THE SELECTION OF HIGH-SKILLED EMIGRANTS

Matthias Parey, Jens Ruhose, Fabian Waldinger, and Nicolai Netz*

Abstract—We measure selection among high-skilled emigrants from Ger-
many using predicted earnings. Migrants to less equal countries are
positively selected relative to nonmigrants, while migrants to more equal
countries are negatively selected, consistent with the prediction in Borjas
(1987). Positive selection to less equal countries reﬂects university qual-
ity and grades, and negative selection to more equal countries reﬂects
university subject and gender. Migrants to the United States are highly
positively selected and concentrated in STEM ﬁelds. Our results highlight
the relevance of the Borjas model for high-skilled individuals when credit
constraints and other migration barriers are unlikely to be binding.

Introducción

I NTERNATIONAL migration of high-skilled individu-

als has risen dramatically in recent decades (Docquier
& Rapoport, 2012). Between 2000 y 2006, los unidos
States attracted 1.9 million and European OECD countries
attracted 2.2 million tertiary-educated migrants (Widmaier
& Dumont, 2011). In the year 2000, high-skilled migrants
represented about 11% of the tertiary-educated population
in OECD countries (Brücker et al., 2012). In the United
Estados, as of 2013, acerca de 19% of the working-age popula-
tion with a bachelor’s degree or higher were foreign born. En
certain ﬁelds such as science, tecnología, engineering, y
matemáticas (STEM), más que 30% were foreign born.1

Access to high-skilled individuals is central to ﬁrms’ suc-
cess and has become even more important in economies
where ideas drive technological progress (Chambers et al.,
1998). When the home-grown pool of high-skilled individu-
als is insufﬁcient, the ability to attract high-skilled migrants
is crucial for improving the quality of a country’s workforce
and its innovative capacity. A deeper understanding of the
selection of high-skilled migrants is therefore important for
sending and receiving countries alike.

While migrant selection has been studied extensively since
Borjas (1987) outlined theoretical predictions for selec-
ción, few papers have studied the selection of high-skilled
migrants. Focusing our analysis on high-skilled migrants
who mostly migrate between developed countries enables
us to investigate a setting where individuals face low legal

Received for publication October 30, 2015. Revision accepted for

publication September 7, 2016. Editor: Amit K. Khandelwal.

* Parey: University of Essex and Institute for Fiscal Studies; Ruhose: Leib-
niz Universität Hannover; Waldinger: London School of Economics; Netz:
DZHW.

We thank the editor Amit Khandelwal and three anonymous referees for
very helpful comments. We also thank Clement de Chaisemartin, tomás
crossley, Christian Dustmann, Tim Hatton, Hans Hvide, Julian Johnsen,
Gordon Kemp, and participants in various seminars. We thank Kolja Briedis,
Christian Kerst, and Gregor Fabian for providing access to the DZHW data.
This study uses data from the Swiss Labour Force Survey (Schweizerische
Arbeitskräfteerhebung, BFS). M.P. gratefully acknowledges the support of
the ESRC Research Centre on Micro-Social Change at the University of
Essex.

A supplemental appendix is available online at http://www.mitpress

journals.org/doi/suppl/10.1162/REST_a_00687.

1 Authors’ calculations based on the 2013 ACS (Ruggles et al., 2010).

barriers to migration and relatively small migration costs.
The economic forces described by the Borjas model should
be particularly relevant in our setup.2

A basic version of the Borjas (1987, 1991) modelo, build-
ing on Roy (1951), predicts that migrants to less equal
countries, such as the United States, should be positively
selected, while migrants to more equal countries, como
Dinamarca, should be negatively selected. Analyzing migra-
tion to both less and more equal countries is therefore
particularly valuable to test the predictions of the model.

We study the selection of high-skilled emigrants by
investigating migration decisions of graduates from Ger-
man universities. Germany exhibits an intermediate level of
inequality for high-skilled individuals (figura 1). By study-
ing selection to less and more equal countries in the same
contexto, we can test both predictions of the Roy/Borjas
modelo. Además, we are able to test whether the pre-
dictions of the Roy/Borjas model hold within the population
of university graduates.3

We use rich survey data on German university gradu-
ates collected by the German Centre for Higher Education
Research and Science Studies (DZHW). German university-
bound students represent a more selective group than their
counterparts in most other economically developed coun-
intentos; this allows us to study migration patterns of the top
11% of the educational distribution.4

To measure selection, we compare predicted earnings of
migrants and nonmigrants. We ﬁrst estimate an augmented
Mincer regression for graduates who work in Germany.
We then use the estimated returns to construct predicted
earnings independent of whether the graduate stays in Ger-
many or migrates abroad. Our data contain a rich set of
personal characteristics including family background, alto
school grades, university education (including the speciﬁc
university, sujeto, and grades), and information on mobility
before enrolling at university. These detailed characteristics

2 See section IIB for a review of empirical papers investigating migrant
selection across the entire skill distribution. The existing papers on migrant
selection mostly focus on low-skilled migration between Mexico and the
United States, where migrants face higher migration costs and legal barriers
to entry. Other papers on high-skilled migrants study other outcomes, semejante
as effects on the receiving economy (Hunt & Gauthier-Loiselle, 2010; Kerr
& Lincoln, 2010; Borjas & Doran, 2012; Moser, Voena, & Waldinger, 2014;
Kerr, Kerr, & Lincoln, 2015; Doran, Gelber, & Isen, 2014) and on source
countries (Docquier & Rapoport, 2012).

3 Since many papers investigate migrant selection between two countries
solo (see online appendix table A.1), they are limited to testing one of the
two predictions of the Roy/Borjas model. While Borjas, Comerciante, y
Poutvaara (2015) study migration from Denmark to multiple destinations,
they focus on positive selection because Denmark has very low levels of
inequality.

4 Administrative data show that about 11% of the cohorts we study in our
paper graduated from a university. En 2012, the stock of university graduates
entre 35- to 44-year-olds was 1,208,000 out of a population of 11,004,000
(DESTATIS, 2013).

La revista de economía y estadística., December 2017, 99(5): 776–792
© 2017 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology. Publicado bajo una atribución Creative Commons 3.0
no portado (CC POR 3.0) licencia.
doi:10.1162/REST_a_00687

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
/
r
mi
s
t
/

a
r
t
i
C
mi
–
pag
d

F
/

9
9
5
7
7
6
1
9
7
4
8
0
9
/
r
mi
s
t
_
a
_
0
0
6
8
7
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
8
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

THE SELECTION OF HIGH-SKILLED EMIGRANTS

777

Figure 1.—Earnings Inequality among the High Skilled: Ratio of 75th
to 25th Percentile in the Earnings Distribution of University Graduates

EE.UU
Francia
Poland
Italia
España
Japón
Canada
Reino Unido
Austria
Luxembourg
Suiza
Bélgica
Alemania
Irlanda
Suecia
Países Bajos
Australia
Norway
Finland
Dinamarca

1.2

1.4

1.6

1.8

75/25 ratio, university graduates

The ﬁgure shows the ratio of the 75th to the 25th percentile in the earnings distribution of university
graduates. Authors’ calculations based on country-speciﬁc earnings surveys (see online appendix table A.3),
showing averages over the period 1998 a 2010. Details on data sources and the construction of inequality
measures are reported in section IIIB and data appendix B.1.

allow us to obtain predicted earnings as a precise measure
of individual earnings potential, so that we can differentiate
between high- and low-productivity graduates. We then com-
pare cumulative distribution functions of predicted earnings
for three groups of graduates: graduates who stay in Ger-
muchos, graduates who migrate to less equal countries, y
graduates who migrate to more equal countries. This allows
us to investigate whether the most or least skilled univer-
sity graduates stay in Germany or select into more or into
less equal destinations. To classify destinations into either
more or less equal countries, we construct new inequality
measures for university graduates, based on individual-level
income surveys from twenty countries.

The selection of university graduates is consistent with
the predictions of the basic Roy/Borjas model. Migrants to
less equal countries have signiﬁcantly higher predicted earn-
ings than nonmigrants. Migrants to more equal countries,
in contrast, have signiﬁcantly lower predicted earnings than
nonmigrants. These ﬁndings hold along the whole distri-
bution of predicted earnings. De hecho, the selection patterns
predicted by the model hold even within subgroups of either
more equal or less equal countries.

The coefﬁcients of the Mincer regression, which form the
basis of our earnings prediction, might be biased if migrants
were nonrandomly selected from the population of graduates
in a way not captured by our observed covariates. We address
potential selection in the augmented Mincer regression using
a sample selection correction (Heckman, 1979). In the selec-
tion equation we use the rollout of ERASMUS, the largest
study-abroad program, as an instrumental variable to pre-
dict whether individuals move abroad or work in Germany.
Changes in the number of ERASMUS places are a good pre-
dictor of international migration (Parey & Waldinger, 2011).
Using the selection-corrected Mincer regression, we conﬁrm
our main results. We also show that our results are not

driven by our particular measure of earnings inequality or
potentially confounding factors that may be correlated with
cross-country inequality.

Además, we show that our results hold for migrants
to European countries. Migration costs to these countries
are low because workers can move freely between Euro-
pean countries without the need for work visas. In further
resultados, we show that migrants to Austria and Switzerland,
two countries with higher earnings inequality than Ger-
muchos, are positively selected, as predicted by the Roy/Borjas
modelo. These countries provide a useful setting to test for
migrant selection because migration costs are particularly
bajo: the two countries share a border with Germany, son
predominantly German speaking, and have broadly similar
labor market institutions, beneﬁt systems, and cultures. Fur-
thermore, Germans do not need visas to work in Austria or
Suiza.

In additional results, we decompose predicted earnings to
identify the characteristics that explain the observed selec-
tion patterns. Migrants to less equal countries have better
university grades, attend better universities, and come from
families with higher socioeconomic backgrounds. Migrants
to more equal countries have studied subjects with lower
returns in the labor market, are more likely to be women,
and attend universities associated with lower labor market
prospects. Curiosamente, migrants to more equal destinations
are in fact positively selected in terms of university grades.
Selection patterns are thus consistent with the model predic-
tions for most, but not all, characteristics.5 Predicted earnings
provide a comprehensive summary measure of expected
productivity that drives migration decisions.

Finalmente, we investigate selection to the United States, uno
of the most important destinations of high-skilled emigrants
from Germany. En los Estados Unidos, earnings inequality
among university graduates is much higher than in Ger-
muchos. As predicted by the Roy/Borjas model, emigrants
from Germany to the United States are positively selected
compared to nonmigrants. We show that migrants to the
United States are positively selected across almost all char-
acteristics compared not only to nonmigrants in Germany but
also to U.S. natives. We also document that migrants from
Germany to the United States are particularly concentrated
in high-paying STEM ﬁelds.

En general, high-skilled individuals form an important group
of potential migrants because of both their relatively high
rates of mobility and their potential contribution to the host
economía. Studying migrant selection in this context is partic-
ularly useful because these migrants face low formal barriers
to migration and are unlikely to be credit constrained. El
observed selection patterns underline the relevance of the
Roy/Borjas model in this setting.

5 A multidimensional extension of the Roy/Borjas model indicates that
focusing on a single characteristic may not reﬂect the overall pattern of
selección, depending on the correlation with other relevant characteristics.
See Dustmann, Fadlon, and Weiss (2011) for a model with two types of
habilidades.

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
/
r
mi
s
t
/

a
r
t
i
C
mi
–
pag
d

F
/

9
9
5
7
7
6
1
9
7
4
8
0
9
/
r
mi
s
t
_
a
_
0
0
6
8
7
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
8
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

778

THE REVIEW OF ECONOMICS AND STATISTICS

II. A Model of Migrant Selection and

Existing Empirical Evidence

A. Roy/Borjas Model of Migrant Selection

In his seminal work, Borjas (1987, 1991) proposes a
theoretical framework for understanding the selection of
international migrants. To motivate our empirical analy-
hermana, we use important insights of the Roy/Borjas model to
highlight the predictions for selection. In our context, univer-
sity graduates decide whether to migrate based on earnings
opportunities abroad (w1) and at home (w0), and migration
costos (C). In this framework, potential log earnings consist of
an observed component (θj, where j = 0 indicates home and
j = 1 indicates abroad) and an unobserved component ((cid:3)j):

log w0 = θ0 + (cid:3)0,
log w1 = θ1 + (cid:3)1.

(1)
(2)

Taking migration costs (C) into account, individuals will
move abroad if the wage gain is larger than the migration
costos:

Migrate = 1 if θ1 + (cid:3)1 > θ0 + (cid:3)0 + C.

(3)

, p2
(cid:3)0

, p2
(cid:3)1

, p2
θ1

The vector of potential outcomes is (θ0, θ1, (cid:3)0, (cid:3)1). Para
tractability, we assume that the outcome vector is jointly
normally distributed with means (μ0, μ1, 0, 0) and variances
(p2
). Mean earnings at home and abroad are
θ0
represented by μj, and the variance of the observed compo-
nent in each country is represented by σ2
. We allow each
θj
type of skill (observables and unobservables) to be correlated
across countries but not across types of skills. σθ0,θ1 is the
covariance in the observed component across countries. Nosotros
refer to the corresponding correlation as ρθ. While our frame-
work incorporates observed and unobserved skills, this does
not affect the underlying economic mechanism developed
by Borjas (1987, 1991).6

We now consider how earnings potential at home, θ0, de
migrants differs from the population mean μ0. Desde
normality assumption, we obtain

mi(θ0|Migrate=1)

(cid:2)

= mi(θ0|θ1 + (cid:3)1 > θ0 + (cid:3)0 + C)
σθ0
σθ1

σθ1
σθ0
σv

= μ0 +

ρθ −

(cid:3)

Fi(z)
1 − Φ(z)

(4)

(5)

where v = θ1+(cid:3)1−θ0−(cid:3)0 is the earnings difference between
abroad and home that has variance σ2
v. z = (μ0 + c − μ1)/σv
is a constant reﬂecting differences in means across destina-
ciones, adjusted for migration costs and normalized by the

(cid:5)

(cid:4)

σθ0/σθ1

variance of the earnings difference. In our empirical analy-
hermana, we investigate how selection on observables relates to
relative inequality
between the two destinations.7
In addition to relative inequality, the theoretical prediction
on selection depends on the cross-country correlation in the
observed component (ρθ). A situation where ρθ is sufﬁciently
high provides a natural benchmark case because we ana-
lyze migration ﬂows between industrialized countries.8 If
the potential destination is less equal than home (σθ1 > σθ0),
migrants will be positively selected: mi(θ0|Migrate=1) > μ0.
Intuitivamente, the positively selected migrants beneﬁt from the
upside opportunities in less equal countries. If the potential
destination country is more equal (σθ1 < σθ0), migrants will be negatively selected: E(θ0|Migrate=1) < μ0. Intuitively, the negatively selected migrants beneﬁt from the insurance of a compressed wage distribution in more equal countries. The model emphasizes the role of earnings inequality for the selection of migrants. Differences in mean earnings between home and abroad have strong effects on migration probabilities (see term z above), but they have no effect on the direction of selection. Borjas (1991) extends the model to include stochastic migration costs, leading to very similar results as long as the migration costs are unrelated to potential earnings; Chiquiar and Hanson (2005) emphasize that selection patterns can change substantially when migration costs vary systemati- cally with earnings potential. Because we are focusing on high-skilled individuals who migrate from an economically developed country to other developed countries, differen- tial migration costs are presumably less important than for lower-skilled migrants who migrate from, for example, Mexico to the United States. B. Empirical Evidence on the Roy/Borjas Model Most empirical papers on international migrant selection study settings where migrants face legal barriers to migration and migration costs are relatively high. Existing papers dif- fer along two main dimensions that affect observed selection patterns. First, different papers use different skill measures to evaluate selection, and second, they study migration ﬂows between a varying set of countries (see online appendix table A.1). A large part of the empirical literature has stud- ied emigration from Mexico to the United States. While some of these papers ﬁnd evidence for negative selection that is consistent with the basic Roy/Borjas model (e.g., Ibarraran & Lubotsky, 2007; Fernandez-Huertas Moraga, 2011; Kaestner & Malamud, 2014, for some characteristics), other papers ﬁnd intermediate selection that suggests that migration costs vary with skills, perhaps driven by poverty constraints (Chiquiar & Hanson, 2005; Orrenius & Zavodny, 2005; Kaestner & Malamud, 2014, for other characteristics). 6 Borjas (1987) develops the original model focusing on the role of unobservables. In the formulation here, this corresponds to the case of σθ0 = 0. Borjas (1991) introduces the distinction between returns to observables and unobservables, focusing on the case where observable skills are perfectly correlated across countries (corr(θ0, θ1) = 1). = σθ1 7 Our data include a rich set of observable characteristics, which allows us to construct an informative measure of skills. See Gould and Moav (2016) for an analysis that investigates selection on unobservable skills. 8 This rules out the case of “refugee sorting” (Borjas, 1987). l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 THE SELECTION OF HIGH-SKILLED EMIGRANTS 779 In their seminal paper, Chiquiar and Hanson (2005) show that a model with skill-varying migration costs provides a better description of migration ﬂows from Mexico to the United States. A number of other papers investigate migrant selec- tion between other pairs of countries. The selection of migrants from Puerto Rico to the United States is consis- tent with the model predictions (Ramos, 1992; Borjas, 2008). Migrant selection from either Norway or Israel to the United States is only partly consistent with the model predictions (Abramitzky, Boustan, & Eriksson, 2012; Gould & Moav, 2016). Finally, a number of papers investigate migrant selection between multiple countries. Some papers ﬁnd support for the model predictions (e.g., Borjas, 1987; Borjas et al., 2015; Stolz & Baten, 2012), while other cross-country studies ﬁnd only partial support for the basic Roy/Borjas model (Belot & Hatton, 2012) or reject the model predictions (Feliciano, 2005; Grogger & Hanson, 2011).9 We are not aware of other papers that focus on the role of inequality for the selection of high-skilled migrants.10 Study- ing these migrants is particularly useful because they face low legal barriers to migration and relatively small migration costs. III. Data A. Data on University Graduates We analyze the selection of high-skilled migrants using survey data on university graduates collected by the Ger- man Centre for Higher Education Research and Science Studies (DZHW). These data come from nationally represen- tative longitudinal surveys of individuals who complete their university education in Germany (for details, see Grotheer et al., 2012). The DZHW sampled university graduates from the graduation cohorts 1992–93, 1996–97, 2000–2001, and 2004–2005.11 We refer to the cohorts by the second year (e.g., 1993 for the 1992–93 cohort). Graduates in each cohort are surveyed twice. The initial survey takes place about twelve months after graduation. The same individu- als participate in a follow-up survey about ﬁve years after graduation (online appendix ﬁgure A.1). The survey is ideal for our purposes because graduates are surveyed even if they move abroad. We focus our analysis on migration decisions that are measured ﬁve years after graduation.12 The surveys are based on a stratiﬁed cluster sampling, with ﬁelds of study, degree types, and universities as strata (Grotheer et al., 2012), and they are representative for the examined pop- ulation. Response rates to the initial surveys range between 30% and 40%, depending on the cohort. We analyze differ- ences in response rates between the initial survey and the follow-up survey according to migration status reported in the initial survey. The follow-up survey response rate is 66% for graduates who have worked in Germany during the ini- tial survey and 59% for graduates who have worked abroad. While this difference is statistically signiﬁcant in a simple t-test, we cannot reject that differences in response rates are uncorrelated to observable characteristics. This suggests that attrition does not change our ﬁndings. We also verify that our results hold when we include the full sample from the initial survey by carrying forward the reponses from the initial sur- vey (see online appendix A.3 for details; results are shown in online appendix ﬁgure A.3 and online appendix table A.6). Five years after graduation, the total number of respon- dents is 6,737 (1993 cohort), 6,237 (1997 cohort), 5,426 (2001 cohort), and 6,459 (2005 cohort). We focus on gradu- ates from traditional universities.13 We restrict the sample to full-time workers because migrating part-time workers are more likely to be tied movers, that is, individuals follow- ing a migrating partner (see Borjas & Bronars, 1991; Junge, Munk, & Poutvaara, 2014). In our data, full-time labor force participation is about 77%. The graduate survey data contain detailed information on graduates’ personal characteristics, family background, study history, and labor market experience (table 1). In addi- tion to the variables summarized in table 1, we also have detailed information on a graduate’s university and ﬁeld of study. Five years after graduation, 5.2% of graduates work abroad. The main destinations are Switzerland, the United States, the United Kingdom, Austria, and France (online appendix table A.2). B. Data on Earnings Inequality We classify destination countries as either more or less equal to Germany using newly constructed measures of earn- ings inequality for university graduates. Existing inequal- ity measures, such as Gini coefﬁcients, typically measure inequality for the overall population, but the decisions of 9 Our focus is on the selection of international migrants. A number of papers investigate the Roy/Borjas model applied to internal migration, including Borjas, Bronars, and Trejo (1992), Dahl (2002), Abramitzky (2009), and Bartolucci, Villosio, and Wagner (2014). 10 Recent papers have highlighted the role of taxes for the migration of inventors and soccer players (Akcigit, Baslandze, & Stantcheva, 2016; Kleven, Landais, & Saez, 2013). 11 Between 1993 and 2005, the majority of German university graduates completed Diplom, Magister, or Staatsexamen degrees. These degrees are usually completed in four to six years and are considered comparable to a master’s degree in other countries in standard international classiﬁca- tions (ISCED 5A, according to the International Standard Classiﬁcation of Education). 12 After graduation, many university graduates enroll in additional training such as legal or teacher traineeships (Referendariat), or Ph.D. programs. Earnings in the initial survey are thus a noisy measure of earnings potential. 13 The German higher education sector consists of traditional universities, universities of applied sciences (Fachhochschulen), specialized universi- ties (focusing on arts, music, or theology), and a small number of private universities. The best students tend to enroll in traditional universities. To estimate the Heckman selection model, we also restrict the sample to grad- uates from universities where at least one graduate works abroad. These sample restrictions reduce the sample by around 30%. Results that include all institutions are very similar to our main ﬁndings (see online appendix ﬁgure A.4). l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 780 THE REVIEW OF ECONOMICS AND STATISTICS Table 1.—Summary Statistics for German University Graduates Job characteristics (after ﬁve years) Working abroad Annual earnings in euro (2001 prices) Potential experience in months Postgraduate education Ph.D. completed Further (non-Ph.D.) degree completed Education ﬁrst degree Final university grade Studied abroad Age at graduation ERASMUS/total students in subject Education before ﬁrst degree Final school grade Apprenticeship Previous mobility Studied in same state as high school Personal characteristics Female Partner Married Child(ren) Parental background Mother’s education (years) Father’s education (years) Mother self-employed Mother salaried employee Mother civil servant Mother worker Mother did not work Father self-employed Father salaried employee Father civil servant Father worker Father did not work Observations Full Sample Mean SD 0.052 43,491 69.201 0.191 0.073 2.018 0.078 26.994 0.040 2.110 0.220 0.659 0.445 0.780 0.416 0.291 13.459 14.852 0.092 0.597 0.108 0.100 0.103 0.194 0.447 0.223 0.113 0.023 11,091 – 19,334 4.161 – – 0.681 – 2.664 0.057 0.639 – – – – – – 3.102 3.065 – – – – – – – – – – Working in Germany Abroad More Equal Mean 0 43,265 69.197 0.182 0.071 2.032 0.072 27.026 0.039 2.119 0.225 0.663 0.444 0.782 0.421 0.297 13.423 14.816 0.093 0.596 0.105 0.103 0.104 0.191 0.448 0.221 0.116 0.024 10,510 Mean 1 39,458 69.719 0.313 0.125 1.698 0.240 26.271 0.052 1.951 0.094 0.583 0.594 0.740 0.281 0.156 14.458 15.458 0.063 0.677 0.177 0.042 0.041 0.188 0.479 0.271 0.063 0.000 96 Abroad Less Equal Mean 1 49,231 69.194 0.371 0.122 1.787 0.169 26.437 0.050 1.959 0.138 0.581 0.445 0.736 0.344 0.184 14.035 15.493 0.091 0.619 0.148 0.049 0.093 0.262 0.406 0.258 0.062 0.012 485 The table shows summary statistics of German university graduates at ﬁve years after graduation. Information on earnings is available for 10,315 of the 11,091 graduates. Average annual earnings of 43,491 euros in 2001 prices correspond to around 79,084 U.S. dollars in 2014 prices. highly skilled migrants will likely depend on the earnings inequality of university graduates. Our main data source is the Luxembourg Income Study (LIS) (2013), which provides access to individual-level earn- ings surveys from several countries. The database covers different years for each country. We use all available sur- vey years for the main destinations of German university graduates (see table A.3 for available survey years in each country). Switzerland and Austria are important destinations for German university graduates but have only relatively lim- ited coverage in the LIS database. We therefore augment the LIS data with additional data for Austria (Microcensus 1999 and EU-SILC 2007, 2008) and Switzerland (Labour Force Survey 1998–2005). To measure earnings inequality for high-skilled individ- uals, we restrict the samples in the individual-level income surveys to university graduates. We further restrict the sam- ples to full-time employees of working age, and we exclude individuals who are self-employed, enrolled in educational institutions, or report negative earnings. Based on the individual-level surveys, we construct earn- ings percentiles for each country and available year using the survey sampling weights (see online appendix table A.3). Some surveys in the (augmented) LIS data report gross earnings, while others report net earnings. To measure cross- country inequality of net earnings, we convert gross into net earnings using the net personal average tax rate of single persons without children from the OECD (2013b).14 Data appendix B.1 provides more detail on the construction of the inequality measures. In our main analysis, we use the ratio of the 75th to the 25th earnings percentile (75/25 ratio) for university gradu- ates to measure earnings inequality across countries. Figure 1 14 The net personal average tax rate is deﬁned as the personal income tax and employee social security contributions net of cash beneﬁts, expressed as a percentage of gross wage earnings. The OECD reports three different tax rates along the earnings distribution: the average tax rate at 67%, 100%, and 167% of average earnings. We apply the tax rate at 67% of average earnings to the 25th percentile and below, the tax rate at 100% of average earnings to earnings between the 25th and the 75th percentile, and the tax rate at 167% of average earnings to the 75th percentile and above. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 THE SELECTION OF HIGH-SKILLED EMIGRANTS 781 shows the ranking of countries according to the 75:25 ratio that we average over 1998 to 2010 to reﬂect the years that correspond to our graduate surveys (online appendix table A.4 reports 75:25 ratios for each country). Inequality is high- est in the United States, followed by France and Poland. The Scandinavian countries and Australia are most equal. Ger- many is ranked in the middle.15 We can therefore investigate the selection of German university graduates into less equal and more into equal countries.16 C. Data on ERASMUS Places As part of our estimation procedure, which we explain below, we use data on the number of ERASMUS places to correct for potential selection bias. We obtain data on the number of study-abroad places in the ERASMUS pro- gram by university, subject, and year from the German Academic Exchange Service (DAAD). The median inter- nationally mobile student studies abroad for one or two semesters about three years before graduation. We assign the number of ERASMUS places in the corresponding aca- demic year, subject, and university to each student. To account for differences in cohort size that affect students’ study-abroad opportunities, we normalize the number of ERASMUS places with the number of students in the cor- responding university and subject (for details, see Parey & Waldinger, 2011). IV. Methods and Results A. The Selection of Migrants to More and Less Equal Destinations For our analysis, we use predicted earnings to measure earnings potential in the home country. This measure of skill represents θ0 in the model we have outlined. We then use predicted earnings to compare the distribution of skills of migrants to less equal countries, migrants to more equal countries, and nonmigrants. To construct predicted earnings, we estimate an aug- mented Mincer regression for nonmigrants only: log w0i = Xiβ0 + ε0i. (6) 15 Recent papers have used large administrative data sets to document a rise in German earnings inequality during the past decades (Dustmann, Ludsteck, & Schönberg, 2009, Card, Heining, & Kline, 2013). In these data sets, earnings are censored at the maximum of social security contributions. For university graduates, 42% (13%) of observations for men (women) are top-coded between 1998 and 2008. Because university graduates are in the top 11% of the educational distribution, we prefer to use earnings surveys in the LIS that are not top-coded. 16 As we measure selection with predicted earnings, an ideal measure of inequality would be based on country-level differences in returns to observed skills. Such a measure would require graduate data sets with com- parable characteristics on each graduate for all major destinations. As these are not available, we use the 75:25 ratio that is based on actual earnings. The empirical results are valid as long as countries with higher 75:25 ratios also exhibit higher returns to observed skills. The estimate of β0 measures returns to skills in the home country. Our data allow us to include a large number of variables Xi to obtain a good prediction of earnings poten- tial. Xi contains variables that measure university experience (ﬁnal university grade, age at graduation, completing univer- sity with a bachelor’s degree, 24 subject ﬁxed effects, and university ﬁxed effects), additional education after gradua- tion (completing a Ph.D. or a non-Ph.D. graduate degree), preuniversity education (ﬁnal high school grades and appren- ticeship before studying), previous mobility (moving to another state between high school and university), potential labor market experience, personal characteristics (gender, marital/partnership status, children), parental background (mother’s and father’s education and occupation), and grad- uate cohort ﬁxed effects. The coefﬁcients of the augmented Mincer regression have the expected signs and magnitudes (table 2, column 1).17 The R2 of about 0.28 is high for a Mincer regression, suggesting that predicted earnings are an informative skill measure for university graduates.18 Next, we predict potential earnings in the home country for migrants and nonmigrants. The predictions are based on the coefﬁcient vector (ˆβ0) and individual characteristics Xi: ˆθ0i = Xi ˆβ0. (7) We then use this measure of skills to compare three groups of interest: migrants to less equal countries, migrants to more equal countries, and nonmigrants. Speciﬁcally, we construct cumulative distribution functions (CDFs) of predicted earn- ings ˆθ0 by migration group, F(ˆθ0 | Migration status), and plot them in the left panel of ﬁgure 2a. The dashed line shows the CDF of nonmigrants. The solid dark line is the CDF of migrants to less equal destinations, such as the United States. This CDF lies to the right of the CDF for nonmigrants, indicating that this group is posi- tively selected in terms of earnings potential. The migrants to these countries have skills that, according to the returns in the Mincer regression, are valued more highly than those of nonmigrants: median log predicted earnings for these migrants are 10.65 compared to 10.61 for nonmigrants. The CDFs of nonmigrants and of migrants to less equal countries do not cross, indicating that these migrants are positively selected over the full range of predicted earnings. We test the statistical signiﬁcance of our ﬁndings in section IVC. The lighter solid line shows the CDF of migrants to more equal destinations, such as Denmark. It indicates that migrants to more equal countries are negatively selected rela- tive to non-migrants. Median log predicted earnings for these 17 Because all graduates are surveyed around ﬁve years after gradu- ation, the variation in potential labor market experience is small, and estimated coefﬁcients are different from the typical pattern observed in Mincer regressions. The omitted degree is a Diplom/Magister degree. Com- pared to graduates with these traditional degrees, graduates sampled after completing a bachelor’s degree have lower earnings. 18 We show that our results are robust to excluding the controls for children and marital/partnership status from the wage regression (online appendix ﬁgure A.5). l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 782 THE REVIEW OF ECONOMICS AND STATISTICS Table 2.—Augmented Mincer Regression for University Graduates in Germany (1) (2) (3) Dependent Variable Labor Earnings Labor Earnings Working in Germany OLS Heckman Selection Model Selection Equation Education ﬁrst degree Final university grade Final grade squared Bachelor’s degree Age at graduation Age squared Postgraduate education Ph.D. completed Further (non-Ph.D.) degree completed Education before ﬁrst degree Final school grade School grade squared Apprenticeship Previous mobility Studied in same state as high school Potential work experience Experience in months Experience squared Personal characteristics Female Partner Married (additionally) Child(ren) Parental background Mother’s education (years) Father’s education (years) Mother self-employed Mother salaried employee Mother civil servant Mother worker Father self-employed Father salaried employee Father civil servant Father worker ERASMUS places/students Mills ratio Graduate cohort FE Subject FE University FE R2/Pseudo-R2 Observations 0.048∗ −0.023∗∗∗ −0.131∗∗∗ −0.026∗∗ 0.000∗ −0.003 −0.024 −0.041 0.009 0.037∗∗∗ −0.010 −0.058∗∗∗ 0.000∗∗∗ −0.131∗∗∗ 0.066∗∗∗ 0.028∗∗∗ −0.040∗∗∗ 0.003∗ 0.003∗ −0.008 −0.012 −0.019 −0.001 0.054∗∗ 0.041∗ 0.027 0.003 YES YES YES 0.282 9,778 (0.027) (0.006) (0.028) (0.011) (0.000) (0.011) (0.015) (0.034) (0.008) (0.010) (0.008) (0.022) (0.000) (0.008) (0.009) (0.009) (0.009) (0.002) (0.002) (0.017) (0.013) (0.018) (0.016) (0.025) (0.024) (0.025) (0.026) 0.046∗ −0.023∗∗∗ −0.132∗∗∗ −0.026∗∗ 0.000∗ 0.000 −0.021 −0.043 0.010 0.037∗∗∗ −0.012 −0.059∗∗∗ 0.000∗∗∗ −0.131∗∗∗ 0.065∗∗∗ 0.028∗∗∗ −0.041∗∗∗ 0.003∗ 0.003∗ −0.009 −0.012 −0.019 −0.003 0.056∗∗ 0.041∗ 0.028 0.003 −0.050 YES YES YES 9,778 (0.027) (0.006) (0.028) (0.011) (0.000) (0.013) (0.016) (0.034) (0.008) (0.010) (0.008) (0.022) (0.000) (0.008) (0.009) (0.009) (0.010) (0.002) (0.002) (0.017) (0.013) (0.017) (0.016) (0.025) (0.024) (0.025) (0.026) (0.095) 0.079 −0.007 0.049 −0.013 0.001 −0.367∗∗∗ −0.251∗∗∗ 0.109 −0.011 0.078 0.131∗∗∗ 0.096 −0.001 −0.047 0.070 0.027 0.210∗∗∗ −0.002 −0.019∗ 0.107 0.010 −0.013 0.194 −0.260 −0.053 −0.132 0.009 −1.197∗∗∗ YES YES YES 0.132 10,315 (0.203) (0.048) (0.158) (0.097) (0.002) (0.065) (0.085) (0.224) (0.052) (0.071) (0.049) (0.138) (0.001) (0.053) (0.058) (0.058) (0.065) (0.010) (0.010) (0.112) (0.086) (0.112) (0.122) (0.195) (0.192) (0.196) (0.209) (0.424) Column 1 reports results from the augmented Mincer regression. Column 2 reports results from the augmented Mincer regression that controls for selection in the decision to work in Germany using a Heckman selection correction. Column 3 reports the corresponding selection equation, which predicts working in Germany with the number of ERASMUS places normalized by the cohort size in a graduate’s university department. Signiﬁcance levels: ***p < 0.01, **p < 0.05, *p < 0.1. migrants are 10.56 compared to 10.61 for nonmigrants. The differences between the CDFs are substantial and in the same order of magnitude as standard estimates for the returns to an additional year of education in the United States (Card, 1999). Inequality varies across potential destination countries. We use this variation to analyze selection to countries with more extreme levels of (in)equality by splitting more and less equal countries into two groups each. Thus, we now compare ﬁve types of destinations: very unequal, somewhat unequal, home, somewhat equal, and very equal countries. We classify the three countries with the most unequal earnings distribu- tions as very unequal and the three countries with the most equal distributions as very equal. Results are shown in the right panel of ﬁgure 2a. Very unequal countries receive the most positively selected migrants, somewhat unequal coun- tries receive somewhat positively selected migrants, some- what equal countries receive slightly negatively selected migrants, and very equal countries receive strongly nega- tively selected migrants. The CDFs are somewhat noisier than in the previous graphs because sample sizes of migrants are relatively small, especially for equal countries. Nonethe- less, the selection pattern follows the theoretical predictions for the ﬁve groups.19 19 In online appendix ﬁgure A.6, we show results where we classify the four, instead of three, most (un)equal countries as most (un)equal. The results are very similar. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 THE SELECTION OF HIGH-SKILLED EMIGRANTS 783 Figure 2.—Predicted Earnings of Migrants and Nonmigrants (a) CDF for Three and Five Groups of Countries (b) CDF for Three and Five Groups of Countries—Earnings Prediction Corrected for Selection l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 (a) CDFs of predicted earnings that are based on returns reported in column 1 of table 2 for three groups: migrants to more equal countries, nonmigrants, and migrants to less equal countries (left panel) and for ﬁve groups: migrants to very equal countries, migrants to somewhat equal countries, nonmigrants, migrants to somewhat unequal countries, and migrants to very unequal countries (right panel). (b) CDFs of predicted earnings that are based on selection-corrected returns reported in column 2 of table 2 for the same groups of countries. Table 3 reports stochastic dominance tests. Online appendix ﬁgure A.2 shows kernel-smoothed versions of the CDFs. B. Controlling for Selection in the Augmented Mincer Regression As our previous analysis has shown, migrants are sys- tematically selected from the home population. Unless this selection is fully accounted for by the observables, the selec- tion could potentially bias the coefﬁcients of the augmented Mincer regression and thus our measure of predicted earn- ings. We use a Heckman selection procedure to control for this potential selection by estimating a selection equa- tion that predicts whether a graduate works in Germany or migrates abroad. We use the introduction and expansion of the ERASMUS student exchange program as an instrumental variable that predicts whether graduates work in Germany. ERASMUS allows students to study abroad in a European country for one or two semesters before they continue their studies in their home country. It was introduced in 1987 and has increased massively since then. In Germany, about 4,925 students participated in ERASMUS in 1990 (the year when the typical graduate of the 1993 cohort had studied abroad); participation rose to 18,482 in 2002 (the year when the typ- ical graduate of the 2005 cohort had studied abroad). The program was introduced at different times and expanded at varying rates, depending on the university and department. Parey and Waldinger (2011) show that the introduction and expansion of ERASMUS signiﬁcantly increased the prob- ability of graduates moving abroad after completing their studies in Germany. The ERASMUS instrument success- fully controls for selection in the Mincer regression if the number of ERASMUS places in a student’s university can be excluded from the Mincer regression. Crucially, we do not use the actual decision to study abroad, but the avail- ability of department-level ERASMUS scholarship places, which predict studying abroad and working abroad later on, to instrument for working in Germany. 784 THE REVIEW OF ECONOMICS AND STATISTICS In our data, the median graduate enrolled in university in 1991–92 and thus before the widespread availability of the Internet. Before the introduction of the Internet, information on the number of ERASMUS places was very difﬁcult to obtain. Even today, few department websites report the exact number of ERASMUS places. It is therefore unlikely that students sorted into certain departments to beneﬁt from more ERASMUS places. To further limit the possiblity of student sorting, we assign the number of ERASMUS places for the subject×university combination where a student ﬁrst enrolls in university. Any potential sorting after the ﬁrst enrollment will therefore not affect the exogeneity of the ERASMUS instrument. Students in certain subjects are systematically more likely to study abroad, and to work abroad later, than students in other subjects. We control for any such subject-speciﬁc differences by including 24 subject ﬁxed effects in the regres- sions. A related concern may be that better universities offer more ERASMUS places and also facilitate working abroad. We control for these university-speciﬁc differences by including a full set of university ﬁxed effects in the regres- sions. We also control for broader trends of studying and working abroad by controlling for cohort ﬁxed effects. Parey and Waldinger (2011) further discuss the exclusion restriction of the ERASMUS instrument. They show that the expansion of ERASMUS in a department is not correlated with a wider push to increase the international outlook of stu- dents. They also show that the probability of studying abroad is completely ﬂat before the introduction of ERASMUS and increases only once ERASMUS has been introduced, sug- gesting that pretrends are not affecting the validity of the ERASMUS instrument. Column 3 of table 2 shows the ﬁrst-stage estimates where we regress whether individuals work in Germany on a measure of ERASMUS scholarship places (normalized by the number of students) in a graduate’s university depart- ment. The availability of ERASMUS signiﬁcantly lowers the probability of working in Germany. Column 2 in table 2 shows that controlling for selection in the Mincer regression has only a small effect on the esti- mated coefﬁcients. In addition to the rich set of observables, this also reﬂects that the share of graduates not migrating (and thus observed in our Mincer regression) is very high and that selection of migrants occurs at both the top and the bottom of the distribution. The coefﬁcient on the Mills ratio is therefore quantitatively small and insigniﬁcant. The resulting CDFs of earnings potential by migration status are presented in ﬁgure 2b. They conﬁrm that migrants to less equal destinations are positively selected, while migrants to more equal destinations are negatively selected. C. Tests for Stochastic Dominance We investigate the statistical signiﬁcance of the substan- tial differences between the CDFs with tests for ﬁrst-order stochastic dominance. As we estimate the Mincer earnings equation in the ﬁrst step of our analysis and construct pre- dicted earnings based on the Mincer regression, we need to account for this additional source of uncertainty when we compute p-values. We therefore apply the bootstrap pro- cedure for stochastic dominance tests developed in Barrett and Donald (2003) and described in further detail in online appendix A.2. We also report p-values from conventional Kolmogorov-Smirnov tests, which do not account for the uncertainty associated with the estimation of parameters in the Mincer regression. The corresponding test results are shown in table 3. The top row of panel A indicates that we can reject the null hypothesis that the more-equal-CDF dominates the CDF of nonmigrants (“Home”) at the 1% level of signiﬁcance (columns 1–3). Similarly, the second row indicates that we can reject that the CDF of nonmigrants dominates the more-unequal-CDF at the 10% level. We also reject that the more-equal-CDF dominates the less-equal-CDF at the 1% level. We even reject these hypotheses when we use the Heckman selection-corrected estimates, as reported in columns 4 to 6. The graphical analysis presented in ﬁgure 2 suggests even more pronounced differences in the CDFs when we limit the comparison to very equal and very unequal countries, respectively. Table 3B indeed shows that the test statistic for the comparison of these more extreme destinations increases substantially. Because the relevant samples become smaller for destinations with more extreme levels of inequality, the p-values do not decrease in all cases. Nonetheless, the test of stochastic dominance now rejects at the 5% level for all three comparisons. Table 3B also reports tests for selec- tion between more similar destinations. The test statistic always has the predicted sign, suggesting that selection fol- lows the basic Roy/Borjas model, even for more similar destinations. As expected, selection patterns to the more sim- ilar destinations are often not statistically signiﬁcant because inequality differences in more similar destinations are much lower and because some country groups attract relatively few graduates. We also test the reverse set of hypothe- ses and cannot reject them. The corresponding p-values are above 0.74 and in most cases above 0.95 (online appendix table A.5). D. Selection of Migrants by Country Our data also allow us to investigate the selection of migrants to each of the nineteen destinations in our sample and thereby go beyond the three or ﬁve groups of countries presented in the previous section. We compute average pre- dicted earnings of migrants to each country and correlate them with the 75/25 ratio (ﬁgure 3). Circle sizes indicate the number of migrants in each country. Apart from a few out- liers, migrants to more equal countries have lower predicted earnings than migrants to less equal countries. We estimate a weighted country-level OLS regression and show the corresponding prediction in ﬁgure 3. In particular, l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 THE SELECTION OF HIGH-SKILLED EMIGRANTS 785 Table 3.—Stochastic Dominance Tests OLS p-Value Heckman Selection Correction p-Value Test Statistic (1) Kolmogorov- Smirnov (2) Barrett- Donald (3) Test Statistic (4) Kolmogorov- Smirnov (5) A. Selection to More Equal and Less Equal Destinations 0.187 0.061 0.220 0.001∗∗∗ 0.031∗∗ 0.000∗∗∗ 0.006∗∗∗ 0.098∗ 0.001∗∗∗ 0.182 0.071 0.218 0.002∗∗∗ 0.009∗∗∗ 0.000∗∗∗ B. Selection to Very Equal, Somewhat Equal, Somewhat Unequal, and Very Unequal Destinations Equal versus home Home versus unequal Equal versus unequal Stochastic dominance tests for very equal and very unequal destinations Very equal versus home Home versus very unequal Very equal versus very unequal 0.258 0.144 0.301 Stochastic dominance tests for more similar destinations 0.179 0.147 0.057 0.133 Very equal versus somewhat equal Somewhat equal versus home Home versus somewhat unequal Somewhat unequal versus very unequal 0.007∗∗∗ 0.004∗∗∗ 0.005∗∗∗ 0.231 0.083∗ 0.109 0.033∗∗ 0.018∗∗ 0.017∗∗ 0.008∗∗∗ 0.379 0.177 0.235 0.101 0.249 0.162 0.301 0.196 0.142 0.065 0.136 0.009∗∗∗ 0.001∗∗∗ 0.005∗∗∗ 0.171 0.099∗ 0.055∗ 0.027∗∗ Barrett- Donald (6) 0.022∗∗ 0.083∗ 0.004∗∗∗ 0.041∗∗ 0.014∗∗ 0.012∗∗ 0.310 0.171 0.173 0.096∗ The table reports one-sided Kolmogorov-Smirnov test statistics and Kolmogorov-Smirnov and Barrett and Donald p-values for CDFs in ﬁgure 2. Barrett and Donald p-values are bootstrapped, following equation (11) in Barrett and Donald (2003, p. 82). In the top row (equal versus home), we test the null hypothesis that the CDF of migrants to more equal destinations stochastically dominates the CDF of nonmigrants, and similarly for other rows. The bootstrap is based on 4,999 replications. See text for details. Signiﬁcance levels: ***p < 0.01, **p < 0.05, *p < 0.1. Figure 3.—Predicted Earnings and Inequality Across Destinations V. Robustness FI AU LU CH IE DE NL . 7 0 1 . 6 0 1 DK US AT ES PL FR IT GB JP CA i s g n n r a e d e t c d e r p i . g v A . 5 0 1 BE SE NO . 4 0 1 1.3 1.5 1.7 75/25 ratio 1.9 The ﬁgure shows average predicted earnings for migrants to each country and the corresponding 75:25 inequality ratio. Circle sizes are proportional to the number of migrants in each destination. The regression line reported in the ﬁgure is estimated in a weighted regression with weights equal to the number of migrants in each country. The slope coefﬁcient is equal to 0.153 with a standard error of 0.081. An unweighted regression has a slope equal to 0.103 with a standard error of 0.101. For country labels, see data appendix table B.2. we regress average predicted earnings ( in each country c: ¯ˆθ0c) on the 75/25 ratio ¯ˆθ0c = γ0 + γ175/25 ratioc + εc. (8) The estimated regression line (ˆγ1) has a slope of 0.153 with a standard error of 0.081 (table 4, column 1, signiﬁcant at the 10% level). This estimate indicates that migrants to destinations with a 75:25 ratio that is higher by 0.4 (the difference between Germany and the United States) have predicted earnings that are 6.1 log points higher. A. Controlling for Possible Confounding Factors The selection pattern described in the previous section is consistent with the theoretical predictions of the Roy/Borjas model. Earnings inequality, however, is not the only factor that differs between home and destination countries. Coun- tries may also differ along other dimensions that could be correlated with migrant selection. We ﬁrst analyze whether confounding factors (Fc) are driving our selection results by controlling for them in the cross-country regression (table 4): ¯ˆθ0c = γ0 + γ175/25 ratioc + γ2 Fc + εc. (9) The Roy/Borjas model predicts that mean earnings should affect the number of migrants to each country but not the direction of selection. Nonetheless, differences in mean earn- ings will affect migration choices and may be correlated with differences in the 75:25 ratios. In our ﬁrst robust- ness check, we therefore control for average log earnings in each country. In this speciﬁcation, the coefﬁcient on the 75:25 ratio increases slightly to 0.180, suggesting an even stronger relationship between inequality and migrant selection (column, 2, signiﬁcant at the 1% level). Migra- tion decisions, especially those of lower-skilled migrants (within the high-skilled population), may also be affected by expected unemployment spells. When we control for unemployment rates of tertiary-educated individuals, the coefﬁcient on the 75:25 ratio is equal to 0.174 (column 3, signiﬁcant at the 5% level). Migration decisions may also be affected by differences in child care provision. When we control for public expenditures on family beneﬁts, the coefﬁcient on the 75:25 ratio is equal to 0.110 (column 4, l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 786 THE REVIEW OF ECONOMICS AND STATISTICS 75/25 ratio Mean earnings Tertiary-educated unemployment share Family expenditure Life satisfaction Constant R2 Observations Table 4.—Cross-Country Regressions (1) 0.153∗ (0.081) (2) 0.180∗∗∗ (0.058) 0.110∗∗∗ (0.033) (3) 0.174∗∗ (0.077) −0.007 (0.007) 10.366∗∗∗ (0.144) 0.183 19 9.161∗∗∗ (0.413) 0.475 19 10.353∗∗∗ (0.138) 0.204 19 (4) 0.110∗ (0.057) (5) 0.247∗∗∗ (0.081) −0.023∗ (0.011) 10.484∗∗∗ (0.104) 0.317 19 0.050∗ (0.024) 9.849∗∗∗ (0.276) 0.282 19 (6) 0.147∗ (0.071) 0.102∗ (0.056) 0.005 (0.009) −0.012 (0.011) 0.003 (0.036) 9.281∗∗∗ (0.531) 0.514 19 The table reports weighted regressions of average predicted earnings of migrants in each country on the corresponding 75/25 ratio and potential confounders. See data appendix B.2 for details on data sources and data appendix table B.2 for country data. Signiﬁcance levels: ***p < 0.01, **p < 0.05, *p < 0.1. signiﬁcant at the 10% level). Migration decisions may also be affected by expectations about general well-being. When we control for a measure of life satisfaction, the coefﬁcient on the 75:25 ratio is 0.247, conﬁrming a strong relationship between earnings inequality and migrant selection (column 5, signiﬁcant at the 1% level). When we control for all poten- tial confounders at the same time, the coefﬁcient on the 75:25 ratio is 0.147, with a p-value of 0.061 (column 6). The previous checks conﬁrm a robust effect of earnings inequality on mean selection levels. In additional tests, we investigate how potential confounders affect selection across the whole distribution of skills. For these tests, we ﬁrst replicate the CDFs from our main results using quantile regressions and then control for possible confounding factors using the quantile regression framework. We regress pre- dicted earnings of each individual i (ˆθ0i) on country group dummies separately for 100 centiles (τ = 0.01, . . . , 0.99) of the predicted earnings distribution: ˆθ0ic = δ0τ + δ1τVery Equalic + δ2τSomewhat Equalic + δ3τSomewhat Unequalic + (cid:3)icτ. + δ4τVery Unequalic (10) is much more equal Very Equalic takes a value of 1 if the individual works in a country that than Germany, Somewhat Equalic takes a value of 1 if the individual works in a country that is somewhat more equal, and so on. The con- stant represents predicted earnings for individuals who work in Germany. Online appendix ﬁgure A.7a shows the quantile regression equivalents of the CDFs in our main results. We then control for potential confounding factors in the quantile regressions by adding country-level controls: ˆθ0ic = δ0τ + δ1τVery Equalic + δ3τSomewhat Unequalic + δ5τFc + (cid:3)icτ. + δ2τSomewhat Equalic + δ4τVery Unequalic (11) From the estimated coefﬁcients, we predict CDFs for each group holding constant the value of the added covariate at the German level. Panels b to f of online appendix ﬁgure A.7 show CDFs that are adjusted for the same confounding factors that we have analyzed in the cross-country regres- sion (table 4). The selection pattern to locations with more extreme levels of (in)equality is robust to controlling for potentially confounding factors. The selection pattern to locations with less extreme levels of (in)equality remains broadly consistent with the predictions of the model (see online appendix table A.7 for stochastic dominance tests). If we control for mean earnings, the CDF for somewhat unequal countries sometimes moves to the left of the CDF for graduates at home. However, the stochastic dominance tests indicate that the two CDFs are not signiﬁcantly different. B. Sensitivity of Results to Alternative Inequality Measures In appendix section A.4, we investigate the sensitivity of our main results to using alternative measures of inequal- ity. We show that the results are very similar for a range of inequality measures that we calculate for the overall popu- lation, including the 90:50 ratio, the 75:25 ratio, the 90:10 inequality ratio, the Gini coefﬁcient, and the Theil index. C. Selection to Europe and to Austria/Switzerland Additionally, we investigate selection to European coun- tries only. German citizens who migrate to these countries face virtually no migration barriers, such as visa require- ments. Germans can settle freely in any country of the European Union and in other European countries, such as Switzerland, Lichtenstein, and Norway.20 Furthermore, migration costs to these countries are relatively low because distances within Europe are small, and travel costs are low. We plot CDFs of predicted earnings of migrants to less equal countries, migrants to more equal countries, and non- migrants (ﬁgure 4a). As for the full sample, migrants to more equal European countries are negatively selected, and migrants to less equal European destinations are positively 20 Graduates from the 1993 to 2001 cohorts who migrated to Poland or Switzerland had minor restrictions to settle in these countries. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 THE SELECTION OF HIGH-SKILLED EMIGRANTS 787 Figure 4.—Predicted Earnings of Migrants to Europe and Austria/Switzerland (a) Europe (b) Austria and Switzerland 1 5 7 . 0 5 . 0 5 2 . 0 0 10 10.2 1 5 7 . 0 5 . 0 5 2 . 0 0 Equal (N=89) Home (N=10,510) Unequal (N=386) 11 11.2 10 10.2 Home (N=10,510) Austria and Switzerland (N=194) 10.4 10.6 Log predicted earnings 10.8 11 11.2 10.4 10.6 Log predicted earnings 10.8 The ﬁgure shows CDFs of predicted earnings (prediction based on selection-corrected returns reported in column 2 of table 2) for migrants to Europe—EU countries (2005), Norway, and Switzerland—and nonmigrants in panel a; and to Austria or Switzerland and nonmigrants in panel b. Online appendix table A.9 (panels A and B) reports stochastic dominance tests. selected, compared to nonmigrants.21 These results suggest that differential migration costs are not driving our main results. We also investigate migrant selection to Austria and Switzerland only. These two countries are very similar to Germany along many dimensions that may affect migration choices. The countries have similar education systems with very similar university graduation rates (OECD, 2013a). The countries also have similar unemployment beneﬁts as mea- sured by replacement rates that ranged between 29% and 33% of gross incomes in 2005 (OECD, 2015). The three countries also share a similar culture. Finally, Austria is German speaking, and in Switzerland, 64% of the popu- lation is German speaking, and more than 90% of German migrants settle in predominantly German-speaking regions of Switzerland (Bundesamt für Statistik, 2010, 2013). While the three countries are similar along many dimensions, they differ in earnings inequality. Both Austria and Switzerland are less equal than Germany. The CDF of predicted earnings of migrants to Austria and Switzerland lies to the right of the nonmigrant CDF (ﬁgure 4b).22 These results indicate that migrants to Austria and Switzerland are positively selected compared to nonmigrants, as predicted by the Roy/Borjas model. VI. Further Results A. Decomposing Migrant Selection Predicted earnings can be considered a summary measure of different skills. To understand the characteristics that 21 We reject that the CDF of migrants to more equal countries dominates the home CDF at the 5% level (online appendix table A.9). As Europe contains few countries with very high inequality, we no longer reject that the home CDF dominates the CDF of migrants to less equal countries ( p-value of 0.19). 22 The test that the home CDF dominates the Austria/Switzerland CDF is rejected at the 10% level (online appendix table A.9, panel B). underlie the observed selection, we use a Blinder-Oaxaca procedure, decomposing the overall difference in predicted earnings into the contribution of each characteristic. For expositional purposes we group characteristics into thirteen categories. The positive selection of migrants to less equal coun- tries mostly reﬂects their university career (ﬁgure 5, panel a1). They have better grades and attend better universities than nonmigrants. The negative selection of migrants to more equal countries reﬂects their university subject, uni- versity quality, and gender (panel a2). They study subjects with lower returns in the labor market, enroll at universities with less favorable labor market prospects, and are more often female. Interestingly, migrants to more equal coun- tries have better grades at university, despite being negatively selected overall. This is consistent with ﬁndings suggesting that migrants are positively selected when skill is measured in terms of education. Decomposition results that use coef- ﬁcients from the selection-corrected Mincer regression are shown in online appendix ﬁgure A.9. Columns 1 to 4 of table 5 summarize how the covariates of the decomposition line up with the overall prediction. For most characteristics, the table shows no signiﬁcant devi- ations from the model predictions. However, there are a number of interesting differences in the relevance of indi- vidual characteristics between less equal and more equal countries. For less equal countries, the pattern of selection in terms of apprenticeship training is not in line with our baseline prediction, and for more equal countries, univer- sity grades show signiﬁcant positive selection among the migrants. Although we do not have the detailed data to investigate these instances, they may reﬂect heterogeneity in returns to characteristics across countries or a correlation of these characteristics with the willingness to move in a way not captured by the model. For example, it is plausible that (former) apprentices may realize a higher return to their l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 788 THE REVIEW OF ECONOMICS AND STATISTICS Figure 5.—Decomposition of Predicted Earnings (a) Migrants to Less Equal and More Equal Countries (a1) Migrants to less equal countries Total University grade University subject University FE Bachelor Further studies School grade Apprenticeship Previous mobility Age/Experience Gender Partner/Children Parental background Graduate cohort Total University grade University subject University FE Bachelor Further studies School grade Apprenticeship Previous mobility Age/Experience Gender Partner/Children Parental background Graduate cohort Total University grade University subject University FE Bachelor Further studies School grade Apprenticeship Previous mobility Age/Experience Gender Partner/Children Parental background Graduate cohort Total University grade University subject University FE Bachelor Further studies School grade Apprenticeship Previous mobility Age/Experience Gender Partner/Children Parental background Graduate cohort (a2) Migrants to more equal countries −0.1 −0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 (b) Migrants to Very Unequal and Very Equal Countries (b1) Migrants to very unequal countries (b2) Migrants to very equal countries (p=0.022)** (p=0.000)*** (p=0.893) (p=0.044)** (p=0.279) (p=0.486) (p=0.831) (p=0.002)*** (p=0.223) (p=0.025)** (p=0.959) (p=0.718) (p=0.001)*** (p=0.210) (p=0.017)** (p=0.000)*** (p=0.006)*** (p=0.460) (p=0.052)* (p=0.441) (p=0.948) (p=0.005)*** (p=0.362) (p=0.137) (p=0.004)*** (p=0.771) (p=0.080)* (p=0.983) (p=0.006)*** (p=0.000)*** (p=0.560) (p=0.188) (p=0.518) (p=0.856) (p=0.735) (p=0.021)** (p=0.257) (p=0.039)** (p=0.109) (p=0.729) (p=0.025)** (p=0.543) (p=0.006)*** (p=0.000)*** (p=0.000)*** (p=0.611) (p=0.115) (p=0.607) (p=0.857) (p=0.049)** (p=0.583) (p=0.244) (p=0.047)** (p=0.115) (p=0.914) (p=0.312) −0.18−0.16−0.14−0.12 −0.1 −0.08−0.06−0.04−0.02 0 0.02 0.04 0.06 0.08 0.10 l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 Panel a1 decomposes the mean difference in predicted earnings between migrants to less equal countries and nonmigrants. The top bar (black) measures the total difference in predicted earnings. The other bars decompose the total difference into the contributions of groups of characteristics (e.g., university grades). More speciﬁcally, the size of the gray bars in panel a1 is obtained by multiplying estimated returns (ˆβHome ) for nonmigrants from column 1 in table 2 (where k indexes a group of characteristics, e.g., all parental background variables or all university ﬁxed effects) with average characteristics of migrants to less equal countries (xLess equal . Panel a2 presents the equivalent decomposition between migrants to more equal destinations and nonmigrants. Panel b presents corresponding results to very unequal and very equal countries. Diamonds indicate 90% conﬁdence intervals. Conﬁdence intervals and p-values are obtained from bootstrapped standard errors (based on 4,999 replications). Signiﬁcance levels: ***p < 0.01, **p < 0.05, *p < 0.1. ) and average characteristics of nonmigrants (xHome ), and then subtracting ˆβHome from ˆβHome xLess equal k xHome k k k k k k THE SELECTION OF HIGH-SKILLED EMIGRANTS 789 Table 5.—Summary of Decomposition Results (1) Less Equal Destinations (2) More Equal Destinations (3) Very Unequal Destinations (4) Very Equal Destinations (5) United States Total University grade University subject University ﬁxed effect Bachelor Further studies School grade Apprenticeship Previous mobility Age/experience Gender Partner/children Parental background Graduate cohort Consistent Consistent – Consistent – – – Reject – Consistent – – Consistent – Consistent Reject Consistent – – – – Consistent – – Consistent – – – Consistent Consistent – – – – – Reject – Consistent – – Consistent – Consistent Reject Consistent – – – – Consistent – – Consistent – – – Consistent Consistent – – Consistent – – Reject – Consistent Consistent – Consistent – The table summarizes results from the Blinder-Oaxaca decomposition shown in ﬁgures 5 and 6a. “Consistent” indicates that the selection along the corresponding characteristic is signiﬁcantly different from 0 at a 5% level of signiﬁcance and in line with the model prediction. “Reject” indicates that the selection along the corresponding characteristic is signiﬁcantly different from 0 at a 5% level of signiﬁcance, in the direction not in line with the model prediction. training in their home labor market and are therefore more attached to their home labor market. It is important to keep in mind that ﬁgure 5 shows the results of a statistical decomposition and that the char- acteristics may be correlated with each other. Predicted earnings provide a natural way of combining the individual characteristics in a summary measure. B. Migration to the United States Migrants to the United States compared to nonmigrants in Germany. In the ﬁnal section, we investigate migrant selec- tion to the United States. The United States is an important destination for university graduates from Germany. In our sample, more than 13% of graduates who go abroad move to the United States; only Switzerland attracts more graduates from Germany. Because U.S. inequality is highest among the major destinations of German university graduates, we expect that German university graduates who migrate to the United States are particularly positively selected. The CDF of migrants to the United States always lies to the right of the nonmigrant CDF (left panel of ﬁgure 6a). The difference between the CDFs of U.S. migrants and nonmigrants is more pronounced than the difference between the CDFs of all migrants to less equal countries and nonmigrants. This highlights the particularly positive selec- tion of migrants to the United States. The test of stochastic dominance is rejected at the 5% level (see online appendix table A.9, panel C). As above, we decompose the difference in predicted earn- ings between migrants to the United States and nonmigrants. U.S. migrants are positively selected according to almost all characteristics, in particular characteristics that relate to the university career and gender. Migrants to the United States study subjects with especially high returns (see the third bar from the top in the right panel of ﬁgure 6a). They are par- ticularly concentrated in STEM ﬁelds. In our sample, about 17.2% of migrants to the United States hold a degree in physics (but only 3.9% of nonmigrants), 9.2% hold a degree in biology (nonmigrants: 2.3%), and 8.1% hold a degree in chemistry (nonmigrants: 3.0%). Furthermore, migrants to the United States are also more likely to hold degrees in com- puter science, economics and management, geography, and engineering, and they are less likely to hold degrees in law, languages, medicine, architecture, and education. Migrants to the United States also obtain higher grades in university than nonmigrants and study in universities where graduates have higher predicted earnings. The decomposition indicates that the United States attracts high-skilled migrants from Germany who have studied in better universities, received higher grades, and are concen- trated in high-paying STEM ﬁelds. Thus, migrants to the United States are precisely those who are considered to be important for innovation and technological progress. Migrants from Germany compared to U.S. natives in the ACS. Finally, we investigate how high-skilled migrants from Germany fare in the U.S. labor market by comparing earnings potential of high-skilled migrants from Germany to high-skilled natives in the United States. For this test, we use data from the American Community Survey (ACS) and identify high-skilled migrants from Germany as individuals born in Germany to non-U.S. parents, migrated to the United States between 1996 and 2010, and were at least 25 years old at the time of migration. These restrictions ensure that our sample of Germans in the United States is as similar as possible to the sample of graduate emigrants from Germany whom we study in our main results. To focus our analysis on the highly skilled, we limit the sample to individuals with a bachelor’s degree or higher, who worked for 50 to 52 weeks per year in full-time jobs, and are 30 to 45 years old (see data appendix B.3 for further details on the ACS data).23 23 Results are similar if we restrict the ACS sample to graduates in more academically oriented subjects to further increase the comparability with graduates from traditional universities in Germany. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 790 THE REVIEW OF ECONOMICS AND STATISTICS Figure 6.—Predicted Earnings of Migrants to the United States (a) Based on Returns German Sample (DZHW Data) 1 5 7 . 0 5 . 0 5 2 . 0 0 1 5 7 . 0 5 . 0 5 2 . 0 0 Total University grade University subject University FE Bachelor Further studies School grade Apprenticeship Previous mobility Age/Experience Gender Home (N=10,510) US (N=87) 11 11.2 Partner/Children Parental background Graduate cohort (p=0.003)*** (p=0.000)*** (p=0.127) (p=0.492) (p=0.000)*** (p=0.857) (p=0.705) (p=0.030)** (p=0.287) (p=0.027)** (p=0.030)** (p=0.456) (p=0.018)** (p=0.475) −0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 10 10.2 10.4 10.6 Log predicted earnings 10.8 (b) Based on Returns U.S. Sample (ACS Data) Total University degree University subject Age Gender Partner/Children Graduate cohort (p=0.000)*** (p=0.000)*** (p=0.000)*** (p=0.000)*** (p=0.000)*** (p=0.169) (p=0.924) 10 10.5 11 11.5 12 US natives (N=289,538) Germans (N=565) Log predicted earnings 0 0.05 0.1 0.15 0.2 0.25 (a) The left panel shows CDFs of predicted earnings (prediction based on selection-corrected returns reported in column 2 of table 2) for migrants to the United States and for nonmigrants. Online appendix table A.9 (panel C) reports stochastic dominance tests. The right panel decomposes the mean difference in predicted earnings between migrants to the United States and nonmigrants. The top bar (black) measures the total difference in predicted earnings. The other bars decompose the total difference into the contributions of groups of characteristics (e.g., university grades). (b) The left panel shows CDFs of predicted earnings in the United States. Prediction based on coefﬁcients of the Mincer regression reported in online appendix table A.10 (column 1) using American Community Survey (ACS) data on U.S. natives. The right panel shows a decomposition of predicted earnings that decomposes the mean difference in predicted earnings between German migrants to the United States and U.S. natives. Diamonds indicate 90% conﬁdence intervals. Conﬁdence intervals and p-values are obtained from bootstrapped standard errors (based on 4,999 replications). Signiﬁcance levels: ***p < 0.01, **p < 0.05, *p < 0.1. We then compare predicted earnings of migrants from Germany to earnings of U.S. natives. We evaluate the skills of German immigrants to the United States using pre- dicted earnings that we construct from returns to skills for U.S. natives (see online appendix table A.10, column 1 for returns to skills for U.S. natives). In terms of the Roy/Borjas model, this test compares the distribution of ˆθ1 of German migrants in the United States to U.S. natives, while our pre- vious results compared distributions of ˆθ0 of migrants and nonmigrants.24 Indeed, our results show that compared to high-skilled U.S. natives, recent migrants from Germany have far higher predicted earnings in the U.S. labor mar- ket. The CDF of predicted earnings of German immigrants lies to the right of the native CDF along the whole earn- ings distribution (left panel of ﬁgure 6b). At the median, log 24 Parallel to equation (5), tion in terms of earnings potential σθ E(θ1|Migrate=1) = μ1 + − ρθ (cid:7) (cid:6) 1 σθ σθ 0 σθ 0 σv 1 φ(z) 1−Φ(z) . the corresponding equation for selec- in the destination country is predicted earnings of migrants from Germany are 11.383, while log earnings of natives are 11.129. At the 25th and 75th percentiles, migrants from Germany have predicted earnings of 11.193 and 11.554, while natives have predicted earnings of 10.937 and 11.334. A back-of-the-envelope cal- culation suggests that the stronger selection in terms of θ1 (relative to our earlier results in terms of θ0) can be recon- ciled with our theoretical prediction, both qualitatively and quantitatively.25 (cid:5) (cid:5) (cid:4) − ρθ σθ1 /σθ0 ρθ − σθ0 /σθ1 25 The selection in terms of θ1 should be stronger than selection in terms (cid:4) of θ0 by a factor of / . Between the United States and Germany, the ratio σθ0 /σθ1 is about 0.8 in our data. While ρθ is unknown, the positive selection in terms of θ0 indicates that ρθ is above 0.8, from equation (5). Suppose ρθ = 0.9; then the factor results in a value of 3.7, which is broadly similar but slightly larger than the observed difference in selection. Because the factor decreases in ρθ, it is straightforward to reconcile the observed difference in selection with a value of ρθ somewhat above 0.9. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 THE SELECTION OF HIGH-SKILLED EMIGRANTS 791 Overall, these results indicate that high-skilled individuals who migrate from Germany to the United States are not only positively selected compared to Germans who do not migrate but also compared to nonmigrants in the United States. To investigate the contribution of different characteristics, we also decompose the difference in predicted earnings between German migrants to the United States and U.S. natives. Because the ACS data are less detailed than our graduate survey data, the decomposition involves fewer characteris- tics. Compared to U.S. natives, German migrants have more advanced degrees (such as professional degrees or Ph.D.s) and graduated in subjects (in particular, STEM subjects) that typically lead to higher-paid employment. German migrants are also less likely to be female than U.S. natives. Over- all, the positive selection compared to U.S. natives reﬂects similar characteristics as the ones we ﬁnd for the positive selection compared to German nonmigrants. VII. Conclusion The seminal work of Borjas has emphasized how migrant selection is driven by inequality in home and destination countries: high-skilled individuals beneﬁt from the upside opportunities in less equal countries, and low-skilled indi- viduals beneﬁt from the insurance of a more compressed wage distribution in more equal countries. This insight has motivated various empirical tests of the Borjas model. In spite of the large differences in inequality across many home-destination country pairs, the empirical evidence is mixed. In this paper, we investigate selection within the group of high-skilled migrants in a setting where migration costs are particularly low. We use predicted wages to measure the skills of migrants and graduates who remain at home. Consistent with the predictions of the basic Roy/Borjas model, we ﬁnd that migrants to more equal countries, such as Denmark, are negatively selected compared to nonmi- grants. Migrants to less equal countries, such as the United States, are positively selected. In further results, we show that migrant selection follows the predictions of the Roy/Borjas model even within subgroups of either more or less equal countries. Our results are robust to controlling for potentially con- founding factors and using alternative measures of inequality in destination countries. We also demonstrate that the selec- tion pattern holds when we study migration within Europe, and migration to Austria and Switzerland, where barriers to migration are particularly low. When we decompose pre- dicted earnings into various skill components, we ﬁnd that selection patterns follow the model prediction for most, but not all, characteristics, suggesting that the choice of the skill measure can affect ﬁndings of migrant selection. Overall, our ﬁndings highlight the importance of the Roy/Borjas model for the selection of high-skilled migrants. REFERENCES Abramitzky, R., “The Effect of Redistribution on Migration: Evidence from the Israeli Kibbutz,” Journal of Public Economics 93 (2009), 498– 511. Abramitzky, R., L. P. Boustan, and K. Eriksson, “Europe’s Tired, Poor, Huddled Masses: Self-selection and Economic Outcomes in the Age of Mass Migration,” American Economic Review 102 (2012), 1832– 1856. Akcigit, U., S. Baslandze, and S. Stantcheva, “Taxation and the Inter- national Mobility of Inventors,” American Economic Review 106 (2016), 2930–2981. Barrett, G. F., and S. G. Donald, “Consistent Tests for Stochastic Domi- nance,” Econometrica 71 (2003), 71–104. Bartolucci, C., C. Villosio, and M. Wagner, “Who Migrates and Why? Evidence from Italian Administrative Data,” working paper (2014). Belot, M. V. K., and T. J. Hatton, “Immigrant Selection in the OECD,” Scandinavian Journal of Economics 114 (2012), 1105–1128. ——— Ständige Wohnbevölkerung nach Kanton und Hauptsprachen (2013), http://www.bfs.admin.ch/bfs/portal/de/index/themen/01/05 /blank/key/sprachen.html. Borjas, G. J., “Self-Selection and the Earnings of Immigrants,” American Economic Review 77 (1987), 531–533. ——— “Immigration and Self-Selection” (pp. 29–76), in J. M. Abowd and R. B. Freeman, eds., Immigration, Trade, and the Labor Market (Chicago: University of Chicago Press, 1991). ——— “Labor Outﬂows and Labor Inﬂows in Puerto Rico,” Journal of Human Capital 2 (2008), 32–68. Borjas, G. J., and S. G. Bronars, “Immigration and the Family,” Journal of Labor Economics 9 (1991), 123–148. Borjas, G. J., S. G. Bronars, and S. J. Trejo, “Self-Selection and Internal Migration in the United States,” Journal of Urban Economics 32 (1992), 159–185. Borjas, G. J., and K. B. Doran, “The Collapse of the Soviet Union and the Productivity of American Mathematicians,” Quarterly Journal of Economics 127 (2012), 1143–1203. Borjas, G. J., I. Kauppinen, and P. Poutvaara, “Self-Selection of Emi- grants: Theory and Evidence on Stochastic Dominance in Observ- able and Unobservable Characteristics,” NBER working paper 21649 (2015). Brücker, H., S. Bertoli, G. Facchini, A. M. Mayda, and G. Peri, “Under- standing Highly Skilled Migration in Developed Countries: The Upcoming Battle for Brains” (pp. 15–188), in T. Boeri, H. Brücker, F. Docquier, and H. Rapoport, eds., Brain Drain and Brain Gain: The Global Competition to Attract High-Skilled Migrants (New York: Oxford University Press, 2012). Bunesamt für Statistik, Ständige und Nichtständige Wohnbevölkerung nach Kanton, Geschlecht, Anwesenheitsbewilligung, Altersklasse und Staatsangehörigkeit (2010), http://www.bfs.admin.ch/bfs/portal/de /index/themen/01/07/blank/data/01.html. Card, D., “The Causal Effect of Education on Earnings” (pp. 1801– 1863), in O. C. Ashenfelter and D. Card, eds., Handbook of Labor Economics, vol. 3A (Amsterdam: Elsevier, North-Holland, 1999). Card, D., J. Heining, and P. Kline, “Workplace Heterogeneity and the Rise of West German Wage Inequality,”’ Quarterly Journal of Economics 128 (2013), 967–1015. Chambers, E. G., M. Foulon, H. Handﬁeld-Jones, S. M. Hankin, and E. G. Michaels, “The War for Talent,” McKinsey Quarterly 3 (1998), 44– 57. Chiquiar, D., and G. H. Hanson, “International Migration, Self-Selection, and the Distribution of Wages: Evidence from Mexico and the United States,” Journal of Political Economy 113 (2005), 239–281. Dahl, G., “Mobility and the Return to Education: Testing a Roy Model with Multiple Markets,” Econometrica 70 (2002), 2367–2420. DESTATIS, Bildungsstand der Bevölkerung. Federal Statistical Ofﬁce of Germany. Docquier, F., and H. Rapoport, “Globalization, Brain Drain, and Develop- ment,” Journal of Economic Literature 50 (2012), 681–730. Doran, K., A. Gelber, and A. Isen, “The Effects of High-Skilled Immigra- tion on Firms: Evidence from H-1B Visa Lotteries,” NBER working paper 20668 (20l4). Dustmann, C., I. Fadlon, and Y. Weiss, “Return Migration, Human Cap- ital Accumulation and the Brain Drain,” Journal of Development Economics 95 (2011), 58–67. l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3 792 THE REVIEW OF ECONOMICS AND STATISTICS Dustmann, C., J. Ludsteck, and U. Schönberg, “Revisiting the German Wage Structure,” Quarterly Journal of Economics 124 (2009), 843– 881. Feliciano, C., “Educational Selectivity in U.S. Immigration: How Do Immi- grants Compare to Those Left Behind?” Demography 42 (2005), 131–152. Fernández-Huertas Moraga, J., “New Evidence on Emigrant Selection,” this review 93 (2011), 72–96. Gould, E. D., and O. Moav, “Does High Inequality Attract High Skilled Immigrants?” Economic Journal 126 (2016), 1055–1091. Grogger, J., and G. H. Hanson, “Income Maximization and the Selec- tion and Sorting of International Migrants,” Journal of Development Economics 95 (2011), 42–57. Grotheer, M., S. Isleib, N. Netz, and K. Briedis, Hochqualiﬁziert und gefragt. Ergebnisse der zweiten HIS-HF Absolventenbefragung des Jahrgangs 2005 (Hanover: HIS GmbH 2012). Heckman, J. J., “Sample Selection Bias as a Speciﬁcation Error,” Econo- metrica 47 (1979), 153–161. Hunt, J., and M. Gauthier-Loiselle, “How Much Does Immigration Boost Innovation?” American Economic Journal: Macroeconomics 2 (2010), 31–56. Ibarraran, P., and D. Lubotsky, “Mexican Immigration and Self-Selection: New Evidence from the 2000 Mexican Census” (pp. 159–192), in G. J. Borjas, ed., Mexican Immigration to the United States (Chicago: University of Chicago Press, 2007). Junge, M., M. Munk, and P. Poutvaara, “International Migration of Couples,” CESifo working paper 4927 (2014). Kaestner, R., and O. Malamud, “Self-Selection and International Migration: New Evidence from Mexico,” this review 96 (2014), 78–91. Kerr, S. P., W. R. Kerr, and W. F. Lincoln, “Skilled Immigration and the Employment Structures of US Firms,” Journal of Labor Economics 33 (2015), S147–S186. Kerr, W. R., and W. F. Lincoln, “The Supply Side of Innovation: H-1B Visa. Reforms and U.S. Ethnic Invention,” Journal of Labor Economics 28 (2010), 473–508. Kleven, H. J., C. Landais, and E. Saez, “Taxation and International Migra- tion of Superstars: Evidence from the European Football Market,” American Economic Review 103 (2013), 1892–1924. Luxembourg Income Study (LIS), “Database (multiple countries) (2013), http://www.lisdatacenter.org. Moser, P., A. Voena, and F. Waldinger, “German-Jewish Émigrés and US Invention,” American Economic Review 104 (2014), 3222–3255. OECD, “OECD Education at a Glance 2013, OECD Indicators (2013a). ——— “OECD Tax Statistics, Taxing Wages” (2013b), http://dx.doi.org /l0.1787/ctpa-twg-data-en. ——— “Beneﬁts and Wages: Statistics: Gross Replacement Rates (with- out Accounting for Taxes), Uneven Years 1961 to 2011,” (2015), http://www.oecd.org/els/beneﬁts-and-wages-statistics.htm. Orrenius, P. M., and M. Zavodny, “Self-Selection among Undocumented Immigrants from Mexico,” Journal of Development Economics 78 (2005), 215–240. Parey, M., and F. Waldinger, “Studying Abroad and the Effect on Inter- national Labor Market Mobility,” Economic Journal 121 (2011), 194–222. Ramos, F. A., “Outmigration and Return Migration of Puerto Ricans” (pp. 49–66), in G. J. Borjas and R. B. Freeman, eds., Immigration and the Workforce: Economic Consquences, for the United States and Source Areas (Chicago: University of Chicago Press, 1992). Roy, A. D., “Some Thoughts on the Distribution of Earnings,” Oxford Economic Papers 3 (1951), 135–146. Ruggles, S., J. T. Alexander, K. Genadek, R. Goeken, M. B. Schroeder, and M. Sobeku, “Integrated Public Use Microdata Series: Version 5.0 [Machine-readable database] (2010). Stolz, Y., and J. Baten, “Brain Drain in the Age of Mass Migration: Does Relative Inequality Explain Migrant Selectivity?” Explorations in Economic History 49 (2012), 205–220. Widmaier, S., and J.-C. Dumont, “Are Recent Immigrants Different? A New Proﬁle of Immigrants in the OECD based on DIOC 2005/06,” OECD Social, Employment and Migration working papers 126 (2011). l D o w n o a d e d f r o m h t t p : / / d i r e c t . m i t . e d u / r e s t / l a r t i c e - p d f / / / / 9 9 5 7 7 6 1 9 7 4 8 0 9 / r e s t _ a _ 0 0 6 8 7 p d . f b y g u e s t t o n 0 8 S e p e m b e r 2 0 2 3
Descargar PDF