretation of the unemployment insurance generosity. Predictions are obtained from the multivariate regression (1) specified in the
column 4 of Table A1 in the Appendix.

Table 3. Unemployment benefit generosity and unemployment insurance generosity,
actual values and different benchmarks, average 2007-2010

Actual value by country -
Actual value by country - Actual value by country - Prediction by country -
actual value at GROUP
actual value at EU average Prediction by country Actual value at EU average
average
UB UI UB UI UB UI UB UI
generosity generosity generosity generosity generosity generosity generosity generosity
1 2 1 2 1 2 1 2
LU -6.7 -0.7 -16.1 -6.3 -18.0 -13.7 -10.7 -8.7 11.2 6.9 10.1 8.1
DK 17.9 23.9 11.5 15.6 10.8 11.3 18.0 17.9 7.1 6.5 5.9 6.0
FI 13.0 2.6 6.6 -5.8 7.0 8.6 -3.7 -2.9 6.1 4.4 6.3 5.5
SE -4.5 1.5 -10.9 -6.8 -13.9 -15.7 -7.4 -8.7 5.9 7.2 7.1 8.1
IE 15.4 -4.7 6.4 2.6 12.4 11.6 -6.2 -6.4 3.0 3.7 1.5 1.7
ES 0.0 6.1 0.3 2.3 -2.8 -2.1 3.0 4.1 2.8 2.1 3.0 1.9
AT 18.9 -5.7 9.5 -11.3 16.2 16.8 -6.7 -6.2 2.6 2.0 1.0 0.6
NL -0.8 5.3 -7.2 -3.1 -3.2 -6.4 4.1 1.7 2.4 5.6 1.2 3.6
UK -4.4 -9.5 -13.4 -2.2 -6.0 -7.9 -8.8 -9.5 1.5 3.5 -0.6 0.1
FR 1.8 5.8 -7.6 0.2 2.4 0.1 5.6 4.2 -0.6 1.7 0.2 1.6
BE 26.1 32.2 16.8 26.6 26.9 25.3 31.2 29.7 -0.8 0.7 1.0 2.4
EE -11.0 -4.9 0.1 0.4 -9.1 -6.9 -4.5 -3.2 -1.9 -4.1 -0.4 -1.7
DE 6.6 -3.7 -2.7 -9.3 8.5 4.4 -4.9 -8.3 -1.9 2.2 1.2 4.6
IT -11.5 -5.4 -11.2 -9.2 -9.5 -7.3 -4.3 -2.9 -2.0 -4.2 -1.1 -2.4
SI -10.5 -4.4 0.6 0.9 -8.1 -6.7 -2.1 -1.2 -2.3 -3.8 -2.2 -3.2
LV -11.1 -5.0 0.0 0.3 -8.5 -7.1 -3.0 -1.4 -2.6 -4.0 -2.0 -3.6
LT -12.9 -6.8 -1.8 -1.5 -10.2 -9.0 -3.9 -2.6 -2.7 -3.9 -2.9 -4.2
MT 20.4 -7.7 11.5 -0.4 24.1 24.7 -4.6 -4.1 -3.6 -4.3 -3.0 -3.6
CZ -13.8 -7.7 -2.8 -2.5 -9.7 -9.5 -4.4 -4.2 -4.0 -4.3 -3.3 -3.5
PT 10.6 10.7 10.9 6.9 16.1 15.3 14.2 13.8 -5.4 -4.7 -3.4 -3.1
HU -11.3 -6.0 -0.3 -0.8 -5.2 -5.0 -1.2 -0.6 -6.1 -6.3 -4.9 -5.4
PL -11.5 -5.4 -0.4 -0.1 -5.3 -5.6 -1.8 -1.9 -6.2 -5.9 -3.5 -3.5
GR -8.0 -4.3 3.1 1.0 -1.3 0.2 2.0 3.6 -6.7 -8.2 -6.3 -7.9
SK -12.6 -6.6 -1.6 -1.3 -0.6 -3.2 -1.7 -4.2 -12.1 -9.4 -4.9 -2.3
Legend: UB = unemployment benefits; UI = unemployment insurance;
Notes: Unemployment insurance generosity is computed for a single, low-wage person, 40 years old and with 22 years of a
contribution period. See Section 3.2. for the definition and interpretation of the unemployment benefit generosity.
Last eight columns in the table take into account predictions obtained from the multivariate regressions specified in Table A1
in the Appendix.

Table 4. Difference between deviations from group average and EU average*

Entitlement Unemployment Inactivity Unemployment Decline in net Job search
conditions trap trap insurance replacement and
duration rate availability
(22 years of conditions
contribution)

AT -0.3 4.1 -4.3 8.6 -18.7 -0.2
BE -0.1 -4.1 -7.9 -8.6 -18.7 -0.2
BG -0.1 1.0 -6.1 -2.2 0.0
CY 0.1 -25.7 0.0
CZ 0.1 1.0 6.1 -5.5 -5.3 0.0
DE -0.1 -3.4 -7.9 8.6 -18.7 0.2
DK -0.4 -6.1 -16.7 -5.9 -12.0 -0.2
EE -0.1 -1.0 -6.1 -0.5 -25.7 0.0
ES -0.5 -0.9 -9.4 -4.9 11.1 -0.1
FI -0.4 0.2 -4.7 -5.9 -12.0 0.2
FR -0.1 4.1 1.0 -8.6 -7.8 -0.2
GR -0.1 1.0 -6.1 -0.2 -25.7 0.0
HU 0.1 1.0 -6.1 -5.5 -8.8 0.0
IE 0.1 10.6 -3.0 1.2 -30.8 0.0
IT -1.2 -0.8 -16.1 4.9 -11.1 -0.3
LT 0.1 1.0 6.1 -5.5 -18.6 0.0
LU 0.3 -4.1 -7.9 8.6 18.7 0.2
LV 1.0 6.1 -5.5 -25.7 0.0
MT 0.4 -12.2 2.8 -6.8 -30.8 0.1
NL 0.4 -4.5 -16.7 -5.9 12.0 -0.2
PL -0.1 -1.0 0.6 -0.2 -25.7 0.0
PT -1.2 -0.2 -4.2 -4.9 11.1 0.3
RO -0.1 -1.0 -6.1 -0.2 0.0
SE 0.1 5.3 6.5 5.9 -12.0 -0.2
SI -0.1 1.0 6.1 -5.5 -21.7 0.0
SK 0.1 -1.0 -6.1 -5.5 -10.4 0.0
UK -0.4 -12.2 3.0 -6.8 -30.8 -0.1
Standard
deviation,
0.74 8.57 14.36 11.59 33.59 0.70
EU
average**
Standard
deviation,
GROUP 0.60 6.72 10.10 9.52 20.17 0.68
average**

* The numbers reported in the table are differences in the absolute of value of deviations of the variables with
respect to the group average and the EU average. Positive numbers are in bold.
**Standard deviation of deviations from benchmark.

Graph 5. Benchmarking unemployment benefit systems with respect to the EU average

Ent_condit Ent_condit
100 100

Job_search 50 Un_trap_avg Job_search 50 Un_trap_avg

0 0

Nrr_decline In_trap_avg Nrr_decline In_trap_avg

Duration Duration

AT EU27 BE EU27

100 100


0 0


Duration Duration

BG EU27 CY EU27

100 100


0 0


Duration Duration

CZ EU27 DE EU27

100 100


0 0


Duration Duration

DK EU27 EE EU27

Graph 5. Benchmarking unemployment benefit systems with respect to the EU average,
continued

100 100


0 0


Duration Duration

ES EU27 FI EU27

100 100


0 0


Duration Duration

FR EU27 GR EU27

100 100


0 0


Duration Duration

HU EU27 IE EU27

100 100


0 0


Duration Duration

IT EU27 LT EU27

31

continued

100 100


0 0


Duration Duration

LU EU27 LV EU27

100 100


0 0


Duration Duration

MT EU27 NL EU27

100 100


0 0


Duration Duration

PL EU27 PT EU27

100 100


0 0


Duration Duration

RO EU27 SE EU27

32

continued
100 100


0 0


Duration Duration

SI EU27 SK EU27

Ent_condit
100

Job_search 50 Un_trap_avg

0

Nrr_decline In_trap_avg

Duration

UK EU27

Graph 6. Benchmarking unemployment benefit systems with respect to the Nordic
group average

100 100


0 0


Duration Duration

DK nordic FI nordic

100 100


0 0


Duration Duration

NL nordic SE nordic

33

Graph 7. Benchmarking unemployment benefit systems with respect to the Continental
group average

100 100


0 0


Duration Duration

AT Continental BE Continental

100 100


0 0


Duration Duration

DE Continental FR Continental

Ent_condit
100


0


Duration

LU Continental

34

Graph 8. Benchmarking unemployment benefit systems with respect to the Anglo-Saxon
group average

100 100


0 0


Duration Duration

CY Anglo-saxon IE Anglo-saxon

100 100


0 0


Duration Duration

MT Anglo-saxon UK Anglo-saxon

Graph 9. Benchmarking unemployment benefit systems with respect to the Southern
group average

100 100


0 0


Duration Duration

ES Southern IT Southern

Ent_condit
100


0


Duration

PT Southern

35

Graph 10. Benchmarking unemployment benefit systems with respect to the Central and
Eastern group average

100 100


0 0


Duration Duration

BG CEE+GR CZ CEE+GR

100 100


0 0


Duration Duration

EE CEE+GR GR CEE+GR

100 100


0 0


Duration Duration

HU CEE+GR LT CEE+GR

100 100


0 0


Duration Duration

LV CEE+GR PL CEE+GR

36

Graph 10. Benchmarking unemployment benefit systems with respect to the Central and
Eastern group average, continued

100 100


0 0


Duration Duration

RO CEE+GR SI CEE+GR

Ent_condit
100


0


Duration

SK CEE+GR

37

Appendix

Estimating benchmarks for unemployment benefit generosity from multivariate
regressions

A1. Specification, sample, estimation
The specification of the regression equation for, respectively overall unemployment benefit
generosity and unemployment insurance generosity are as follows:
UBgenerosity i ,t = α t + β 1 y i ,t −1 + β 2 u i ,t −1 + β 3 ltu i ,t −1 + β 4 fbalancei ,t −1 + β 5 lmpi ,t −1 + eit , (A1)

UIgenerosity i ,t = α t + β1 yi ,t −1 + β 2 u i ,t −1 + β 3ltu i ,t −1 + β 4 fbalancei ,t −1 + β 5 lmpi ,t −1 + eit , (A2)

where the term α t refers to time fixed effects.
The definition of variables is provided in the following table.

Variable Definition
UBgenerosity Total available unemployment benefit support (unemployment insurance +
unemployment assistance) for a low wage single person, in months of
previous full net income, joint EC-OECD tax and benefits project
UIgenerosity Total available unemployment insurance support for a low wage single
person, in months of previous full net income, joint EC-OECD tax and
benefits project
y Real GDP per capita at constant prices from 2005, AMECO
u Unemployment rate, Labour Force Survey, Eurostat
ltu Long term unemployment rate measured as a share of long-term unemployed
in labour force, Labour Force Survey, Eurostat
fbalance Government budget balance, as % of GDP, AMECO
lmp Expenditures on labour market services (activation policies), as % of GDP,
Eurostat

The sample includes 24 EU countries over the period 2001-2010. Bulgaria, Romania and
Cyprus are excluded from the sample due to short available time series.
The estimation is carried out across countries, controlling for time effects, by means of OLS
with standard errors robust with respect to heteroscedastic errors.

A2. Results
Regression results are estimated in the table below. As estimates are not fully robust to the
inclusion of the active labour market policy variable (lmp), specifications excluding this
variable are also presented.
The benchmarks are obtained as the prediction from the regressions estimated below.

38

Table A1 – Determinants of unemployment benefit generosity
24 EU countries, 2001-2010

Dependent variable: UB generosity Dependent variable: UI generosity
Explanatory (1) (2) (1) (2)
variables
Real GDP per 0.1956405* 0.1757816+ 0.170726** 0.1672543**
capita [0.0954045] [0.0922541] [0.0546395] [0.0562539]

Unemployment 0.7726355+ 0.8328556+ 0.7913636* 0.7128282+
rate [0.4089653] [0.4795027] [0.3047878] [0.3666056]

Long term -2.196717** -1.679218* -0.9264094+ -0.2675044
unemployment [0.638547] [0.7149512] [0.5282451] [0.6642391]
rate
Government 0.4436693+ 0.3652645 0.7276251** 0.806174**
budget balance [0.261754] [0.2908932] [0.1998956] [0.2497218]

Expenditures on 21.89311* 14.39714*
labour market [10.68627] [7.208663]
services
Constant 13.10658* 23.00918** 6.251502+ 8.957961
[5.277544] [4.858395] [3.410164] [5.955565]

Observations 216 174 216 174
R-squared 0.2186 0.2856 0.1562 0.2143
Number of 24 24 24 24
countries
Estimation method: evidence from year-by-year cross-section regression. Robust standard errors in brackets.
Clustering of standard errors by country: ** p<0.01; * p<0.05; + p<0.1.

39

Une étude de la Commission européenne compare la générosité des systèmes de chômage

More Related Content

What's hot

Viewers also liked

Similar to Une étude de la Commission européenne compare la générosité des systèmes de chômage

More from lesoirbe

Recently uploaded

Une étude de la Commission européenne compare la générosité des systèmes de chômage