MASARYK UNIVERSITY
FACULTY OF ECONOMICS AND ADMINISTRATION
Migration and Labor Market Imbalances:
Do Immigrants Move to Where
Their Labor is Needed?
HABILITATION THESIS
MARTIN GUZI
BRNO, 2018
Declaration
I hereby declare that the Habilitation Thesis titled Migration and Labor Market
Imbalances: Do Immigrants Move to Where Their Labor is Needed? is my own work
and that I have cited all of the literature and other expert resources therein in accordance
with applicable legal regulations, the Internal Rules and Guidelines of Masaryk
University, and the internal managing acts of Masaryk University and the Faculty of
Economics and Administration.
Brno, April 2, 2018
Abstract
The economic literature suggests that immigrants are more fluid than natives in
responding to changing shortages in the labor market. Drawing on the EU LFS and EU
SILC datasets, we study the relationship between residual wage premia as a measure of
skill shortages in different occupation-industry-country cells and the shares of
immigrants and natives working in these cells. We confirm that immigrants from the
EU-12 and the rest of Europe (outside the EU) are more responsive to skill shortages in
comparison with the natives, and we find statistically significant positive results for
African and American immigrants in specific economic, institutional, and policy
contexts. Immigrants appear to be more responsive to skill shortages in relatively poorer
and high-unemployment countries relative to the natives. This implies that immigration
can help these economies with fluid labor. The relative responsiveness of some
immigrant groups declined during the crisis years (those from the rest of Europe),
whereas other groups of immigrants became particularly fluid during the Great
Recession, such as low-skilled immigrants from EU-12 and the high-skilled African and
American immigrants. Our results suggest that immigrants may play an important role
in labor adjustment during times of asymmetric economic shocks, and support the case
for well-designed immigration policy and free movement of workers within the EU.
Key words: labor supply, migration, skill shortage, institutions, policy, integration
Acknowledgments
As my family has given me invaluable support with their love and
incessant belief in my abilities, they certainly deserve most of the
credit. Earlier versions of chapters of this habilitation thesis appeared
in papers and reports I have previously published together with my coauthors
Lucia Mýtna Kureková and Martin Kahanec. I would like to
thank them for the positive cooperation during our research projects
and for many fruitful discussions we had along the way. I am also
thankful to my colleagues and staff of Masaryk University for
providing an outstanding research environment.
Contents
1 Introduction.........................................................................................................11
1.1 Can migrants fill labor shortages? ............................................................12
1.2 Migrants‘ responsiveness to skill shortages..............................................14
2 Methodology and empirical framework..............................................................17
2.1 Measuring skill shortages .........................................................................19
2.2 Measuring immigrant-native relative supply............................................24
2.3 Empirical strategy - the baseline specification .........................................25
2.5 Endogeneity ..............................................................................................26
3 Data and Sample Characteristics ........................................................................29
3.1 Distribution of workers in occupations and industry................................30
3.2 Composition of immigrants by origin and length of stay.........................34
3.3 Labor market outcomes and working conditions of immigrants ..............36
3.4 Summary of descriptive evidence.............................................................41
4 The relative responsiveness of immigrants to skill shortages.............................43
4.1 Baseline estimates.....................................................................................43
4.2 Estimates by education .............................................................................47
4.3 Estimates by the presence of children.......................................................48
4.4 Estimates by age and gender.....................................................................50
4.5 Summary of findings ................................................................................52
5 Context-dependence of migrants’ responsiveness.............................................. 53
5.1 Augmented model .................................................................................... 53
5.2 The behavior of immigrants under different economic conditions .......... 54
5.3 Relative responsiveness to skill shortages by immigration context......... 58
5.4 Summary of findings................................................................................ 60
6 Migration as an adjustment mechanism in the Great Recession ....................... 63
6.1 Changes in the relative responsiveness of immigrants over time ............ 65
6.2 Estimates by education............................................................................. 66
6.3 Summary of findings................................................................................ 67
7 Conclusion.......................................................................................................... 71
8 References .......................................................................................................... 77
11
1 Introduction
In the economic literature, labor mobility is seen as a way to improve the
flexibility of the labor market, facilitating its adjustment to shifts in employment
demand between sectors of the economy, demographic shocks on the supply side, and
other external influences (see e.g. Kahanec and Zimmermann, 2016, and Ritzen and
Zimmermann, 2014). Labor mobility can be low in segmented labor markets with
barriers to mobility between submarkets, which results in slower adjustment to
structural shocks, leading to the contemporaneous presence of skill shortages and
unemployment. When immigrants fill jobs that have been vacant due to skill shortages
or when immigrants move out of sectors suffering from negative economic shocks to
take up jobs in booming sectors, they facilitate the adjustment of the economy to
structural shocks.
Migration costs faced by workers born in the EU may be substantial and prevent
people from moving across regions. Slow labor market adjustment can lead to
significant economic costs, which include the costs of forfeited production due to vacant
positions, unemployment and loss of earnings, the social costs of unemployment
benefits, and other indirect costs. According to Lucifora and Origo (2002), these costs,
including short- and long-term as well as direct and indirect components, were in the
range of 6.9 – 7.1% of GDP across selected EU countries in 1999. Bennet and
McGuinness (2009) find that hard-to-fill and unfilled vacancies are related to
substantially lower firm productivity. Moreover, skill gaps contribute to skill mismatch
in filling vacancies, with both under- and over-skilling lowering labor productivity
(Tang and Wang, 2005; Quintini, 2011). With many unfilled high-skilled positions, an
12
economy performs below its potential in terms of innovation. Furthermore, skill
shortages can lead to a broadening of wage differentials across sectors and skills, with
wages for sought-after skills rising and wages for skills in excess supply dropping
(Neugart and Schömann, 2002). This can contribute to increasing inequalities between
the wages of skilled and unskilled workers (Lucifora and Origo, 2002).
1.1 Can migrants fill labor shortages?
Labor shortages and skill mismatches cause involuntary company adjustments, reduced
productivity levels, and diminished economic growth prospects. They also lead to
overqualification among workers (working in a job below one’s education level), which
can cause productivity loss for companies and wage penalties for workers. The extent to
which migration and migration policy can address the issue of labor shortages and skill
gaps is a widely debated issue. Martin and Ruhs (2011) suggest several methods to be
used by employers to eliminate labor shortages, e.g. increasing wages, training less
skilled workers, changing production processes, increasing imports, and admitting the
migrant labor force. The latter can lead to over-dependence on migrants in some sectors.
Several studies suggest that managed migration can solve the labor shortage problem
(Boswell et al., 2004; Kahanec and Zimmermann, 2010).
A study by Ruhs and Anderson (2008) on labor market shortages and
immigration policy attempts to analyze the key question of how to connect the
admission of migrants with the labor market needs in the destination country, while
emphasizing the controversy in approaches toward filling labor and skill shortages with
the migrant labor force. Employers often claim that labor shortages should be filled by
migrants as native workers are either not willing or qualified to perform certain jobs,
13
while labor unions often propose that it is mere exploitation of cheap labor. Ruhs and
Anderson (2008) emphasize that there is no universal definition for skills shortage and
no optimal policy to address this issue, as shortage and skill evaluation depends on
employers. According to the authors, employers may prefer migrants because they tend
to have higher skills and a better “attitude” to work, they have lower wage and
employment expectations, and because of migrant characteristics and immigrant status
restrictions or a readiness to accept jobs below their skills. The authors claim that in
some industries, existing shortages and an overreliance on migrants could have been
caused by a low level of labor market regulation and low level of vocational training
(e.g. in the UK construction sector), or low wages and poor working conditions (social
care in the UK).
A study by Kahanec et al. (2013) analyzes existing structural inefficiencies in
labor markets characterized by labor and skill shortages and mismatches. The authors
suggest that improved labor market flexibility could improve labor market adjustment,
particularly in sectors with low resident labor mobility. In this case, the foreign labor
force could be an option to mitigate labor shortages. A study on the labor market
assimilation of immigrants in Spain by Fernández and Ortega (2006) finds that
immigration was important for satisfying the labor demand in many sectors, such as
services and construction, where the native force supply would not meet the demand.
They also find that immigration helps to reduce wage pressures. However, the analysis
shows that there might be challenges for Spain to retain the immigrant flows.
14
1.2 Migrants‘ responsiveness to skill shortages
Studying the effects of immigration on the US labor market, Borjas (2001) finds
that newly arrived immigrants tend to settle in states with higher wages (paid for the
skills migrants possess) and consequently help to equalize economic opportunities
across states. This could be explained by the fact that newly-arrived immigrants, after
deciding to depart their home country and actually leaving it, have already incurred a
significant proportion of the costs of mobility. This proportion of migration costs does
not influence their future location decisions within the destination labor market. In
contrast, the costs of mobility within the domestic labor market influence the decisions
of the native employees and of earlier immigrants.
The theoretical model conceptualized by Borjas (2001) illustrates how
immigration can facilitate the adjustment of labor markets after temporary or structural
shocks, thereby contributing to economic efficiency. This is how immigration can be
seen as grease on the labor market wheels (Borjas 2001). The impact of immigration
grease may diminish with the assimilation of immigrants into local environments as
their costs of departing their current positions may become more similar to those of
natives. In addition, immigration grease may be impeded by policies that restrict
geographic or job mobility directly or indirectly linked to immigrant status. This is how
the economic, institutional and policy context can have a role in determining how
responsive low- and high-skilled immigrants are to labor market imbalances as
compared to natives. The empirical analysis can shed light on the country characteristics
that influence how immigration can reduce labor market imbalances and contribute to
its flexibility.
15
Empirical findings confirm that immigrants are more responsive to wage
differentials in the labor market than natives. This is found in the US labor market
(Borjas, 2001), in Spain (Amuedo-Dorantes and De la Rica, 2010, who use a measure of
employment prospects rather than wages), in Norway (Røed and Schøne 2012), the UK
(Dustmann, Frattini and Preston, 2012) and in the EU (Guzi et al. 2015).
In this study, we contribute to the literature by exploring the diversity across EU
member states to analyze how immigrants’ relative responsiveness to skill shortages
varies across institutional contexts, for different types of immigrant groups, and over the
business cycle. On this basis, five research questions emerge that are important both for
our understanding of immigrants' labor market mobility and in understanding
implications for migration and integration policies.
First, are immigrants more or less responsive than natives to skill shortages across
and within EU labor markets?
Second, under what economic, institutional, and policy contexts do immigrants
respond to skill shortages more (or less) fluidly than natives?
Third, how does responsiveness to labor market imbalances vary across different
immigrant groups in terms of their origin and time since immigration?
Fourth, how does the responsiveness of immigrants to skill shortages vary with
demographic characteristics?
Fifth, how does responsiveness to skill shortages vary over the business cycle?
We address these questions by investigating the labor markets of the EU-15,
primarily using data from the EU Labor Force Survey (EU-LFS) and the EU Statistics
on Income and Living Conditions (EU-SILC). The Member States that joined the EU in
2004, 2007, and 2013 are not included in the analysis due to the relatively more limited
16
inflow of immigrants into these countries and resulting small sample sizes (Kahanec
and Zaiceva, 2009).
In this work, we build on the approach of Guzi et al. (2015, 2018), who are the
first to measure the responsiveness to skill shortages of immigrants relative to natives in
the EU-15. We contribute novel elements to the literature on labor mobility between
different types of immigrants and their responsiveness to labor market imbalances. We
further contribute to the policy debate by investigating the labor market fluidity of
immigrants across economic, institutional, and policy contexts. In particular, we
investigate if the mobility of immigrants differs between countries with different levels
of GDP, economic growth, unemployment rate, generosity of welfare spending,
openness to admitting immigrant workers, and the scale of immigration. Finally, we
study whether and how various low- and high-skilled immigrant groups in the EU-15
responded to asymmetric economic shocks across sectors, occupations, and states
before and during the Great Recession.
This study is structured as follows. In the second chapter, we introduce a
theoretical framework to formalize the location decisions by natives and immigrants.
We develop a measure of skill shortages and describe the estimation strategy to identify
the average responsiveness of immigrants to skill shortages as compared to natives. In
the third chapter, we introduce the data and compare the labor market outcome of
immigrants and natives. The fourth chapter reports the results of the baseline finding
and measures the variation in immigrants’ responsiveness to labor and skill shortages
across demographic characteristics. In the fifth chapter, we show how immigrants'
responsiveness to skill shortages varies in economic, institutional, and policy contexts.
In the sixth chapter, we present results over the business cycle.
17
2 Methodology and empirical framework
The neoclassical theory of migration asserts that migration is stimulated primarily by
rational economic considerations of relative benefits and costs, mostly financial but also
psychological (Todaro and Smith 2006). The origins of the theory date back to works of
Hicks (1932), Lewis (1954), and Harris and Todaro (1970), who explained migration in
the process of economic development and argued that migration is driven by geographic
differences in labor supply and demand and the resulting differentials in wages between
labor-rich versus capital-rich countries. Under the assumption of full employment, it
predicts a linear relationship between wage differentials and migration flows (Bauer and
Zimmermann 1999; Massey et al. 1993; Borjas 2008). In the extended neoclassical
models, migration is determined by expected rather than actual earnings and the key
variable is earnings weighted by the probability of employment (Bauer and
Zimmermann 1999; Massey et al. 1993; Czaika 2015).
Human capital theory enriches the neoclassical framework by incorporating
individual socio-demographic characteristics as an important determinant of migration
at the micro-level (Bauer and Zimmermann 1999, Sjaadstad 1962). Human capital
endowments, skills, age, marital status, gender, occupation, labor market status, as well
as preferences and expectations, strongly affect migration choices. Heterogeneity
between individuals is an important factor and different individuals in the same sending
country demonstrate different propensities to migrate and preferences for various
destination countries (Bonin et al. 2008). The propensity to migrate generally decreases
with age and typically increases with education level (Bauer and Zimmermann 1999).
18
The theoretical framework of this work is based on the model first presented by
Borjas (2001). The basic intuition of his model is that immigrants can be expected to be
more fluid than natives in responding to changing skill and labor imbalances in the
labor market. This is because immigrants’ costs of departing their home, job, and
networks of family members and friends in the country of origin are sunk upon arrival
to a new country.1
Therefore, it costs immigrants little to choose one place above
another as a destination. It follows that this should be especially true for newly-arrived
immigrants, not yet deeply amalgamated in their new milieu in the receiving country. In
contrast, natives’ decision to move entails pecuniary and non-pecuniary costs of parting
with their home, job and networks. As a result, the location decisions of immigrant
workers should be more responsive to wage differentials than those of natives.
Following Borjas (2001), our model assumes that natives and immigrants are
allocated across states ( ) and occupation-industry groups ( ). Let denote the
wage of worker ∈ 	in country 	 ∈ . We assume that is net of any adjustment
costs, such as re-qualification costs, which are fixed and equal for all workers in a given
country and occupation-industry group. A worker moves to a new country or to a new
occupation-industry group from their original country ( ) and occupation-industry
group ( ) if
∈ , ∈ 	−	 	− > 0 (1)
where 	stands for the costs of moving between countries, occupations, and industries,
and, more specifically, it is assumed to include all pecuniary and non-pecuniary
(psychological) costs of leaving behind the country, occupation, and industry of origin.
1
Sunk costs include monetary and nonmonetary (e.g. social costs) investments made by
immigrants that are nonrefundable (lost) after movement.
19
It is reasonable to assume that for natives, moving costs can be substantial, and thus
moves occur only when the incentives (e.g. wage differential) are large. In contrast, for
immigrants these costs are sunk upon arrival and thus assumed to be zero.2
Following
this argument, new immigrants can be expected to be the most responsive to wage
differentials among all other immigrant groups, who are in turn likely to be less fettered
by D than the natives. The more time passes from initial immigration, the more similar
the migration decisions of immigrants become compared to those of natives, due to the
gradually increasing attachment to the location and employment.3
It is clear that as long as > 0, the cost induces friction into labor market
adjustment, as there can be situations in which wage differentials persist across
countries and occupation-industry groups without inducing relocation. This can lead to
inefficiencies in the allocation of workers across industries and occupation groups. In
principle, capital flows could lead to the equalization of wage differentials, but with
substantial lags and adjustment costs.
2.1 Measuring skill shortages
The term skill or labor shortage has no universally accepted definition. It is used to
denote a shortfall of individuals in the workforce (labor shortage), but also to refer to a
2
Immigrants may come to the EU not only for economic reasons, but also as refugees
or as relatives or dependent minors of economic immigrants. The model assumes that
immigrants make a decision under perfect information. The costs of migration may be assumed
to differ by the country of origin and destination, by individual skills or other circumstances.
Such cost differences could be, for example, due to the distance between the home and the new
language. Nevertheless, for the purposes of this work, it suffices to assume this simplified cost
structure. The model could be extended to a more complex cost structure in a straightforward
way. In addition, the model could be presented in a dynamic version, with wages and costs as
stochastic variables. In this case, the key relationships would hold in terms of expected present
values.
3
On the process of integrating immigrants into destination economies, see Constant,
Gataullina and Zimmermann (2009).
20
gap between workers and jobs in the economy (mismatch) (Trendle, 2008). Labor
shortage is generally conceptualized as a disequilibrium state of excess demand
(Zimmermann, Bonin, Fahr, & Hinte, 2007, p 69). Defined in the most basic sense,
labor shortages arise where the demand for workers in a particular occupation exceeds
the supply of workers who are qualified, available and willing to do that type of work
(Veneri, 1999). It may arise in the short term from sudden shifts in consumption
patterns, trade patterns or have a more long-lasting nature due to rapid structural
changes in the economy typically spurred by technological progress.
Skill shortage is a more specific way to approach the concept of labor shortages.
Skill shortages may have a number of forms that can be proxied through different
parameters, and also require different policy interventions (Rutkowski 2009), e.g. a skill
shortage of workers might occur in occupations in which the specific skills cannot be
easily substituted. This might happen in the aftermath of economic restructuring biased
towards production that requires specific skills. A second type of skill shortage is
possible when there are workers with occupational skills that are in demand, but they
lack other essential skills, such as IT skills or soft skills (soft-skills gap) (see also
Anderson and Ruhs 2008; Trendle 2008). Striestka-Ilina (2007) uses the term skill gaps
to describe a qualitative mismatch between the supply or availability of human
resources and the requirements of the labor market. This term is used with respect to
employers who are not satisfied with the overall skill-set of the workforce – existing or
labor market entrants – who might be lacking a variety of skills despite having acquired
education.
The mismatch is generally defined as the occurrence of a shortage of some
skills/occupations and a simultaneous excess of other skills (Arratibel et al. 2007;
21
Rutkowski, 2009). Different ‘types’ of mismatch on the labor market have been
proposed (Boswell, Stiller, and Straubhaar 2004; Anderson and Ruhs 2008) as follows:
Qualitative mismatch appears when the qualifications of workers and the
qualification profiles of vacancies are not matched;
Regional mismatch occurs when unemployed persons seeking work and firms
offering suitable jobs are located in different regions and the jobs and/or the
labor are immobile; such a mismatch can be exacerbated by information deficits
when supply does not meet labor demand due to a lack of information;
Preference/expectation mismatch arises when unemployed or labor market
entrants are unwilling to take up a certain type of work, even if it matches their
qualifications, due to inadequate remuneration, working conditions, or status.
Labor market mismatch demonstrates itself in the form of skill shortages in the
economy which, in reality, are often an outcome of all three types of mismatches
described above: qualitative, regional, and preference. The existence of skills or labor
shortages is therefore defined as a situation in which employers have considerable
difficulty filling vacancies for an occupation – or specialized skill needs within that
occupation – at current levels of remuneration and conditions for employment, and a
reasonably accessible location (Trendle 2008). Quintini (2011) suggests that skill
shortages can be caused by labor shortages, geographical mismatch, or insufficiency of
workers with the necessary skills.
Measuring skill shortages is a non-trivial task, and a range of different
approaches can be identified in the literature. A common approach to measuring skills
or occupational shortage is computing the unemployment-to-vacancies ratio (U/V ratio),
which shows the number of jobseekers per one job opening (Obadic, 2006; Padoa-
22
Schioppa, 1991; Rutkowski, 2009). Education (ISCED) and occupation (ISCO) are the
most frequently used dimensions for the analysis. The larger the index for a particular
skill group, the larger the mismatch or the shortage for that category. A large variation
in indexes across occupational (or educational) groups signals a large skill mismatch in
the economy. A high U/V ratio across occupations means that there is a lack of jobs
rather than skill mismatch.
According to Quintini (2011), skill shortages can be measured by employers’
assessment (through surveys), vacancy rates or wage growth. The vacancy rate
approach is dependent on the quality of vacancy data, which may be focused on lowskilled
positions and thus underestimate more skilled segments (see also Boswell et al.
2004). The wage-growth approach is also limited as it does not include non-financial
incentives (e.g. training and attracting the foreign labor force).
Various studies argue that in order to identify occupations with shortages, it is
necessary to look at multiple indicators, such as unemployment and vacancy rates,
employment growth, wage growth, and their changes over time (Zimmermann et al.
2007; Martin and Ruhs, 2011; Veneri, 1999). However, synthesizing these different
measures into one all-encompassing indicator to be used in quantitative analysis is not a
trivial matter. Cohen (1995) considered six top-down indicators for 193 occupational
groups: occupational unemployment rate and change in occupational employment;
change in wages; expected long-run growth of occupation; total replacement demand
for the occupation; number of certifications to employ foreign workers; and vocational
specifications required. He projected demand for a number of highly-skilled positions,
but according to Martin (2010), his efforts were largely unsuccessful. Veneri (1999)
used three criteria to identify the presence of labor shortages in 68 occupations in the
23
1990s: employment growth at least 50% above the average, median wage growth at
least 30% above the average, and the occupational unemployment rate at least 30%
lower than the average of all occupations. This method has not proved successful as the
presence of shortage in occupations identified by employers was not confirmed by these
criteria. Zimmermann, et al. (2007) made another attempt to merge indicators that may
signal labor shortages into a single summarizing indicator. They developed several
procedures, but found too much variation between the different indicators to
deliver convincing conclusions for immigration policy.
Employers’ own assessment represents the most direct way of estimating the
existence of shortages. The length of time it takes for an employer to fill a vacancy, or
the share of employers who report difficulty in recruitment are possible measures
(Constant and Tien, 2011; Quintini, 2011). Lucifora and Origo (2002) define external
and internal skill shortages and use the number of hard-to-fill vacancies as a measure of
shortage. However, problems with international comparability and the frequency of
surveys represent two major disadvantages of using this type of survey data.
In this study, we use unexplained wage premiums as an indicator of skill
shortage, following Borjas (2001), Dustmann et al. (2012) and Guzi et al. (2015, 2017,
2018). We proxy skill shortages at the level of occupation-industry-country cells by
wage premiums as the part of wages that remains unexplained after compositional
differences across cells is netted out. In particular, for each year separately we estimate
a log-wage regression of the form
= + + , (2)
where is the log wage of worker i who belongs to occupation-industry group k in
country c, is a vector of individual characteristics including gender, education,
24
work experience and work experience squared, and is the error term.4
The wage and
all variables in vector are normalized to have zero means. The vector can be
then interpreted as the (adjusted) percent wage differential between the average wage of
individuals in the particular occupation-industry-country cell and the mean wage for a
given year in the EU-15. The indicator of skill shortage is used in the analysis to
measure immigrants' responses to changes in residual wage premia, in relation to the
natives' response.
2.2 Measuring immigrant-native relative supply
The second key variable in the analysis is the measure of the relative supply of
immigrants and natives in each occupation-industry-country cell. Following Borjas
(2001), the index is expressed by the ratio of immigrants belonging to a given
occupation-industry group in a given country and year, and the total number of
immigrants in the EU-15, relative to a similarly defined relative supply of natives in the
cell and year. Formally
=
!"#$ !$%
&"#$ &$%
, (3)
where ' is the number of immigrants belonging to occupation-industry group
k, and country c in year t. The total number of immigrants in the EU-15 in year t is
denoted as	' . The denominator similarly indicates the relative supply of natives
( (% 	in the particular cell and year. The index equals 1 when immigrants and natives
have the same distribution across occupation-industry groups and countries. The index
4
By including education as an independent variable, we eliminate any wage
differentials that arise due to the educational attainment of workers, but we assume that the
residual wage premia are invariant across skill groups.
25
is greater than 1 if immigrants in a given occupation-industry-country cell were
overrepresented relative to the natives, and it equals 0 if no immigrants are present in a
particular occupation-industry-country cell. In the empirical analysis, the index is
calculated for different immigrant groups distinguished by their origin or time since
immigration in order to test their differences in the responsiveness to skill shortages in
relation to natives. We further calculate the index for workers with and without tertiary
education, and with and without children to test the variation in immigrants’
responsiveness to skill shortages across various contexts.
2.3 Empirical strategy - the baseline specification
To measure the relative responsiveness of immigrants to changing shortages in the labor
market, we adopt a first-difference regression model as follows:
∆ = * + ∆ +, + -,. / +, + -012 +, + - + - + - + 3 , (4)
where ∆ , our measure of skill shortage, is lagged by one year. This is because the
reaction of workers to changes in the labor market is likely to be delayed. The
dependent variable is the relative supply of immigrants	 . The model also includes an
occupation-industry cell, and country and year fixed effects (- , - and - ), which act as
controls for any specific factors that might affect the relative supply of immigrants.
Additionally, the model includes lagged values of country-level unemployment rate and
GDP growth to account for variation in economic conditions between countries and
over time. We estimate this model using the Ordinary Least Squares method with robust
(Eicker-Huber-White) heteroskedastic-consistent standard errors, and every observation
26
is weighted by the total number of individuals in the cell.5
We do not adjust for
differences in living costs in the construction of in Equation 2; this is done in the
regression model of Equation 4 through the inclusion of country-year fixed effects.
2.5 Endogeneity
The measure of skill shortage captures any increase in the price of labor that cannot be
explained by the changing composition of workers in terms of gender, education, and
work experience in the given occupation-industry-country cell. Although estimating the
model in first differences, including country and year fixed effects, controls for a range
of additional factors specific to these categories, there may still be changes in residual
wages that need not reflect increased shortage of labor in the given group or country,
but are rather due to changing wage bargaining, regulation, or other factors that change
the price of labor beyond the variation captured by the compositional and fixed effects.
Whereas this potentially introduces measurement error in the link between measured
and actual skill shortages, it does not affect our key argument, measuring immigrants'
and natives' relative responsiveness to changing residual wages (whether due to
shortages or other factors).
Another obvious issue in this type of models is that any measure of wages, such
as , may be endogenous with respect to any measure of the labor force in the given
labor market, such as . We acknowledge this issue, and note that as immigrants
constitute a relatively small share in most labor markets and 	is a measure of
composition and not size of labor supply in the given labor market (industry-occupation-
5
Analytic weights (aweights in Stata) are typically appropriate when analysis is based
on data containing averages.
27
country cell), the possible endogeneity of may be less problematic than in other
models linking wages and labor supply. Moreover, we lag by one year, as is
common in the literature (e.g. Borjas, 2001; Dustmann, et al. 2012; and Guzi et al.,
2017, 2018). Finally it is possible that immigrants increase the relative supply of labor
in a given skill group, which should cause wages to decrease for that group and
therefore the estimated coefficients can be interpreted as a lower bound.
28
29
3 Data and Sample Characteristics
The empirical part of this study draws from representative samples of households in
fifteen EU member states6
obtained from the European Union Labour Force Survey
(EU-LFS) and EU Statistics on Income and Living Conditions (EU-SILC). Both data
sets are representative household surveys conducted annually in all member states of the
EU, and follow the international standard classification of economic activity (coded
according to NACE) and occupation (coded according to ISCO). Both datasets include
information on respondents' personal circumstances (including nationality and country
of birth), their labor market status, and job characteristics during the reference period.
The sampling structure of the surveys focuses primarily on permanent residents and
therefore does not capture short-term and seasonal migration.7
Thanks to its large
sample size, the EU-LFS provides reasonably reliable information about the share of
foreign-born and native population across occupation and industry cells in each country.
We use the information on workers' earnings from the EU-SILC to measure skill
shortages across occupation-industry-country cells, as explained above. The empirical
analysis exploits samples from 2004-2016 of individuals aged 15 to 64. For the
estimation of skill shortage (Equation 2), we rely on a sample of 0.8 million individuals
in the EU-SILC dataset. The index of relative supply (Equation 3) is calculated based on
14.3 million employed individuals in the EU-LFS dataset. Additional variables, such as
6
The sample includes Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland,
Italy, Luxembourg, the Netherlands, Portugal, Spain, Sweden and the United Kingdom.
Unfortunately, as the regional classification is not consistently included in the data, an analysis
at the regional level is not possible.
7
The EU-LFS has been used in several studies that analyze immigration in Europe, as it
uniquely provides both cross-country and longitudinal dimensions (e.g. Dustmann and Frattini,
2011; D'Amuri and Peri, 2014).
30
national GDP annual growth, and total unemployment rate were obtained from
Eurostat.8
The terms 'immigrant population' or 'immigrant individuals' are used in the broad
context of immigration, and the origin of immigrants is based on the country of birth.
One exception is Germany, for which immigrant origin can be determined only by
nationality. The native population refers to people residing in the country of birth. The
EU-LFS allows us to distinguish immigrants by the year of immigration and their origin
(the choice of broader regions of origin is determined by the respective variable in the
dataset). In the analysis, we distinguish six groups of immigrants which aggregate
several geographic regions: EU-15 (includes EU-15 and EFTA), EU-12 (includes
countries which joined the EU in 2004 and 2007), Europe (includes European countries
outside the EU-15 and EU-12), Africa (Africa and Middle East), Asia, and America
(includes both Americas, Australia and Oceania).
3.1 Distribution of workers in occupations and industry
In the analysis, we estimate the sensitivity of the relative labor supply of
immigrants to wage differentials across occupation-industry-country groups. For each
of the EU-15 countries, we partition the labor force into occupation-industry-country
cells defined by four occupation levels based on the ISCO classification, and nine
industry groups based on the NACE classification (see Table 3.1). This categorization
generates 36 occupation-industry groups, for each of which we calculated the skill
shortage and the index of the relative supply of immigrants in each country and year. In
8
Data from the Eurostat database accessed in February 2018: GDP growth rate expressed in the
percentage change on the previous year (table tec00115), and total unemployment rate (table
tsdec450).
31
the analysis, we allowed only occupation-industry-country cells of sufficient size in all
years.9
Table 3.1 Definition of occupation and industry groups
Occupation category ISCO-1
coding
Occupation group
Legislators, senior officials and managers 1 High
Professionals 2 High
Technicians and associate professionals 3 High
Clerks 4 Intermediate general
Service workers and shop and market sales workers 5 Intermediate general
Skilled agricultural and fishery workers 6 Intermediate specific
Craft and related workers 7 Intermediate specific
Plant and machine operators and assemblers 8 Intermediate specific
Elementary occupations 9 Low
Economic activity NACE
coding
Industry
group
Manufacturing, mining and quarrying and other industry C,D, E 1
Construction F 2
Wholesale and retail trade G 3
Transportation and storage, accommodation and food service H,I 4
Information and communication, financial and insurance
activities J,K 5
Education M 6
Human health N 7
Public administration, defence, and social work activities O, P, Q 8
Agriculture, forestry and fishing A, B 9
9
In each country, cells are selected if they include at least 20 observations in the EU-SILC
database, and the weighted size of cells in the EU-LFS database is 50,000. Unfortunately, in the
first year, data for several countries are not available. In the 2004 EU-SILC dataset, Germany,
the Netherlands, and the United Kingdom are missing. In the EU-LFS data, we cannot identify
the origin of respondents in Italy in 2004.
32
Tables 3.2 and 3.3 illustrate the distribution of immigrants across occupation and
industry groups. Interestingly, the occupational structure of EU-15 immigrants is almost
identical to that of the native workers, while the other immigrant groups are overrepresented
in occupations requiring lower qualifications. One striking observation is
that immigrant workers from the EU-12 and rest of Europe (outside the EU) are
substantially more frequently employed in lower-ranked occupations (intermediate
specific or low group) relative to not only the natives, but also to all the other immigrant
groups. This points to a high degree of over-qualification. This may be due to the
relatively recent arrival of these immigrants in the receiving countries, but may also
signify patterns of temporary migration, whereby down-skilling (and saving on the costs
of acquiring country-specific human capital) may be an optimal strategy for temporary
immigrants (Kahanec and Shields, 2013). The distributional patterns observed for
immigrants from Africa and America are very similar. In contrast, Asian immigrants are
more employed in occupations requiring higher qualifications than all the other
immigrant groups except the EU-15.
With respect to distribution across sectors, immigrant workers are primarily
concentrated in construction (except for the Asian group); transportation,
accommodation and food; and they are least concentrated in the education sector. Some
immigrant groups are concentrated in other industries, e.g. EU-15 group in finance,
EU-12 in agriculture, European in manufacturing, African in health, Asian in wholesale,
and American in public administration and social work.
33
Table 3.2 Distribution of native and immigrant workers across occupation groups
Occupation group Native EU15 EU12 Europe Africa Asia America
High 0.43 0.46 0.19 0.17 0.32 0.35 0.31
Intermediate general 0.27 0.24 0.23 0.23 0.26 0.31 0.28
Intermediate specific 0.22 0.19 0.31 0.35 0.21 0.15 0.17
Low 0.08 0.11 0.27 0.25 0.21 0.19 0.24
Source: Own rendering based on EU-LFS 2004-2016 data
Note: Occupation groups are described in Table A1 in the Appendix. The sample
includes individuals aged 15 to 64 in EU-15 countries. Immigrants are recognized by
the country of birth or nationality (Germany). Population weights are applied.
Table 3.3 Distribution of native and immigrant workers across industry groups
Industry group Native EU15 EU12 Europe Africa Asia America
Manufacturing 0.17 0.16 0.17 0.22 0.13 0.12 0.10
Construction 0.07 0.08 0.15 0.14 0.08 0.03 0.08
Wholesale and trade 0.14 0.12 0.11 0.13 0.13 0.16 0.11
Transportation and food 0.11 0.15 0.18 0.16 0.16 0.25 0.17
Communication and
financial 0.14 0.17 0.11 0.12 0.16 0.13 0.15
Education 0.08 0.08 0.03 0.03 0.06 0.05 0.06
Human health 0.12 0.10 0.08 0.07 0.13 0.12 0.09
Public administration 0.13 0.12 0.13 0.12 0.12 0.12 0.21
Agriculture and fishing 0.03 0.01 0.04 0.02 0.02 0.01 0.02
Source: Own rendering based on EU-LFS 2004-2016 data
Note: Industry groups are described in Table A2 in the Appendix. The sample includes
individuals aged 15 to 64 in EU-15 countries. Immigrants are recognized by the country
of birth or nationality (Germany). Population weights are applied.
34
3.2 Composition of immigrants by origin and length of stay
This section provides the characteristics of the immigrant population in EU-15
countries. Evidence is based on EU-LFS data that includes 12% of immigrant in the
population aged 15 to 64. According to the data, the share of immigrants increased from
8.9% in 2004 to 14.7% in 2016. Figure 1 presents the composition of immigrant
population by origin and by the years since immigration. The relative size of
immigrants born in another EU-15 country has been decreasing from more than 25% in
2004 to less than a fifth of immigrant population in recent years. The inflow of
immigrants from outside EU-15 countries intensified after the EU enlargements in 2004
and 2007 (Kahanec, Pytlikova and Zimmermann 2016). Figure 1 illustrates that the
development of the EU-12 group was particularly dynamic and more than doubled from
7 to 17% over the studied period. In general, the foreign-born population in the EU-15
is dominated by individuals from Africa, who comprise one fourth of the immigrant
population. Immigrants from Asia also increased their relative share during the studied
period. The relative number of immigrants from rest of Europe and America has
changed less, with each group comprising approximately one sixth of the immigrant
population.
The composition of the immigrant population in the EU-15 also varies by time
since arrival. The period after the European enlargement in 2004 was characterized by
rising inflows of new immigrants that declined with the outbreak of the Great Recession
(see Figure 1). In 2016, almost two thirds of the immigrant population comprises
established immigrants who reside in the host country for more than ten years. New
immigrants with less than six years since arrival and immigrants with six to ten years
since immigration have an equal share.
35
Figure 1 Composition of immigrant population by origin and years since arrival in the
EU-15
By origin
By years since immigration
Source: Own rendering based on EU-LFS 2004-2016 data
Note: Shares add up to unity. Immigrants are recognized by the country of birth or
nationality (Germany). YSM indicates the group of immigrants by years since
immigration. Population weights are applied.
0.05
0.10
0.15
0.20
0.25
0.30
2004 2006 2008 2010 2012 2014 2016
EU15
EU12
Europe
Africa
Asia
America
0.10
0.20
0.30
0.40
0.50
0.60
2004 2006 2008 2010 2012 2014 2016
YSM 11+
YSM 6-10
YSM 1-5
36
3.3 Labor market outcomes and working conditions of immigrants
This section reports differences in labor market outcomes between immigrants
and natives. In Figures 2-5, we first show employment and unemployment probabilities
and then show the incidence of a long-term unemployment spell, on-the-job search
behavior, and job mobility across economic sectors.
Figure 2 depicts the ratio of employment and unemployment rates, respectively,
for immigrants of different origins relative to native workers. Ratios closer to one
indicate that immigrants’ position in the labor market is similar to that of natives.
Immigrants from the EU-15 and EU-12 show employment to population ratio on par
with the natives and both groups even outperform the natives in recent years. With
respect to unemployment, EU-12 immigrants are more unemployed relative to natives
and their incidence of unemployed increased around the inception of the Great
Recession, but has been decreasing since then. The unemployment of immigrants from
the EU-15 remains at levels similar or better relative to natives (the ratio is close to
unity or below). Immigrants from Asia and rest of Europe show rapid improvement in
their labor attachment, with increasing participation and falling unemployment over the
studied years, although the Great Recession slowed down the adjustment process. The
opposite trend is visible for immigrants born in Africa and America, as their
employment and unemployment rates have deteriorated since 2008 relative to natives.
37
Figure 2 Immigrant-to-native ratio of employment and unemployment rates
in the EU-15 for various immigrant groups
Employment rate
Unemployment rate
Source: Own rendering based on EU-LFS 2004-2016 data
Note: Ratio. Population weights are applied.
0.70
0.75
0.80
0.85
0.90
0.95
1.00
1.05
1.10
2004 2006 2008 2010 2012 2014 2016
EU15
EU12
Europe
Africa
Asia
America
0.80
1.00
1.20
1.40
1.60
1.80
2.00
2.20
2.40
2.60
2004 2006 2008 2010 2012 2014 2016
EU15
EU12
Europe
Africa
Asia
America
38
In a similar fashion, we demonstrate the differences in job search behavior,
unemployment duration, and mobility across economic sectors. Figure 3 compares the
share of workers in the immigrant and native workforce who seek other employment
when they already have a job. The relatively high rates of on-the-job searches among
immigrants compared to the natives may suggest a lower attachment to their current job,
but also an increased risk of involuntary job-to-job transitions or unemployment. In
particular, immigrants from Africa and America exhibit high search intensity, which
increased after 2008 when their incidence of unemployment remained high (see Figure
2). We generally observe higher job search intensity for all immigrants than for natives.
The only exception is workers from the EU-15 who exhibit the same intensity of search
behavior as natives.
Figure 4 reports the immigrant-to-native ratio of long-term unemployment,
defined as a period without a job for more than eleven months. This reflects the ability
of workers to find employment and avoid long spells of unemployment. It appears that
the inception of the Great Recession hit immigrants from Africa and America, and to
lesser degree from the EU-12, who show prolonged unemployment spells relative to
natives. Interestingly, Asian immigrants managed to avoid long unemployment spells
throughout the period and particularly during the years of the Great Recession. The ratio
for immigrants from rest of Europe was one of the highest throughout; converging to
unity before 2009, but further increasing afterwards. The incidence of long-term
unemployment for EU-15 immigrants converged to unity and remained on par with
natives in the most recent years. The EU-12 immigrants show a significantly high
ability to avoid long-term unemployment relative to natives and other immigrant
groups.
39
Figure 3 Immigrant-to-native ratio of the proportion of workers who look for another
job
Source: Own rendering based on EU-LFS 2004-2016 data
Note: The sample includes employed workers. Population weights are applied.
Figure 4 Immigrant-to-native ratio of workforce without a job for more than 11 months
Source: Own rendering based on EU-LFS 2004-2016 data
Note: The sample includes employed and unemployed people. Population weights are
applied.
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
2004 2006 2008 2010 2012 2014 2016
EU15
EU12
Europe
Africa
Asia
America
0.5
0.7
0.9
1.1
1.3
1.5
2004 2006 2008 2010 2012 2014 2016
EU15
EU12
Europe
Africa
Asia
America
40
Finally, Figure 5 compares the share of immigrant and native workers who
changed economic sectors (according to NACE classification) during the previous year.
We generally observe higher inter-sectoral mobility among immigrants than among
natives. Interestingly, immigrants with high on-the-job search (especially those from
America and to a lesser degree also those from Africa and the EU-12) are also among
the more mobile across sectors in the labor market. Before 2009, immigrants from
America and the EU-12 were substantially more likely to change economic sectors
within one year than natives. Their inter-sectoral mobility converged in later years but
remained high relative to natives.
Figure 5 Immigrant-to-native ratio of proportion of workers who changed industry
during the previous year
Source: Own rendering based on EU-LFS 2004-2016 data
Note: The sample includes employed people in the current and previous year.
Population weights are applied.
0.7
0.9
1.1
1.3
1.5
1.7
1.9
2.1
2.3
2.5
2.7
2.9
3.1
2004 2006 2008 2010 2012 2014 2016
EU15
EU12
Europe
Africa
Asia
America
41
3.4 Summary of descriptive evidence
Overall, the descriptive evidence presented in this chapter indicates the favorable
position of EU-15 immigrants in the European labor market. The working conditions of
this group are highly comparable to the native-born workforce. For this reason, in all
models we treat EU-15 countries as one entity, and study the relative responsiveness of
immigrants to skill shortages in relation to the EU native group (i.e. EU-15 natives,
including those living in their EU-15 member state of origin, as well as those living in
another EU-15 member state). This scenario is compatible with Borjas (2001), who uses
the group of US-born as a reference analysis.10
The findings presented in this chapter provide further evidence that immigrants
from the EU-12 differ from the other immigrant groups in several respects. They exhibit
high attachment to the labor market, but also relatively low risk of unemployment.
Unemployment spells for EU-12 immigrants are short, their job search intensity is high,
and they are able to change employment across economic sectors in the economy.
Along with immigrants from rest of Europe, EU-12 immigrants are most
concentrated in elementary occupations and exhibit the highest degree of down-skilling.
This is consistent with the hypothesis that being recently-arrived immigrant in the
receiving countries enables these immigrants to be more flexible workers, intensely
seeking employment opportunities and responding to changing economic conditions.
The responsiveness of the other immigrant groups is likely to be restrained by
10
As a formal check of descriptive evidence, the responsiveness of EU-15 immigrants is
compared with natives by estimating Equation 4. The estimate on skill shortage is not
significant to confirm the similarity between these two groups.
42
institutional barriers, as many such immigrants do not enjoy all the provisions of free
intra-EU mobility of workers.
43
4 The relative responsiveness of immigrants to
skill shortages
Using the data and methodology outlined above, we test the relative responsiveness of
immigrants, from various origins and length of stay, to skill shortages across sectors and
occupations.11
In all models, the dependent variable is the supply of immigrants relative
to the EU workforce in occupation-industry-country cells expressed in first difference.
The relative responsiveness of immigrants to skill shortages is assessed in comparison
with the EU workforce that represents workers born and residing anywhere in the
EU-15.
4.1 Baseline estimates
Table 4.1 shows the baseline results. The top panel in the table presents estimates from
the specification that includes the occupation-industry cell, country, and year fixed
effects. In the bottom panel, the model is augmented to include unemployment rate and
GDP growth to account for variation in economic conditions across countries and over
time. The key finding is that all coefficients on skill shortage presented in Table 1 are
positive, which is in line with the theoretical expectations. The positive estimates
obtained for skill shortage in the baseline model indicate that the relative supply of
immigrants in a particular occupation-industry-country group rose in those cells where
the wage premium (indicating a skill shortage) also rose. However, the statistical
significance of these results differs across immigrant groups.
11
This chapter extends the results published in Guzi et al. (2015, 2018).
44
Statistically significant estimates are obtained for EU-12 and European
immigrants, whereas immigrants from Asia and America, in the statistical sense, behave
similarly to the EU natives. The estimates for African immigrants fell short of statistical
significance (p-values are 0.125 and 0.16 in the top and bottom panel of Table 4.1,
respectively). These findings imply that immigrants are, in general, moving to
occupations and industries and countries that exhibit growing skill shortages at least as
much as, and for some immigrant groups more flexibly than, the natives. The inclusion
of GDP growth and unemployment rate only slightly changes the effect, and
importantly, the significance remains mostly unchanged.
Table 4.1 The relative responsiveness of immigrants to skill shortage
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.468 *** 0.65 *** 0.11 -0.05 0.114 0.236 0.461 *** 0.189 **
shortage (0.168) (0.225) (0.071) (0.126) (0.107) (0.151) (0.152) (0.096)
R2 0.027 0.035 0.035 0.026 0.016 0.035 0.049 0.097
N 2369 2281 2828 2031 2633 2474 2745 3170
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.367 ** 0.611 *** 0.102 -0.08 0.046 0.073 0.402 *** 0.177 *
shortage (0.167) (0.226) (0.073) (0.126) (0.109) (0.148) (0.154) (0.098)
GDP gr 0.002 0.013 -0.002 -0 0.004 0.015 *** -0.021 *** 0.002
(0.006) (0.010) (0.002) (0.007) (0.004) (0.006) (0.006) (0.002)
UR -0.027 *** -0.01 -0.002 -0.01 -0.013 ** -0.034 *** -0.019 *** -0.002
(0.009) (0.008) (0.004) (0.007) (0.006) (0.007) (0.006) (0.003)
R2 0.032 0.036 0.035 0.029 0.02 0.053 0.056 0.097
N 2369 2281 2828 2031 2633 2474 2745 3170
Source: Own calculations based on EU-SILC, EU-LFS, and Eurostat data.
Notes: The dependent variable is the supply of immigrants relative to EU native in the
particular occupation-industry-country cell expressed in first difference. The skill
shortage for the same cell is also expressed in first difference and lagged. All models
include cell, year, and country fixed effects. Models in the bottom panel include GDP
growth (GDP gr) and unemployment rate (UR). YSM indicates the group of immigrants
by years since immigration. The number of observations in the model varies because we
allow only occupation-industry-country cells of sufficient size in all years. Regressions
are weighted by the number of observations for the occupation-industry-country cell.
Heteroskedastic-consistent standard errors are in parentheses, *, **, *** identifying
significance at 10, 5, 1 per cent levels, respectively.
45
To test whether our results are driven by a particular country, we have replicated
the analysis using samples without Germany (to test if identifying immigrants by
nationality can potentially bias the results), and without Luxembourg (the country with
by far the highest share of immigrants). In both cases, the baseline estimates presented
in Table 4.1 change only minimally and retain statistical significance.
The estimated coefficients on skill-shortage can be interpreted in terms of the
relative elasticity of supply of immigrants and natives: ε=(dln(Z))/(dln(W)). This
elasticity gives the percentage change in the relative number of immigrants who choose
to reside in a particular occupation-industry-country cell for a given percentage change
in the wage. The wage index measures the average log-wages in each occupationindustry-country
cell, so that ε=β/Z. As the mean value of Z is 1.66 and 1.87 for EU-12
and European immigrants, respectively, the coefficients on skill shortage in the bottom
panel of Table 4.1 imply an elasticity of supply of 0.22 and 0.33 relative to EU natives.
For comparison, Borjas (2001) estimates an elasticity of 1.3 for new immigrants in the
US. The estimated elasticity for all immigrants cannot be calculated based on
information in his paper, but it is likely to be below one. In the UK, Dustmann et al.
(2012) estimates the elasticity of 2.0 for immigrants with less than 10 years in the UK.
Based on the information in their paper, the calculated elasticity for all immigrants is
0.88. Given the barriers to mobility across (and within) EU member states, we do not
find it surprising that our estimate of elasticity is somewhat lower than those estimated
for the US and UK.
The point estimates on skill shortages for the three groups of immigrants by
years since migration in Table 4.1 indicate a nonlinear nature of the relationship
between immigrants’ assimilation and their relative mobility. The magnitude of
46
estimates follows an inverse U-shaped pattern. The coefficient on skill shortages is the
highest for the middle group of immigrants, i.e. those with 6-10 years since migration,
and this finding is consistent with the notion of two competing forces affecting
immigrants’ relative mobility in comparison with the natives: with more years since
migration, immigrants learn to cope with mobility barriers, but become more firmly tied
to their local contexts in host labor markets.
Significant estimates on GDP growth and unemployment rate in Table 4.1
indicate that economic circumstances influence the supply of immigrants more than that
of the EU natives (e.g. Dustmann, Glitz and Vogel 2010; De la Rica and Polonyankina
2013). Interestingly the estimate for GDP growth is positive and significant for recent
immigrants (YSM 1-5) and negative and significant for the middle group of immigrants
(YSM 6-10). This implies that recently-arrived immigrants respond to skill shortages
very fluidly in growing economies, though economic growth reduces the labor supply of
the middle group of immigrants (YSM 6-10). In Chapter 6, the responsiveness of
immigrants to skill shortages during the business cycle is examined in detail.
The visible drop in our estimates on skill shortage for EU-12 and recentlyarrived
immigrants (YSM 1-5) after inclusion of economic variables in the bottom panel
of Table 4.1 signals a higher sensitivity of these immigrants to general labor market
conditions. This could indicate that informational asymmetries play a greater role for
these recently-arrived immigrant groups, who may be better able to discern the
aggregate unemployment rate, rather than labor market imbalances across occupationindustry-country
cells. The estimates for the other immigrant groups are only slightly
smaller in magnitude and retain their significance, compared with the model without
economic controls. These findings provide further support for the hypothesis that
47
immigrants respond to market incentives and opportunities more fluidly than natives. In
the further analysis, we proceed with the specification that includes controls for GDP
growth and the unemployment rate.
4.2 Estimates by education
In the next step, we replicate the analysis distinguishing workers with and without
tertiary education. The measure of skill shortage remains the same as above, but the
relative supply of immigrants and EU natives (Equation 3) is calculated separately for
low- and high-skilled workers. Estimates presented in Table 4.2 confirm that the higher
responsiveness of EU-12 and European immigrants, relative to the EU natives, is driven
by low-educated workers. The estimated effects of skill shortage for tertiary educated
workers are statistically not different from zero, meaning that these immigrant groups
respond to skill shortages similarly to the corresponding EU-natives. The only
exception is the group of high-skilled African immigrants, for which the estimate is
positive and significant.
An interesting pattern emerges in the results by time since immigration. Among
low-educated workers, the effect is significant for the group of immigrants with some,
but not extensive, experience in the host country (6-10 years). On the other hand, in the
group of high-educated workers, it is established immigrants (10+ years) who are more
responsive to skill market opportunities than EU natives. This could mean that
established high-educated immigrants are very responsive to skill shortages as they are
better equipped to overcome institutional barriers to their mobility than their loweducated
counterparts. The estimated effects of skill shortage for other immigrants are
statistically not different from zero, meaning that those groups respond to skill shortages
48
similarly to the corresponding EU natives. Particularly, the mobility of recently-arrived
low-educated immigrants may be constrained by institutional barriers linked to their
inexperience and short time in the country.
Table 4.2 Relative responsiveness to skill shortages of low- and high-skilled workers
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.447 *** 0.562 ** 0.047 -0.08 -0.027 0.054 0.428 ** 0.168
shortage (0.170) (0.229) (0.092) (0.155) (0.140) (0.162) (0.173) (0.113)
R2 0.037 0.042 0.026 0.036 0.018 0.054 0.074 0.091
N 1945 1862 2464 1672 2137 1830 2117 2930
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill -0.124 0.203 0.304 ** -0.19 0.164 0.115 -0.011 0.274 **
shortage (0.369) (0.398) (0.136) (0.202) (0.156) (0.268) (0.202) (0.117)
R2 0.039 0.031 0.032 0.029 0.022 0.067 0.025 0.026
N 1166 1006 1463 1021 1400 1199 1290 1817
Workers with tertiary education
Workers with less than tertiary education
Source: Based on EU-SILC, EU-LFS, and Eurostat data.
Note: See note to Table 4.1. Estimates are obtained from separate regressions. All
models include lagged GDP per capita and unemployment rate (not reported in the
table) and cell, year, and country fixed effects.
4.3 Estimates by the presence of children
A key conjecture of the theoretical model presented above is that migration costs D are
lower for immigrants than for natives. The results above indicate that these costs depend
on immigrants’ country of origin and years since migration. Although measuring D is
beyond the scope of this work, we can test the presence of children, an important
determinant of D, for immigrants’ relative responsiveness to labor market imbalances.12
12
The role of housing tenure would be another interesting aspect to test.
According to the EU SILC, immigrants are up to three times less likely to be
homeowners, indicating lower migration costs. However, the EU LFS does not contain
49
Table 4.3 Relative responsiveness to skill shortages of workers with and without
children
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.232 -0.224 -0.055 -0.06 -0.346 * 0.034 -0.318 -0.017
shortage (0.301) (0.421) (0.117) (0.181) (0.178) (0.276) (0.238) (0.119)
R2 0.052 0.021 0.022 0.03 0.021 0.041 0.084 0.039
N 1278 1437 1945 1177 1478 1286 1662 2318
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.3 * 0.679 *** 0.063 -0.15 0.165 0.051 0.478 *** 0.308 **
shortage (0.176) (0.247) (0.102) (0.166) (0.131) (0.169) (0.144) (0.129)
R2 0.055 0.044 0.039 0.035 0.027 0.046 0.054 0.086
N 1935 1751 2387 1626 2099 2059 2156 2865
Workers with children
Workers without children
Source: Based on EU-SILC, EU-LFS, and Eurostat data.
Note: See note to Table 4.1. Estimates are obtained from separate regressions. All
models include lagged GDP per capita and unemployment rate (not reported in the
table) and cell, year, and country fixed effects.
The EU LFS data includes information on whether the children of the individual
are living in the same household. We use that information to calculate the index of the
relative supply of immigrants and EU natives (Equation 3) separately on the sample of
workers with and without children. In this way, we can test whether the presence of
children in the household influences the relative responsiveness of immigrants to
changing wage premia in the labor market. The reference group is composed of EUnatives
as in the baseline model. The comparison of estimates for workers with and
without children in Table 4.3 indicates that the presence of children reduces the
mobility of immigrants in the labor market. The labor market mobility of immigrants
with children is statistically not different from that of EU-natives, with the only
exception being Americans with children, who are found to be less mobile than
comparable EU-natives. The results obtained from the baseline model are confirmed
housing tenure information, which makes it impossible to execute such test with this
data.
50
and imply that higher responsiveness to skill-shortages is found particularly among
immigrants from EU-12 and rest of Europe living without children.
4.4 Estimates by age and gender
We further test how the relative responsiveness of immigrants to skill shortages varies
by age and gender. The analysis is replicated with the index of the relative supply of
immigrants and EU natives calculated for three age cohorts (15-29, 30-44, and 45-65)
and separately for male and female workers. Table 4.4 presents results. Not
surprisingly, the youngest group of immigrants is found to be more responsive to wage
premia relative to their young native counterparts (the estimates are significant for
immigrants from EU-12 and rest of Europe, and larger in magnitude in comparison to
the baseline estimates). The labor mobility of immigrants in the middle- and old-aged
groups is confirmed to be similar to EU-natives with two exceptions: Asian immigrants
aged 30-44 are found to be less mobile, while immigrants in the group from rest of
Europe older than 45 are more mobile as comparable to EU-natives.
The results by gender display a peculiar pattern; whereas the effect for
immigrants from EU-12 is driven by male workers, the effect for Europe immigrants is
driven by female workers. Interestingly, the negative effect is estimated for male
immigrants of Asian origin and also in the middle-age cohort. This may imply barriers
to mobility in the EU states specific to this group.
51
Table 4.4 Relative responsiveness to skill shortages by age and gender
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.452 * 0.763 ** 0.319 0.121 -0.067 0.294 0.637 *** 0.311
shortage (0.254) (0.325) (0.232) (0.250) (0.202) (0.243) (0.208) (0.259)
R2 0.06 0.043 0.029 0.047 0.016 0.054 0.074 0.062
N 1251 1014 1313 864 1176 1535 1368 1686
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.195 0.326 0.096 -0.43 ** -0.198 -0.045 0.039 0.091
shortage (0.245) (0.313) (0.118) (0.212) (0.210) (0.216) (0.289) (0.164)
R2 0.035 0.048 0.043 0.026 0.023 0.038 0.054 0.086
N 1823 1787 2427 1537 2011 1799 2126 2813
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.01 1.243 *** 0.077 -0.04 0.277 -0.564 0.424 0.178
shortage (0.302) (0.356) (0.100) (0.186) (0.169) (0.414) (0.277) (0.111)
R2 0.035 0.055 0.026 0.041 0.011 0.039 0.061 0.059
N 1343 1250 2058 1371 1624 709 880 2712
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.526 *** 0.36 0.031 -0.28 * -0.099 0.064 0.412 *** 0.063
shortage (0.188) (0.246) (0.098) (0.147) (0.129) (0.192) (0.145) (0.104)
R2 0.042 0.021 0.022 0.03 0.02 0.047 0.08 0.053
N 1580 1684 2505 1634 2066 1729 1977 2793
EU12 Europe Africa Asia America YSM 1-5 YSM 6-10 YSM 11+
Skill 0.154 1.036 *** 0.204 0.087 0.109 0.11 0.346 0.36 **
shortage (0.229) (0.372) (0.132) (0.195) (0.160) (0.212) (0.242) (0.154)
R2 0.029 0.061 0.04 0.021 0.014 0.054 0.034 0.089
N 1803 1560 1933 1379 1951 1627 1855 2528
Workers of age 15-29
Workers of age 30-44
Workers of age 45-65
Male workers
Female workers
Source: Based on EU-SILC, EU-LFS, and Eurostat data.
Note: See note to Table 4.1. Estimates are obtained from separate regressions. All
models include lagged GDP per capita and unemployment rate (not reported in the
table) and cell, year, and country fixed effects.
52
4.5 Summary of findings
Our findings corroborate the notion developed in the theoretical section of this paper
that immigrant workers are expected to be more responsive to the changing economic
environment than native workers. The results imply that it is primarily low-skilled
migrants from EU-12 and rest of European countries and high-skilled African
immigrants who are more responsive to skill shortages relative to EU-natives. Other
immigrant groups respond to skill shortages similarly to the corresponding EU natives.
As conjectured by the theoretical model, the presence of children increases migration
costs and defers the labor mobility of immigrants. The series of additional analysis
reveals that it is young cohorts of immigrants who are primarily driving our results.
With respect to gender, the results are less clear and vary by the origin of immigrants.
53
5 Context-dependence of migrants’ responsiveness
The responsiveness of immigrants and natives to skill shortages may depend on a
number of contextual factors that are shaped by local economic, institutional, and policy
contexts.13
The theoretical model also implies that institutions and policies lowering the
costs of adjustment, requalification, or occupational mobility should, in general,
increase workers' responsiveness to skill shortages, although lock-in effects may set in
with the time spent in one destination. In this chapter, we extend the model to explore
how immigrants' responsiveness to skill shortages varies across various economic and
institutional conditions.
5.1 Augmented model
The baseline model (Equation 4) can be augmented to study how immigrants, relative to
natives, respond to skill shortages with respect to GDP level, economic growth,
unemployment rate, the generosity of welfare spending, immigration context, and the
restrictiveness of immigration policies. Specifically, we adopt a variation of the firstdifference
model of Equation (4) as follows:
∆ = ,∆ +, + 0∆ +,4 + 54 + - + - + - + 3 , (5)
where 4 is a dummy variable indicating the presence of a certain type of economic,
institutional and policy context in country and year 6. When the underlying contextual
information used to calculate this dummy variable is continuous, such as in the case of
the share of total social expenditure in GDP, the value of 4 is obtained by setting
13
This chapter extends the results published in Guzi et al. (2018).
54
4 = 1 for countries above the median value for all countries and zero otherwise. Thus,
in the case of social expenditure, 4 equals 1 for countries that are among the more
generous welfare states. While dichotomizing these variables results in some loss of
variation in the data, it enables us to capture institutional, policy, and economic
variation across the countries and work with interaction effects in a tractable way.
Adding the interaction term to the model changes the interpretation of the key
coefficients. In a model without the interaction term, , can be interpreted as the direct
effect of a skill shortage on the relative supply of immigrants. The interaction term
reflects the fact that immigrants' responsiveness to skill shortages may be different in
different contexts. Hence, in a model with the interaction term, the effect of a skill
shortage on the relative supply of immigrants is not limited to ,, but is equal to
, + 04 . , is then interpreted as the effect of a skill shortage on immigrants'
responsiveness when 4 = 0 (e.g. in countries with below-the-median welfare spending)
and , + 0 is the effect of a skill shortage when 4 = 1 (e.g. in countries with abovethe-median
welfare spending). The introduction of interaction terms thus enables us to
shed light on the heterogeneity of immigrants' relative responsiveness to skill shortages
across occupation-industry-country cells under different contexts.
5.2 The behavior of immigrants under different economic conditions
We first examine the sensitivity of results to the local economic conditions. To this end,
we introduce dummy variables indicating whether a country has, through our sample, an
above-the-median level of GDP, economic growth, unemployment rate, or share of
welfare spending in GDP (4 = 1; zero otherwise) and examine their interactions with
55
skill shortage.14
Thus, we pick up the medium- to long-term economic differentials
between the countries. Table 5.1 illustrates the separation of countries according to
these variables. It shows that economic conditions in the EU-15 are diverse and that
each of these three economic variables picks up unique dimensions of their economic
development.
Table 5.1 Economic and migrant context in the EU-15
Country GDP
pc
GDP
growth
Unempl.
rate
Welfare
spending
Migration
rate
Immigration
policy
AT 33277 1.42 5.21 14.23 0.18 -3
BE 30938 1.48 7.99 16.20 0.15 1
DE 31431 1.44 7.10 16.23 0.10 60
DK 32769 1.07 5.85 17.69 0.09 4
ES 24908 1.15 17.27 12.59 0.15 16
FI 29923 1.03 8.11 16.58 0.04 5
FR 27992 1.12 9.27 16.71 0.12 37
GR 21354 -1.14 16.38 9.48 0.09 25
IE 37700 4.62 9.88 13.63 0.17 8
IT 26933 -0.20 9.24 10.98 0.11 15
LU 67489 2.92 5.20 13.53 0.45 18
NL 34708 1.31 5.56 15.58 0.13 20
PT 20608 0.25 11.47 11.23 0.08 16
SE 32592 2.25 7.49 16.57 0.17 18
UK 28638 1.52 6.22 15.36 0.15 -50
Source: EU-LFS, Eurostat and DEMIG (2015).
Note: Reported values are average values across the period 2004-2016. GDP per capita,
total unemployment, and welfare spending are taken from the Eurostat database. Based
on EU-LFS, we calculate migration rate as the share of foreign-born individuals in the
working age population and the share of immigrants from non-EU27 countries in the
immigrant working age population. Migration policy indicator is based on 558 policy
changes identified in the DEMIG POLICY database in the EU15 countries over the
2004-2014 period which relate to border/land control and legal entry/stay. The indicator
of policy liberalization is calculated as the sum of policy changes coded as -1, 0 or 1
and weighted by the level of policy change (on the scale from 1 to 4). Figures in bold
indicate values above the median in the respective category.
14
Data from the Eurostat database accessed in February 2018: GDP in current prices
expressed in PPS per capita (table nama_10_pc), expenditure on social protection
excluding pensions in percentage of GDP (table tps00098).
56
Results are reported in Table 5.2 separately for immigrants of different origins
and time since arrival in the host country. The responsiveness to skill shortages for
immigrants from rest of Europe, as well as more established immigrant groups (YSM6-
10 and YSM11+), is statistically significant in economically weaker countries as
measured by GDP per capita or economic growth. Estimates for the EU-12 immigrants
are positive and fell short of statistical significant with p-value 0.14. Thus, it appears
that immigrants who have acquired the right to freely move across the EU (EU-12
immigrants) or have learned how to overcome barriers to mobility (with at least six
years since migration) are particularly instrumental in increasing labor market
efficiency in economically weaker countries.
The interactions with the unemployment rate indicate that the immigrant groups
that are particularly fluid in lower-GDP and slowly-growing countries (immigrants from
rest of Europe, as well as those with at least six years since migration) are also those
whose relative responsiveness to skill shortages is positive and statistically significant in
countries with high unemployment rates. On the other hand, the point estimates are, in
many cases, statistically not different in low- and high-unemployment countries, and
also in low- and high-GDP countries.
Remarkably, all the estimated coefficients on skill shortages are statistically zero
or positive. Thus, the finding that immigrants’ responsiveness to skill shortages equals
or exceeds that of the natives is robust to a country’s economic performance as
measured by countries’ GDP, economic growth, and unemployment rates.
57
Table 5.2 Relative responsiveness to skill shortages by economic conditions
EU12 Europe Africa Asia America YSM1-5 YSM6-10 YSM11+
Low GDP 0.404 0.928 ** 0.087 0.061 0.006 0.137 0.519 ** 0.245
(0.275) (0.366) (0.104) (0.214) (0.178) (0.189) (0.264) (0.153)
High GDP 0.326 0.264 0.131 -0.221 0.103 0.066 0.224 0.082
(0.221) (0.282) (0.095) (0.152) (0.101) (0.222) (0.148) (0.117)
Low GDPgr 0.426 1.005 ** 0.131 0.124 0.112 0.115 0.759 ** 0.351 **
(0.311) (0.400) (0.114) (0.242) (0.196) (0.203) (0.298) (0.171)
High GDPgr 0.322 0.255 0.085 -0.225 -0.004 0.09 0.057 0.012
(0.202) (0.262) (0.086) (0.143) (0.092) (0.203) (0.133) (0.104)
Low UR 0.227 0.252 0.079 -0.174 -0.096 -0.131 0.124 0.055
(0.161) (0.244) (0.081) (0.130) (0.089) (0.146) (0.122) (0.107)
High UR 0.57 1.113 ** 0.138 0.053 0.207 0.416 0.697 ** 0.301 *
(0.356) (0.451) (0.122) (0.283) (0.195) (0.294) (0.309) (0.171)
Low SOCX 0.486 * 1.034 *** 0.118 -0.002 0.069 0.194 0.485 ** 0.263 *
(0.263) (0.374) (0.093) (0.208) (0.153) (0.222) (0.231) (0.139)
High SOCX 0.167 -0.016 0.089 -0.208 0.02 -0.056 0.203 0.002
(0.146) (0.21) (0.11) (0.137) (0.112) (0.154) (0.133) (0.115)
Source: Own calculations based on EU-SILC, EU-LFS, and Eurostat data.
Notes: See note to Table 4.1. Estimates are obtained from separate regressions.
Presented results are calculated effects , and , + 0 from interactions with skill
shortage. The considered dimensions are GDP growth (GDP gr), unemployment rate
(UR), and the share of welfare spending in GDP (SOCX). See Table 5.1 for the
breakdown of countries.
The findings reported in Table 5.2 further imply that immigrant workers from
the EU-12 and rest of Europe are more responsive to skill shortages in those EU-15
countries that are less generous in terms of welfare spending. It appears that high
welfare spending makes immigrants less responsive to wage premia. This result
indicates a lock-in effect of welfare generosity on the relative responsiveness to labor
market imbalances. We also find that immigrants with a longer immigration history
(YSM 6-10 and YSM 11+) are particularly responsive to skill shortages in low-welfare
countries. In welfare-generous countries, the flexibly of immigrants in responding to
labor market opportunities is statistically no different from native workers.
58
5.3 Relative responsiveness to skill shortages by immigration context
European countries differ greatly in the characteristics of their immigrant population
and their immigration policies. We further explore how the relative responsiveness of
immigrants to skill shortages is affected by the scale of immigration and the
composition of immigrant population. Based on the EU-LFS, we calculate (i) the share
of foreign-born individuals in the working age population, denoted migration rate; and
(ii) the proportion of each immigrant group distinguished in our analysis in the
immigrant working age population (proxy for the size of immigrant network).
In addition, we construct an indicator of liberalization of migration policies from
the DEMIG POLICY database (DEMIG 2015). Migration policy indicator is based on
558 policy changes identified in DEMIG data in the EU-15 countries over the 2004-
2014 period which relate to border/land control and legal entry/stay. The indicator is
constructed as the sum of policy changes coded as -1, 0, or 1 (higher number implies
policy liberalization) and weighted by the level of policy change (on the scale from 1 to
4). Table 5.1 presents the country mean values of these variables. In the analysis, we
split the countries into two groups, with the median as the threshold, and introduce
interaction variables with the skill shortage variable.
Our findings presented in Table 5.3 show that immigrants from rest of Europe
and immigrants with a longer immigration history (YSM 6-10 and YSM 11+) are more
responsive to skill shortages than natives in countries with a below-the-median
migration rate. This indicates that a smaller immigrant population provides for greater
mobility of immigrant workforce, possibly due to less competition between immigrant
groups in the labor market. The opposite pattern emerges for EU-12 immigrants who
exhibit higher mobility in countries with an above-the-median migration rate.
59
Our estimates further corroborate the importance of social networks, i.e.
immigrants are more responsive to the changing economic environment when their
network is larger. One channel through which the social network decreases the
adjustment costs from mobility is by effectively transmitting information about the
economic environment. The immigrants originating from North and South Americas
and Australia exhibit higher mobility in countries where their network is smaller, which
may be attributed to the labor market competition within this immigrant group.
Table 5.3 Relative responsiveness to skill shortages by immigration context
EU12 Europe Africa Asia America YSM1-5 YSM6-10 YSM11+
Low MR 0.297 0.793 ** 0.092 0.084 0.029 0.06 0.607 ** 0.26 *
(0.260) (0.340) (0.103) (0.202) (0.180) (0.178) (0.276) (0.154)
High MR 0.411 * 0.333 0.122 -0.249 0.072 0.133 0.156 0.067
(0.228) (0.302) (0.096) (0.164) (0.112) (0.225) (0.153) (0.116)
Small
network 0.394 ** 0.216 0.064 -0.026 0.237 ** 0.207 0.365 *** 0.109
(0.165) (0.221) (0.102) (0.193) (0.096) (0.143) (0.125) (0.100)
Large
network 0.506 * 0.715 ** 0.154 -0.092 0.071 0.364 0.542 * 0.277
(0.275) (0.288) (0.096) (0.159) (0.143) (0.339) (0.32) (0.18)
Restrictive
policy 0.346 0.491 0.09 0.05 -0.155 -0.207 0.436 0.211
(0.288) (0.375) (0.091) (0.202) (0.197) (0.203) (0.286) (0.180)
Liberal
policy 0.387 ** 0.721 ** 0.112 -0.208 0.2 ** 0.306 0.371 ** 0.15 *
(0.193) (0.287) (0.102) (0.146) (0.101) (0.202) (0.162) (0.088)
Source: Own calculations based on EU-SILC, EU-LFS, and Eurostat data.
Notes: See note to Table 4.1. Estimates are obtained from separate regressions.
Presented results are calculated effects , and , + 0 from interactions with skill
shortage. The considered dimensions are migration rate (MR), the migration network,
and the immigration policy. See Table 5.1 for the breakdown of countries.
The openness of immigration policy affects the costs of immigration and
therefore the type of migrants who enter the country. A restrictive migration policy
effectively reduces the inflow of immigrants in the labor market and increases the
selectivity of immigrants. Consequences are difficult to predict with regard to
60
immigrants' potential to react to imbalances in the labor market. In contrast, an open
migration policy possibly intensifies competition in the host labor market, and may
therefore assist the labor market by removing shortages. Our results in Table 5.3 imply
that immigrants are particularly responsive to skill shortages relative to natives in
countries that have, over time, liberalized their immigration programs. It is also possible
that countries with an open migration policy also operate favorable integration policies
that enable immigrants to be more mobile across occupation-industry groups. Those
policies may include unrestricted access to jobs, the possibility to leave and re-enter the
country more easily, and the right to reside anywhere in the host country.
5.4 Summary of findings
We find that immigrants appear to be more responsive to skill shortages in relatively
poorer and high-unemployment countries relative to the natives. This implies that
immigration can help these economies with fluid labor.
Our results indicate that immigrant workers are particularly responsive in
countries with less generous welfare spending. Guzi et al. (2018) further explore the
role of welfare state institutions in determining the responsiveness of immigrants and
natives to skill shortages, and conclude that more generous welfare states do not
necessarily inhibit immigrants’ mobility. Rather, the authors find that welfare
generosity is more complex and may be shaped by various institutional
complementarities, and further affected by the characteristics of different immigrant
groups that mediate their access to welfare.
Finally, our results imply that immigrants are particularly responsive to skill
shortages relative to natives in countries that have introduced more liberal migration
61
policies, and that have a lower scale of immigration.. Our findings further corroborate
the importance of social networks; immigrants are more mobile in countries when their
network is larger.
62
63
6 Migration as an adjustment mechanism
in the Great Recession
In this chapter, we study how immigrants have responded to asymmetric economic
shocks across sectors, occupations, and states in the pre- and post-crisis periods in
Europe.15
The Great Recession that started in 2008 profoundly affected European
economies and labor markets, although the magnitude of impact differed considerably
between states.
The limited fiscal capacity and lack of room for monetary adjustment between
the member states of the eurozone make labor mobility an important vehicle of
adjustment to asymmetric economic shocks. In a country experiencing a positive labor
demand shock, workers are initially drawn from the unemployment pool and more
inactive workers start entering the labor force. If the shock persists, real wages rise and
the labor force starts growing as a result of the inflow of workers from other
geographical locations. Similar dynamics may be observed in the opposite direction in
the case of a negative shock. Arpaia et al., (2016) find that population movements in the
EU in response to economic shocks have almost doubled since the introduction of euro,
and the increase in mobility is triggered more by country than by region-specific shock.
The importance of mobility as an adjustment mechanism to asymmetric labor demand
shocks has therefore increased over time in the EU.
The responsiveness of migration to macroeconomic fluctuations was particularly
important for Europe during the recent Great Recession. The mobility of workers can
15
An earlier version of this chapter was published as Guzi and Kahanec (2017).
64
help to offset some of the unemployment increase. Jauer et al. (2014) estimate that
about a quarter of the unemployment increase during the Great Recession in the EU was
absorbed by population movements. The authors find that labor market adjustment in
Europe during the Great Recession of 2008 was primarily driven by immigrants from
recent EU accession countries.
Migrants originating from the member states that joined the EU in 2004 and
2007 (EU-12) have increased the adjustment capacity of the labor markets to cope with
asymmetric shocks (Kahanec, 2013). These immigrants are relatively young and well
educated, and adjust to changing economic conditions more flexibly. Kahanec and
Zimmermann (2016) document various adjustment channels at the EU-wide and
national levels, through which post-enlargement mobility helped to cushion some of the
economic shocks during the Great Recession. On the other hand, groups of immigrants
from outside the EU may face institutional, regulatory, or socio-economic constraints
limiting their adjustment potential. Findings from the literature also report that, relative
to natives, the labor market situation of immigrants is more sensitive to economic
shocks. Using data from Germany and the UK, Dustmann et al. (2010) show that
immigrants’ risk of unemployment is more sensitive to the economic cycle than that of
natives. De la Rica and Polonyankina (2013) find increased competition for jobs among
immigrants during the Great Recession in Spain. Relative to natives, immigrants moved
into jobs more intensive in manual skills during the recession years.
In the analysis in this chapter, we explore how the labor mobility of immigrants
from different origins evolved in the pre-crisis years and in the most recent, post-crisis,
years.
65
6.1 Changes in the relative responsiveness of immigrants over time
To understand how the responsiveness to skill shortages changes over the business
cycle, we use the following empirical strategy: We estimate the baseline model
(Equation 4) on seven consecutive subsamples between 2004 and 2016, each for a
seven-year interval.16
As before, in all models we treat EU-15 countries as one entity
and study the relative responsiveness of immigrants to skill shortages in comparison
with the EU native group.
The results reported in Table 6.1 show the responsiveness of immigrants from
different origins (in rows) and in different time intervals (in columns). The significant
estimates for EU-12 immigrants are obtained also in the samples containing more
recession years. The magnitude of effect obtained from different time periods follows a
U-shaped pattern and the effect becomes insignificant around the onset of the Great
Recession (2008-2014). The estimates for the European immigrant group suggest higher
responsiveness compared with the EU natives in the years before the Great Recession.
The responsiveness of other immigrant groups to skill shortage shows less definite
patterns: The responsiveness to skill shortages vis-à-vis the EU natives for the African
group is significant in the sample, only consisting of recession years, Asians’
responsiveness becomes negative in the recession years, and Americans’ responsiveness
is statistically indistinguishable from that of the EU-natives throughout the studied
periods.
16
The results obtained from subsamples of shorter or longer length lead to the same
conclusions. The results for alternative subsamples are available from the authors upon request.
Splitting the sample into seven-year intervals is preferred to provide a sufficient sample size.
66
With respect to time since immigration, the estimates for recently-arrived
immigrants (YSM 1-5) are higher in the later periods, but the effect becomes
statistically significant only in the 2009-2015 period. The estimate of the coefficients on
skill shortage for the immigrant group with arrival 6-10 years ago follows a hump
shape, with the peak around the inception of the Great Recession. The estimate for
established immigrants (YSM 11+) is strongest before the Great Recession and
decreases to near-zero in later periods.
6.2 Estimates by education
We further investigate how the responsiveness to skill shortages varies according to the
qualification of a worker. To this end, we replicate the estimation in Table 6.1 for
workers with and without tertiary education (we apply the same strategy in section 4.2).
The results are presented in Tables 6.2 and 6.3.
The positive estimates for EU-12 immigrants during the Great Recession and
immigrants from rest of Europe before the Great Recession are confirmed, arising
primarily in the low-educated segment of the labor market. High-skilled immigrants
from rest of Europe were particularly responsive before the Great Recession and around
the inception of the Great Recession (2006-2012). High-educated immigrants from
Africa and America exhibit higher responsiveness relative to comparable EU natives
during the recession years. The effect for the group of low-educated Americans is
peculiar, as it changes from positive (but the estimate fell short of statistical
significance) before the Great Recession to negative and significant in later periods. The
negative effect found for Asian immigrants in the recession years in Table 6.1 is driven
by high-skilled workers.
67
The observed patterns for immigrants by the years since immigration indicate
that the effect for immigrants with six to ten years since immigration is driven by loweducated
workers. The significant positive effects are present around the onset, but also
before, the Great Recession. The results obtained for established immigrants (11+ years)
in Table 6.1 are confirmed for both low- and high-educated workers – the effect is
present only in the years before the Great Recession for the low-skilled, and peaks
around the onset of the Great Recession for the high-skilled.
6.3 Summary of findings
In this chapter, we explore how our results vary over the business cycle during the Great
Recession. Low-educated immigrants from the EU-12 in particular have responded to
changing wage premia more fluidly than the EU natives also at the onset of the Great
Recession. The high-educated workers from African and American origins exhibit
particularly high responsiveness to skill shortages during the Great Recession.
We also find that low-skilled immigrants’ responsiveness to skill shortages
peaks for those with 6-10 years since migration. This may point to their (still) low
attachment to their specific location in the host labor market, but an already sufficient
adjustment to the conditions, and hence, ability to overcome barriers to migration in the
host labor market. For established immigrants (YSM 11+), responsiveness to skill
shortages is higher in the period before the Great Recession, during which the effect is
reduced. However, the effect for high-skilled established immigrants in this category
peaks around the onset of the Great Recession.
Overall, our results suggest that immigrants may play an important role in labor
adjustment during times of asymmetric economic shocks.
68
Table 6.1 Relative responsiveness of immigrants in different periods
2004-2010 2005-2011 2006-2012 2007-2013 2008-2014 2009-2015 2010-2016
EU12 0.388 0.812 ** 0.723 ** 0.573 ** 0.434 0.618 ** 0.099
(0.267) (0.320) (0.292) (0.262) (0.299) (0.302) (0.213)
R2 0.107 0.041 0.057 0.048 0.058 0.058 0.037
N 1085 1111 1090 1070 1070 1070 1070
Europe 1.254 *** 0.690 ** 0.744 ** 0.393 0.282 0.049 0.054
(0.346) (0.349) (0.361) (0.354) (0.396) (0.397) (0.280)
R2 0.092 0.051 0.079 0.040 0.025 0.032 0.039
N 1051 1069 1047 1025 1025 1025 1025
Africa 0.029 0.127 0.193 * 0.226 ** 0.180 0.246 ** 0.012
(0.101) (0.104) (0.103) (0.114) (0.129) (0.122) (0.115)
R2 0.052 0.035 0.042 0.043 0.028 0.037 0.050
N 1287 1316 1300 1285 1285 1285 1284
Asia 0.202 0.195 0.216 -0.095 -0.228 -0.322 -0.378 *
(0.174) (0.186) (0.195) (0.211) (0.233) (0.239) (0.219)
R2 0.090 0.057 0.074 0.055 0.063 0.056 0.049
N 915 951 940 930 930 930 930
America 0.108 0.051 -0.040 -0.200 -0.207 -0.100 -0.008
(0.113) (0.128) (0.129) (0.190) (0.204) (0.218) (0.232)
R2 0.044 0.044 0.035 0.031 0.046 0.028 0.040
N 1195 1227 1213 1200 1200 1200 1198
YSM 1-5 -0.100 0.235 0.066 0.282 0.284 0.713 ** 0.062
(0.189) (0.250) (0.235) (0.243) (0.287) (0.307) (0.223)
R2 0.078 0.068 0.087 0.099 0.114 0.112 0.103
N 1130 1156 1138 1120 1120 1120 1120
YSM 6-10 0.572 *** 0.617 *** 0.641 *** 0.281 0.243 0.197 0.083
(0.192) (0.208) (0.230) (0.247) (0.300) (0.312) (0.259)
R2 0.105 0.050 0.073 0.068 0.071 0.069 0.093
N 1257 1283 1261 1240 1240 1240 1240
YSM 11+ 0.439 *** 0.365 *** 0.360 *** 0.116 0.024 -0.052 -0.144
(0.122) (0.129) (0.131) (0.172) (0.190) (0.193) (0.176)
R2 0.151 0.102 0.163 0.138 0.125 0.108 0.101
N 1450 1479 1457 1435 1435 1435 1433
Source: Own calculations based on EU-SILC, EU-LFS, and Eurostat data.
Notes: The dependent variable is the supply of immigrants relative to EU native
expressed in first difference. All models include cell, year, and country fixed effects.
YSM indicates the group of immigrants by years since immigration. Each cell includes
the coefficient on skill shortage estimated from a separate model for different immigrant
groups (in rows) and on the sample limited to the period specified in the header row.
The number of observations in the model varies because we allow only occupationindustry-country
cells of sufficient size in all years. Regressions are weighted by the
number of observations for the occupation-industry-country cell. Heteroskedasticconsistent
standard errors are in parentheses, *, **, *** identifying significance at 10, 5,
1 per cent levels, respectively.
69
Table 6.2 Relative responsiveness of immigrants (workers with less than tertiary educ.)
2004-2010 2005-2011 2006-2012 2007-2013 2008-2014 2009-2015 2010-2016
EU12 0.456 0.861 *** 0.659 ** 0.514 ** 0.475 * 0.623 ** 0.136
(0.286) (0.321) (0.274) (0.241) (0.283) (0.291) (0.206)
R2 0.102 0.043 0.062 0.054 0.063 0.068 0.041
N 883 913 899 885 885 885 885
Europe 1.104 *** 0.411 0.428 0.056 0.097 -0.079 0.140
(0.353) (0.378) (0.353) (0.348) (0.376) (0.376) (0.288)
R2 0.098 0.060 0.088 0.057 0.036 0.040 0.037
N 860 871 853 835 835 835 835
Africa -0.067 0.048 0.125 0.133 0.129 0.183 -0.041
(0.127) (0.132) (0.129) (0.136) (0.159) (0.151) (0.138)
R2 0.038 0.033 0.039 0.037 0.026 0.033 0.042
N 1115 1146 1135 1125 1125 1125 1124
Asia 0.193 0.126 0.192 -0.138 -0.203 -0.332 -0.372
(0.202) (0.212) (0.213) (0.231) (0.257) (0.281) (0.312)
R2 0.094 0.100 0.112 0.085 0.093 0.068 0.037
N 754 786 775 765 765 765 765
America 0.216 0.075 -0.175 -0.392 * -0.505 ** -0.367 -0.223
(0.139) (0.149) (0.145) (0.222) (0.244) (0.267) (0.318)
R2 0.035 0.043 0.040 0.035 0.051 0.028 0.029
N 967 995 984 975 975 975 975
YSM 1-5 -0.070 0.225 -0.035 0.085 0.055 0.351 -0.022
(0.216) (0.274) (0.246) (0.238) (0.290) (0.314) (0.244)
R2 0.078 0.077 0.101 0.111 0.119 0.098 0.113
N 834 856 843 830 830 830 830
YSM 6-10 0.591 *** 0.657 *** 0.649 *** 0.264 0.290 0.217 0.115
(0.205) (0.208) (0.227) (0.260) (0.321) (0.351) (0.328)
R2 0.127 0.057 0.079 0.078 0.075 0.071 0.103
N 967 992 976 960 960 960 958
YSM 11+ 0.454 *** 0.322 ** 0.282 * -0.013 -0.063 -0.111 -0.190
(0.146) (0.146) (0.146) (0.185) (0.204) (0.212) (0.204)
R2 0.138 0.103 0.151 0.121 0.115 0.094 0.096
N 1342 1369 1347 1325 1325 1325 1323
Source: Own calculations based on EU-SILC, EU-LFS, and Eurostat data.
Notes: See notes to Table 6.1.
70
Table 6.3 Relative responsiveness of immigrants (workers with tertiary education)
2004-2010 2005-2011 2006-2012 2007-2013 2008-2014 2009-2015 2010-2016
EU12 0.080 0.886 0.821 0.341 -0.198 -0.337 -0.370
(0.612) (0.542) (0.581) (0.655) (0.677) (0.619) (0.541)
R2 0.087 0.059 0.092 0.112 0.089 0.068 0.036
N 531 547 537 528 529 530 529
Europe 1.349 ** 0.878 1.351 ** 0.458 -0.572 -0.405 -0.862
(0.640) (0.626) (0.611) (0.663) (0.736) (0.715) (0.544)
R2 0.051 0.056 0.109 0.081 0.077 0.066 0.024
N 466 474 462 450 450 450 450
Africa 0.262 0.406 ** 0.544 *** 0.730 *** 0.505 ** 0.545 ** 0.256
(0.188) (0.194) (0.199) (0.244) (0.251) (0.231) (0.234)
R2 0.059 0.038 0.040 0.042 0.044 0.047 0.045
N 665 684 673 664 665 665 665
Asia 0.254 0.147 0.054 -0.183 -0.505 -0.505 -0.655 **
(0.339) (0.303) (0.288) (0.333) (0.354) (0.332) (0.324)
R2 0.060 0.081 0.066 0.057 0.066 0.070 0.087
N 447 480 479 478 478 478 478
America -0.074 0.073 0.338 0.612 ** 0.811 *** 0.554 * 0.426
(0.206) (0.234) (0.242) (0.296) (0.292) (0.295) (0.281)
R2 0.041 0.036 0.043 0.076 0.121 0.052 0.082
N 629 655 650 644 643 643 642
YSM 1-5 0.450 0.662 0.561 0.616 0.229 0.695 0.037
(0.443) (0.436) (0.430) (0.505) (0.531) (0.486) (0.363)
R2 0.099 0.137 0.184 0.178 0.147 0.152 0.068
N 545 561 552 545 545 545 545
YSM 6-10 0.091 0.167 0.278 0.367 0.364 0.199 -0.095
(0.330) (0.310) (0.306) (0.361) (0.404) (0.353) (0.292)
R2 0.041 0.034 0.045 0.031 0.059 0.044 0.071
N 582 604 596 590 590 590 590
YSM 11+ 0.517 *** 0.511 ** 0.695 *** 0.692 *** 0.507 ** 0.250 0.030
(0.185) (0.198) (0.186) (0.222) (0.224) (0.208) (0.146)
R2 0.056 0.033 0.048 0.068 0.047 0.081 0.036
N 822 852 841 830 830 830 829
Source: Own calculations based on EU-SILC, EU-LFS, and Eurostat data.
Notes: See notes to Table 6.1.
71
7 Conclusion
The topic of migration and labor mobility has been at the center of discussion in
Europe, particularly since the 2004 and 2007 EU enlargements and the European
migration crisis of 2014-16. In addition to security concerns, economic arguments have
played a major role in these debates, primarily concerning the economic impacts of
labor mobility on the receiving labor markets. Fundamental economic arguments
suggest that immigrants can be expected to be more fluid than natives in responding to
changing skill and labor imbalances, and hence contribute to a more efficient allocation
of labor in the EU. This is because immigrants’ costs of departing their home, job, and
family and friend networks in the country of origin are sunk upon arrival to a new
country, as their decision to part with them has already been made. It follows that this
should be particularly true for newly-arrived immigrants, not yet deeply integrated in
their new milieu in the receiving country. In contrast, natives’ decision to move entails
pecuniary and non-pecuniary costs of parting with their original home, job, and
networks.
The migration literature documents important contributions of the mobile
immigrant workforce in the labor adjustment process (Borjas, 2001; Dustmann et al.,
2012; and Guzi et al., 2015; Jauer et al., 2014). This work extends the methodology of
Borjas (2001), Guzi and Kahanec (2017), and Guzi et al. (2018) by studying the
responsiveness of immigrants to changing skill shortages. The analysis is performed in
the context of the EU-15 countries, which differ in many economic, institutional, and
policy variables. We use data from EU-LFS, EU-SILC, and several other auxiliary
72
datasets, to exploit this diversity and study how it interacts in the way in which natives
and different types of immigrants respond to labor market imbalance.
We find that immigrants’ responsiveness to skill shortages equals or exceeds that
of the natives across all the studied contexts. In particular, cases indicating that
immigrants would be less flexible than natives in responding to skill shortages are rare
in this analysis. We confirm that immigrants from the EU-12 and the rest of Europe
(outside the EU) are more responsive in comparison with the natives, and we find
statistically significant positive results for African and American immigrants in specific
economic, institutional, and policy contexts. Asian immigrants remain the only group
that does not exhibit higher mobility in the labor market under any contexts.
Conversely, lower labor market mobility is identified for male immigrants of Asian
origin and in the middle-age cohort. This may imply barriers to mobility in the EU
states specific to this group.
Our results bring additional insights to the literature, i.e. that it is low-skilled
immigrants from the EU-12 and the rest of Europe and high-skilled African immigrants
who are more responsive to skill shortages relative to EU-natives. Other immigrant
groups respond to skill shortages similarly to the corresponding EU natives. This
finding is consistent with initial observations that immigrants from the EU-12 exhibit
high attachment to the labor market, their job search intensity is high, and they are able
to change employment across economic sectors in the economy. Immigrants from the
EU-12 and the rest of Europe are most concentrated in elementary occupations and
exhibit the highest degree of down-skilling. This is consistent with the hypothesis that
recently-arrived immigrants are more flexible workers, intensely seeking employment
opportunities and responding to changing economic conditions. The responsiveness of
73
the other immigrant groups is likely to be restrained by institutional barriers, as many
such immigrants do not enjoy all the provisions of free intra-EU mobility of workers.
As conjectured by the theoretical model, the presence of children increases
migration costs and defers the labor mobility of immigrants. The series of additional
analysis reveals that it is young cohorts of immigrants who are primarily driving our
results. With respect to gender, the results are less clear and vary by the origin of
immigrants.
The diversity across EU member states enables us to study how immigrants'
responsiveness interacts with economic conditions and institutional factors. Immigrants
appear to be more responsive to skill shortages in relatively poorer and highunemployment
countries relative to the natives. This implies that immigration can help
these economies with fluid labor. An important finding is that high welfare spending
may disincentivize the flexibility of immigrants relative to natives in responding to skill
shortages. However, the role of the welfare state may involve various institutional
complementarities beyond the impact of welfare generosity. Our results further show
that immigrant workers are particularly fluid in countries with a generally lower scale of
immigration, and in countries that have introduced a more open immigration policy.
In the analysis of immigrants’ relative responsiveness to labor shortages over the
business cycle during the Great Recession, we follow the labor market using moving
seven-year windows. We find varying patterns for different immigrant groups.
Immigrants from the EU-12 exhibited high responsiveness to labor shortages relative to
the EU natives, which peaked and became significant during the Great Recession.
However, the effect is concentrated in the low-skilled segment of the labor market.
Results for immigrants from the rest of Europe imply their relative responsiveness is
74
positive and statistically significant only in the periods covering mostly the pre-crisis
years. Again, the effect is stronger when measured among low-skilled workers. The
high-skilled African and American immigrants are confirmed more mobile relative to
EU natives during the Great Recession. Low-educated immigrants originating from
America exhibit a peculiar pattern, however. The higher responsiveness in the prerecession
period (2004-2010) gradually diminishes and these immigrants became less
responsive relative to natives during the recession years. Whether this could reflect a
tightening of the migration policy with respect to non-EU groups during the Great
Recession is a topic for further investigation. We also find that low-skilled immigrants’
responsiveness to skill shortages peaks for those with 6-10 years since migration. This
may point to their (still) low attachment to their specific location in the host labor
market, but an already sufficient adjustment to the conditions, and hence, ability to
overcome barriers to migration in the host labor market. For established immigrants
(with more than 10 years since migration), responsiveness to skill shortages is higher in
the period before the Great Recession, during which the effect is reduced. However, the
effect for high-skilled established immigrants in this category peaks around the onset of
the Great Recession.
As for the limitations of our study, the analysis presented does not permit causal
interpretations, since the studied economic, institutional, and policy contexts cannot
always be seen as fully exogenous. Similarly, although skill shortages are lagged by one
period and are measured regardless of the immigrant status of workers in a cell, due to
some serial correlation, the immigrant-native relative labor supply could still affect skill
premia across cells. However, it may be argued that such reverse channels reinforce our
results, as they tend to attenuate the studied effects (Dustmann et al., 2012).
Furthermore, our data are not capturing irregular migrants who are typically even more
75
responsive to labor market changes. From the perspective of immigrant populations, we
are unable to consider the aspect of the quality of employment that immigrants attain
when flexibly responding to labor market opportunities across the EU countries.
Our study shows that the coefficient of labor-shortage is significantly positive in
many, but not all economic, institutional, and policy contexts in the EU-15. This is an
important result that deserves further study. Our results also indicate that policies
matter; while some policies seem to enable immigrants to respond to changing labor
market conditions, others may be inhibiting immigrant workers’ mobility. As
immigrants’ labor market mobility provides for a more efficient allocation of labor in
host labor markets, policies that inhibit their mobility are costly in terms of forgone
GDP and forgone economic opportunities.
76
77
8 References
Amuedo-Dorantes, C., & de la Rica, S. (2010). Immigrants’ responsiveness to labor
market conditions and their impact on regional employment disparities: Evidence
from Spain. SERIEs, 1(4), 387–407. http://doi.org/10.1007/s13209-009-0002-y
Anderson, B., & Ruhs, M. (2008). A need for migrant labour? The micro-level
determinants of staff shortages and implications for a skills based immigration
policy. COMPAS: Oxford University Press.
Arpaia, A., Kiss, A., Palvolgyi, B., & Turrini, A. (2016). Labour mobility and labour
market adjustment in the EU. IZA Journal of Migration, 5(21), 1–21.
http://doi.org/10.1186/s40176-016-0069-8
Arratibel, O., Heinz, F. F., Martin, R., Przybyla, M., Serafini, R., Zumer, T., &
Rawdanowicz, L. (2007). Determinants of growth in the central and eastern
European EU member states: A production function approach. ECB Occasional
Paper No. 61, (61).
Bauer, T. K., & Zimmermann, K. F. (1999). Assessment of Possible Migration Pressure
and Its Labor Market Impact following EU Enlargement to Central and Eastern
Europe. IZA Research Report No. 3. Bonn: IZA.
Bennett, J., & McGuinness, S. (2009). Assessing the impact of skill shortages on the
productivity performance of high-tech firms in Northern Ireland. Applied
Economics, 41(6), 727–737.
Bonin, H., Eichhorst, W., Florman, C., Hansen, M. O., Skiöld, L., Stuhler, J.,
Zimmermann, K. (2008). Geographic Mobility in the European Union: Optimising
its Economic and Social Benefits. IZA Research Report No. 19. Bonn: IZA.
Borjas, G. J. (2008). Labor Economics (4th ed.). Dublin: McGraw-Hill
International/Irwing.
Borjas, G. J. (2001). Does Immigration Grease the Wheels of the Labor Market?
Brookings Papers on Economic Activity, 32(1), 69–134.
Boswell, C., Stiller, S., & Straubhaar, T. (2004). Forecasting labour and skills shortages:
How can projections better inform labour migration policies. Hamburg Institute of
International Economics (HWWA), Paper Prepared for the European Commission,
DG Employment and Social Affairs, (July), 1–54.
Cohen, M. (1995). Labor Shortages as America approaches the twenty first century.
Ann Arbor: University of Michigan Press.
Constant, A., & Tien, B. (2011). Germany’s Immigration Policy and Labour Shortages.
IZA Research Report No. 41. Bonn: IZA.
78
Constant, A. F., Gataullina, L., & Zimmermann, K. F. (2009). Ethnosizing immigrants.
Journal of Economic Behavior and Organization, 69(3), 274–287.
http://doi.org/10.1016/j.jebo.2008.10.005
Czaika, M. (2012). Migration in times of uncertainty. On the role of economic prospects
(No. 52). International Migration Institute.
D’Amuri, F., & Peri, G. (2014). Immigration, jobs, and employment protection:
Evidence from europe before and during the great recession. Journal of the
European Economic Association, 12(2), 432–464.
http://doi.org/10.1111/jeea.12040
De la Rica, S., & Polonyankina, T. (2013). The impact of Immigration on Occupational
Specialisation among Natives in Spain: Does the Business Cycle Matter? Revista
de Economía Aplicada, 21(63).
Dustmann, C., & Frattini, T. (2011). Immigration: the European experience. Centro
Studi Luca d’Agliano Development Studies Working Paper, (326).
Dustmann, C., Frattini, T., & Preston, I. (2012). Can immigration constitute a sensible
solution to sub-national and regional labour shortages? Report for the Migration
Advisory Committee.
Dustmann, C., Glitz, A., & Vogel, T. (2010). Employment, wages, and the economic
cycle: Differences between immigrants and natives. European Economic Review,
54(1), 1–17.
Fernandez, C., & Ortega, C. (2008). Labor market assimilation of immigrants in Spain:
Employment at the expense of bad job-matches? Spanish Economic Review, 10(2),
83–107. http://doi.org/10.1007/s10108-007-9032-4
Guzi, M., & Kahanec, M. (2017). How immigrants helped EU labor markets to adjust
during the Great Recession. International Journal of Manpower, 38(7), 996–1015.
http://doi.org/10.1108/IJM-08-2017-0205
Guzi, M., Kahanec, M., & Mýtna Kureková, L. (2018). How Immigration Grease is
Affected by Economic, Institutional and Policy Contexts: Evidence from EU Labor
Markets. Kyklos, in press.
Guzi, M., Kahanec, M., & Mýtna Kureková, L. (2015). How Immigration Grease Is
Affected by Economic, Institutional and Policy Contexts: Evidence from EU Labor
Markets. IZA Discussion Paper No. 9188. Bonn: Institute of Labor Economics.
Hicks, J. (1932). The Theory of Wages. London: Macmillan.
Jauer, J., Liebig, T., Martin, J. P., & Puhani, P. A. (2014). Migration as an Adjustment
Mechanism in the Crisis? A Comparison of Europe and the United States. OECD
social, employment and migration working papers.
79
Kahanec, M. (2013). Labor mobility in an enlarged European Union. International
Handbook on the Economics of Migration. Cheltenham: Edward Elgar.
Kahanec, M., Biavaschi, C., Mýtna Kureková, L., & Zimmermann, K. W. (2013).
Labour Migration from EaP Countries to the EU – Assessment of Costs and
Benefits and Proposals for Better Labour Market Matching. IZA Research Report
No. 56. Bonn: IZA.
Kahanec, M., & Shields, M. (2013). The working hours of immigrants in Germany:
temporary versus permanent. IZA Journal of Migration, 2(1), 1–15.
Kahanec, M., & Zaiceva, A. (2009). Labor market outcomes of immigrants and noncitizens
in the EU: An East-West comparison. International Journal of Manpower,
30(1), 97–115.
Kahanec, M., & Zimmermann, K. F. (2016). Labor Migration, EU Enlargement, and
the Great Recession. Berlin, Heidelberg: Springer.
Kahanec, M., & Zimmermann, K. F. (2010). Migration in an Enlarged EU: A
Challenging Solution? In Five Years of an Enlarged EU: A Positive Sum Game
(pp. 63–94). http://doi.org/10.1007/978-3-642-12516-4_5
Kogan, I. (2007). Working through barriers. Dordrecht: Springer Netherlands.
Kogan, I. (2011). New Immigrants - Old Disadvantage Patterns? Labour Market
Integration of Recent Immigrants into Germany. International Migration, 49(1),
91–117.
Lewis, W. (1954). Economic Development with Unlimted Supplies of Labour.
Manchester School of Economic and Social Studies, 22, 139–191.
http://doi.org/10.1111/j.1467-9957.1954.tb00021.x
Lucifora, C., & Origo, F. (2002). The economic cost of the skill gap in Europe. Istituto
per La Ricerca Sociale, Milan, Italy.
Martin, P. (2010). A need for migrant labour? UK–US comparisons. In M. Ruhs & B.
Anderson (Eds.), Who needs migrant workers. Oxford: Oxford University Press.
Martin, P., & Ruhs, M. (2011). Labor Shortages and U.S. Immigration Reform:
Promises and Perils of an Independent Commission. International Migration
Review, 45(1), 174–187.
Massey, D. S., Arango, J., Hugo, G., Kouaouci, A., Pellegrino, A., & Taylor, J. E.
(1993). Theories of International Migration: A Review Migration and Appraisal.
Population and Development Review, 19(3), 431–466.
Neugart, M., & Schömann, K. (2002). Employment Outlooks: Why Forecast the Labour
Market and for Whom? WZB Discussion Paper, FS I 02-20.
80
Obadić, A. (2006). Theoretical and empirical framework of measuring mismatch on a
labour market. Zbornik Radova Ekonomskog Fakulteta U Rijeci: Časopis Za
Ekonomsku Teoriju I Praksu, 24(1), 55–80.
Padoa-Schioppa, F. (1991). Mismatch and labour mobility. Cambridge: Cambridge
University Press.
Quintini, G. (2011). Over-Qualified or Under- Skilled? A Review of Existing Literature.
Paris: OECD Publishing.
Ritzen, J., & Zimmermann, K. F. (2014). A vibrant European labor market with full
employment. IZA Journal of European Labor Studies, 3(1), 1–24.
http://doi.org/10.1186/2193-9012-3-10
Røed, M., & Schøne, P. (2012). Does immigration increase labour market flexibility?
Labour Economics, 19(4), 527–540.
Ruhs, M. (2011). Openess, skills and rights: An empirical analysis of labour
immigration programmes in 46 high- and middle-income countries. Policy and
Society Working Paper No. 88. Oxford: Center on Migration.
Rutkowski, J. (2009). Ukraine. Labor Demand Study. Washington: World Bank.
Sjaastad, L. A. (1962). The Costs and Returns of Human Migration. Journal of Political
Economy, 70(S5), 80.
Striestka-Ilina, O. (2007). Skills shortages. Contribution to the Agora conference
“Building a European VET area.” Thessaloniki: Cedefop.
Tang, J., & Wang, W. (2005). Product market competition, skill shortages and
productivity: Evidence from Canadian manufacturing firms. Journal of
Productivity Analysis, 23(3), 317–339.
Todaro, M. P., & Smith, S. C. (2006). Economic Development. Melbourne: Pearson
Addison Wesley.
Trendle, B. (2008). Skill and labour shortages - definition , cause and implications.
Labour Market Research Unit, 1(54), 1–34.
Veneri, C. M. (1999). Can Occupational Labor Shortages be Identified Using Available
Data? Monthly Labor Review, 122(3), 15–21.
Zimmermann, K., Bonin, H., Fahr, R., & Hinte, H. (2007). Immigration Policy and the
Labor Market. Berlin, Heidelberg: Springer-Verlag.