Content uploaded by Aleida Cobas-Valdés
Author content
All content in this area was uploaded by Aleida Cobas-Valdés on Apr 08, 2019
Content may be subject to copyright.
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
CUBAN MIGRATION TO THE UNITED STATES
AND THE EDUCATIONAL SELFSELECTION
PROBLEM
Aleida Cobas Valdés and Ana Fernández Sainz
University of the Basque Country (UPV/EHU), Spain
Abstract
The aim of this article is to study the educational self-selection problem of Cuban
migration to the US. For this analysis, we specify and estimate a binary logit model to
analyse the observable covariates that explain migration probability. The data used in
the study came from the 2010 Census of Population and Housing in the US and from
the 2002 Cuba Census of Population and Housing, both data set have been provided by
IPUMS (2010) and IPUMS International (2011). The results indicate that education, age
and occupational covariates explain migration probability. Moreover, there is a positive
educational self-selection problem, that is, those people with a higher education migrate.
The principal contribution of this article is to demonstrate how high-level education
increases the probability of Cubans emigrating. The positive educational self-selection
problem has significant negative consequences, for example, loss of human capital.
Keywords: migration, Cuba, self-selection, education, binary logit model
Introduction
Until the early twentieth century, Cuba was considered a country of immigrants.
Cuban people have been shaped by three major migration flows: European
(mainly Spanish), African and Chinese, the most important being the Spanish
power, involving around one and a half million people (Pérez de la Riva 2000).
In the second half of the nineteenth century, one-third of the Cuban population
was born outside the island.
From 1850 to 1899, 900,000 immigrants entered Cuba, primarily Spanish
immigrants, representing 90 per cent of European immigration: mainly men
working in the sugar and tobacco industry (Pérez de la Riva 2000). In 1899,
10.97 per cent of the Cuban population was born abroad, 81 per cent of which
IJCS6_1 41 06/06/2014 11:35
42 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
were male (Demographic Yearbook of Cuba 1961 [1965]). From 1902 to 1932,
1.25 million immigrants entered Cuba, of which 800,000 were Spanish.
After 1926, immigration declined until becoming insignificant in 1930. The
global crisis of 1929 and the subsequent collapse of the sugar industry in the
early years of the 1930s resulted in the loss of the immigration country status
that characterised Cuba up until then (Aja Díaz 2002). In 1953, the proportion
of people born abroad dropped to 3.95 per cent (Demographic Yearbook of
Cuba 2010).
Cuba’s external migration rate, defined as the ratio of the difference in the
number of immigrants and emigrants with respect to average population, per
1,000 population has been negative for several decades. In the last 30 years,
it reached its lowest level in 1980 and 1994. In 1980, the figure reached 14.6
per 1,000 and in 1994 4.4 per 1,000 (Demographic Yearbook of Cuba 2010),
coinciding with two major waves of migration from Cuba to the US, the first
known as the Puente Marítimo del Mariel and the second as the Crisis de
los Balseros.
US has been for Cuba, and for other Latin American countries, the main
destination of migration. The US Census for 2010 revealed that 50.5 million
people (16.36% of the entire population) in the US are Hispanics, and this
number rose from 35.3 million in 2000 to 50.5 million in 2010 (US Census
Bureau 2010). Of these Hispanics, 1.12 million were born in Cuba, representing
3 per cent of foreigners living in the US (Motel and Patten 2012).
Based on this data, the aim of this article is to analyse the characteristics,
mainly educational, of Cubans who have emigrated to the US and compare them
with those of Cubans who have remained in Cuba. In this way, we intend to
address the self-selection problem among Cuban emigrants to the US in terms
of educational levels and analyse the importance of educational levels on the
probability of Cuban migration.
The self-selection problem means that rational agents optimise their decision
to participate in different markets, work, education, migration, etc. (Hotz
2011), and therefore, the migrants choose markets that offer more attractive
expectations. Roy’s (1951) model was the first to address this problem, analysing
how individuals optimise their decision to belong to one group or another in a
given market, depending on their skills.
Self-selection not only depends on the unobservable characteristics of an
individual such as ability, motivation, relatives or friends in the US (Borjas 1987)
or access to financial resources (Chiquiar and Hanson 2005) but mainly depends
on observable characteristics such as education.
If there is a positive relationship between migration and education, that is,
more educated persons migrate, we could be talking about the existence of
IJCS6_1 42 06/06/2014 11:35
CUBAN MIGRATION TO THE US AND EDUCATIONAL SELFSELECTION 43
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
human capital flight. If the migration of highly skilled labour is permanent, this
process results in an increase in the growth potential of the receiving country
of migrants and may represent a loss to the country of origin (Albo and Ordaz
Díaz 2011).
The self-selection problem of Cuban migrants has not been addressed since
Borjas (1991) in the context of migration to the US from different countries,
including Cuba. The article is structured as follows: in the next section, we analyse
the methodology commonly used to treat self-selection; in the third section, the
data used are described; in the fourth section, we present the estimation results;
and in the fifth section, the major conclusions are listed.
Methodology
Consider individuals, who will always choose the alternative that is in their
best interests. The individual decision will be based on comparing the utility of
living in their origin country with the expected utility in the destination country,
including the disutility of moving to that destination (Sjaastad 1962).
Let Ue be the utility that reports Cuban migration to the US and Uno the
no-migration utility, so that:
(1)
The X vector consists of a set of individual, observable characteristics, such as
education, age, gender, professional category, etc.
The parameters vector b reflects the impact that covariate X has on the
individual utility, ee and eno are disturbances or error terms and are considered
independent of vector X, and it is assumed that they follow a logistic distribution.
Error terms, ee and eno may be related to each other with correlation coefficient r.
An individual will migrate if the utility from emigrating is higher than the
utility from not migrating, that is, if:
(2)
Taking account that the utility (to migrate or not migrate) is unobservable,
what we observe is the decision taken by the individual. We assume Y = 1 when
the individual selects the alternative to emigrate and Y = 0 when the individual
selects the alternative of not migrating, so that:
UX
UX
´
´
e
ee
no
no no
βε
βε
=+
=+
XX´´
eeno no
>
βε
βε
+> +
UU
eno
IJCS6_1 43 06/06/2014 11:35
44 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
where
(3)
is the self-selection condition.
Emigrating implies costs (C) associated with migration formalities (both in
the origin country in case of legal migration and in the destination country)
and transport. Most illegal immigrants have to advance money to enter the
destination country (Orrenius and Zavodny 2005), and therefore, they must
also consider costs related to the decision to emigrate.
Let the migration cost in ‘time-equivalent’ units (number of working hours
required to migrate). Borjas (1987, 1991) assumes that p is constant, whereby
implying that all individuals require the same number of working hours to cover
the migration costs.
Without loss of generality, this article assumes that p is constant. Borjas
(1991) demonstrated that considering p as a random variable does not lead to
particularly different results compared to those obtained when it is considered
to be constant.
Now for the probability that an individual migrates the vector conditional on
observable characteristics that are taken into account, will be raised as follows:
(4)
Pr Pr
Pr
Pr
βεβε
εε ββ
== >
=′+−
′−>
=−>− −′
ob YXob UU X
ob XX X
ob XX
[( 1) /] [( )]
[( 0) /]
[(( )( ))/]
eno
eeno no
enoeno
X/X[(( )( )) ]
enoeno
εε ββ
−>−− ′
Pr Pr
Pr
Pr
π
βεβεπ
εε πββ
β
Λβ
== −−>
=′+−
′−−>
=−>− −′
=′
=′
ob YXob UU X
ob XX X
ob XX
FX
X
[( 1) /] [( 0) /]
[( 0) /]
[(( )( ))/]
()
()
eno
eeno no
enoeno
IJCS6_1 44 06/06/2014 11:35
CUBAN MIGRATION TO THE US AND EDUCATIONAL SELFSELECTION 45
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
where L(b′X) is the Cumulative Logistics Distribution Function and b is the
parameter vector.
Emigration of those with a high level of education will be more likely if the
education has a greater performance in the US than in Cuba, that is, if the
difference between the parameter associated with educational level in the US
and the parameter associated with educational level in Cuba in the regression
equation is positive. Hence, the most qualified individuals find incentives
to migrate. This implies positive selection of individuals depending on their
educational level.
Emigration of Cubans with low education levels will be more likely if the
difference between the parameter associated to the educational level in the US
and the parameter associated to educational level in Cuba in the regression
equation is negative, which implies that individuals with a high educational level
will have little incentive to migrate and therefore results in negative selection
(Borjas 1991).
Data
The data used in this article come from the random sample of 1 per cent of
the 2010 US Population and Housing Census, provided by Integrated Public
Use Microdata Series (IPUMS 2010). This sample includes only individuals who
entered the US at the age of 17 years or over.
This approach intends to avoid people who have completed their training in
the US (Lowell et al. 2008). In addition, we have considered only individuals
under the age of 50 years, since we assume that the group aged between 16
and 49 years consists of those who will be most likely to migrate for economic
reasons (Bertoli et al. 2010).
The sample above is completed with a sample of Cubans living in Cuba in
2002 provided by Integrated Public Use Microdata Series-International (IPUMS-
International 2011) which corresponds to a 10 per cent random sample of the
Population and Housing Census of Cuba. As in the sample of Cubans in the US,
we considered only individuals between 17 and 49 years.
In both samples, we have only considered working individuals. Thus, we have
a total of 12.176 observations for Cubans in the US and for Cubans in Cuba
81.641 observations. Table 1 describes the characteristics of the samples used.
In the Cubans in Cuba sample, the most commonly observed age group are
individuals between 33 and 40 years (34.7%), while for the Cubans in the
US sample, the most observed group are individuals between 41 and 49 years
(43.2%).
IJCS6_1 45 06/06/2014 11:35
46 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
The proportion of Cubans in Cuba between 17 and 24 years is 3 times higher
than that of Cubans in the US in this age group. Figure 1 shows that 50 per cent
of Cubans in Cuba consist of those aged 35 years and above and that 50 per
cent of the Cubans in the US sample are 39 years old or more. The mean age for
Cubans in the US is 38 years, while in the case of those who have not migrated,
it is 35 years. It is therefore younger Cubans who do not emigrate.
In the case of the Cubans in the US sample, the percentage of individuals with
more than 12 years of education is higher (41%), whereas in the case of the
Cubans in the Cuba sample, this group represents 15 per cent of individuals,
whereby the proportion of Cubans with higher education in the US is almost 3
times higher than the proportion of people with the same educational level in
Cuba. This fact suggests that those who have received more education decided
to migrate.
Table 1 Sample Description
Cubans in CubaaCubans in USa,b
2002 2010
Absolute
Frequencies
Relative
Frequencies
Absolute
Frequencies
Relative
Frequencies
Educational level
0–8 years 7892 0.097 479 0.039
9–12 years 61643 0.755 6690 0.549
13 or more years 12106 0.148 5007 0.411
Observations 81641 1.000 12176 1.000
Age
17–24 years 9415 0.115 477 0.039
25–32 years 22843 0.280 2028 0.167
33–40 years 28309 0.347 4417 0.363
41–49 years 21074 0.258 5254 0.432
Observations 81641 1.000 12176 1.000
Occupational Category
Category 1 16274 0.199 3202 0.263
Category 2 12335 0.151 774 0.064
Category 3 53032 0.650 8200 0.673
Observations 81641 1.000 12176 1.000
Source: Based on data from the US Census of Population and Housing 2010 and the Population and Housing
Census of Cuba 2002, both databases provided by Integrated Public Use Microdata Series (IPUMS) and
IPUMS-International.
a Individuals between 17 and 49 years
b Only considered individuals who were 17 years or older at the time of entering the US.
IJCS6_1 46 06/06/2014 11:35
CUBAN MIGRATION TO THE US AND EDUCATIONAL SELFSELECTION 47
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
Only 4 per cent of the Cubans in the US have less than 9 years of education.
This would indicate that these individuals do not have incentives to emigrate to
the US since, possibly, they are under-valued in the US labour market.
In both samples, we are more likely to find individuals with between 9 and
12 years of education, 76 per cent of Cubans in Cuba have reached this level
of education and 55 per cent in the case of the Cubans in the US sample. In the
sample of Cubans in the US, the mean education years figure is 12.91, and in the
sample of Cubans living in Cuba, this mean is 11.15. In both samples, about 50
per cent of the individuals observed have 12 or more years of education.
Figure 2 shows the distribution of years of education in Cubans in Cuba and
in Cubans in US samples, respectively.
These results are consistent with the data provided by the 2010 US Census,
indicating that 20.8 per cent of the Cuban population in the US have studied
between 9 and 12 years and 30.2 per cent have received 13 or more years of
education (US Census Bureau 2010).
Regarding professional category, in both samples, the highest percentage of
individuals fall in what we have called Category 3 (Skilled Workers). In Cuba,
they represent 65 per cent of the observations, while in the US, they represent
67 per cent of the sample. For Category 2, (Technical Level) the proportion of
individuals in Cuba that work in this category is 2.35 times higher than the
proportion of individuals in this category in the US sample. For Category 1,
Figure 1 Boxplot of the age
Source: Authors’ calculation based on data from the Cuban Census of Population and Housing 2002 and the
US Census of Population and Housing 2010, both databases provided by Integrated Public Use Microdata
Series (IPUMS).
IJCS6_1 47 06/06/2014 11:35
48 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
(University-Directors-Executive), the percentage of individuals in the US sample
is 26 per cent, while in the Cuba sample, they represent 20 per cent.
Empirical Results
To study the self-selection problem in terms of levels of education of Cubans
who emigrate to the US, we estimate a logit model to calculate the migration
probability using the sample described in the previous section. Our interest is to
analyse the impact of variables such as age, level of education and professional
category on the migration probability. In particular, we have calculated the
impact of education, so as to conclude what kind of self-selection problem we
are facing.
The variables used are described in Table 2 with the main estimation results
figuring in Table 3. According to the obtained results, individuals with higher
educational levels are more likely to migrate. Keeping all other variables constant,
the migration probability of an individual with 9 to 12 years of education would
Figure 2 Histogram of the years of education
Source: Authors’ elaboration based on data from the Cuban Census of Population and Housing 2002
and the US Census of Population and Housing 2010, both databases provided by Integrated Public Use
Microdata Series (IPUMS).
IJCS6_1 48 06/06/2014 11:35
CUBAN MIGRATION TO THE US AND EDUCATIONAL SELFSELECTION 49
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
be multiplied by 0.165, while the probability of an individual with less than
9 years of education would be multiplied by 0.067.
A decrease in the probability of migration is higher if the individual’s years
of education is less. Thus, if we only take level of education into consideration,
in the model, the opportunity to emigrate for an individual with 13 or more
years of education is 14.93 times higher than for an individual with 8 years or
less of education and 6.06 times than for an individual having 9 to 12 years of
education.
Figure 3 shows the behaviour of the migration probability according to
the years of education of the individual, keeping all other variables constant.
It peaks when the individual has 13 or more years of education. This group
contains individuals whose level of education is above the mean of years of
studies in Cuba (11.15 years, as shown in the previous section). Hence, we can
conclude that according to the results obtained with the maximum likelihood
estimation, Cubans show positive self-selection, in terms of their educational
level to immigrate to the US.
Table 2 Model Variables Description
Age 1 Dummy variable, takes value 1 if the individual is between 17 and 24.
Age 2 Dummy variable, takes value 1 if the individual is between 25 and 32.
Age 3 Dummy variable, takes value 1 if the individual is between 33 and 40.
Age 4 Dummy variable, takes value 1 if the individual is between 41 and 49. Is
the variable reference or indicator of the age.
Level 1 Dummy variable, takes value 1 if the individual has between 0 and 8 years
of education.
Level 2 Dummy variable, takes value 1 if the individual has between 9 and
12 years of education.
Level 3 Dummy variable, takes value 1 if the individual is 13 years of education or
more. (Reference group for educational level).
OCCAT 1 Dummy variable, takes value 1 if the individual belongs to the
professional status of university, Directors, Executives.
OCCAT 2 Dummy variable, takes value 1 if the individual belongs to the
professional status of Technical Professional.
OCCAT 3 Dummy variable, takes value 1 if the individual belongs to the professional
status of Qualified Workers. (Reference group for professional category).
Source: Author’s own elaboration based on the variables used in the maximum likelihood estimate binary
logit model.
IJCS6_1 49 06/06/2014 11:35
50 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
Table 3 Estimation Results: Logit Model
Estimation (B) Typical Error
Mean
Wald Statistic Degrees of
Freedom
Significance
Level
Exp(B) I.C. 95% para EXP(B)
Inferior Superior
Step 1aLevel 4494.432 2 0.000
Level 1 -1.919 .050 1475.023 1 0.000 .147 .133 .162
Level 2 -1.338 .021 3995.165 1 0.000 .262 .252 .274
Constant -.883 .017 2760.832 1 0.000 .414
Step 2bLevel 3824.577 2 0.000
Level 1 -1.931 .050 1466.998 1 0.000 .145 .131 .160
Level 2 -1.231 .022 3259.769 1 0.000 .292 .280 .304
Age 1677.654 3 0.000
Age 1 -1.336 .050 711.606 1 0.000 .263 .238 .290
Age 2 -1.003 .028 1237.456 1 0.000 .367 .347 .388
Age 3 -.473 .023 420.939 1 0.000 .623 .595 .652
Constant -.462 .021 486.706 1 0.000 .630
Step 3cLevel 4525.688 2 0.000
Level 1 -2.697 .055 2434.497 1 0.000 .067 .061 .075
Level 2 -1.803 .029 3930.423 1 0.000 .165 .156 .174
Age 1662.855 3 0.000
Age 1 -1.369 .050 737.059 1 0.000 .254 .231 .281
Age 2 -1.004 .029 1213.903 1 0.000 .367 .346 .388
Age 3 -.463 .023 390.160 1 0.000 .629 .601 .659
OCCAT 1632.515 2 0.000
OCCAT 1 -1.082 .031 1187.121 1 0.000 .339 .319 .360
OCCAT 2 -1.112 .041 749.559 1 0.000 .329 .304 .356
Constant .335 .030 123.197 1 0.000 1.398
Source: Author’s own elaboration based on the results of the maximum likelihood estimate of the binary logit model.
aVariable (s) entered (s) in step 1: LEVEL. bVariable (s) entered (s) in step 2: AGE. cVariable (s) entered (s) in step 3: OCCAT.
IJCS6_1 50 06/06/2014 11:35
CUBAN MIGRATION TO THE US AND EDUCATIONAL SELFSELECTION 51
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
It must be taken into account that the individual’s choice affects both the
origin country and the host country (Fernández-Huertas Moraga 2011).
Migration of better educated individuals has a negative impact on countries of
birth since the educational investment of these people is unrecoverable (Didou
2009). However, some authors claim that the emigration of skilled individuals
has a positive impact on the origin country through remittances or investments
made by migrants (Durand et al. 2001).
In terms of benefits to the host country, migration increases its production
and technological capacity, and it implies, from an economic point of view, no
significant cost in terms of social assistance, which would be required if the
majority of those who migrate were people with less training (Cuecuecha 2005).
In terms of age, if we only consider this variable in the model, the opportunity
to migrate for an individual aged between 41 and 49 years is 1.59 times higher
than that of an individual aged between 33 and 40 years, 2.73 times higher than
that of an individual aged between 25 and 32 years and 3.94 times higher than
that of an individual under 25 years. This result is consistent with the sample
descriptive analysis we performed – the Cubans that migrated to the US are older
than Cubans who remained in Cuba.
Individuals with professional category 3 (Skilled Workers) have a migration
opportunity of 2.95 times that of individuals in category 1 (University-Directors-
Figure 3 Predicted migration probability and years of study
Source: Authors’ own elaboration based on the binary logit model estimates.
IJCS6_1 51 06/06/2014 11:35
52 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
Executive) and 3.04 times that of individuals with category 2 (Technical Level),
if the model only considers this variable. Our future goal is to explain how the
fact that the individual work in a job commensurate with their qualifications
influences the migration probability.
Conclusions
In this article, we have analysed the self-selection problem of Cuban emigrants
to the US in terms of individual, observable characteristics: age, professional
category and educational level. We have used the Population and Housing Census
of the United States (2010) and the Census of Population and Housing for Cuba
(2002). In both samples, we have only considered workers aged between 17 and
49 years.
For the analysis, we have proposed a binary logit model that explains the
choice of the individual, at the time of emigration, depending on the variables
considered. The main conclusion obtained, of the maximum likelihood
estimation of the model is that Cubans positively self-select in their migration
decision to move to the US, in terms of educational level those with more years
of study than the mean of the distribution of years of study in Cuba.
In terms of age, the most who migrate are people aged between 41 and 49 years.
The qualified workers category also contributes to the migration probability,
which is contrary to the fact that most migrants are the most qualified people.
However, we will not consider this result to be conclusive given the fact that
we have not considered whether individuals in both Cuba and the US have a
professional category in line with their educational level.
The positive educational self-selection problem has negative consequences on
Cuba, not only in terms of the non-recoverable educational investment but also,
and more importantly, in terms of an important loss of human capital.
Acknowledgements
The authors acknowledge financial support from the Econometrics Research
Group (Basque Government grant IT-642-13).
Bibliography
Aja Díaz, A. (2002) La emigración cubana. Balance en el siglo XX (Colección: Centro
de Estudios de Migraciones Internacionales). La Habana, Cuba: Red de Bibliotecas
Virtuales de Ciencias Sociales de América Latina y el Caribe. Available at http://bib-
liotecavirtual.clacso.org.ar/ar/libros/cuba/cemi/emig.pdf. (Accessed at 25 April 2012.)
IJCS6_1 52 06/06/2014 11:35
CUBAN MIGRATION TO THE US AND EDUCATIONAL SELFSELECTION 53
IJCS Produced and distributed by Pluto Journals www.plutojournals.com/ijcs/
Albo, A. and J.L. Ordaz Díaz. (Agosto, 2011) Migración mexicana altamente calificada en
EEUU y Transferencia de México a Estados Unidos a través del gasto en la educación
de los migrantes. Documentos de Trabajo, Análisis Económico, Servicio de Estudios
Económicos. Madrid: Fundación BBVA.
Bertoli, S., J. Fernández-Huertas and F. Ortega. (2010) Crossing the Border: Self-
Selection, Earnings and Individual Migration Decisions. Discussion paper series No.
4957, Econstor. Available at http://www.econstor.eu/dspace/bitstream/10419/36837/1/
626944201.pdf. (Accessed 22 May 2012.)
Borjas, G.J. (1987) ‘Self Selection and the Earnings of Immigrants’, American Economic
Review 77(4): 531–53.
Borjas, G.J. (1991) ‘Immigration and Self-Selection’, in J.M. Abowd and R.B. Freeman
(eds) Immigration, Trade and the Labor Market. Chicago, IL: The University of
Chicago Press.
Chiquiar, D. and G.H. Hanson. (2005) ‘International Migration, Self-Selection, and
the Distribution of Wages: Evidence from Mexico and the United States’, Journal of
Political Economy 113(2): 239–81.
Cuecuecha, A. (2005) ‘The Immigration of Educated Mexicans: The Role of Informal
Social Insurance and Migration Costs’. Canadian Economics Association. Available at
http://economics.ca/2005/papers/0298.pdf. (Accessed 23 June 2012.)
Demographic Yearbook of Cuba. (1961 [1965]) Republic of Cuba: ONE. La Habana:
Junta Central de Planificación.
Demographic Yearbook of Cuba. (2010) Republic of Cuba: ONE. Available at http://
www.one.cu/publicaciones/cepde/anuario_2010/anuario_demografico_2010.pdf.
Didou, S. (2009) ‘¿Pérdida de cerebros y ganancia de saberes?: La movilidad internacional
de recursos humanos altamente calificados en América latina y el Caribe’, in S. Didou
Aupetit y E. Gérard (eds) Fuga de cerebros, movilidad académica, redes científicas.
Perspectivas latinoamericanas. México, D.F.: Centro de Investigación y de Estudios
Avanzados del Instituto Politécnico Nacional, pp. 25–63.
Durand, J., D.S. Massey and R.M. Zenteno. (2001) ‘Mexican Immigration in the United
States: Continuities and Changes’, Latin American Research Review 36(1): 107–27.
Fernández-Huertas Moraga, J. (2011) ‘New Evidence on Emigrant Selection’, Review of
Economics and Statistics 93(1): 72–96.
Hotz, V.J. (2011). The Roy Model of Self-Selection: Simple Case. Topics in Labor Economics
Duke University. Avaliable at: http://public.econ.duke.edu/~vjh3/e262p_07S/readings/
RoyModelSimpleCase.pdf.
Integrated Public Use Microdata Series-International (IPUMS-International). (2011)
Version 6.1 (Machine-Readable Database). Minneapolis, MN: Minnesota Population
Center and University of Minnesota.
Integrated Public Use Microdata Series (IPUMS: Steven Ruggles, J., A. Trent, K. Genadek,
et al.). (2010) Version 5.0 (Machine-Readable Database). Minneapolis, MN: University
of Minnesota.
Lowell, B.L., C. Pederzini and J. Passel. (2008) ‘The Demography of Mexico/US
Migration’, in A. Escobar y S.F. Martin (eds) Mexico-US Migration Management: A
Binational Approach. Lanham, MD: Lexington Books.
Motel, S. and E. Patten. (2012) Hispanic of Cuban Origin in the United States.
Statistical Profile, Pew Hispanic Center. Available at http://www.pewhispanic.org/
files/2012/06/2010-Cuban-Factsheet.pdf. (Accessed 10 July 2012.)
IJCS6_1 53 06/06/2014 11:35
54 ACADEMIC ARTICLE ALEIDA COBAS VALDÉS AND ANA FERNÁNDEZ SAINZ
I J C S 6.1 S 2014
Orrenius, P.M. and M. Zavodny. (2005) ‘Self Selection among Undocumented Immigrants
from Mexico’, Journal of Development Economics 78(1): 215–40.
Pérez de la Riva, J. (2000) Los culíes chinos en Cuba. La Habana, Cuba: Editorial Ciencias
Sociales.
Roy, A.D. (1951) ‘Some Thoughts on the Distribution of Earnings’, Oxford Economic
Papers 3(2): 135–46.
Sjaastad, L.A. (1962) ‘The Costs and Returns of Human Migration’, Journal of Political
Economy 70(5, Part 2: Investment in Human Beings): 80–93.
US Census Bureau. (2010) Geographical Mobility/Migration Main. Washington, DC: US
Department of Commerce.
IJCS6_1 54 06/06/2014 11:35
