The Skill Matching Challenge - CEDEFOP

The skill matching
challenge
Analysing skill mismatch
and policy implications

Luxembourg: Publications Office of the European Union, 2010

A great deal of additional information on the European Union
is available on the Internet.

It can be accessed through the Europa server (http://europa.eu).

Cataloguing data can be found at the end of this publication.

Luxembourg:
Publications Office of the European Union, 2010

ISBN 978-92-896-0485-7

© European Centre for the Development of Vocational Training, 2010
All rights reserved.

Designed by Fotone - Greece
Printed in the European Union

The European Centre for the Development
of Vocational Training (Cedefop) is the European Union’s
reference centre for vocational education and training.
We provide information on and analyses of vocational
education and training systems, policies, research and practice.
Cedefop was established in 1975
by Council Regulation (EEC) No 337/75.

Europe 123, 570 01 Thessaloniki (Pylea), GREECE
PO Box 22427, 551 02 Thessaloniki, GREECE
Tel. +30 2310490111, Fax +30 2310490020
E-mail: info@cedefop.europa.eu
www.cedefop.europa.eu

Aviana Bulgarelli, Director
Christian Lettmayr, Deputy Director
Peter Kreiml, Chair of the Governing Board

Foreword

Cedefop analyses of skill needs and supply provide a core input to debates on
future developments in European labour markets. Policy-makers increasingly
recognise the added value of anticipating skill trends to support evidence-
based policy-making.
At the time of writing this report, Europe is battling a major recession.
Matching skills and jobs is gaining importance with rising unemployment
and increasing difficulties for people entering the labour market to find
jobs matching their potential. At the same time, structural trends challenge
Europe’s ability to provide the skills that its economies need to grow and
prosper. The European population is ageing rapidly, leading to a large
outflow of experienced workers in the coming decades while those nearing
retirement will need to adapt to the new skills that many jobs require. The
dynamics of today’s labour markets not only create new jobs while others
shrink or disappear, but also contribute to changing skill requirements in
many existing jobs.
The analysis of skill mismatch is, therefore, a crucial pillar of Cedefop
work on skills. Assessing mismatch and its effects involves much more
than straightforward comparisons between skill needs and skill supply.
Skill mismatch is a complex phenomenon affecting citizens, enterprises,
economies and societies. It refers not only to skill gaps and shortages, but
also to skills exceeding job requirements.
The prominence of skill mismatch as a core challenge on many policy
agendas, such as the new skills of jobs initiative, inspired Cedefop to explore
and analyse skill mismatch systematically. After setting the priorities for
future work on skill mismatch research in 2008 (Cedefop, 2009), this report
is the first in a series bridging the worlds of research and practice, aiming
to provide insights that enable evidence-based policy-making. The report
gives a broad overview of skill mismatch and the factors that contribute; it
analyses economic and social costs and discusses why mismatch should
be a concern for national and European policy-makers. As cross-country
information on skill mismatch is limited, the report also discusses how skill
mismatch can be better approached and measured in surveys to give more
detailed and sound analyses in the coming years.

2 The skill matching challenge

I believe this report serves a clear need by structuring the currently scattered
information on skill mismatch. Organising research findings to answer seven
important questions, the report is a direct input to evidence-based policy-
making. I am convinced this publication will add to the understanding of
skill mismatch in Europe and help national and European policy-makers to
make choices to aid matching skills and jobs in Europe.

Aviana Bulgarelli
Director of Cedefop

Acknowledgements 3

Acknowledgements

This report is the result of a team effort with contributions from those working
on Cedefop’s skill mismatch research. In particular, Jasper van Loo, who
supervised and edited the publication, Manfred Tessaring and Pascaline
Descy. This report was prepared under contract number 2008-FWC11/
AO/RPA/PDE-GUTCH-JVLOO/VETBenefits-skillmismatch/007/08 Lot 7.
It is largely based on the contributions of Nigel O’Leary and Peter Sloane
(Welmerc, School of Business and Economics, Swansea University), Seamus
McGuinness and Philip O’Connell (ESRI – Economic and Social Research
Institute, Dublin) and Kostas Mavromaras (NILS – National Institute of
Labour Studies, Flinders University, Adelaide, Australia). Thanks also go to
Alena Zukersteinova and Torsten Dunkel from Cedefop for their valuable
comments and suggestions.

Table of contents

Foreword 1
Acknowledgements 3
Executive summary 6

1. Why be concerned about skill mismatch? 9
1.1. Framing the issue and policy background 9
1.2. Glossary of terms 12
1.3. Aim and format of this report 14

2. What do we know about skill mismatch in Europe? 16
2.1. Vertical mismatch and impact on wages 16
2.2. Horizontal mismatch and impact on wages 21
2.3. What is the likely duration of skill mismatches? 22
2.3.1. Individual overeducation and overskilling 22
2.3.2. Overeducation and overskilling at firm level 23
2.3.3. Undereducation and underskilling in individuals/firms 24
2.3.4. Skill shortages 24
2.3.5. Skill gaps 25
2.3.6. Skills obsolescence 25

3. What causes skill mismatch? 27
3.1. Causes of overeducation 27
3.2. Underlying causes of skill shortages 32
3.3. Causes of skills obsolescence 32
3.4. Mismatch and individual characteristics 34

4. What are the links between the different types
of skill mismatch? 36
4.1. Skill shortages, skill gaps, undereducation and underskilling 36
4.2. What is the relationship between overeducation,
overskilling and skill shortages? 40
4.3. Combinations and durations of mismatch 43

Table of contents 5

5. Business cycle impact on skill mismatch 46
5.1. Skill mismatch and business cycles 46
5.2. The impact of the economic crisis 47
5.3. Evidence from Australia 50

6. The implications of skill mismatch 51
6.1. Skill mismatch and productivity 51
6.1.1. Overeducation and productivity 52
6.1.2. Skill shortages and productivity 53
6.1.3. Skill gaps and productivity 53
6.2. Skill mismatch and job satisfaction 54
6.3. Possible positive impacts of skill mismatch 55
6.4. Skill mismatch costs for society 56

7. Appropriate skill mismatch measuring 59
7.1. Measuring overeducation and overskilling 59
7.2. Focusing on overskilling 62
7.3. Measuring skills obsolescence 64

8. Skill mismatch data needed in Europe 65
8.1. Improving data on skill mismatch in Europe 65
8.2. What information should be collected and how? 66
8.3. The way forward 68

Annex: comprehensive schemata for skill mismatch
questions in surveys 69
References 72

List of tables
1. Glossary of terms 13
2. Reported incidence of overeducation 18
3. Types and causes of skills obsolescence 33
4. Combination of mismatch for firms 43
5. Data sets containing questions to individual employees on skill
mismatch 60

Boxes
1. Theoretical approaches to overeducation 29
2. The Beveridge curve 39

Executive summary

With global competition increasing, demographic change unfolding and
rapid technological change intensifying, skill mismatch has come to the
forefront of Europe’s policy debate. Skill mismatch refers not only to skill
shortages or gaps, but also to qualifications, knowledge and skills exceeding
job requirements. Shortages in some sectors may occur simultaneously
with overeducation in others. When skill mismatch takes time to resolve
it imposes real costs on individuals, enterprises and societies. Given the
relevance of matching skills and jobs and how information on mismatch is
scattered, this report synthesises evidence and provides new insights to
promote comprehensive policies that prevent and address skill mismatch
and its consequences. An overview of the meaning and characteristics of
different types of skill mismatch, such as overeducation and overqualification,
are provided in the glossary of terms in Section 1.2.
Skill mismatch is a widespread phenomenon in Europe, with overeducation
incidence averaging around 30 % and a substantial share of the population
undereducated. Whether skill mismatch is stable or increasing over time
differs between countries. The overeducated suffer a wage penalty, but
earn more than their matched colleagues, while the reverse applies to the
undereducated. Mismatch between the type of skills someone has and
the type of skills required is also prominent and has a negative impact on
earnings. There is, however, some disagreement on the extent to which
mismatch is a temporary or a long-term phenomenon, but both overeducation
and overskilling may persist over long periods. The overeducated may be
trapped in low-level jobs while the undereducated or underskilled are paid
more than those with the same level of education in matching jobs and,
therefore, have no incentive to move. The duration of skill shortages will
be a function of their level and complexity, while the duration of skill gaps
is linked to retraining. Skills obsolescence can occur over a relatively short
number of years compared to a working lifetime, which means that lifelong
learning is essential for successful careers.
To understand the causes of mismatch, one needs to look at its various
forms. A core reason explaining overeducation is the asymmetry of labour-
market information. Better-informed individuals are better prepared to avoid
overeducation than less-informed ones. Part of what is described in the

Executive summary 7

literature as ‘overeducation’ reflects the differences in individual abilities and
of skills within particular education categories. Light can be cast on this issue
by considering the job satisfaction of workers in this state or the extent to
which their skills are used. Several factors are responsible for skill shortages,
such as insufficient training, education and training systems responding
slowly to market changes, skill-biased technological progress and business
cycles. Skill shortages also emerge when people accept a low-skilled job
while continuing to search for a better match, for instance when options
are limited for personal or family reasons. Skills obsolescence, which is a
process rather than a state, occurs when skills depreciate due to ageing or
lack of use at work; or when economic developments change the skills that
firms need. Mismatch is more relevant for specific groups, such as young
people entering the labour market, older workers, females, ethnic minorities
and the disabled. Certain racial minorities suffer from mismatch to a greater
degree than the majority population. Whether women suffer more than men
is less clear. Older workers are likely to suffer most from skills obsolescence
but little evidence links skills obsolescence to long-term problems.
What are the links between the different types of skill mismatch?
Overeducation and overskilling are expected to be negatively related to
labour shortages. Whenever they coexist it is probably because skills or
education are of the wrong type or because the two phenomena refer to
different occupations in the same enterprise, industry or economy. It seems
there is no strong correlation between the occurrence of skill shortages
and skill gaps, but the latter may be closely linked to underskilling. Various
types of mismatch can be present at the same time in firms but probably
not simultaneously within specific occupations.
Different phases of the business cycle result in different imbalances
between labour demand and supply. These imbalances have profound
implications on how skill mismatch arises and how it influences the economy.
Studying the dynamics of skills creation and destruction is at the heart of
understanding mismatch driven by the business cycle. Currently, there is
insufficient empirical evidence to address the complex policy issues involved.
In the present downturn, the following policy issues need to be addressed
urgently: the role of training; the role of obsolescence; and how they may
affect different parts of the economy, in particular vulnerable groups of
people and vulnerable industries, such as those under intense international
competition.
Mismatched individuals may suffer from lower wages and are less satisfied
with their jobs than if they were properly matched; employers may suffer from


lower productivity; and the economy may suffer from a loss of output. This is
true for overeducation, overskilling and skill shortages. Both skill shortages
and skill gaps might lead to a loss of competitiveness; this is supported by
empirical evidence. Similar arguments apply to over/undereducation and
over/underskilling; this is confirmed by studies of US firms. However, if
skilled workers are hired anticipating future manpower requirements this
could have a positive effect for firms and be favourable for society through
increased innovation.
How to measure skill mismatch is a core concern. Three alternative
methods have been used: systematic job evaluation (objective measure),
worker self-assessment (subjective measure) and empirical method (where
data sets do not contain a direct question on the phenomenon). Each method
has weaknesses. The report argues that over and underskilling may be
superior measures to over and undereducation, especially if we are concerned
with potential welfare losses. Skills obsolescence can be measured in several
ways, but little consensus exists on which method is the most appropriate,
while few data sets contain questions that can be used to assess skills
obsolescence.
Good data are probably the most crucial prerequisite to supporting
timely, effective and evidence-based skill mismatch policies. The current
shortcomings of data sets may take years to remedy. Ideally, new matched
employer-employee panel data, with information on labour demand and
supply, should be developed. Collecting these data is very expensive and
several years of data are needed for full research potential. Individual or
household data can be used instead, but with a comprehensive coverage of
the various elements of mismatch, which is not currently the case. Remedying
existing data sets is cheaper than developing new ones. An alternative cost-
effective method for collecting Europe-wide data on mismatch could be a new
module containing questions on mismatch, possibly introduced in several
existing European large panel surveys at regular but infrequent intervals
(such as once every three years). We support the view that alternative
means of data collection should also be considered for issues that may be
too detailed to be informed through conventional interviews.

CHAPTER 1

Why be concerned about skill
mismatch?

Skill mismatch should be of concern to all citizens, but especially to policy-
makers, employers’ associations and trade unions. Because of skill-
biased technological change, its incidence will increase over time, though
overeducation could have some positive effects. As Europe’s population
is ageing rapidly, skills obsolescence is also likely to grow in importance.
Matching skills and jobs is, therefore, crucial for Europe.
It is unrealistic to assume that labour markets can work without temporary
imbalances but skill mismatch that takes time to resolve or is even permanent
implies real economic and social losses. Aiming to provide insights supporting
evidence-based policy-making, this report considers all types of skill
mismatch and their interrelationships, recognising that their measurement
and interpretation is not always straightforward. The report focuses on
seven research questions, on some of which relatively little is known. In this
chapter, we examine why policy-makers and other stakeholders should be
concerned about skill mismatch. Subsequently, we provide a glossary of
terms and elaborate the aim and the set-up of this report.

1.1. Framing the issue and policy background
Skill mismatch has come to the forefront of the policy debate in Europe,
with global competition increasing, demographic change progressing and
rapid technological change intensifying (European Commission, 2008a).
Skill problems might arise, with shortages in some sectors combined with
the problem of overeducation in others. Skilled people frequently working
in elementary jobs is a pervasive international phenomenon. The challenge
facing Europe is to provide businesses with the right skills to encourage
innovation and growth, to address unemployment and to ensure that the
potential of Europe’s citizens is fully utilised.
As the European Commission Communication on new skills for new jobs
(European Commission, 2008) observes, skill matching should be improved
to confront rising unemployment. It is also suggested that the composition of
skills emerging from both European Union universities and training systems
needs to be better adapted to innovation-driven economies. Further, skill-


biased technological change results in a need for education upgrading as
demand for more highly-educated workers increases at the expense of
the less educated (1). The recent Cedefop report on Future skill needs in
Europe (Cedefop, 2008) notes that out of a net employment increase of
over 13 million jobs between 2006 and 2015 almost 12.5 million will be at
the highest qualification level (ISCED 5 and 6).
Skill mismatch is a crucial policy issue, not only for policy-makers but
also for social partners including employers associations and trade unions.
Reducing skill mismatch is likely to generate social benefits with higher
job satisfaction for employees and possibly reduced stress from work,
increasing health and wellbeing. At macro level, skill shortages and skill
gaps can potentially lead to a loss of competitiveness as wage rates are
bid up and productivity lowered within industries where skill problems exist.
Productivity may also suffer if firms are forced to place lower-skilled workers
in skilled positions and/or if where a skills shortage exists, workers use their
position to alter their employment terms and conditions in a way that harms
productivity.
As the European Commission Communication on new skills for new
jobs (European Commission (2008b) states, skills upgrading is important
for Europe’s future growth potential and for employment equity, given the
vulnerability of those lacking skills. It goes on to state that, due to imperfect
information and structural rigidities, the lack of appropriate skills harms
competitiveness, particularly for small enterprises. The communication
suggests that skills upgrading is the most relevant solution to current and
anticipated imbalances in the labour market.
Skill-biased technological and organisational change has not only increased
the demand for highly-skilled manpower but has also been encouraged itself
by the availability of such manpower. An increasing pace of technological
change is likely, however, to result in growing skill shortages and skill gaps
if appropriate education and training is not provided. At the same time the
routine tasks of the middle-skilled may be eliminated by new technologies
(hollowing out).
Mismatches are usually explained by a combination of information
asymmetry between employers and employees, incomplete information
in the labour market, differences between people, and transactions costs.

(1) For a recent theory paper on the implications of on-the-job search in a matching model, which traces
out the effects on the composition of jobs, unemployment, wages and labour-market reactions to
shifts in the supply of and demand for skills, see Dolado et al. (2009).

CHAPTER 1
Why be concerned about skill mismatch? 11

Therefore, tackling mismatch by providing labour-market information and
efficient job placement services should be a priority for policy-makers. This
holds for both vertical and horizontal mismatch.
Vertical mismatch, which is commonly referred to as overeducation,
occurs when an individual is hired for a job advertised as requiring a lower
level of education than the individual possesses, or alternatively, when an
individual is employed in a job which requires a lower level of education.
When overeducation is partly a consequence of lower ability at a given
education level or choice of employment by the individual concerned, this
does not, in itself, imply a particular role for policy. If overeducation is merely
a temporary phenomenon, policy-makers should not be too concerned but,
for many employees, overeducation lasts a long time. Yet, if employers hire
the overeducated, anticipating a future need, their action may have positive
rather than negative repercussions.
Along these lines, for Schlotter (2008), overeducation can be a desirable
feature of education and training systems and, following Sloane (2003), low-
skilled workers may compensate for their lack of skill/ability by additional
education to gain employment which they might otherwise not obtain.
However, overskilling has negative effects on the wages earned by individuals
experiencing it and on their job satisfaction. Greater care in job matching
may, therefore, not only improve the welfare of overskilled employees, but
also have positive effects on productivity and the growth in the economy.
Finally, skills obsolescence is likely to grow in significance as the pace
of economic and technological change intensifies and the labour force is
ageing. Here there is a need for policies which encourage employers to
retain older workers, as it is much more difficult to find new employment
as workers age.
Graduates sometimes face horizontal skill mismatch, when not the level,
but the type of education or skills are inappropriate for the job. Those with
specific degrees do better than those with more general degrees (arts and
humanities). This suggests that better information should be provided for
potential students to address and prevent mismatches of this type.
A particular policy issue is whether genuine mismatch requires action from
education providers or employers. For skill mismatch, whether horizontal
or vertical, Weststar (2009) suggests that attention should be shifted away
from education reform towards considering the impact of workplace and
job design. This could be best achieved by giving workers greater task
discretion on both social and technical control. The former includes taking
part in decision-making, while the latter includes the ability to plan certain


aspects of the work and elements of choice in the way work is performed.
Training needs of mismatched workers should also be considered.
Data on skill mismatch in Europe shows that it is pervasive, but evidence
is scattered and there is much confusion about what is meant by different
aspects of skill mismatch. Poor understanding of skill mismatch implies poor
understanding of interactions between types of skill imbalances, thereby
preventing coordinated actions. Addressing skill mismatch problems is
especially relevant for vulnerable groups on the labour market, such as
migrants, older workers and young people transiting from school to work.

1.2. Glossary of terms
Literature is not always clear on what is precisely meant by different types
of skill mismatch. For example, it is often not straightforward to assess
whether mismatch refers to qualitative or quantitative aspects and what is
meant by occupational or education levels in a context where the notion of
levels is blurring in many workplaces. Despite this, it is useful to have some
initial understanding of various types of mismatch before examining more
detailed issues. Therefore, Table 1 presents a glossary of terms. Although
the definitions are based on a loose interpretation of literature and sometimes
combine qualitative and quantitative dimensions of mismatch, they allow
better understanding of different types and dimensions of mismatch. As an
example, overeducation is defined in terms of years and overqualification
in terms of credentials. While these terms are used interchangeably in
literature and not all questions in various surveys allow a clear distinction
between the two, they are not equivalent. In fact, Brynin and Longhi (2009),
using data from the e-Living project showed that, while a qualification higher
than required mostly generates a wage premium, time spent in education
does not. The latter may involve repeating a year of study and reflects
variations in ability, motivation or institutional arrangements which differ
between countries.

CHAPTER 1

Table 1. Glossary of terms

Overeducation A situation in which an individual has more education than the current job requires (measured in years).

Undereducation A situation in which an individual has less education than the current job requires (measured in years).

Overqualification A situation in which an individual has a higher qualification than the current job requires.

Underqualification A situation in which an individual has a lower qualification than the current job requires.

Overskilling A situation in which an individual is not able to fully utilise his or her skills and abilities in the current job.

Underskilling A situation in which an individual lacks the skills and abilities necessary to perform on the current job to
acceptable standards.

Credentialism A situation in which the level of education required to obtain the job exceeds the level of education
required to perform the job adequately. It relies upon a belief of the employer that the possession of
certificates and diplomas implies higher productivity on the part of the individual.

Real A situation in which an individual possesses more education than the current job requires and also in
overeducation which current skills and abilities are underutilised.

Formal A situation in which an individual possesses more education than the current job requires, but in which
overeducation current skills and abilities are fully utilised.

Genuine A situation in which an individual has more education than the current job requires and this state has a
overeducation negative effect on the level of job satisfaction.

Apparent A situation in which an individual has more education than the current job requires, but this does not
overeducation adversely affect the level of job satisfaction.

Skill shortage A situation in which the demand for a particular type of skill exceeds the supply of available people with that skill.

Skill surplus A situation in which the supply of available people with a particular skill exceeds the demand for it.

Skill gap A situation in which the level of skills of the currently employed is less than that required to perform the
job adequately or the type of skill does not match the requirements of the job.

Economic skills A situation in which skills previously utilised in a job are no longer required or have diminished in
obsolescence importance.

Physical Physical or mental skills and abilities deteriorate due to atrophy or wear and tear.
(technical)
obsolescence
Vertical mismatch A situation in which the level of education or skills is less or more than the required level of education or skills.

Horizontal A situation in which the level of education or skills matches job requirements, but the type of education or
mismatch skills is inappropriate for the current job.

Subjective meas- The mismatch estimate is obtained by self assessment in employee responses to a questionnaire.
ures of mismatch
Objective The mismatch estimate is obtained by evaluating job requirements.
measures of
mismatch
The empirical The mismatch estimate is derived from differences in the actual education of an individual within an
mismatch method occupation relative to the mean or modal level of education of all people employed in that occupation
(where there are no direct data on mismatch obtained from either of the above methods).
Crowding out/ When better qualified workers are hired to do jobs that less qualified workers could also do, they are crowding
bumping down out the less qualified workers from what would be their traditional employment possibilities for that level of
skill. Bumping down refers to the fact that this process works from the top to the bottom and pushes the less
qualified workers to even lower level jobs. At the extreme some lower level workers may lose employment
altogether when crowding out occurs.


1.3. Aim and format of this report
Of all types of skill mismatch, the concept of overeducation has received
most attention in literature. This concept was first introduced in 1976 by
Richard Freeman in The overeducated American, an influential book based
on US experience (Freeman, 1976). Since then literature on overeducation
has grown substantially and points to the 30-40 % of the working population
falling into this category and suffering from a substantial wage penalty
compared to properly-matched workers. In most countries a smaller, though
often substantial, number of workers are mismatched in the sense of being
undereducated for their job. Interpretation of these phenomena is, however,
far from straightforward. First, measurement issues arise from the diversity of
information sources, such as the subjective responses to questions directed
at employees, ‘objective’ estimates of the education requirements of jobs
obtained by job analysis, or estimates derived by the empirical method
from the spread of education qualifications within occupations. Second, the
relationship between occupations and their education requirements may
be imprecise. Thus, some jobs may merely specify a minimal education
requirement, while education requirements may be rising over time as jobs
become more complex.
In recent years, a more limited strand of literature, encouraged by new
data on the issue, has focused on both overskilling and underskilling (e.g.
Allen and van der Velden, 2001; Green and McIntosh, 2007). This has raised
questions about the relationship between overeducation and overskilling.
There is also much older literature on skill shortages, i.e. insufficient numbers
of suitably-qualified applicants to fill available posts, and the associated
concept of skill gaps, i.e. when management judges that existing employees
lack appropriate skills to maximise their contribution to output. Further, there
is distinctive, but more limited, literature on the related question of skills
obsolescence.
Although there has been a lot of attention on different types of skill
mismatch, little attempt has been made to synthesise and integrate these
disparate strands in literature. By drawing together consistently the various
strands of research literature on skill mismatch and by synthesising empirical
evidence, this study provides new insights to promote comprehensive policies
to prevent and address skill mismatch and its consequences.

CHAPTER 1

While conducting a literature survey, this study also reviews concepts,
theories and empirical evidence on skill mismatch. To analyse and evaluate
skill mismatch in a broad and comprehensive way, it focuses on the following
research questions:
(a) what do we know about skill mismatch in Europe?
(b) what causes skill mismatch?
(c) what are the links between the different types of skill mismatch?
(d) what impact do business cycles have on skill mismatch?
(e) what are the implications of skill mismatch?
(f) what is the most appropriate method of measuring skill mismatch?
(g) what types of data are needed to assess and analyse skill mismatch
in Europe?
As some of these questions have been relatively neglected in literature,
any conclusions drawn must be tentative.

CHAPTER 2

What do we know about skill
mismatch in Europe?

This chapter examines the pervasiveness of skill mismatch in Europe.
Ideally, we would wish to know the extent of various types of mismatch and
trends over time but, as comprehensive data on skill mismatch is limited,
we focus instead on discussing the incidence of vertical and horizontal skill
mismatch and their impact on wages. Though incidence varies substantially
across countries, overeducation is found to be substantial across Europe
with significant negative effects on earnings. The incidence of overskilling
is lower, but higher for men than for women and increases with the level
of education. In several countries, where evidence is available, horizontal
mismatch is also substantial and impacts on earnings in a similar way. In
Section 2.3, we review evidence on the duration and persistence of mismatch.
The available information supports the hypothesis that much of it persists
over the medium to the long term and may differ by level of education.
Evidence on skills obsolescence suggests that many individuals, particularly
the more educated in high-tech industries, may require substantial lifelong
learning to remain productive.

2.1. Vertical mismatch and impact on wages
One of the most recent surveys (McGuinness, 2006), summarised in Table 2,
in line with Groot and Maassen van den Brink’s (2000) meta analysis, found
that subjective measures of overeducation yield higher mean figures (29 %)
than objective measures (22 %). This suggests that employees do not have, as
some have argued, a clear tendency to exaggerate their own qualifications. The
lowest incidence is recorded in the Netherlands and the highest in the US. For
McGuinness (2006) it is reasonable to conclude that overeducation has remained
relatively stable over time, although it is not always the case. Green and Zhu
(2008) reported that in the UK overeducation increased for men from 21.7 % in
1992 to 33.2 % in 2006, while for women the comparable figures were 23.8 %
and 32.1 %. However, they also split the overeducated into two groups – ‘real’
and ‘formal’ according to whether or not it is accompanied by underutilisation
of skill. Real overqualification has been steady or rising only slowly.

CHAPTER 2
What do we know about skill mismatch in Europe? 17

Gottschalk and Hansen (2003) classified occupations in the US as graduate
and non-graduate on a different basis: whether 90 % or more of workers
in that occupation were graduates or, failing that, there was a significant
graduate pay premium of 10 % or more in that occupation. They found that
the probability of a graduate being employed in a non-graduate job actually
fell between 1983 and 1996, a result ‘which stands in stark contrast to those
in previous studies’ (Gottschalk and Hansen, p. 450). Similar findings have
been obtained by Cardoso (2007) for Portugal and by Grazier et al. (2008)
for the UK. It is clear, therefore, that conventional data on the incidence of
overeducation should be treated with caution.
Ignoring the above distinctions about the precise nature of overeducation,
Sicherman (1991) suggested that there were two stylised facts in relation to
the effect of mismatch on pay. First, there is a pay penalty for overeducation,
but the overeducated earn more than their matched colleagues. Second,
the undereducated suffer a pay penalty relative to their matched colleagues
but earn more than if they were properly matched. Later studies tend to
confirm these results. Groot and Maassen van der Brink (2000) in their
meta analysis summarised 25 studies as producing unweighted average
rates of return of 5.6 % for attained years of education, 7.8 % for required
years of education, 3 % for years of overeducation and -1.5 % for years of
undereducation.
There is, however, little or no information on how employers differentiate
the pay of their employees in this way. Is it, for instance, based on some
form of assessment of performance? There is evidence that wage benefits
of overeducation decline with additional labour-market experience. This
is consistent with ‘sheepskin effects’, whereby initial pay is determined by
certification, but eventually workers are sorted into their most appropriate jobs
on the basis of performance, so that qualifications become less significant
over time.
Chevalier (2003) found a pay penalty of 26 % for genuinely overeducated
graduates, but only 7 % for those who were only apparently so. The wage
penalties for overeducation appear to be greater than for overskilling, but
this may reflect a failure to correct for unobserved differences between
individuals in most studies. Wage effects may also vary by level of education or
occupation. Mavromaras et al. (2009d) found that there were wage penalties
associated with overskilling in both Australia and the UK, but these were
larger for men than for women and increased with education level.


Table 2. Reported incidence of overeducation

Name of study Year Definition employed Country Data collection Incidence of
overeducation (%)

Groot and Maassen 2000 Meta-analysis 26.2 (a)
van den Brink

Galasi 2008 Subjective job requirements 25 Countries 2004-06 33.00

Groot 1996 Objective standard deviation-based UK 1991 11.00
measure
Alpin 1998 (1) Objective occupational 1995 27.00
dictionary-based measure
UK
(2) Objective standard deviation 1995 37.70
modal measure
Sloane et al. 1999 Subjective – job requirements UK 1986-87 30.63

Green et al. 1999 (1) Subjective UK 1995 27.40
(2) Subjective UK 1997 32.00

Dolton and Vignoles 2000 Self assessment – subjective job UK 1986 30.00
requirement
(1) Subjective – satisfaction UK 1996 40.40 (a)
1996 40.70 (a)
Battu et al. 2000 (2) Objective occupational UK 1996 21.75 (a)
dictionary-based measure 1996 22.15 (a)

(3) Subjective – degree requirements UK 1996 33.65
1996 38.00 (a)
Dolton and Siles 2003 Subjective – not clear which one UK 1998 22.00
they used

(1) Objective occupational UK 1996 17.00 (a)

Chevalier 2003 (2) Subjective – job requirement UK 1996 32.40 (a)

(3) Subjective – satisfaction UK 1996 16.20 (a)

Green and Zhu 2008 Subjective job requirements UK 1992 26.40
1997 26.50
2001 32.50
2006 37.30
Grazier et al. 2008 Occupational classification UK 1994 40.00
2004 30.00
Hannan et al. 1998 Objective standard deviation Ireland 1992 20.00

McGuinness 2003a Subjective – job requirements Northern 2000 20.00
Ireland
McGuinness 2003b Subjective – job requirements Northern 1999 24.00
Ireland

CHAPTER 2

overeducation (%)

Hartog and 1988 Subjective – required education Netherlands 1982 16.00
Oosterbeek objective 1960 7.00
1971 13.60
1977 25.70
Groot 1993 Objective standard deviation-based Netherlands 1983 16.10
measure
Groot and Maassen 2000 (1) Objective standard deviation- Netherlands 1994 11.85 (a)
van den Brink based measure
(2) Objective occupational Netherlands 1994 15.90 (a)
(3) Subjective Netherlands 1994 11.15 (a)

Allen and van der 2001 Subjective – required education Netherlands 1998 14.00
Velden
Dekker et al. 2002 Objective occupational dictionary- Netherlands 1992 30.60
based measure
(1) Objective standard deviation Portugal 1991 9.40
mean measure
(2) Objective standard deviation Portugal 1991 25.50
Kiker et al. 1997 modal measure
(3) Objective occupational Portugal 1991 33.10

1986 5.00
Cardoso 2007 Occupational classification Portugal
1999 1.00
Alba-Ramirez 1993 Subjective – job requirements Spain 1985 17.00
Bauer 2002 Objective standard deviation-based Germany 1984-98 11.50 (a)
measure
Büchel and van Ham 2002 Subjective – job requirements Germany 1998 15.80 (a)
Patrinos 1997 Objective standard deviation-based Greece 1977 16.00
measure
Duncan and Hoffman 1981 Subjective – job requirements US 1976 42.00
Burris 1983 Objective occupational dictionary- US 1977 21.70
based measure

1969 35.00
Rumberger 1987 (1) Subjective – job requirements US 1973 27.00
1977 32.00
(2) Objective occupational US 1973 57.00

Verdugo and 1989 Objective standard deviation-based US 1980 10.90 (b)
Verdugo measure


overeducation (%)

Sicherman 1991 Subjective – job requirements US 1976 and 1978 40.80

(1) Subjective – required education US 1969 35.00
1973 27.00

Tsang et al. 1991 1977 32.00
(2) Objective occupational US 1973 57.00
(1) Objective standard deviation- US 1985 13.00
based measure
Cohn and Khan 1995 (2) Subjective measure – Sicherman US 1985 33.00
definition

Robst 1995 Subjective – required education US 1976, 1978 44.68
and 1985

Hersch 1995 Subjective – job requirements US 1991 21.00
(1) Subjective – job requirements US 1985 50.00 (a)

McGoldrick 1996 (2) Objective occupational US 1985 30.80 (a)
and Robst dictionary-based measure
(3) Objective standard deviation- US 1985 12.55 (a)
based measure

Daly et al. 2000 Subjective – required education US 1976 37.75 (a)
US 1985 32.65 (a)
Germany 1984 17.50 (a)
Gottschalk and 2003 Occupational classification US 1983 9.00
Hansen
1996 5.00
Vahey 2000 Subjective – required education Canada 1982 31.00 (a)
1986 35.00 (a)
Cohn and Ng 2000 Objective modal measure Hong Kong
1991 34.00 (a)

(a) Average, (b) Males only Source: Adapted from McGuinness (2006).

There are also differences between countries. Galasi (2008) used the
European Social Survey to analyse the effects of education mismatch across
25 countries. Overall, 33 % of workers were found to be overeducated, but this
ranged from 15 % in the Netherlands to 79 % in Estonia; these data suggest
a higher proportion of undereducated, with a mean of 59 %. Otherwise most
of the empirical regularities found in earlier literature seem to hold.

CHAPTER 2

2.2. Horizontal mismatch and impact on wages
We must also consider horizontal mismatch. Robst (2008) suggested the
following four hypotheses in relation to this type of mismatch:
hypothesis 1: the wage effects of overeducation are larger when workers
are also mismatched the basis of type of degree and level of
occupational requirements;
hypothesis 2: among workers mismatched on the basis of level and type of
education, the wage effect depends on the transferability of
skills between type of degree and occupation;
hypothesis 3: individuals who face discrimination (such as coloured people,
the disabled and single women) or who have lower ability are
more likely to be mismatched due to demand-related reasons.
Supply-related mismatches are more likely among married
men and women than among single men and women;
hypothesis 4: the wage effects of mismatch are likely to be larger when
workers have accepted a job due to demand-side reasons
(i.e. genuine mismatch) compared to supply-side reasons
(i.e. apparent mismatch).
Robst finds that overeducated men, whose work and field of study are related,
earn 2.4 % less than well matched men. However, if the job is only somewhat
related the penalty rises to 7.1 % and to 21.6 % if it is completely unrelated. For
women the corresponding figures are 3.6 %, 5.3 % and 19.1 %. The income
penalty for horizontal mismatches appears to exceed that for vertical mismatches.
Further, while most workers report supply-side reasons as the main reason
for being in a particular job, it also appears that wage effects are large where
workers are mismatched for demand-side reasons. A Swedish study (Nordin et
al., 2008) found that being horizontally mismatched was associated with 38 %
lower income for men and 26 % for women. An Irish study (Kelly et al., 2008)
found, in contrast, a much smaller pay penalty of around 5 % for those who
were mismatched by field, though this varied by field of study. In this case the
vertical mismatch penalty (14.4 %) exceeded the horizontal.


2.3. What is the likely duration of skill
mismatches?
The extent to which a particular form of mismatch persists over time
determines the costs to the individual, firm or economy. While evidence on
persistence of overeducation is rich, it is limited to insights based on cross-
sectional data. Meaningful tests for mismatch persistence require longitudinal
data on individuals and jobs over a longer period of time. Evidence on the
persistence of other forms of mismatch is scarce, but some conclusions
can be drawn.

2.3.1. Individual overeducation and overskilling
There has been considerable debate on whether individual overeducation
will persist over time. Sicherman (1991) found a positive and significant
relationship between job mobility and overeducation which was taken as
evidence of non-persistence. However, later research produced more mixed
results.
Some studies have reported evidence of higher rates of job mobility (e.g.,
Alba-Ramirez, 1993; Sloane et al. 1999; Alba-Ramirez and Blázquez, 2003),
higher rates of promotion within firms (Hersch, 1995; Dekker et al. 2002;
Alba-Ramirez and Blázquez 2003; Groeneveld and Hartog, 2004), or greater
levels of quit intention (Hersch, 1995; Robst, 1995) among the overeducated.
Sloane et al. (1999) also reported that, while overeducated workers were
more likely to have short job tenure with their current employer, they were
also more likely to experience involuntary job separations and to exit from
employment into unemployment. Further, for McGuinness and Wooden (2007)
increased job mobility cannot be taken as evidence of falling overeducation
given that workers are most likely to move from one overeducated job to
another. The most convincing evidence to support the view that overeducation
can be a long lasting phenomenon comes from studies of graduate labour
markets, but these studies are typically based on cross-sectional data or short
time-series. Dolton and Vignoles (2000) found that 38 % of a large sample
of UK graduates in 1980 was overeducated in their first job; six years later
this proportion still stood at 30 %. Similarly, McGuinness (2003c) reported
that among a sample of graduates in Northern Ireland, 31 % indicated that
a university degree was not a requirement for their first job, and after two to
four years this proportion was still at 24 %. Finally, Frenette (2004) reported
on longitudinal data for Canadian graduates, which showed little evidence

CHAPTER 2

of any decline in the incidence of overeducation over a three-year window
(between two and five years after graduation).
Research on overskilling persistence is much more limited as panel data
on overskilling in Europe are lacking. McGuinness and Wooden (2007)
examined durations using the first four waves of the household income
and labour dynamics for Australia (HILDA) panel. The study concludes that
overskilling tends to be a non-transitory phenomenon with less than 25 %
of employees who were severely overskilled in wave 1 likely to be matched
three years later. The study also found that, as for overeducation, increased
job mobility cannot be evidence of reduced overskilling. Mavromaras et al.
(2009b) using HILDA found that overskilling varies by education level: it is
rare among university graduates, but highly persistent; it is low but significant
among school leavers from general education; and none at all is found
among vocational education and training graduates.
Graduate studies focusing on overeducation and overskilling are typically
based on short time-series. This hampers the ability to examine whether
overeducation in the early stages of a career are reversible later on, when
people acquire more training or experience. Also, these studies only provide
a partial picture of mismatch persistence, as they focus on the transition
from school to work. Far less attention is on transition from unemployment
to work, mainly because suitable data is lacking, but also due to insufficient
analysis of the incentives and preferences, decisive factors in people’s
education and occupation choices. Many research questions remain open,
but we can tentatively conclude that overeducation and overskilling are
non-transitory phenomena.

2.3.2. Overeducation and overskilling at firm level
Although overeducation and overskilling can be a long lasting phenomenon for
individuals, it does not mean firms retain overeducated or overskilled workers
for long periods. High rates of voluntary and involuntary separations among
the overeducated and overskilled (2) suggest that their average job tenure
is likely to be some fraction of their total mismatch duration. However, the
fact that the overeducated earn a premium relative to matched colleagues,
coupled with the finding that the probability of involuntary separation tends
to be much lower than the probability of voluntary separation (3), suggests
that employers hiring overeducated and overskilled workers have some

(2) Sloane et al., 1999; McGuinness, 2003; McGuinness and Wooden, 2007.
(3) McGuinness and Wooden, 2007.


productivity advantage. This is confirmed by Jones et al. (2009) using
British workplace employment relations survey (WERS) data to demonstrate
higher productivity in firms employing overeducated workers. However,
the consistent finding that the overeducated and overskilled have lower
job satisfaction (4) also indicates that retaining such workers may not be
in the firm’s long-term interest. Therefore, the extent to which firms retain
overskilled or overeducated workers will depend on the magnitude of any
short-term benefit relative to the long-term cost. From an economy point of
view, demand for skills and educated labour rising over time should reduce
the incidence of overskilling and overeducation, provided that the surplus
are in areas of demand within the economy and that workers are mobile.

2.3.3. Undereducation and underskilling in individuals/firms
Generally, the undereducated have no incentive to become matched, as
they earn a premium relative to matched colleagues with similar education
levels (5). Little research exists to date for the underskilled, so for now we
will assume that their situation reflects that of the undereducated.
Accepting that earnings reflect workers’ marginal products, the finding
that the undereducated are paid less than matched individuals in the same
job who are matched (6) suggests that firms employing undereducated
workers may have some productivity penalty. The extent to which a firm will
continue to hire undereducated workers will vary according to the underlying
causes of the undereducation itself. If undereducation relates to an inability
to hire suitably qualified workers, i.e. a skill shortage, then the duration of
undereducation within the firm will be highly correlated with the persistence
of the skill shortage within the economy generally.

2.3.4. Skill shortages
Duration of skill shortages depends on the level at which the problem is
assessed. For firms, the shortage will persist as long as the vacancy remains
unfilled or until the firm adopts an alternative strategy for dealing with the
problem. Duration will be longer the more complex and skilled the unfilled
vacancy so duration should vary across occupations and sectors. For the
economy more generally, duration of skill shortages will depend on the time
elapsing before it is communicated through the market mechanism and the
period required to induce and train individuals in the specific skill area.

(4) Battu et al., 1999; Fleming and Kler, 2007.
(5) Kiker et al., 1997; Sloane et al., 1999.
(6) Kiker et al., 1997.

CHAPTER 2

2.3.5. Skill gaps
Given that skill gaps relate to the firms’ current employees and that they are
specific to the firm’s requirements, dealing with skill gaps largely remains the
responsibility of firms. Generally, skill gaps can only be eradicated through
firm-sponsored training or by providing individuals with relevant incentives
to participate in upskilling. Duration depends on the period necessary to
retrain individuals in the relevant skill areas, as for skill shortages, and on
costs involved. Given the difficulties and costs associated with equipping
workers with basic and advanced skills, durations are likely to be shorter for
skill gaps related to intermediate skills. However, when skill gaps arise as a
result of firms adopting a second best strategy in the face of skill shortage,
duration will again be related to the duration of skill shortage.

2.3.6. Skills obsolescence
Here we are interested in how long on average it takes for a worker’s skill
to become obsolete. In an early study, Bosworth (1978) noted that little was
known about the obsolescence of technical knowledge. However, using a
restrictive definition – the time when a particular technology was no longer
embodied in goods currently being sold – he was able to use patent renewal
data to measure obsolescence. Evidence pointed to annual rates of decay
of well over 10 % in the post-war period. Neuman and Weiss (1995) showed
that obsolescence was more evident in high-tech than low-tech industries
and greater for more educated workers. Allen and van der Velden (2007),
using a Dutch survey, reported that 30 % of skills in their sample had become
obsolete with a half-life for competences in the range of 10 to 15 years.
This implies that many individuals would have to renew their competences
regularly over their career, giving justification for lifelong learning.
In the 1970s several studies examined the rate of obsolescence among
engineers. Notably, Kaufman (1974; 1975) pointed out that the introduction
of computer technology was critical and ironically threatened most those
introducing it: engineers, computer scientists and other technical professionals.
He defined obsolescence as ‘the degree to which organisational professionals
lack the up-to-date knowledge of skills necessary to maintain effective
performance in either their current or future work roles’ (Kaufman, 1974
p. 23). He developed an open systems approach (Kaufman, 1989) to
analyse obsolescence, based on extensive questionnaires sent to technical
professionals in a single organisation. The approach considered four elements:
environmental change (e.g. rapid changes in technology); organisational
climate (determined largely by managerial policies and practices); the nature


of work (in terms of job assignments); and individual characteristics such as
the cognitive and personality factors related to keeping up-to-date. The last of
these is important in determining how individuals cope with rapid technological
change (Trimmer et al., 1998; Pazy, 1998; 2006). The US National Academy
of Sciences concluded in 1985 that engineers could work productively
over a longer period if they had access to effective continuing education
regardless of the state of the business cycle. This required interdisciplinary
approaches and non-technical skills that are not imparted by the formal
training of engineers.

CHAPTER 3

What causes skill mismatch?

This chapter analyses the causes of different types of skill mismatch. Based
on evidence, we discuss the causes of overeducation, skill shortages and
skills obsolescence, and reflect on the role of individual characteristics.
Overeducation can result from the asymmetric nature of the supply of
labour-market information. However, literature has devoted little attention to
explaining overeducation by considering the decision processes of firms and
individuals. A key issue is whether perceived mismatches represent labour
market sorting on the basis of the lower ability of the individuals concerned; it
is perhaps necessary to disaggregate overeducation according to the level of
job satisfaction or use of skills of overeducated workers. Insufficient training,
markets reacting slowly to changes, skill-biased technological progress and
business cycles are core factors explaining skill shortages. When skills
depreciate due to ageing or when economic developments change the skills
that firms need, this process is referred to as skills obsolescence. Young
people entering the labour market, older workers, females, ethnic minorities
and the disabled are disproportionately affected by skill mismatch.

3.1. Causes of overeducation
Education type and level may influence the likelihood of mismatch on entering
employment. As Stigler (1961) observed, the higher the level of education
the better defined the set of jobs a worker can obtain and the better the
information individuals have about the nature of the labour market. Further,
individuals in narrower fields of study (such as law or medicine) are likely
to have a better defined set of job alternatives than those in more broadly
defined ones (such as arts and humanities). Thus, Kucel and Byrne (2008)
suggested that those from broader educational backgrounds are likely to
be less well informed about labour-market opportunities. Using the 2003-05
British labour force survey, which contains information on type of degree
and the nature of the job finding process, they found that graduates in hard
disciplines are significantly less likely to apply for a job through a private
employment agency, by relying on an advertisement or by a method other


than through personal contacts relative to those from broader educational
backgrounds. Thus, overeducation can result from the asymmetric nature
of the supply of labour-market information.
Büchel (in Cedefop; Büchel, 2001) advocated regarding overqualification as
a result of a decision-making process by both employers and employees. For
employees, the choice between unemployment and work, or different types
of jobs, precedes overeducation. Employer decisions may involve dealing
with problems in recruiting workers with the right skills. Very few studies have
focused on the decision-making processes of both firms and individuals to
explain the occurrence and persistence of overeducation. Research on the
role of employers in hiring mismatched workers is also rare, although this is
a possible strategy for fluctuating demand for skilled employees and rising
demand for skilled manpower. Sloane et al. (1996) also pointed out that
undereducated workers compensated for lack of qualifications by having
more experience, tenure and training (the substitutability hypothesis), so one
cannot necessarily infer that mismatch is wasteful. In their matching model
for UK graduates, Battu et al. (1999) included variables capturing the nature
of the job (e.g. occupations, sector of work and size of firm). Unsurprisingly,
higher occupations increase the likelihood of matching, with individuals in the
education sector and private practice more likely to be matched and size of
firm being positively and strongly related to the probability of a match.
A key issue is to what extent perceived mismatches represent labour-
market sorting on the basis of varying quality of labour within education
levels. This is particularly so as the number of graduates has increased
substantially in many countries. If less able individuals have entered the
pool of graduates, quality may have fallen over time which may increase
the incidence of overeducation. For example, Battu et al. (1999) controlled
for the degree class and institution type in the UK, finding that the class of
degree has a significant effect on graduate earnings, with a first class degree
having a 8-13 % premium over a lower-second class one, and graduates
from more established universities earning 8-11 % more than graduates from
former polytechnics, holding constant a measure of overeducation. Dolton
and Vignoles (2000) and Green and Zhu (2008) also found that those with
better degree classifications and those from more established universities
are less likely to be overeducated.
There may also be cohort effects if individuals qualify for entry into
the labour market in periods when there is an excess supply of qualified
applicants. This will happen because it takes some time for workers to accept
a lower-skilled job and then continue searching for a more complex one

CHAPTER 3
What causes skill mismatch? 29

which pays more (Gautier, 2000). This may be reinforced when the individual
faces short-term constraints on the search process through personal, family,
location or other reasons. As Dolton and Silles (2003) also suggested, the
process of finding a suitable job may take some time, so we would expect
in this case, and in contrast to what was stated earlier, that overeducation
is a temporary phenomenon. Most studies of skill mismatch, however,
focus on the state of mismatch rather than on the process of becoming
mismatched; available data does not allow for analyses that follow the same
individuals in their employment careers, which makes a valid assessment
of persistence difficult.
Most data sets do not contain the data necessary to correct for individual
heterogeneity. One exception is the British national child development study.
Using this data set, Green et al. (1999) found that individuals who scored
higher on a maths test administered at age 16 were significantly less likely to
be overeducated later in their working lives, and this effect was substantial.
Their analysis of the International adult literacy survey (IALS) also suggested
that individuals who lacked good quantitative skills were more likely to be
overeducated. Büchel and Pollmann-Schult (2001) found that in Germany
a poor grade in the school leaving certificate had a strong effect on the
later risk of overeducation for graduates with a vocational degree and after
controlling for selectivity effects.
Box 1. Theoretical approaches to overeducation

There is no generally-accepted theory of overeducation (McGuinness, 2006), but it is possible
to consider how the phenomenon relates to various models of the labour market as illustrated
in the table below. The standard Becker/Mincer human capital model suggests in its simplest
form that wages are determined by endowments of human capital and particularly years of
education and experience. There is no role for demand side factors, where more years of
education imply higher earnings, and no distinction is made between actual and required
years of education. Thus, any year of education is just as valuable as another. In contrast, in the
Thurow job competition model, marginal productivity resides in the job rather than the worker,
and in the extreme, education simply serves to place the individual at the front of the queue
for jobs, as it signals that the employer will be required to invest less in training. Hence, the
return on overeducation is zero, as all workers in a given job are paid the same. In contrast,
the Sattinger job assignment model (Sattinger, 1993) considers both supply and demand to be
important determinants of wages, as there is an allocation problem in assigning heterogeneous
workers to heterogeneous jobs. At any time it is unlikely that there will be a perfect match and
overeducation can be a persistent problem. Earnings will be a function of both worker and job
characteristics and this model can encompass the human capital and job competition models
in a more general equation which is sometimes referred to as the Duncan and Hoffman model
(1981) or in other cases as the ORU model.


log Wt= β0 + β1qr+ β2qs β3qu+ε
In this model, actual educational qualifications (q) of individuals are decomposed into the
required or normally accepted level (qr), surplus or more than required qs), and deficit or less
than required (qu), qualifications in relation to those necessary to do the job. The human capital
specification implies that β1=β2=β3 and the job competition model that β2=β3=0. Tests of
these models, using an F test on the residual sum of squares, tend to suggest that generally
the job assignment model outperforms the other two. The theory of career mobility associated
with Sicherman and Galor (1990) suggests that individuals may be prepared to accept jobs
with low returns on education provided that this is accompanied by a higher probability of
promotion. In this context overeducation is a purely temporary phenomenon and additional
returns on education will be obtained later. Finally, Frank’s theory of differential overqualification
(1978) is based on the proposition that, for married couples, the search for a pair of jobs will be
constrained geographically and only infrequently will the best job offer for both spouses occur in
the same location. If the husband’s job choice dominates because of the wife’s responsibilities
for children, married women will face the greater constraints, whether as tied stayers or tied
movers and we expect married women to experience greater overeducation than either married
men or single men or women.

Theory Main proponent(s) Main determinants of wage Predictions of
overeducation (overskiling)
Human Becker (1964), Supply (personal Return on years of
capital Mincer (1974) characteristics) overeducation = return
model on years of required
education
Job Thurow (1975) Demand (job queues) Return on overeducation
competition is zero
model
Assignment Sattinger (1993) Supply and demand Return on years of
model overeducation < return
on required education
Career Sicherman and Galor Training costs are general • in the short-run
mobility (1990) and paid for by the return on years of
theory individual overeducation < return
on required education
• in the long-run
return on years of
overeducation < return
on years of required
education
Theory of Frank (1978) Gender and family status Return on years of
differential overeducation < return
overqualifi- on required education
cation

CHAPTER 3

Where data sets lack the appropriate variables, indirect methods are
necessary. Thus, Chevalier (2003) distinguished between apparently and
genuinely overeducated graduates in the UK based on their answers to
questions on job satisfaction. Graduates in graduate jobs are defined as
‘matched’, whatever their level of satisfaction. Those satisfied but in non-
appropriate (or non-traditional graduate) jobs are described as ‘apparently
overeducated’ and those in non-graduate jobs and dissatisfied as ‘genuinely
overeducated’. A similar approach was adopted by Allen and van der Velden
(2001), who used a 1991 Dutch data set which allowed job levels to vary
for given education qualifications. They distinguished between a formal
education mismatch and a skill mismatch based on two key questions:
whether the current job offered sufficient scope to use acquired knowledge
and skills (utilisation); and whether they would perform better in their current
job if additional knowledge and skills had been acquired (skill deficit). While
about 15 % of their sample experienced underutilisation, a much higher
proportion experienced skills deficits and the latter was only weakly related
to education mismatches.
These results cast doubt on one of the key assumptions of assignment
theory, namely that education mismatches imply skill mismatches. One
interpretation is that overeducation in part masks unmeasured ability
differences and is, therefore, spurious. Taking these results together, it
seems clear that a substantial part of what is described in literature as
overeducation merely reflects the heterogeneity of individual abilities and
of skills within particular education categories. Far from implying inefficient
allocation of labour, this suggests that the labour market is functioning
effectively in allocating workers to jobs which match their skills and abilities
(Sloane, 2003).
Brunello et al. (2007) argued that employment protection legislation might
increase the extent of skill mismatch by making it harder for individuals
to obtain their first job and more difficult for firms to reduce staff because
of restrictions on firing. Using the European Community household panel
1994-2001 they found a positive association between skill mismatch and
employment protection legislation. Their regression analysis for the five
largest countries suggests that men are more likely to be mismatched than
women (7), that overqualification increases with years of education (apart from
Germany), and declines with increased labour experience, which, in contrast
to evidence cited earlier, suggests that overqualification is a temporary

(7) This is in contrast to other findings reported elsewhere.


phenomenon. There are also important differences between countries, with
the suggestion that for Germany and the UK some workers may choose to
become overqualified to avoid unemployment. Adopting a fourfold classification
of mismatch in terms of combinations of overeducation, overskilling and
matching, it appears that skill mismatch rather than overqualification causes
pay penalties.

3.2. Underlying causes of skill shortages
Skill shortages can be caused by underinvestment in training. The idea of
market underprovision of skills can be traced back to Pigou in 1912 (cited
by Haskel and Holt, 1999) who argued that the additional transferable skills
associated with training will raise the probability that a worker will be poached,
thus reducing the likelihood that training will take place. Individuals may
underinvest due to perceived high training costs, coupled with uncertain
returns (Booth and Snower, 1996). Information gaps or inadequately structured
education provision may also lead to incorrect investments on the part of
students, while insufficient labour-market mobility can also lead to imbalances
(Neugart and Schomann, 2002).
Institutional factors may also prevent the labour market from clearing; for
instance, centralised wage bargaining systems may be unable to consider
sectoral or regionally-based changes in demand. A sluggish supply-side
may also contribute to the problem. For instance, Haskel and Martin (2001)
argued that skill shortages in the UK have arisen primarily as a consequence
of education supply failing to keep pace with skill-biased technical progress.
Finally, the incidence of skill shortages will vary according to the economy’s
position in the business cycle. Specifically, skill shortages tend to have a
strong negative correlation with the claimant count (Frogner, 2002). Hogarth
and Wilson (2001) go so far as to suggest that skill shortages are actually a
significant driver of the business cycle itself, placing a productivity constraint on
the economy which in turn leads to a downturn in activity. Further, the factors
listed above may also act as drivers for undereducation and underskilling.

3.3. Causes of skills obsolescence
Skills obsolescence, which may lead to skill gaps, is a process rather than a
state. De Grip and van Loo (2007) distinguished between physical (technical)
and economic skills obsolescence as outlined in Table 3. Technical skills

CHAPTER 3

obsolescence results from depreciation of manual skills or physical strength
through ageing, while economic obsolescence occurs as a result of changes
in production technology which alter the optimal input mix for firms. If there
are new skill requirements as a result of changes in production, this may
have a major impact on job specific skills. If employment declines across
an industry or occupation, this is referred to as sectoral skills obsolescence.
If firm reorganisation or closures occur more job losses may result in firm
specific skills obsolescence. Physical (technical) obsolescence and job loss
combined may result in organisational knowledge loss. Thijssen (2005) has
identified a third type of skills obsolescence: perspectivistic obsolescence.
This refers to outdated views and beliefs on work and the work environment.
Skills obsolescence can manifest itself through unemployment, crowding out
to lower-skilled jobs, lower wages or through problems in performing the job
properly. As workers become more educated and skill biased technological
change remains important, while at the same time the workforce is ageing,
skills obsolescence is likely to become even more significant.

Table 3. Types and causes of skills obsolescence

Types Forms of human capital depreciation

Physical (technical) obsolescence
Wear and tear • deterioration of mental and physical capacities through
ageing, illness or injury
• decreasing capacity to learn and adapt to new situations
Atrophy Insufficient use of skills as a result of overeducation, increased
specialisation or unemployment
Economic skills obsolescence
Job specific skills New skill requirements due to changes in production
obsolescence
Sectoral skills obsolescence Employment declines either by industry or occupation
Firm-specific skills Voluntary or involuntary quits resulting from firm reorganisation
obsolescence or closure
Organisational knowledge loss
Combining wear and tear of skills of individual workers or
quitting by workers with firm specific skills
Perspectivistic obsolescence
Outdated views or beliefs concerning the nature of work and
the work environment
Source: Adapted from de Grip et al., 2007.


Weinberg (2007) showed that skills obsolescence arises in two ways.
First, the diffusion of new technology may increase the obsolescence of
existing skills. Second, experienced workers who have accumulated greater
stocks of skills relevant to the old technology lose more human capital when
switching to new tasks. Contrary to expectations, skills obsolescence occurs
not only in strongly field-specific and technical fields of study, but also in
more generic and non-technical fields (Allen and van der Velden, 2007).
The traditional vintage human capital model emphasises the imperfect
transferability of human capital across technologies, implying that younger
workers will be the primary adopters of new technology. New technologies
may also complement existing skills and there is some evidence that this
applies in particular to more educated men with substantial (30-40 years)
experience.

3.4. Mismatch and individual characteristics
Are there particular personal characteristics which make certain individuals
more prone to mismatch than others? Particular groups may be more subject
to discrimination in the labour market – young workers, older workers,
females, ethnic minorities and the disabled – and this feature may also
manifest itself in labour-market mismatch.
Young workers are likely to be particularly prone to skill mismatch as new
entrants into the labour market, and the finding that overeducation is linked
to a lack of work experience supports this hypothesis. One Dutch study by
Dekker et al. (2002), for example, found that the proportion of overeducated
decreased from 41.7 % for the 15-19 age group, to 27 % for the 30-44 age
group and 18 % for the 49-64 age group. Many other studies report similar
results. Older workers may suffer from mismatch as a consequence of skills
obsolescence, which is more likely in industries subject to rapid technological
change or in decline through shifting patterns of demand.
In an early US study, Frank (1978) examined the link between gender and
overqualification. He suggested that, in a dual earner household, there would
be greater constraints on married women than married men on account of
child-rearing causing intermittent labour-force participation on the part of
the wife. Married women would tend to be tied stayers or tied movers, thus
limiting their ability to obtain optimal matches in the labour market. However,

CHAPTER 3

McGoldrick and Robst (1996), who attempted to test whether married women
in small local labour markets were more prone to overqualification than those
in larger labour markets, could find no evidence to support this proposition.
Büchel and Battu (2003) extended this analysis using German data by
making allowance for commuting distances between home and work and
found that longer commuting distances tended to reduce the likelihood of
overqualification, though not differentially between men and women.
An alternative reason why one might expect gender differences in mismatch
is the presence of fixed costs of employment together with higher quit rates
for women than men. This may lead employers to require higher ability from
women relative to men when hiring for particular jobs. Thus, van Ours and
Ridder (1995) found that women in their Dutch data set required almost
six months more work experience than men to be hired for the same job,
implying that they were overqualified in this sense. Yet, overall there is
mixed evidence of skill mismatch being a more serious problem for women
than for men across countries, with counter examples. Why there are such
differences between countries seems to be an important issue for further
research.
Studies of skill mismatch across ethnic minorities are limited in number.
Battu and Sloane (2004) used the Fourth national survey of ethnic minorities
1993/94 to investigate whether overeducation was greater for ethnic minorities
in Britain than for the white majority using a modal measure of required
education across 60 occupations. They found that overeducation was higher
for non-whites than for whites, while undereducation was lower for non-
whites, but within the non-white population the African Asian group was
more likely to be overeducated than Indians. The consequences of labour-
market mismatch in terms of the wage penalty were more severe than for
the indigenous white population. Lindley (2009) examined the extent of
overeducation among UK immigrants using labour force survey data for
1993-2003, again estimating over- and undereducation in terms of deviations
from the mode. The results were largely consistent with those of Battu and
Sloane (2004), though in this case Indian immigrants were more likely to
be overeducated compared to indigenous whites.
Studies of labour-market mismatch among the disabled are notable by their
absence, and often a disabled dummy is not even included in the estimating
equation. Where it is, there is sometimes a significant effect on the probability
of overeducation. This is clearly an area for future research.

CHAPTER 4

What are the links between
the different types of skill
mismatch?

This chapter analyses to what extent, how and why different types of skill
mismatch occur simultaneously. First, we look at skill shortages, skill gaps,
undereducation and underskilling. We then examine whether overeducation,
overskilling and skill shortages can occur simultaneously. In the third part of
the chapter we focus on the duration of particular combinations of mismatch.
There does not appear to be a strong correlation between the occurrence
of skill shortages and skill gaps, but the latter may be closely linked to
underskilling. Both overeducation and overskilling may be a consequence
of excess labour supply, which can be illustrated by the Beveridge curve.
Skills obsolescence is different from the above as it is a process rather than
a state. Levels of skill are important as it is open to a skilled worker to obtain
an unskilled job, while it is not normally open to an unskilled worker to obtain
a skilled job. This is reflected in the concept of skill bumping (or crowding
out). Various types of mismatch can be present at the same time in individual
firms but not often in the same occupations. Ideally we require matched
employer-employee data sets to cast light on all these phenomena.

4.1. Skill
shortages, skill gaps, undereducation
and underskilling
Skills related problems can arise because of skill shortages or skill gaps.
Skill shortages refer to a skills related labour-market shortage, i.e.
where there are not enough individuals with the required skills within the
economy to fill existing vacancies. This is the most simplistic definition to
skill shortages. However, measurement issues are clearly important (e.g.
Green et al., 1998) and some recent attempts have been made to move
towards a more refined system of definition that incorporates training lead
times. For example, Richardson (2006) proposed the following system for
defining skill shortages:
(a) level 1 shortage: there is a shortage of individuals with the relevant
technical skills who are not already using them and it takes a long
training time to develop these skills;

CHAPTER 4
What are the links between the different types of skill mismatch? 37

(b) level 2 shortage: there are few people who have sufficient skills and are not
already using them, but such skills can be developed relatively quickly.
Skill gaps occur when skill levels in the workforce are below those desired
by employers or when job requirements do not match precisely the content
of the knowledge and abilities of individuals.
The 2001 UK employer skills survey (UKESS) concluded that skill
shortages were more commonly associated with skilled occupations with
more substantial education and training requirements. In terms of skill
gaps, the UKESS study found that these were more strongly related to poor
human resource management practices such as lack of staff training and
poor labour retention. Therefore, it is not at all evident that skills gaps and
skill shortages are necessarily strongly correlated.
Each of these concepts points to the difficulties employers are faced with
in attracting sufficient labour of the right quality. Skill shortages are not easy
to measure, but indicate that the demand for labour exceeds the (available)
supply at the given market wage. However, if employers are not willing to pay
the market wage to attract additional labour, this should not be considered
a labour shortage. In a classic paper Arrow and Capron (1959) focused
on the possibility that lags in decision-making could result in a dynamic
shortage. Such lags may occur through the time taken to recognise that
there is a need to raise salaries, in obtaining approval for the increase and
in putting the new pay policy into action. The reaction speed of an employer
will be influenced by institutional arrangements such as the presence of
long-term labour contracts and partly by the speed with which information
about salaries, vacancies and the availability of workers becomes generally
available throughout the market.
Wage adjustments are, however, only one possible means of adjustment
and non-wage adjustments appear to be more common in the short-term.
Common examples are reducing output, adjusting inventories, increasing
overtime, using more vigorous recruitment methods or reducing hiring
standards. It is in the last of these circumstances that skill gaps may arise
in the sense that employees hired under such circumstances are less likely
to possess the skills necessary to complete tasks with maximum efficiency.
Skill gaps seem to equate with underskilling and it is difficult to differentiate
between the two, apart from the fact that skill gaps may reflect the views of
the employer and underskilling the views of the employee. In both cases
the remedy seems to lie with the provision of additional training.
In one study of Northern Ireland, Bennett and McGuinness (2009) found
that 58 % of employers in the IT and engineering sectors adjusted to the


problem of skill shortages by training and upskilling their existing staff,
and 47 % chose to recruit staff from other backgrounds and train them up.
Under such circumstances workers may find themselves undereducated
in relation to existing workers in the occupation to which they have been
upgraded. Though individuals in this category may lack formal qualifications,
they may have previous experience of the type of work, so that they will not
necessarily be underskilled.
All the above scenarios imply that demand exceeds supply. Where supply
exceeds demand, a reverse scenario will apply. If there has been a rapid
increase in the numbers graduating with certain qualifications skill surpluses
may arise, resulting in the downgrading of some individuals who find it more
difficult to obtain a job matching their qualifications. This crowding out may
result in skill bumping in which workers with lower qualifications are also
subject to downgrading or lose their jobs altogether. Both overeducation and
overskilling may result in these circumstances. If overeducation is caused
by too many qualified workers chasing too few jobs we would expect a high
rate of unemployment among those workers with lower qualifications who
have to compete with more educated workers, as well as a low vacancy rate.
Overeducation can also result from institutional regulations, if unemployed
persons are required by employment officers to accept lower-skilled jobs than
those for which they are qualified to avoid losing unemployment benefits,
as occurs under workfare policies.
Underskilling may arise as a result of skills obsolescence, which occurs
when skills become less valuable in the production process. This particularly
applies to older workers. While they may be retrained in new production
techniques, such as computer technology, this may be uneconomic because
of the shorter pay-back period compared to that of younger workers. A
further problem is that a mismatch between the level of education and/or
skills and the job lack of use of capabilities may lead to a decline in workers
cognitive abilities (the ‘use it or lose it’ phenomenon as described by de
Grip et al., 2008).
Suitable employee-employer linked data are undoubtedly required (see
Chapter 8) if we are to understand fully and quantify the relationships
that exist between skill mismatches for the firm and skill mismatches for
individuals. However, in the absence of such data, it is possible to develop
some reasonable and ultimately testable hypotheses based on existing
knowledge. First, evidence suggests that little overlap between skill shortages
and skill gaps exists. For instance, data from the 2001 UK employer skills
survey revealed that less than 1 % of establishments experienced both skill

CHAPTER 4

shortages and skill gaps. Such a weak occurrence supports the view that firms
experiencing skill gaps have little difficulty in recruiting staff, suggesting that
skill gaps tend to be more heavily associated with skill deficiencies, while skill
shortages are more commonly associated with higher level functions (8).

Box 2. The Beveridge curve

Both cases of supply exceeding demand and demand exceeding supply can be portrayed on the
Beveridge curve (see below) which relates unemployment to unfilled vacancies. This assumes
that there is labour-market equilibrium when the number of unemployed (U) equals the number of
unfilled vacancies (V). When U does not equal V there is an imbalance at the level of the economy,
region, industry, occupation or individual firm. However, this pays no attention to the quality of
employed labour relative to the quality of filled jobs. Those in employment may be under- or
overutilised as reflected in levels of education qualifications, skills or abilities.
Vacancies

B

UV2
A

UV1

0
Unemployment

(8) Haskel and Martin (2001) showed that skill shortages are greater in the high-tech sectors of the
economy.


Both A and B represent points of equilibrium but, in the latter case, vacancies equal
unemployment with a higher level of both vacancies and unemployment, representing increased
structural unemployment. In this sense there is a greater degree of mismatch either because
labour is immobile or it is equipped with insufficient or the wrong sort of skills.
When an economy is subjected to shocks, such as rapid technological change, skill mismatch
will occur if workers cannot readily learn new skills or are unwilling to accept jobs that do not
utilise their existing skills, and employers are unable or unwilling to redesign jobs so that they
can be performed by people with a different set of qualifications from those originally intended.
Geographic mismatch may arise if neither workers nor jobs are fully mobile because of the
costs of mobility. Other considerations are on the employer-side, whether minimum hiring
standards can be met by applicants or wage offers are acceptable to them. This suggests that
labour market dynamics are likely to play an important role. In a recent study Ochsen (2009)
estimated the Beveridge curve for nine countries, eight of them European, over the period
1960-99. He found that mismatch is not generally lower in those countries considered to have
more flexible labour markets, though it has decreased in some countries in which the labour
market has become more flexible.

Frogner (2002) reported on evidence from an occupational disaggregation
of the UK employer skills survey data that tends to verify this view (9). However,
this is not to say that skill gaps cannot occur among highly qualified staff.
For instance, McGuinness and Bonner (2002) and Bennett and McGuinness
(2009) uncovered skill gaps in soft skill areas such as business awareness
and communication among IT and engineering professionals.

4.2. What is the relationship between
overeducation, overskilling and skill
shortages?
It is important to determine whether overeducation results from supply-side
or demand-side reasons. The former may include accepting jobs for their
amenities, location advantages or family reasons. The latter may include a
lack of jobs with appropriate education requirements or discrimination on
the grounds of gender, race, disability or other group characteristics.

(9) Professional, associate professional and skilled trades were found to account for 56% of all skill
shortage vacancies despite accounting for only 34% of all jobs. With respect to skill gaps, according
to the 2001 UK employer skills survey, firms employing sales/service staff, personal service staff
or workers in elementary occupations, were most likely to report skill gaps.

CHAPTER 4

If, when advertising a vacancy, an employer requires a higher level of
education than is necessary to perform the job, it is referred to as credentialism.
This is a term widely used in sociological literature to describe a situation
in which employers require diplomas or certificates for jobs to screen and
identify personal traits that are desirable in the workplace. It is assumed
that people with higher level diplomas or certificates are inherently more
productive. This may also be justified on the grounds that more educated
individuals learn faster or change the nature of the job in such a way that
their productivity eventually becomes higher than that of less educated
workers. Credentialism is more likely to occur when there are skill surpluses.
Employers hiring workers with lower education than required is likely to be
exceptional and reflects skill shortages.
We would expect a negative relationship between overeducation and/
or overskilling and labour shortages. Since shortages are an increasing
function of excess labour demand over labour supply, it becomes necessary
for employers to make maximum use of existing labour and we would expect
it to be easier for previously overeducated or overskilled employees to find
better matches. Where skill shortages and/or overeducation/overskilling
coexist this is symptomatic of individuals possessing the wrong skills or type
of education or preferring their current ‘mismatch’ because of compensating
job differences.
Few studies have examined the relationship between these variables
over time because most data sets containing appropriate questions tend
to be single cross-sections and only panel data extending over reasonably
long-time periods can consider changes in employment status. We would
expect a continuing shortage of, say, engineers to be reflected in reduced
rates of overeducation and overskilling among engineering graduates. In
contrast, a surplus of engineers resulting from a sudden increase in the
supply of engineering graduates would be expected to lead to a bumping
down phenomenon in which non-graduate engineers are crowded out of
higher engineering jobs. Skott (2006) has noted an asymmetry between the
options facing high- and low-skill workers. It is open to high-skill workers who
are unable to obtain high-skill jobs to accept low-skill jobs for which they are
overeducated. An equivalent option is not open to low-skill workers since
they are not qualified for high-skill jobs. This could give rise to a decline in
both the relative employment and the relative wage of low-skilled workers
when faced with a transfer of high-skill workers into low-skill jobs. Skott
observed that this occurred between the early 1970s and the mid-1990s
in the UK and the US. Any tendency of employers to keep skilled labour in


the recession because they fear it will be difficult to recruit skilled labour in
the recovery phase will tend to weaken the above tendency.
The idea that, as the proportion of the labour force with qualifications
increases, some of the educated may be forced into lower-prestige jobs
(generating overeducation) has been examined by Battu and Sloane (2000).
If the less educated are bumped down this implies that the mean level of
education will rise across all occupations and some of the uneducated may
be forced into unemployment. This, in turn, implies that wage differentials
overall will widen, thus offsetting any tendency for the wage differential of
the educated to fall relative to the wage for uneducated workers with an
increase in the supply of qualifications.
Using data from the social change and economic life initiative and the
British household panel survey, both of which use the Hope-Goldthorpe scale
to measure occupational prestige, Battu and Sloane (2000) found evidence
of skill bumping for Britain between 1986, 1991 and 1995, but this declines
as one moves down the occupational hierarchy and does not extend to the
lowest education qualification categories (10). In a similar study for Belgium,
Nicaise (2000) found evidence of bumping down for men, but not for women
(11). However, if crowding out is important we would expect to find that it is
concentrated in recessionary periods when skilled workers find it more difficult
to obtain jobs at their own level. Further, it if is a temporary phenomenon
only new workers are likely to be subject to it, as existing employees may
have job protection. Using Dutch data for the period 1992 to 1995, Gautier
(2000) found relatively little evidence of crowding out, but where it does
happen it seems to be a temporary phenomenon and to be driven either
by short-term variations in the supply of high skilled workers or by shocks
to the relative productivity of high skill jobs. Asplund and Lilja (2000, for
evidence of bumping down in Finland) found that the deep recession in
the early 1990s led to a substantial decline in the employment of the least
educated and this boosted the upgrading of occupational skills, consistent
with bumping down. It remains true, however, that there are relatively few
studies which have attempted to assess the extent of crowding out and skill
bumping across countries and over time.

(10) The Hope-Goldthorpe scale has 36 categories ranked by ‘social desirability’ of male occupations.
The categories are assumed to provide a high degree of differentiation in terms of both occupational
function and employment status. It is important to highlight that this class schema was devised for men,
but the scores are commonly used for both men and women (Goldthorpe and Hope, 1974).
(11) It is suggested that women may be more driven by family considerations than are men. Thus,
non-employment of women is more frequently determined by supply-side constraints than is the
case for men and this tends to offset the bumping down effects on wages.

CHAPTER 4

4.3. Combinations and durations of mismatch
Some general predictions can be developed about the likely relationships
between worker and firm mismatches. Following Cedefop (Cedefop, Descy
and Tessaring, 2001) under and overeducation (or more strictly under and
overqualification) occurs when qualifications of workers do not compare
with job entry requirements; over and underskilling occur when job content
does not compare with workers’ skills and competences. We wish to draw
out potential interrelationships between measures of skill shortages at
individual and firm levels. Table 4 summarises some potential connections
for firms.

Table 4. Combination of mismatch for firms

Skill shortages Skill gaps

Overeducation Overeducation and skill shortages There is a potential correlation as the
are not likely to occur together overeducated may still not possess
within occupations within firms. the correct mix of skills required by
But they can coexist across the firm. However, overeducation will
occupations within firms. not be correlated with basic skill gaps.

Overskilling Overskilling and skill shortages These can coexist within occupations
are not likely to occur together within firms but only in different
within occupations within firms. dimensions whereby the skills surplus
But they can coexist across to the individual are not the skills
occupations within firms. demanded by the employer.

Undereducation Undereducation and skill Within firms and occupations, while
shortages will exist within we do not expect both to coexist in
occupations within firms if firms basic occupations with low entry
are reacting to skill shortages by requirements it is possible for them to
upskilling existing staff. coexist as the education requirement
of the occupation rises.

Underskilling Underskilling and skill shortages We might expect some correlation
will exist within occupations within occupations and within firms
within firms if firms are reacting if both the worker and the employer
to skill shortages by upskilling recognise they are deficient in the
existing staff and some deficits required skills.
remain.


In terms of overeducation and skill shortages, the occupations within
which the firm employs overeducated workers will not correspond with the
occupations where skill shortages are being experienced. However, this is
not to say that the firm cannot simultaneously experience overeducation
and skill shortages in different occupations. Nevertheless, there is nothing
to suggest a strong correlation is probable at firm level. At macroeconomic
level, a high incidence of both overeducation and skill shortage would provide
a strong indication that the composition of education supply is poorly aligned
with labour demand. Turning to skill gaps, it is possible that overeducated
workers within the firm may also be those identified as possessing skill gaps:
the overeducation may be indicative of the fact that such workers possess
the incorrect mix of skills required by the firm. However, the consistent
finding that the overeducated earn more than matched individuals doing the
same job would tend to suggest the overeducated do not have skill gaps.
Further, the overeducated, by nature of their education attainment, will not
tend to be deficient in basic competences, where skill gaps are most likely
to occur, leading us to the general conclusion that overeducation and skill
gaps are also likely to be poorly correlated at firm level.
Overskilling arguably represents a much more direct measure of individual
underutilisation than overeducation. Nevertheless, the relationship between
overskilling and skill shortages remains unchanged from that outlined for
overeducation. While both phenomena may occur across occupations within
firms they will not tend to be highly correlated within occupations. Once again,
at economy level, a high incidence of both overskilling and skill shortages
suggests that some rebalancing of post-compulsory education and training
supply would benefit both workers and firms. Overskilling and skill gaps
can, however, occur both within occupations and within firms, but only in
different dimensions, whereby the skills that are surplus to the individual
are not the skill areas in which the firm is experiencing a shortfall. Unlike
overeducation, where the relationship between skill gaps can be discounted,
no study has yet demonstrated that the overskilled earn a premium relative
to matched workers carrying out the same job, leaving open the possibility
that the overskilled may, nonetheless, lack essential skills.
Evidence suggests that many employers attempt to overcome the problem
of skill shortages by upskilling their workers to fill vacancies. It may well be
that reallocated workers do not possess the credentials required to fill the
post; therefore, some correlation between undereducation and skill shortages
might be expected. However, evidence from firm-level surveys suggests that
skill shortages relate to a lack of individuals with the necessary skills and

CHAPTER 4

experience, as opposed to qualifications, and so the within-firm correlation
is likely to be weak. In terms of skill gaps, if the deficiencies relate to basic
skills then undereducation is unlikely, given that entry requirements to the
occupations in question will be relatively low. However, the likelihood of
correlation will be higher the greater the entry requirement of the positions
where skill gaps are being reported.
Finally for underskilling, we might expect a relatively strong relationship
between this phenomenon and skill shortages in instances where firms
seek to relieve the problem of unfilled vacancies by training and reallocating
existing staff. Specifically, if upskilling deficiencies remain in terms of the
individual worker’s ability to carry out the job in question, then upskilling
will result in skill gaps. So, there are circumstances where we might expect
strong correlation between underskilling, skill shortages and skill gaps at
firm level.
Table 4 neglects skills obsolescence, which is different in nature as it is
strongly related to technical change, to rapid changes in work organisation
and to shortcomings in higher education in particular. It is only weakly
related to current skill shortages and not at all to the prospects for future skill
acquisition (Allen and van der Velden, 2007). In fact, skills obsolescence
should be considered a process rather than a state.

CHAPTER 5

Business cycle impact on skill
mismatch

This chapter considers the impact of business cycles on the occurrence
of skill mismatch. It is difficult to assess what impacts the economic crisis
Europe is facing at the time of writing will have on skill mismatch. Little is
known about the relationship between skill mismatch and the business cycle
because of the lack of extensive panel data, but there are reasons to believe
that such a relationship exists. Some general assessments are possible:
there is also evidence that skill mismatch is a persistent phenomenon in
difficult times and that overskilled workers who change their jobs do not
always improve their position.

5.1. Skill mismatch and business cycles
Few, if any, studies have examined the relationship between various dimensions
of skill mismatch and the business cycle because most data sets which
contain appropriate questions tend to be single cross-sections. To understand
what happens over the business cycle requires panel data extending over a
sufficiently long period to pick up the various stages of the cycle.
By their very nature, business cycles result in a continual change in the
balance between labour demand and supply. Change happens in relation to
the match between the stock of job vacancies and the skills of the unemployed,
and to the flow of jobs and skills into and out of the labour market through
job destruction, job creation, training and skills obsolescence. The impact of
business cycles varies by sector and occupation. Volatility in cyclical excess
product demand or supply is likely to influence labour-market outcomes in
terms of more or less under- and overeducation and under- and overskilling.
In the first instance, business cycles will impact on skill mismatches through
changing the level and composition of the demand for skills and the returns
on the way in which these skills are matched in the workplace (reduced in
a downturn, increased in an upturn). It is also likely that the response of the
labour market itself could influence the duration and depth of a business
cycle, especially through the ability of the market to preserve existing skills
and offer training in new ones.

CHAPTER 5
Business cycle impact on skill mismatch 47

Training is crucial in this context. The way the costs of training end up
being split between the worker, the employer and possibly the State could
be influenced by the business cycle. It could also influence the resulting
skills composition and ability of the labour market to match jobs and workers
during a recession and the recovery that follows. The distinction between
(firm or sector) specific and more general training and education matters
because of the financial incentives they imply regarding who covers the
cost of training. Firm/sector specific skills are not transferable to other firms/
sectors. It follows that firms have a stronger incentive to contribute to the
cost of generating specific skills rather than general skills, as the former
cannot be poached as easily.
Recent research by Mavromaras et al. (2009d; 2009c) using UK and
Australian data suggested that skill mismatches are less prevalent and
their negative consequences are less severe where education is more
specific to the needs of the employer (i.e. a higher proportion of ‘specific’
versus ‘general human capital’). However, this finding was based on data
that included only years in which the economy was doing well. The recent
downturn indicates that, what can be an advantage during a buoyant economy,
can also become a problem during a downturn, because workers with
more (sector, occupation) specific skills are less flexible in shifting to new
jobs if and when their own sector/occupation is adversely affected by the
downturn. In a recession, the best skills for a displaced worker are general
ones, as they allow broader search in sectors that may have been less
affected by the downturn. For example, highly specialised vehicle industry
workers may suffer a larger drop in pay during the current recession than
less specialised but equally well qualified service industry workers. There
is little research in this area and generating evidence using conventional
large survey data sets will not be feasible for some time: the data at hand in
2009 just touches on the starting phase of the current downturn and so all
data series refer to very different general economic conditions. Further, the
data from previous recessions may not be as useful because considerable
institutional changes have taken place in the last two to three decades in
most Western economies.

5.2. The impact of the economic crisis
Despite the lack of relevant research and appropriate data, some developments
are easy to forecast for the recent downturn. The (re-)training effort will have


to be intensified before the upturn has even started because of its lagged
effects. If we wait until an upturn is clearly visible, we will experience a slower
recovery handicapped by skill shortages. However, there is a natural reduction
in the willingness of employers to pay for training during a recession. A
prime reason for this is that employers contribute to the cost of training and
recessionary pressures have a negative effect on such investment. There
is also a natural increase in the willingness of individuals, especially the
unemployed and new labour-market entrants, to purchase more education
and training as this may be the only way to obtain a new job.
Further, there is a considerable reduction in the social opportunity cost of
the State paying for the worker training during the current recession, which
will be the stronger the deeper the recession proves to be. This applies in
particular to displaced workers, but also to young labour-market entrants.
There may be problems, however, with these hypotheses. First, the displaced
workers will probably be the least productive ones and their training may not
be the most cost-effective intervention (12). Although equity arguments may
shift the balance in favour of training the less able displaced workers, as
may the presence of a strong unemployment support and welfare net, such
considerations may not reduce the level of skill mismatch when the upturn
occurs. Second, it is not clear where and how training should be provided.
A mix of classroom and workplace training seems appropriate, as is a mix
of specific and general training. However, there is little research to guide
specific choices and the danger is always present in subsidising skills and
sectors that have little economic future but a strong lobbying presence in
the current industrial scene.
Skill shortages are also related to both the business cycle and the extent
of skill mismatches. In the upturn, skill shortages will lead to increases in
undereducation and underskilling as labour with appropriate education and
skill is in short supply. Conversely, in the downturn new entrants to the labour
market may experience difficulties in locating jobs with the appropriate levels
of education and skill.
There is some suggestion from skills shortage literature that some
shortages persist over most stages of the business cycle in some skilled
occupations: what varies is their extent over different stages of the business
cycle. Such shortages are also often evident in industries and occupations
with modest employment growth or even employment decline. Shortages

(12) The same will not apply to young labour-market entrants, where long-term adverse consequences
will influence a much wider segment of ability distribution.

CHAPTER 5
Business cycle impact on skill mismatch 49

can coexist with relatively high overall unemployment or in the occupation
under analysis. One reason is that, in cyclical downturns, employers often
reduce training places and apprenticeships, which can affect the number or
trained workers for many years to come, including the upturn years when
demand for skills may have fully recovered. Further, qualified workers who
lose their job during downturns often do not return to the same industry or
occupation. This can result from demand or supply changes in industry
restructuring, making some of the older skills obsolete, or individual workers
having experienced long-term unemployment/non-participation during the
downturn, with adverse effects on their ability to work when the upturn
occurs. In these situations undereducation and underskilling may also exist
regardless of the stage of the business cycle, and it can play a role in the
intensity of downturns and the speed at which an upturn in product demand
can be translated into a wider economic upturn. The role of mature industries
and mature workforce (sometimes the two go together) may intensify the
effect of mismatches and may indicate the need to train younger workers
and retrain older workers.
There may also be cohort effects in how graduates are matched with jobs
related to the business cycle. If graduates enter the labour market in a period
of recession they may be more likely to experience overeducation; the effect
of this may persist over the longer term. Further, taking the employment
of an overeducated graduate (presumably to a non-graduate job) as an
example, this employment is likely to cause bumping down as a side effect,
displacing a non-graduate worker who may be less qualified generally, but
perfectly well qualified for the (non-graduate) job the overqualified graduate
obtained. It is well established in literature that increasing unemployment
in the less qualified part of the labour market may cause severe scarring
effects. The longer the lead time needed for training any specific type of
graduate, the more likely it is that cyclical volatility may have detrimental
scarring effects. Similar phenomena may be observed where there may be
sudden sharp increases in the number of graduates even if the economy is
not in recession, as absorption is more difficult. This is again accentuated
where there are long training lead times.
Some of the more general literature on the nature of dynamic labour-
market adjustment is relevant. Although worker quits are often associated
with both wage gains and higher job satisfaction, many workers quit again
within one or two years. There are two rival explanations. First, Burdett
(1978) assumed that job quality in a new job is perfectly observable ex
ante, so that repeated quit behaviour results from workers using new


jobs as a stepping stone to even better jobs. Second, Jovanovic (1979)
assumed the complete absence of such information, so that repeated job
moves result from disappointment over job quality. Gielen (2008) observed
that an appropriate test is whether any improvement in job satisfaction is
permanent. The British household panel survey shows that the stepping
stone model can explain up to 80 % of repeated job behaviour and 20 %
learning about job quality. Using Finnish data, Bockerman and Ilmakunnas
(2009) suggested that heterogeneity of workplaces and wage compression
hinder creating sufficiently compensating wage differentials, so that adverse
working conditions encourage workers to quit. The implication is that, at any
time, labour markets will be in disequilibrium.

5.3. Evidence from Australia
In one of the few attempts to measure the dynamics of overskilling,
McGuinness and Wooden (2007) used the first four waves of the household
income and labour dynamics for Australia (HILDA) data set. The results
suggest that, while the overskilled express a desire to quit in order to obtain
a job which offers a better match, they are not confident of their ability to
achieve this. Further, though they have a higher propensity to be mobile
than matched workers, some of this higher mobility is due to involuntary
separations; even where job separations are voluntary, most job moves do
not result in improved skill matches. Also, many overskilled workers move into
unemployment or inactivity. This adds to evidence that policy should focus
on skill mismatch rather than education mismatch. Using the first six waves
of HILDA, Mavromaras et al. (2009b) examined the extent of persistence
defined as the ability of the state to perpetuate among the overskilled. The
strongest persistence was found for those with higher education. Lower
persistence, but still strong and statistically significant, was found for those
with up to year 12 schooling. There is no trace of overskilling persistence for
those with vocational education and training, indicating that mismatches are
much easier not only to avoid but also to correct for those with vocational
training. A caveat for this, and many other contemporary findings, is that
they have been derived from data drawn exclusively from Australia in an
economic upturn. It is not clear that these findings can be generalised and
extrapolated to study today’s recession climate in Europe without careful
consideration. Despite this caveat, evidence suggests that, at least for some
groups, overskilling is not a purely temporary phenomenon.

CHAPTER 6

The implications of skill
mismatch

This chapter discusses the implications of skill mismatch. The next section
examines the impact of mismatch in terms of productivity and discusses
overeducation, skills shortages and skills gaps. The following section reviews
the impact of skill mismatch on job satisfaction. We then turn to possible
positive impacts of overeducation or overskilling for employers and end by
reflecting on the overall costs of mismatch for society. Skill shortages have
a significant impact on productivity, GDP growth and earnings, though less
is known about the effects of skill gaps or skills obsolescence. We know that
overeducation reduces job satisfaction for individual workers while, for firms,
hiring overeducated workers can be a deliberate strategy to benefit from
extra skills in the long term, despite negative impacts in the short term.

6.1. Skill mismatch and productivity
We might expect that both skills shortage and skills gap can lead to a loss
of competitiveness as wage rates are bid up and productivity lowered within
industries where skill problems exist. Productivity may also be harmed as
firms may be forced to place lower-skilled workers in skilled positions and/
or, if in an area with skills shortage, workers use their position to alter their
terms and conditions inefficiently from a productivity perspective.
Finegold and Soskice (1988), and Haskel and Holt (1999), take the analysis
a step further by describing a scenario whereby skills shortage can lead to
low-skill equilibrium within the economy. They described a situation where
firms react to skills shortage by investing in technologies that complement the
low-skilled element of the workforce, resulting in an economy characterised
by low wages and a low demand for high-skilled labour. As a consequence of
the low demand for high wage/high-skilled workers within the labour market,
individuals will no longer invest in their own human capital, thus perpetuating
the low-skill equilibrium position. Tether et al. (2005) argued that once a low-
skill equilibrium has become established, there are no obvious policy levers
for correcting the situation. However, despite some extensive discussion
of the theoretical concept in literature, there is little actual evidence linking
skills shortage with the emergence of a low-skill equilibrium.


6.1.1. Overeducation and productivity
While the wage effects of overeducation, overskilling, undereducation and
underskilling are well known (Hartog, 2000; Sloane, 2003; McGuinness,
2006; Mavromaras et al., 2009a; 2009d) their implications for firm productivity
are less clear cut. In imperfect labour markets, firms may obtain a rent from
the services of their employees, which means that the link between the
wage and productivity is imprecise. There may also be differences between
short-term and long-term effects, since overeducation is known to reduce
job satisfaction, which in turn may negatively influence performance, both
directly and indirectly, through its effects on absence and quit rates. Tsang
and Levin (1985) suggested that overeducated workers often exhibit counter-
productive behaviour in the workplace, but this might also be influenced by
their lower ability, health problems or a different attitude towards work.
The effects of overeducation on productivity can be measured in two ways.
If wages equal marginal productivity, the finding that overeducated workers
receive a wage premium over matched colleagues, though less than matched
workers with the same education as themselves, suggests that such work-
ers are somewhat more productive than matched colleagues. An alternative
and preferable approach, given the presence of imperfect labour markets, is
to measure productivity directly through analysis of output per head in indi-
vidual companies or industries. The expectation is that, since overeducation
reduces job satisfaction, there will also be a negative effect on productivity,
through higher shirking, absence and quit rates. As Hartog (2000) observed,
this leaves a puzzle since, along this line of reasoning, overeducation should
reduce wages rather than increase them as the wage evidence suggests.
If workers are genuinely mismatched, in the sense that they are not fully
utilising their education qualifications or skills in their current jobs, reallocating
labour appropriately should increase productivity and GDP/GNP at macro
level. However, as Borghans and de Grip (2000) pointed out, to assess this
properly we require better understanding of the skills that people currently
possess and the precise ways in which they utilise these skills.
Unfortunately, much less is known about this than about the allocation
of different types of workers in the labour market. Also, what is known is
largely confined to single case studies. In one of the few direct attempts to
measure the effect of overeducation on productivity, Tsang (1987), applying
a Cobb-Douglas production function approach and using data on companies
within the US Bell Corporation, found evidence of negative productivity
impact. Unfortunately, Tsang’s (1987) study is the only attempt to measure
productivity effects using company data.

CHAPTER 6
The implications of skill mismatch 53

6.1.2. Skill shortages and productivity
Skill shortages are important for their impact on productivity, GDP/GNP
growth and earnings. Frogner (2002) cited descriptive evidence from the
2001 European social survey to support this in terms of the productivity
impact, although evidence on earnings was more limited. A study by the
National Institute of Economic and Social Research (Mason and Wilson,
2003) reported that relative to German plants, output per worker in British
firms was over 60 % lower, with the authors arguing that this productivity gap
was a consequence of lower-skill levels. Further, Haskel and Martin (1996)
suggested that skill shortages reduced annual productivity growth in the UK
by 0.4 percentage points over the period 1983-99. On productivity corollaries,
Nickell and Nicolitsas (1997) estimated that a permanent 10 % rise in firms
reporting skill shortages lowers investment proportionately, and R&D by a
temporary 4 %. Evidence from more specific sectoral studies include that
of Bennett and McGuinness (2009), who reported that, after controlling for
the influence of selection bias, output per worker was lower in high-tech
firms experiencing both hard-to-fill and unfilled vacancies. Finally, Forth and
Mason (2004) found among UK firms that a quantity shortfall in the number
of graduate engineers and scientists led to lower output per worker.
But not all empirical evidence points in the same direction. For instance,
McGuinness and Bonner (2002) and McGuinness and Doyle (2003; 2005)
found no evidence that unfilled vacancies substantially lowered the productivity
of Northern Ireland IT and construction firms. Further, for Forth and Mason
(2004) skill shortages arising as a result of the poor quality of applicants
had no productivity impact.

6.1.3. Skill gaps and productivity
While the productivity effects of skill shortages in firms have been relatively
well researched, much less is known about the productivity effects of skill gaps.
Nevertheless, some studies tested the impact of skill gaps on productivity-
related variables (McGuinness and Bennett, 2006; McGuinness and Bonner,
2002), but no significant relationships were found. Some descriptive evidence
suggests, however, that firm performance across a range of indicators
is severely constrained as a consequence of such gaps (e.g. Hogarth
and Wilson, 2001). A series of Northern Ireland sectoral studies (NIERC,
2000; 2001; 2002; Erini, 2005) reported descriptive evidence that skill gaps
substantially hampered firm performance on several fronts. Further, the UK
national employment survey (2003) also found that over a fifth of employers
reporting skill gaps delayed introducing new products, while a third stated


that skill gaps represented a barrier to implementing new work practices
(Tether et al., 2005).
Further, de Grip et al. (2008) found, using a Dutch data set, that job
worker mismatch induces cognitive decline (use it or lose it) with respect to
immediate and delayed recall abilities, cognitive flexibility and verbal fluency.
This adds to the relevance of preventing overeducation and also provides a
reason why overeducation should have a negative effect on earnings.
For skills obsolescence, contrary to expectations, Allen and van der
Velden (2001) could find little evidence of long-term problems and no effect
at all on current wage levels. This may be a result of informal on-the-job
training or the fact that job loss is an important outcome for workers with
obsolete skills.

6.2. Skill mismatch and job satisfaction
The overeducated and/or overskilled earn less than if properly matched, but
more than properly matched colleagues. There is, therefore, an incentive
to move to a job reflecting better their education and skills. In contrast, the
undereducated or underskilled earns more than if properly matched, but less
than properly matched colleagues. As in this case there is no incentive to
move to a job reflecting better education and skills we expect the mismatch
to last longer than in the previous case. This should be reflected in lower
job satisfaction for the overeducated/overskilled and higher job satisfaction
for the undereducated/underskilled.
There is limited, but increasing, literature on the effects of skill mismatch on
job satisfaction. For the US, Tsang and Levin (1985) were the first explicitly
to model this relationship. They suggested that overeducation may result in
reduced work effort, increased production cost and lower productivity. Tsang
(1987) found evidence for this in the Bell Telephone company and Tsang et al.
(1991) confirmed this, using a broader cross-industry data set. Hersch (1991,
1995) found similar results for Oregon manufacturing and warehouse firms.
For Britain, Battu et al. (1999) found that job satisfaction among graduates
was significantly adversely affected by mismatch in non-graduate jobs. Yet,
though Belfield and Harris (2002) found a similar relationship, using the same
data set, this did not appear to be associated with education quality, which
might be expected to raise job satisfaction through increasing the job-offer rate.
In contrast, Büchel (2002) found no significant difference in job satisfaction
between overeducated and adequately educated workers in Germany.

CHAPTER 6

Allen and van der Velden (2001) differentiated between education and skill
mismatches and found that there was only a significant negative relationship
between mismatch and job satisfaction for skill mismatch. Similarly for Britain,
Green and Zhu (2008) found that overqualification is not a problem for job
satisfaction in itself if it is not accompanied by skill mismatch. For Spain,
Badillo-Amador and Vila (2008) also found that skill mismatches are a better
predictor of job satisfaction than education mismatches, which are not only
related to differences between individuals but also influenced by workers’
previous perceptions, suggesting that a dynamic structure is appropriate
for analysing job satisfaction.
The above studies neglect the possibility that an overlap exists between
observed and unobserved characteristics which determine both overeducation
and job satisfaction. For this reason one Australian study (Fleming and Kler,
2007) used a different statistical technique and considers various facets of
job satisfaction rather than just overall satisfaction. It was found that the
impact of overeducation on job satisfaction is positively biased for three of
the six facets of job satisfaction considered when using a traditional statistical
technique. However, even with the improved estimation, the significant
negative relationship between overeducation and job satisfaction remains.
Similar results for facets of job satisfaction using Australian data were found
by Linsley (2005) and Kler (2006) (13), the latter finding that overeducated
graduates are more satisfied with job security than matched graduates. Finally
Mavromaras et al. (2009d) used the panel element of the household income
and labour dynamics for Australia and distinguish between overskilling and
overeducation, finding that it is only overskilling on its own or jointly with
overeducation which has a negative effect on job satisfaction. This parallels
the findings in relation to the effect of mismatch on wages.

6.3. Possible positive impacts of skill mismatch
Employers will hire overeducated or overskilled workers if they believe that
such workers are more productive and the extra productivity is not outweighed
by the higher wages required to attract such workers. They will also use the
undereducated and underskilled if the lower wage costs of such workers
compensate for lower productivity.

(13) The facets considered by Linsley are satisfaction with work and with level of responsibility; in the
case of Kler they are satisfaction with pay, job security, hours, work and flexibility.


Dupuy and de Grip (2002) suggested that overeducation is related to
the hiring policy of large firms. By hiring overeducated workers when the
supply of highly educated workers exceeds demand for their services,
large firms increase their opportunities to substitute high-skill for low-skill
workers in times when high-skill workers are in short supply. High elasticity
of substitution between high-skill and low-skill workers in large firms explain
their higher rates of labour productivity. Further, this also encourages skill
biased technological change (Acemoglu, 2002). However, to determine
precisely the relative efficiency of using the overeducated instead of matched
workers is difficult. As Jochmann and Pohlmeier (2003) pointed out, the
traditional Mincer human capital earnings function approach may reflect
decreasing returns on schooling or the non-random selection of individuals
into overeducation status. Using the 1998 wave of the German socioeconomic
panel they presented Bayesian estimates of treatment effects to overcome
these problems and found, contrary to earlier studies which neglect selectivity
issues, no evidence of overeducation depressing earnings.
Jones et al. (2009) included an overskilling variable, using the British
workplace employment relations survey 2004 (which is a matched employer-
employee data set), in equations explaining five measures of performance.
He found that the variable is positive and significant in at least some of the
equations. This may explain why employers hire overskilled workers and
pay them more, but the effect is likely to be short-term as such workers are
more likely to quit. For Germany Büchel (2002), using the German socio-
economic panel, found that overeducated workers in low-skill jobs (where
most of the overeducated are found) tend to be more productive than their
matched colleagues, which he also took to explain employer motivation in
hiring the overeducated. Clearly, however, more studies among firms are
required to understand better this phenomenon and to examine its short-
term versus long-term implications.

6.4. Skill mismatch costs for society
The economy will lose the output that could have been generated by real-
locating genuinely mismatched workers to higher productivity jobs. There is
no loss of output implied, however, for apparently mismatched workers (as
defined in Chevalier, 2003) as their current jobs reflect their actual levels of
ability or skills. Further, whether overeducation imposes real costs is challenged
by some researchers who view overeducation or overskilling as an inherent

CHAPTER 6

phenomenon of market economies, as education and training systems cannot
adapt instantly to labour-market needs. Others have argued that the notion
of overeducation is not valid, as people are able to shape their own jobs or
employers may tailor jobs to the skills of workers (Halaby, 1994).
It is reasonable to conclude that skill shortages are likely to prove costly
to the economy by restricting productivity growth in key sectors. Accepting
that, for many reasons, the market will fail to clear and skill shortages persist,
another question arises on the extent to which the wage mechanism is
sufficient to provide an effective signal of changing labour-market conditions.
Evidence is mixed. Finegold and Soskice (1988) found that individuals in
Britain are less likely than in other countries to increase training in reaction
to a rise in labour price. However, Connolly (1996), using data from the
national child development study, suggested that the demand for training is
responsive to changes in price. Nevertheless, the principal problem arises
from the often substantial training time-lag associated with producing skilled
labour, implying that the market signal usually occurs too late to aid a timely
market response. Further, the price mechanism itself may be a poor indicator
of labour-market shortage, given that increasing wages is one of a range of
strategies adopted by firms to tackle the problem of skill shortages (Bennett
and McGuinness, 2009). Frogner (2002) confirms this, reporting that less
than half of firms, claiming to experience skill shortages in the 2001 European
social survey, indicated a willingness to raise the wage.
Thus, job mismatch may be associated with social costs, which can be
defined as private costs plus negative externalities. Alternatively, social
benefits may be higher than private benefits if it were not for mismatch.
For instance, at macroeconomic level national output, and hence welfare,
is potentially lower than if workers and firms were fully matched in the
economy. The loss will be a function of the number of workers who are
mismatched and the difference in productivity if they were reallocated. Such
a loss is likely to be greater for skill mismatch than education mismatch,
given the fact that the latter is not likely to incorporate lower ability effects.
In addition, tax revenues may be wasted on equipping individuals with non-
productive education, suggesting less than optimal allocation of resources
by governments. However, education may also enable certain individuals
to gain employment who otherwise may find it difficult to do so.
Most studies that have examined the relationship between mismatch and
job satisfaction have found a significant negative relationship (for British
graduates see e.g. Battu et al., 1999). This would represent a social cost.
However, recent work using the household income and labour dynamics in


Australia panel suggests that such a negative relationship only applies to
overskilling, not to overeducation, and that it is not uniform across all employed
individuals. This needs to be examined in relation to other countries.
Finally, there may be a link between labour-market mismatch and
discrimination, which makes certain groups more susceptible to mismatch.
Battu and Sloane (2004) and Lindley (2009) have, for example, found that
relative to whites, individuals from some ethnic groups are significantly more
likely to be overeducated, suggesting that they find it harder to locate optimal
matches. In addition, migrants may be faced with lower recognition of their
qualifications, if obtained in country of origin, than is warranted on economic
grounds. For gender, differences in mismatch appear not to be marked in
many countries. However, Johansson and Katz (2007) found that, in Sweden,
women were more likely to be overeducated and men more likely to be
undereducated. It is also possible that gender-based differences in subject
choice may further exacerbate the problem through increased horizontal
mismatch. Again in Sweden, Nordin et al. (2008) found that there were clear
gender gaps in the share of mismatched jobs in certain disciplines. Men were
more likely to be mismatched in female-dominated occupations and females
more so in male-dominated occupations. The gender gap in income declines
from 12 % to 2 % when type of degree is included in their analysis, implying
that men are more likely to use education to select into well-paid fields than
are women. Last, for older workers we have to consider the problem of skills
obsolescence and the possibility of a subsequent mismatch arising from
structural factors. A particular problem arises from the fact that there is a
relatively short payback period in relation to any investment in retraining and
several studies suggest that retraining older workers may be uneconomic
for this reason. However, if there is substantial skills obsolescence there
may be no alternative to retraining older workers. As employers are less
likely to hire older workers than they are to continue to employ them, this
emphasises the need for measures to encourage employers to retain older
workers rather than to dismiss them.
Rising levels of education may also affect pay dispersion. Increasing
dispersion in the returns on graduate education in the UK has been detected
by Green and Zhu (2008) over the period 1994 to 2006, when the numbers
of graduating increased sharply. While the benefits of graduate education
improved at the top end of the distribution for both men and women, at the
bottom end they sharply decreased. This means that the financial risks
associated with investment in higher education have increased, since
individuals cannot be certain where they will end up in terms of the overall
earnings distribution. This could result in insufficient supply of individuals
coming forward to meet the needs of the economy.

CHAPTER 7

Appropriate skill mismatch
measuring

Consistent with the considerable attention to overeducation in literature, we
begin by giving an overview of data sets that include items on overeducation
or overskilling. Five European and two UK data sets extending beyond 2000
contain relevant questions, but only the household income and labour dynamics
for Australia data set is a panel. We discuss three alternative measures:
the objective method, the subjective method and the empirical method
can be used to measure overeducation and reflect on their weaknesses.
Subsequently, we explain why overskilling offers a better picture of skill
surpluses. Few data sets, however, contain questions on it. In the final
section we argue that skills obsolescence can be measured in different
ways, but that there is little consensus on what is the most appropriate way
of measuring; few data sets contain questions that can be used to assess
skills obsolescence.

7.1. Measuring overeducation and overskilling
In Table 5 we list the main recent cross-country data sets which contain relevant
questions on overeducation and overskilling, plus several single-country
data sets which contain questions on overskilling and have advantages.
We focus on the concept of overeducation and overskilling, where the main
measurement issues seem to lie. Three alternative measures have been
used to measure overeducation:
(a) systematic job evaluation, used to determine the precise level of
qualifications required to perform a particular job. It is referred to as
an ‘objective’ measure, but fails to allow more than one education
level to be appropriate for particular occupations, particularly if they
are broadly defined. It also fails to allow for the possibility that the
appropriate education level may change over time;
(b) worker self-assessment, referred to as a ‘subjective’ measure, depends
on the precise question asked, such as ‘how much education is required
to obtain a job like yours?’ or ‘what level of education is required to
perform this job?’. When firms specify a level of education higher than


necessary to perform a job, it suggests the use of credentialism and
may well result in overskilling or skill underutilisation;
(c) most data sets do not contain a direct question on overeducation,
in which case reliance has to be placed on the empirical method.
Conventionally, mismatching occurs when the level of education
is more than one standard deviation above or below the mean for
education within an occupation. This cannot be directly compared
with the above two measures as it ignores minor differences between
actual and mean education.
Several studies have sought to assess the correlation between different mea-
sures of mismatch. McGoldrick and Robst (1996) showed that the incidence of
male overeducation varied from 50 % under the objective measure to 30 %

Table 5. Data sets containing questions to individual employees on
skill mismatch

Data set Type Dates Questions
European Panel 1994-2001 overskilling
Community (three questions)
household panel
European social Cross-section over 30 Biennial 2002/03, over- and
survey countries 2004/05, 2006/07, undereducation
2008/09
European working Cross-section 1996, 2000, 2005 over- and underskilling
conditions survey
Flexible Mail questionnaire, nine 2005 overskilling,
professional in the countries (graduates underskilling,
knowledge society only) horizontal mismatch,
(REFLEX) overeducation and
undereducation
EU e-living project Household survey six 2001 and 2002 over- and
countries underqualification
Workplace Matched employer- 2004 over- and underskilling
employment employee data
relations survey
UK skills survey Cross-section 1992, 1997, 2001, 2006 over- and
underqualification
overskilling
Household income Panel 2001 overskilling
and labour
dynamics in
Australia (HILDA)

CHAPTER 7
Appropriate skill mismatch measuring 61

under the subjective measure and 16 % under the empirical method. Battu et
al. (2000) found that the various measures tended to identify different people
as being overeducated with correlations under two subjective measures being
around 0.2 to 0.3. Verhaest and Omey (2006) considered the reliability of five
different measures of overeducation on the basis of encompassing tests. They
concluded that objective job evaluation and subjective data on the level of
education required to do the job should always be preferred over the empirical
method or to subjective measures of the education level required to get the
job. Allen et al. (2006) suggested that analysis of the wage consequences
of education mismatch needed to be complemented by including the effects
of job-worker skill mismatches, expecting that the wage effects of educa-
tion mismatches would be absorbed by the skill mismatch measure. In fact,
there is little correlation between the two. Green and McIntosh (2007) found
a correlation between overeducation and overskilling of 0.2. Further, when
an overskilling variable is included in a wage equation already containing an
overeducation variable, the coefficient on overeducation only falls from -0.164
to -0.144, a statistically insignificant difference.
An obvious problem with measures for overeducation is that they treat dif-
ferent degrees of overeducation in the same way. A graduate in a job for which
only a higher school leaving certificate is required is treated as a graduate in
a menial job, while the policy implications of the two cases might be different.
It is less likely this occurs under the empirical method, but using the standard
deviation criterion is purely arbitrary. Further, there is another ambiguity ac-
cording to whether the mean or the mode is used as the standard. In many,
though not all countries, the incidence of overeducation is higher than that
of undereducation, so the mode, rather than the mean, would appear to be
the appropriate yardstick. Where the number of workers in an occupation is
limited or the dispersion of actual qualifications wide, any measure of central
location is likely to be unreliable. Thus, De Oliveira et al. (2000) suggested
generally limiting the analysis to cases where the modal number of years of
education is shared by at least 60 % of the workers in that occupation.
The measures discussed above are examples of vertical mismatch but
none deal with the case of horizontal mismatch, where the level of education
is appropriate but the type is not. For example, some degree fields focus on
occupation-specific skills that may not easily transfer to other occupations,
while others focus on more general human capital (e.g. arts). But, perhaps, the
biggest problem of overeducation as a measure of mismatch is the frequent
omission of any measure of ability in empirical work on the phenomenon,
so that we simply do not know whether overeducated workers are in this
state simply for low ability given the level of education.


7.2. Focusing on overskilling
Such reasoning suggests that overskilling might be regarded as a superior
measure of mismatch, particularly where the question makes direct references
to ability. The Australian HILDA panel asks respondents whether they use
many of their skills and abilities in their current job, with a response of 1
corresponding to strongly disagree and 7 to strongly agree. This, however,
requires a cut-off point for being overskilled and does not allow clear
delineation of the underskilled (14). In contrast, the question in the British
workplace employment relations survey 2004 simply asks individuals how
well the skills they personally have match the skills they need to do their
present job. Here, however, there is a five point scale with respondents
answering that their own skills are much higher, a bit higher, about the same,
a bit lower or much lower than needed. This enables a clear distinction to be
made between those severely overskilled or underskilled or only moderately
so. Neither of these data sets contains a direct question on overeducation,
though one can be derived using the empirical method.
The European Community household panel, which ran for eight years
between 1994 and 2001, also contains questions on overskilling. In 2001
respondents were asked whether they had formal training or education
that had given them the skills needed for the present type of work and how
much this training and education had contributed to their present work. A
further question asked whether the respondents felt that they had the skills
or qualifications to do a more demanding job than they now have. Brunello
et al. (2007) have examined these questions for five countries using both
probit and multinomial analysis, and Badillo-Amador and Vila (2008) have
done so for Spain. The latter use the modal procedure to estimate education
mismatch using years of education. They defined the overskilled as those
answering positively to the first and third questions and the underskilled
as answering negatively. Those who answer yes to first question and no
to the third are defined as adequately skilled; those who answer no to the
first question and yes to the third are, debatably, defined as wrongly skilled.
These questions are also not ideal as they refer to education as well as
skills; it is impossible to disentangle the two.
The European working conditions survey conducted in 1996, 2000 and
2005 contains relevant questions on skill mismatch. In 2005 respondents

(14) The same will not apply to young labour-market entrants, where long-term adverse consequences
will influence a much wider segment of ability distribution.

CHAPTER 7
Appropriate skill mismatch measuring 63

were asked which of the following alternative would best describe their skills
in their own work:
(a) I need further training to cope well with my duties (underskilled);
(b) my duties correspond well with my present skills (matched);
(c) I have the skills to cope with more demanding duties (possibly
overskilled).
The 2005 survey was not analysed in relation to these questions.
The REFLEX graduate survey 2005 contains data on graduates in nine
European countries and has the advantage of asking separate questions on
skill and education mismatches. Skill mismatches are identified from workers’
responses on a scale from 1 (not at all) to 5 (to a very large extent) to the
questions: to what extent does your current work demand more knowledge
and skills than you can actually offer; and to what extent are your knowledge
and skills utilised in your current work? For overeducation the equivalent
question is ‘what type of education do you feel is most appropriate for the
current work?’. This can be compared directly with reported education
levels.
The European social survey 2002/03-2008/09 has information on 25
countries and contains two questions relevant to education matching: first ‘if
someone was applying nowadays for the job you do now, would they need
any education or vocational schooling beyond compulsory education?’;
and ‘about how may years of education and vocational schooling beyond
compulsory education would they need?’ The 2001/02 e-Living project
contains questions on both qualifications and years of education but for a
more restricted range of countries.
It is necessary to consider all possible combinations of skill mismatch.
Ignoring undereducation and underskilling, which are less common
phenomena, there are four possible matching combinations:
(a) matched education and matched skills;
(b) overeducation, but skill matching;
(c) matched education, but overskilling;
(d) overeducation and overskilling.
The first case is more likely where there are specific qualifications, as in
law or medicine or skilled craft apprenticeships. The second is more likely
when the individual is at the lower end of the ability distribution for a given
qualification, so that he or she is unable to obtain a job which matches
that level of qualification. However, the job obtained may make full use of
that person’s skills or abilities. Alternatively, this combination may result
through choice, if the job has non-pecuniary benefits which offset the lower


wage offered. The third case might occur when employers raise the level of
education required for the job, as they realise that the supply of individuals with
a particular level of education is now higher than it was. If the requirements
of the job have not changed this amounts to grade inflation. The final case
may be more likely for married women returning to work after childbirth or who
are constrained by a job move on the part of the husband (tied movers). The
constraints on job search make it more difficult to find jobs commensurate
with their qualifications and abilities.
In the context of the above four measures, recent work by Mavromaras
et al. (2009d), using the HILDA panel, indicates the need to run separate
equations for each of the four groups. Only those subject to overskilling, either
on its own or jointly with overeducation, suffer from lower job satisfaction,
and for men when controlling for unobserved individual heterogeneity there is
no clear wage penalty for overeducation. This is not the case for overskilled
men. This emphasises the need for data sets which have separate questions
on overeducation and overskilling, which are panel in nature and ideally
which enable one to identify horizontal and vertical mismatches.

7.3. Measuring skills obsolescence
Finally, we should consider measuring skills obsolescence, which is far from
straightforward. In principle it can be measured:
(a) objectively, by testing the deterioration of workers’ skills over time;
(b) through subjective methods such as asking workers (or employers)
whether they face skills obsolescence;
(c) by the individual measurement of productivity (e.g. tests of workers’
skills deterioration or levels of wages);
(d) by the rate of innovation which should indicate the speed of
obsolescence.
Survey data questions include Blechinger and Pfeiffer (2000) ‘how much
of the occupational knowledge and skills that you acquired during your
apprenticeship can you still apply in your current work?’ and Van Loo et al.
(2001) ‘have your qualifications depreciated due to changes in your current
job?’ and ‘what is the extent to which your qualifications are suited to your
current job?’. Relatively few data sets include such questions.

CHAPTER 8

Skill mismatch data needed in
Europe

Given the problems that current data sources pose, this final chapter considers
data needed to analyse skill mismatch comprehensively in Europe. We
would learn most from a new matched employer-employee panel containing
appropriate questions but this would be very expensive. We suggest to add
to panel data sets several questions in the form of modules, to be asked
once every two or three years; to use several European panel surveys to
allow comparative analyses; to consider alternative methods for information
collection of a more detailed, but also revealing nature.

8.1. Improving data on skill mismatch in Europe
One of the current problems in analysing mismatch in Europe is the absence
of appropriate questions in regular series such as labour force surveys and
household panels. A panel with enough observations to observe changes
in mismatch over the business cycle would be ideal and, if a matched
employer/employee panel, this would enable control for unobserved individual
heterogeneity both in relation to personal characteristics of individuals and
establishments.
While this would provide substantial gains in understanding the mismatch
problem, developing a new data set would be very costly. To collect this new
information cost-effectively would demand funding and support for adding
a set of pertinent labour-market mismatch questions to several existing
large data sets. This can be difficult because most data sets are already
overburdened and interviewee fatigue is a serious problem but it would be a
lot cheaper than starting a new data set. Given the rather long-term nature
of training and education, yearly collection of the relevant information may
not be essential.
The cost-effectiveness of the data collection could be improved by dis-
pensing with annual information and starting a special module, collected
once every two or three years, as part of several existing, long-running data
sets across Europe. The questions asked should preferably be the same
in all data sets. Where we judge that there may be a requirement for some


additional questions for purposes of equivalisation, these should be asked
as necessary. Appropriate momentum could be assured by starting with
collection in years 1, 3, 5 and then moving to a longer three-year interval,
8, 11, etc. The information collected in this way can be almost as useful
as that collected annually and will be supported by the annual information
as well. Where specific detailed information cannot be obtained through
conventional survey interview methods, alternative methods, such as quali-
tative analysis and discrete choice experiments, should be considered as
complements that could enhance our understanding obtained from large
survey data sets.

8.2. What information should be collected and
how?
For new information useful within a wider survey context, separate ques-
tions should be asked in relation to skill shortages, skill gaps, over- and
undereducation, over- and underskilling, and skills obsolescence. Some of
the information could be obtained from employers, some from employees
and some from both to check for consistency (15). This would be much
cheaper and would provide real gains in identifying links between these
different concepts.
Only the employer has the relevant information on skill shortages and
should be asked whether there are any occupations they are unable, or have
difficulty, to fill. This could be supplemented by questions on the duration
of vacancies and the wage competitiveness of the employer. Further,
employers could be asked what they perceive to be the reasons for the skill
shortages. For skill gaps, employers should be asked to cite cases where
existing employees do not have the required qualifications, experience and/
or specialised skills to meet the firm’s skill needs for an occupation. This
needs to be supplemented by questions on training provision and why this
has not solved the problem or has not taken place. In case of a matched
data set, employees could be asked the same question on underskilling to
check for consistency in responses from the two parties.
Questions on over- and undereducation should ideally focus on both
qualifications and time spent on being educated. Further they should focus

(15) A major project including a study and network of European labour-market analysts using firm panel
data and linked employer-employee data was launched in 2008 and is due to report this year.

CHAPTER 8
Skill mismatch data needed in Europe 67

on the education level required to do the job rather than to obtain the job,
though the latter could be asked as a supplementary question. Objective
and subjective approaches are to be preferred to the empirical method.
Information on the objective approach could be obtained from the employer,
while the subjective approach requires questions addressed to the employee.
If both are asked in a matched panel, the responses could be checked
for consistency. In particular, employers need to be asked why they hire
overeducated workers and how they differentiate their pay from that of
properly matched colleagues. Similarly, employees need to be asked why
they accepted a job for which they were overeducated. Any panel element
could be used to ascertain time trends. Questions also need to be asked
on horizontal and vertical mismatch.
Questions on over- and underskilling seem to be more informative than
those on over- and undereducation and seem to be less contaminated by
unobserved individual heterogeneity. It is necessary to ensure that any
scale used for employees to rank the extent to which they use their skills
and abilities allows the underskilled and the overskilled to be classified
clearly. It is important to include ‘ability’ and ‘skill’ in the question. Questions
on the extent and type of training should be included to establish whether
or not the provision of training diminishes the extent of overskilling and
underskilling.
Skills obsolescence is often hard to measure using currently available data
sets. Questions on skills obsolescence should allow differentiation between
economic and technological change explanations. The latter will be illustrated
by reductions in overall employment and relative employment by occupation,
on which the employer is more likely to have information, while the former is
more related to the individual. This needs to be linked to age and the provision
of retraining. Using a panel, workers can be asked at different times to rate
the extent to which their skills are suitable for their current job; the reduction
in the score will then indicate the degree of obsolescence. Induction time
(the time the respondent expects a new recruit to learn how to do the job)
has been found to be a strong determinant of skills obsolescence, and so
it is useful to have a question on this in any survey.
An important but overlooked aspect is the role of past training on present
overeducation and overskilling levels. It should not be assumed that the
provision of education and training always reflects accurately the preferences
of the trainees. Little is known, in the economic context of mismatches,
of the role that education providers may play in generating mismatches.
How provision is funded and organised, as well as how it tries to address


the demand by trainees and their future employers, is as important as it is
inadequately researched. For example, the space at the borders between
vocational and university level can be riddled with problems related to
entrenched education positions. Key aspects are curriculum design and
permeability between VET and general education. Such problems appear
in different guises in many countries and their presence can distort the
efficient provision of education in a fast changing economy. The collection
of information on the type of provision and how this meets the needs of
trainees and the industry should become standard in related data sets.

8.3. The way forward
What is the way forward? First, comprehensiveness of relevant questions on
skill mismatch in surveys (see Annex) should be ensured to understand the
relationship between different forms of mismatch. For example, do vertical and
horizontal mismatches tend to have a positive, negative or no relationship?
Does skills obsolescence tend to intensify the problem of labour shortages?
Only if individual data sets contain a comprehensive list of questions are we
likely to be in a position to answer such questions. Then it will be necessary
to establish which European panel data sets can be used to include these
questions alongside their wider survey questions.

ANNEX

Comprehensive schemata for
skill mismatch questions in
surveys

1. Qualifications (employee questions)
(a) How long did it take you to obtain the highest qualification you currently
hold?
(b) Was it obtained on a part-time or full-time basis?
(c) What is the normal time taken to obtain this qualification?
(d) From which institution did you obtain this qualification?
(e) Was the qualification classified and if so what classification did you
obtain?
(f) Is a particular qualification required to obtain the job you currently
hold?
(g) If so, what is it?
(h) Has the level of required qualification changed since you obtained
this job?
(i) What level of qualification do you consider to be necessary to perform
the job adequately?
(j) What field of study do you consider to be most appropriate for this
work?
• exclusively own field;
• own or related field;
• a completely different field;
• no particular field.
(k) Has your employer offered you training to enable you to perform your
current job better or to perform another job?

2. Skills and abilities (employee question)
(a) How well do the skills and abilities you have match the skills needed
to do your current job?
My own skills are:
• much higher;
• a bit higher;
• about the same;
• a bit lower;
• much lower than needed.


3. Skill shortages and skill gaps (employer questions)
(a) Are there any occupations that you are unable to fill or have difficulty
in filling?
• please name them
(b) On average how long does it take before a vacancy is filled?
(c) How do your wage rates for these shortage occupations compare with
those of your main competitors in the labour market?
• above average;
• average;
• below average.
(d) What are the main reasons for these labour shortages?
• lack of qualified applicants;
• the job is unattractive;
• location;
• other, please specify.
(e) Are there any occupations in which existing employees do not have the
required qualification, experience and/or specialised skills to perform
the job to acceptable or most efficient levels?
• please name them
(f) What are the main reasons for these skill deficiencies?
• skill shortages;
• changing job requirements;
• labour turnover;
(g) Which skills are inadequate?
• technical/ICT;
• numeracy;
• literacy/communication skills;
• problem solving;
• teamwork;
(h) Why does the company not provide off-the-job training to remedy this?
• costs;
• time constraints;
• no suitable training available;

Comprehensive schemata for skill mismatch questions in surveys 71

4. Skills obsolescence (employee questions)
(a) Are there any skills you once possessed no longer required in your
current job?
• if so, please name them
(b) Are there new skills which have emerged in your current job that you
do not possess?
• if so, please name them
(c) Do you consider that your ability to perform your current job has
diminished over time?
(d) How long does it take a new employee to become fully proficient in
the job?
(e) Has this duration increased over time?
(f) Do you think that you will be able to continue in your present job until
the normal age of retirement?
(g) Has training been provided by your employer to update your skills?
(h) If so, when and how often has this been provided?

References

Acemoglu, D. (2002). Technical change, inequality and the Labour Market.
Journal of Economic Literature, XL, 7-72.
Alba-Ramirez, A. (1993). Mismatch in the Spanish labor market. Journal of
Human Resources, Vol. 28, No 2, p. 259-278.
Alba-Ramirez, A.; Blázquez, M. (2003). Types of job match, overeducation
and labour mobility in Spain. In Büchel, F. et al. Overeducation in
Europe. Cheltenham: Edward Elgar, p. 65-92.
Allen, J. et al. (2006). Wage effects of job-worker mismatches: heterogeneous
skills or institutional effects. Maastricht: ROA – Research Centre for
Education and the Labour Market.
Allen, J.; van der Velden, R. (2001). Educational mismatches versus skill
mismatches: effects on wages, job satisfaction, and on-the-job search.
Oxford Economic Papers, Vol. 53, No 3, p. 434-452.
Allen, J.; van der Velden, R. (2007). When do skills become obsolete, and
when does it matter? In De Grip, A. et al. (eds). The economics of
skills obsolescence, p. 27-50.
Alpin, C. et al. (1998). Over and undereducation in the UK graduate labour
market. Studies in Higher Education, Vol. 1, p. 17-34.
Arrow, K.J.; Capron, W.M. (1959). Dynamic shortages and price rises: the engi-
neer-scientist case. Quarterly Journal of Economics, Vol. 73, p. 292-308.
Asplund, R.; Lilja, R. (2000). Has the Finnish labour market bumped
the least educated? In Borghans, L.; de Grip, A. (eds). The over-
educated worker: the economics of skill utilisation. Cheltenham:
Edward Elgar, p. 57-76.
Badillo-Amador, L.; Vila, L.E. (2008). Education and skill mismatches:
wage and job satisfaction consequences. Cartagena: Polytechnic
University of Cartagena.
Battu, H. et al. (1999). Overeducation among graduates: a cohort view.
Education Economics, Vol. 17, No 1, p. 21-38.
Battu, H. et al. (2000). How well can we measure graduate overeducation
and its effects? National Institute Economic Review, No 171, January
2000, p. 82-93.

References 73

Battu, H.; Sloane, P.J. (2000). Overeducation and crowding out in Britain.
In Borghans, L.; de Grip, A. (eds). The overeducated worker.
Cheltenham: Edward Elgar, p. 157-174.
Battu, H.; Sloane, P.J. (2004) Over-education and ethnic minorities in
Britain. Manchester School, Vol. 72, No 4, p. 535-559.
Bauer, T. (2002). Educational mismatch and wages: a panel analysis.
Economics of Education Review, Vol. 21, p. 221-229.
Becker, G. (1964). Human capital: a theoretical and empirical analysis with
special reference to education. Irvington, NY: Columbia University
Press.
Belfield, C.R.; Harris, R.D.F.(2002). How well do theories of job matching
explain variations in job satisfaction across education levels?
Evidence for UK graduates. Applied Economics, Vol. 34, p. 535-
548.
Bennett, J.; McGuinness, S. (2009). Assessing the impact of skill shortages
on the productivity performance of high-tech firms in Northern
Ireland. Applied Economics, Vol. 41, No 6, p. 727-737.
Blechinger, D.; Pfeiffer, F. (2000). Technological change and skill
obsolescence: the case of German apprenticeship training. In
Heijke, H.; Musken, J. (eds). Education and training in a knowledge-
based economy. London: MacMillan, p. 243-276.
Bockerman, P.; Ilmakunnas, P. (2009). Job disamenities, job satisfaction,
quit intentions and actual separations; putting the pieces together.
Industrial Relations, Vol. 48, No 1, p. 73-96.
Booth, A.; Snower, D. (1996). Acquiring skills: market failures, their
symptoms and policy responses. Cambridge: University Press.
Borghans, L.; de Grip, A. (2000). The debate in economics about skill
utilisation. In Borghans, L.; de Grip, A. (eds). The overeducated
worker. Cheltenham: Edward Elgar, p. 3-26.
Bosworth, D.L. (1978). The rate of obsolescence of technical knowledge: a
note. Journal of Industrial Economics, Vol. XXVI, No 3, p. 273-279.
Brunello, G. et al. (2007) (eds). Education and training in Europe. Oxford:
Oxford University Press.
Brynin, M.; Longhi, S. (2009). Overqualification: major or minor mismatch?
Economics of Education Review, Vol. 28, p. 114-121.
Büchel, F. (2002). The effects of overeducation on productivity in Germany:
the firm’s viewpoint. Economics of Education Review, Vol. 21, p.
263-275.


Büchel, F.; Battu, H. (2003). The theory of differential overqualification:
does it work? Scottish Journal of Political Economy, Vol. 50, No 1,
p. 1-16.
Büchel, F.; Pollmann-Schult, M. (2001). Overeducation and skill endow-
ments: the role of school achievement and vocational training qual-
ity. Bonn: IZA – Institute for the Study of Labour (Discussion paper,
No 337).
Büchel, F.; van Ham, M. (2002). Overeducation, regional labour markets
and spatial flexibility. Bonn: IZA – Institute for the Study of Labor
(IZA Discussion papers, No 424).
Burdett, K. (1978). A theory of employee job search and quit rates. American
Economic Review, Vol. 68, No 1, p. 212-220.
Burris, V. (1983). The social and political consequence of overeducation.
American Sociological Review, Vol. 48, No 4, p. 454-469.
Cardoso, A. (2007). Jobs for young university graduates. Economics
Letters, 94, 271-277.
Cedefop (2008). Future skill needs in Europe: medium term forecast:
synthesis report. Luxembourg: Publications Office.
Cedefop (2009). Skill mismatch: identifying priorities for future research.
Luxembourg: Publications Office (Working paper, No 3).
Cedefop; Büchel, F. (2001). Overqualification: reasons, measurement
issues and typological affinity to unemployment. In Cedefop; Descy,
P.; Tessaring, M. Training in Europe: second report on vocational
training research in Europe 2000: background report volume 2.
Luxembourg: Publications Office.
Cedefop; Descy, P.; Tessaring, M. (2001). Training and learning for
competence: second report on vocational training research in
Europe: synthesis report. Luxembourg: Publications Office (Cedefop
Reference series).
Chevalier, A. (2003). Measuring over-education. Economica, Vol. 70, No 3,
p. 509-531.
Cohn, E.; Khan, P. (1995). The wage effects of overschooling revisited.
Labour Economics, Vol. 2, Issue 1, p. 67-76.
Cohn, E.; Ng, C.Y. (2000). Incidence and wage effects of underschooling
and overschooling in Hong Kong. Economics of Education Review,
Vol. 19, p. 159-168.
Connolly, S. (1996). Does training choice depend on prospective returns?
East Anglia: Economics Research Centre (ERC Working paper,
9604).

References 75

Daly, A. et al. (1985). Productivity, machinery and skills in a sample of British
and German manufacturing plants. National Institute Economic
Review, Vol. 111, No 1, p. 48-61.
Daly, M. et al. (2000). Premiums and penalties for surplus and deficit
education: evidence from the United States and Germany. Economics
of Education Review, Vol. 19, p. 169-178.
De Grip, A. et al. (2008). Job-worker mismatch and cognitive decline.
Oxford Economic Papers, No 60, p. 237-253.
De Grip, A.; van Loo, J. (2007). The economics of skills obsolescence:
a review. In de Grip, A. et al. (eds). The economics of skills
obsolescence: theoretical innovations and empirical applications.
Bingley: Emerald, p. 1-26 (Research in Labor Economics, Vol. 21).
De Oliveira, M. (2000). The role of human capital and technological change
in overeducation. Economics of Education Review, Vol. 19, p. 199-
206.
Dekker, R. et al. (2002). The effects of training and overeducation on
career mobility in a segmented labour market. International Journal
of Manpower, Vol. 23, No 2, p. 106-125.
Dolado, J.J. et al. (2009). On-the-job search in a matching model with
heterogenous jobs and workers. Economic Journal, No 119, p. 200-
228.
Dolton, P.; Siles, M. (2003). The determinants and consequences of
overeducation. In Büchel, F. et al. (eds). Overeducation in Europe.
Dolton, P.; Vignoles, A. (2000). The incidence and effects of overeducation
in the UK graduate labour market. Economics of Education Review,
Vol. 19, No 2, p. 179-198.
Duncan, J.; Hoffman, S. (1981). The incidence and wage effects of overe-
ducation. Economics of Education Review, Vol. 1, No 1, 7p. 5-86.
Dupuy, A.; de Grip, A. (2002). Do large firms have more opportunities to
substitute between skill categories than small firms? Aarhus: Centre
for Labour Market and Social Research (Working paper 02-01).
Erini, J. (2005). An examination of the Northern Ireland labour market for
construction skills. Belfast: DEL – Department for Employment and
Learning.
European Commission (2008a). The Bordeaux communiqué on enhanced
European cooperation in vocational education and training.
Communiqué of the European Ministers for Vocational Education
and Training, the European social partners and the European


Commission, meeting in Bordeaux on 26 November 2008 to review
the priorities and strategies of the Copenhagen process. Brussels:
European Commission.
European Commission (2008b). Communication from the Commission to
the European Parliament, the Council, the European Economic and
Social Committee and the Committee of the Regions: new skills for
new jobs; anticipating and matching labour market and skills needs.
Luxembourg: Publications Office (COM(2008) 868 final).
Finegold, D.; Soskice, D. (1988). The failure of training in Britain: analysis
and prescription. Oxford Review of Economic Policy, Vol. 4, No 3,
p. 21-52.
Fleming, C.M.; Kler, P. (2007). I’m too clever for this job: a bivariate probit
analysis on overeducation and job satisfaction in Australia. Applied
Economics, i-First, p. 1-16.
Forth, J.; Mason, G. (2004). The impact of high-level skill shortages on
firm-level performance: evidence from the UK technical graduate
labour market. London: NIESR – National Institute of Economic and
Social Research (Discussion paper, 235).
Frank, R.H. (1978). Why women earn less: the theory and estimation of
differential overqualification. American Economic Review, Vol. 68,
No 3, p. 360-373.
Freeman, R.B. (1976). The overeducated American. New York: Academic
Press.
Frenette, M. (2004). The overqualified Canadian graduate: the role of the
academic program in the incidence, persistence, and economic re-
turns to overqualification. Economics of Education Review, Vol. 23,
No 1, p. 29-45.
Frogner, M.L. (2002). Skill shortages. Labour Market Trends, January
2002, p. 17-27.
Galasi, P. (2008). The effect of educational mismatch on wages for 25
countries. Budapest: Corvinus University (Working papers on the
Labour Market, BWP – 2008/8).
Gautier, P. (2000). Do more high skilled workers occupy simple jobs during
bad times? In Borghans, L.; de Grip, A. (eds). The overeducated
worker. Cheltenham: Edward Elgar, p. 203-230.
Gielen, A.C. (2008). Repeated job quits: stepping stones or learning about
quantity. Bonn: IZA – Institute for the Study of Labour (Discussion
paper, No 3838).

References 77

Goldthorpe, J.H.; Hope, K. (1974). Social grading of occupations: a new
approach and scale. Oxford: Clarendon Press.
Gottschalk, P.; Hansen, M. (2003). Is the proportion of college workers in
non-college jobs increasing? Journal of Labor Economics, Vol. 21,
p. 449-471.
Grazier, S. et al. (2008). Graduate employment in the UK: an application of
the Gottschalk-Hansen model. Bonn: IZA – Institute for the Study of
Labour (Discussion paper, No 3618).
Green, F. et al. (1998). The meaning and determinants of skills shortages.
Oxford Bulletin of Economics and Statistics, Vol. 60, No 2, p. 165-87.
Green, F. et al. (1999). Overeducation and skills: clarifying the concepts.
London: Centre for Economic Performance.
Green, F.; McIntosh, S. (2007). Is there a genuine under-utilisation of skills
amongst the over-qualified? Applied Economics, Vol. 39, p. 427-439.
Green, F.; Zhu, Y. (2008). Overqualification, job dissatisfaction, and
increasing dispersion in the returns to graduate education. Kent:
University of Kent, Department of Economics (Discussion papers
KDPE 08/03).
Groeneveld, S.; Hartog, J. (2004). Overeducation, wages and promotions
within the firm. Labour Economics, Vol. 11, No 6, p. 701-714.
Groot, W. (1993). Overeducation and the returns to enterprise related
schooling. Economics of Education Review, Vol. 17, No 4, p. 299-
309.
Groot, W. (1996). The incidence and returns to overeducation in the UK.
Applied Economics Vol. 28, p. 1345-1350.
Groot, W.; Maassen van den Brink, H. (2000). Overeducation in the labor
market: a meta analysis. Economics of Education Review, Vol. 19,
Issue 2, p. 149-158.
Halaby, C.N. (1994). Overeducation and skill mismatch. Sociology of
Education, Vol. 67, p. 47-59.
Hannan, D.F. et al. (1998). Trading qualifications for jobs: overeducation
and the Irish youth labour market. Dublin: ESRI.
Hartog, J. (2000). Overeducation and earnings: where are we, where should
we go? Economics of Education Review, Vol. 19, No 2, p. 131-148.
Hartog, J.; Oosterbeek, H. (1988). Education, allocation and earnings in
the Netherlands: overschooling? Economics of Education Review,
Vol. 7, No 2, p. 185-194.


Haskel, J.; Holt, R. (1999). Anticipating future skill needs: can it be done?
Does it need to be done? Sudbury: DfEE (Skills task force research
group, Research paper 1).
Haskel, J.; Martin, C. (2001). Technology, wages and skill shortages: evi-
dence from UK micro data. Oxford Economic Papers, Vol. 53, No 4,
p. 642-658.
Haskel, J.; Martin, C. (1996). Skill shortages, productivity growth and
wage inflation. In Booth, A.; Snower, D. (eds). Acquiring skills:
market failures: their symptoms and policy responses. Cambridge:
University Press, p. 147-174.
Hersch, J. (1991). Education and job match. Review of Economics and
Statistics, No 79, p. 140-144.
Hersch, J. (1995). Optimal mismatch and promotions. Economic Inquiry,
No 33, p. 611-624.
Hogarth, T.; Wilson, R. (2001). Skills matter: a synthesis of research on the
extent, causes and implications of skills deficiencies. Nottingham:
Department for Education and Skills (DFES report SMSI).
Jochman, M.; Pohlmeier, W. (2003). The causal effect of overqualification
on earnings: evidence from a Bayesian approach. In Büchel, F. et al.
Overeducation in Europe. Cheltenham: Edward Elgar, p. 93-108.
Johansson, M.; Katz, K. (2007). Wage differences between women and
men in Sweden: the impact of skill mismatch. Uppsala: IFAU –
Institute for Labour Market Policy Evaluation (Working paper 2007-
13).
Jones, M.K. et al. (2009). Training, job satisfaction and workplace
performance in Britain: evidence from WERS 2004. Labour, Vol. 23,
No s1, p. 139-175.
Jovanovic, B. (1979). Job matching and the theory of turnover. Journal of
Political Economy, No 87, p. 972-990.
Kaufman, H.G. (1974). Obsolescence and professional career development.
New York: Amacom.
Kaufman, H.G. (1975) (ed.). Career management: a guide to combating
obsolescence. New York: IEEE Press/Wiley Interscience.
Kaufman, H.G. (1989). Obsolescence of technical professionals: a measure
and a model. Applied Psychology: an International Review, Vol. 38,
No 1, p. 73-85.
Kelly, E. et al. (2008). The economic returns to field of study and competencies
among higher education graduates in Ireland. Dublin: ESRI – European
Studies Research Institute (Working paper, No 242).

References 79

Kiker, B. et al. (1997). Overeducation and undereducation: evidence for
Portugal. Economics of Education Review, Vol. 16, p. 111-125.
Kler, P. (2006). The impact of overeducation on job satisfaction among
tertiary educated Australians. St. Lucia: University of Queensland
(LERG – Labour Economics Research Group, Discussion paper,
No 10).
Kucel, A.; Byrne, D. (2008). Are overeducated people insiders or outsiders?
A case of job search methods and overeducation in the UK. Dublin:
ESRI – European Studies Research Insitute (Working paper, No
258).
Lindley, J. (2009). The over-education of UK immigrants and minority
ethnic groups: evidence from the labour force survey. Economics of
Education Review, No 28, p. 80-89.
Linsley, I. (2005). Overeducation in the Australian labour market: its
incidence and effects. Victoria: University of Melbourne, Department
of Economics Research Group (Discussion paper, No 939).
Mason, G.; Wilson, R. (2003). Employer skill survey: new analysis and
lessons learned. London: NIESR – National Institute of Economic
and Social Research.
Mavromaras, K. et al. (2009a). Assessing the incidence and wage effects
of over-skilling in the Australian labour market. Economic Record,
Vol. 85, No 268, p. 60-72.
Mavromaras, K. et al. (2009b). Overskilling dynamics and education
pathways. Melbourne: Institute of Applied Economic and Social
Research (Working paper, No 22/09).
Mavromaras, K. et al. (2009c). The incidence and wage effects of
overskilling among employed VET graduates. Monograph prepared
for the National Centre for Vocational Education Research, Adelaide,
Australia.
Mavromaras, K. et al. (2009d). The problem of overskilling in Australia and
Britain. Manchester School (forthcoming).
McGoldrick, K.M.; Robst, J. (1996). Gender differences in overeducation: a
test of the theory of differential overqualification. American Economic
Review, No 86, p. 280-284.
McGuinness, S. (2003a). Graduate overeducation as a sheepskin effect:
evidence from Northern Ireland. Applied Economics, Vol. 35, No 5,
p. 597-608.
McGuinness, S. (2003b). University quality and labour market outcomes.
Applied Economics Vol. 35, Issue 18, p. 1943-1955.


McGuinness, S. (2006). Overeducation in the labour market. Journal of
Economic Surveys, Vol. 20, No 3, p. 387-418.
McGuinness, S.; Bennett, J. (2006). Examining the link between skill
shortages, training composition and productivity levels in the
Northern Ireland construction industry, International Journal of
Human Resource Management, Vol. 17, No 2, p. 265-279.
McGuinness, S.; Bonner, K. (2002). Employer characteristics and practices
as drivers of unfilled IT vacancies. Service Industries Journal, Vol. 22,
No 4, p. 137-153.
McGuinness, S.; Doyle, J. (2003). Three years on: reassessing the Northern
Ireland labour market for IT skills. Belfast: DEL – Department for
Employment and Learning.
McGuinness, S.; Doyle, J. (2005). An examination of the Northern Ireland
labour market for construction skills. Belfast: DEL – Department for
Employment and Learning.
McGuinness, S.; Wooden, M. (2007). Overskilling, job insecurity and career
mobility: evidence from Australia. Bonn: IZA – Institute for the Study
of Labour (Discussion paper, 2938).
Mincer, J. (1974). Schooling, experience and earnings. New York: Columbia
University Press.
Neugart, M.; Schomann, K. (2002). Employment outlooks: why forecast
the labour market and for whom? Berlin: WZB – Social Science
Research Center Berlin (Discussion paper FS I 02-206).
Neuman, S.; Weiss, A. (1995). On the effects of schooling vintage on
experience – earnings profiles: theory and evidence. European
Nicaise, I. (2000) The effect of bumping down on wages: an empirical
test. In Borghans, L.; de Grip, A. (eds). The overeducated worker.
Nickell, S.; Nicolitsas, D. (1997). Human capital, investment and
innovation: what are the connections? London: Centre for Economic
Performance (Discussion paper, No 20).
NIERC (2000) – Northern Ireland Economic Research Centre. A study of
the Northern Ireland labour market for IT skills. Belfast: DHFETE
– Department of Higher and Further Education Training and
Employment.
labour market conditions in the Northern Ireland electronics industry.
Belfast: DEL – Department for Employment and Learning.

References 81

the Northern Ireland labour market for mechanical engineering skills.
Belfast: DEL – Department for Employment and Learning.
Nordin, M. et al. (2008). Education-occupation mismatch: is there an income
penalty? Bonn: IZA – Institute for the Study of Labour (Discussion
paper, No 3806).
Ochsen, C. (2009). On the measurement of mismatch. Applied Economics
Letters, No 16, p. 405-409.
Pazy, A. (1998). Concept and career-stage differentiation in obsolescence re-
search. Journal of Organisational Behaviour, Vol. 17, No 1, p. 59-78.
Pazy, A. (2006). The threat of professional obsolescence: how do
professionals at different career stages experience it and cope with
it? Human Resource Management, Vol. 29, No 3, p. 251-269.
Pigou, A.C. (1912). Wealth and welfare. London: Macmillan.
Richardson, S. (2006). What is a skill shortage? Adelaide: National Institute
of Labour Studies.
Robst, J. (1995). Career mobility, job match, and overeducation. Eastern
Economic Journal, Vol. 21, No 4, p. 539-550.
Robst, J. (2008). Overeducation and college major: expanding the defini-
tion of mismatch between schooling and jobs. Manchester School,
Vol. 76, No 4, p. 349-368.
Rumberger, R. (1987). The impact of surplus education on productivity and
earnings. Journal of Human Resources, Vol. 22, No 1, p. 24-50.
Sattinger, M. (1993). Assignment models of the distribution of earnings.
Journal of Economic Literature, No 31, p. 831-880.
Schlotter, M. (2008). Origins and consequences of changes in labour/
market skill needs. Analytical report for the European Commission.
Munich: EENEE – European Expert Network on Economics of
Education.
Sicherman, N. (1991). Overeducation in the labour market. Journal of
Labor Economics, Vol. 9, No 2, p. 101-122.
Sicherman, N.; Galor, O. (1990). A theory of career mobility. Journal of
Political Economy, Vol. 98, No 1, p. 169-192.
Skott, P. (2006). Wage inequality and over-education in a model with
efficiency wages. Canadian Journal of Economics, Vol. 39, No 1,
p. 94-123.
Sloane, P.J. (2003). Much ado about nothing? What does the overeducation
literature really tell us? In Büchel, F. et al. (eds). Overeducation in
Europe: current issues in theory and policy. Cheltenham: Edward


Elgar, p. 11-48.
Sloane, P.J. et al. (1996). Overeducation and the formal education/
experience and training trade-off. Applied Economics Letters, No 3,
p. 511-515.
Sloane, P.J. et al. (1999). Overeducation, undereducation and the British
labour market. Applied Economics, No 31, p. 1437-1453.
Stigler, G.J. (1961). The economics of information. Journal of Political
Economy, No 69, p. 213-225.
Tether, B. et al. (2005). A literature review on skills and innovation: how
does successful innovation impact on the demand for skills and how
do skills drive innovation? Manchester: ESRC – Centre for Research
on Innovation and Competition.
Thijssen, J.G.L. (2005). Niet meer up-to-date: de betekenis van (perspec-
tivische) obsoletie nader verkend [No longer up-to-date: the mean-
ing of (perspectivistic) obsolescence]. BBV-Bulletin, Vol. 12, No 1,
p. 4-10.
Thurow, L.C. (1975). Generating inequality. New York: Basic Books.
Trimmer, K.J. et al. (1998). An evaluation of factors affecting professional
obsolescence of information technology professionals. Proceedings
of the 31st annual Hawaii international conference on system
sciences. New York: IEEE, p. 406-415.
Tsang, M.C. (1987). The impact of the underutilisation of education on
productivity: a case study of the US Bell companies. Economics of
Education Review, Vol. 6, No 3, p. 239-254.
Tsang, M.C. et al. (1991). The impact of surplus schooling on worker
productivity. Industrial Relations, Vol. 30, No 2, p. 209-228.
Tsang, M.C.; Levin, H.M. (1985). The economics of overeducation.
Economics of Education Review, Vol. 4, No 2, p. 93-104.
Vahey, S. (2000) The great Canadian training robbery: evidence on the
returns to educational mismatch. Economics of Education Review,
Vol. 19, p. 219-227.
Van Loo, J. et al. (2001). Skills obsolescence: causes and cures.
International Journal of Manpower, Vol. 22, No 1/2, p. 121-137.
Van Ours, J.C.; Ridder, G. (1995). Job matching and job competition:
are lower educated workers at the back of job queues? European
Verdugo, R.; Verdugo, N. (1989). The impact of surplus schooling on
earnings. Journal of Human Resources, Vol. 24, No 4, p. 629-643.
Verhaest, D.; Omey, E. (2006). Discriminating between alternative

References 83

measures of overeducation. Applied Economics, Vol. 38, p. 2113-
2120.
Weinberg, B.A. (2007). New technologies, skills obsolescence, and skill
complementarity. In De Grip A. et al. (eds). The economics of skills
obsolescence: theoretical innovations and empirical applications.
Bingley: Emerald, p. 101-118 (Research in Labor Economics series,
Vol. 21).
Weststar, J. (2009). Worker control as a facilitator in the match between
education and jobs. British Journal of Industrial Relations, Vol. 47,
No 4, p. 723-740.

The skill matching
challenge
Analysing skill mismatch
and policy implications

Luxembourg:
Publications Office of the European Union

2010 - 83 pp. - 17 x 24 cm
ISBN 978-92-896-0485-7
Cat. No: TI-31-10-301-EN-C
Free of charge - On request from Cedefop

No of publication: 3056 EN

The Skill Matching Challenge - CEDEFOP

More Related Content

Viewers also liked (6)

Similar to The Skill Matching Challenge - CEDEFOP (20)

More from SHR Moisio (12)

Recently uploaded (20)

The Skill Matching Challenge - CEDEFOP