Journal of Rural Studies 35 (2014) 123e132

Contents lists available at ScienceDirect

Journal of Rural Studies

journal homepage: www.elsevier.com/locate/jrurstud

The selective migration of young graduates: Which of them return to

their rural home region and which do not?
Patrick Rérat a, b, *
a
Institute of Geography, University of Neuchâtel, Espace Louis-Agassiz 1, 2000 Neuchâtel, Switzerland
b
Institute of Geography and Sustainability, University of Lausanne, Geopolis, 1015 Lausanne, Switzerland

a b s t r a c t
Keywords: This paper addresses the migration behaviours of young university graduates from a rural region in
Graduate migration Switzerland. Based on a questionnaire survey, it compares graduates’ current place of residence (i.e.
Internal migration
whether or not they returned to their home region) with characteristics related to their socio-familial,
Life course
Rural region
migration and professional trajectories. The propensity to return varies not only according to labour
Brain drain market variables (employment opportunities), but also to other factors, some of which have even more
influence than job opportunities. The graduates’ life course position (kind of household), their partners’
characteristics (level of education and home region) and their family background (socio-economic status
and history of migration) all play a central role. On the whole, results show that migration appears as a
selective and complex process embedded in the life course of graduates.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction
In many countries, the proportion of young adults accessing
higher education is growing (Smith et al., 2014). This “human
capital” is seen as essential for regional growth and development in
the context of the knowledge economy (Corcoran et al., 2010). Thus
attracting or keeping highly qualified youth is a salient issue,
particularly for rural regions, many of which experience the net
out-migration of this population group (Thissen et al., 2010),
described as “brain drain”.
Research into this matter has tended to focus on the migration of
young adults moving away from rural regions in order to go to
university, but less is known about what happens after graduation
(Smith and Sage 2014). This paper addresses the migration of young
graduates, starting with two general observations that can be
traced back as far as Ravenstein’s seminal work on migration
(1876): firstly, each current of migration produces a movement in
the opposite direction, although usually not of the same volume.
Secondly, migration is a selective process. In other words, migration
does not concern each population group with the same intensity
(Ravenstein identified differences in terms of age and gender, and
between urban or rural regions, for example).
On the basis of these observations, this paper addresses two
under-researched dimensions of the migration of young graduates.
* Institute of Geography, University of Neuchâtel, Espace Louis-Agassiz 1, 2000
Neuchâtel, Switzerland.
E-mail address: Patrick.rerat@unine.ch.
It first takes into account not only dominant flows but also counter
flows, i.e. the graduates who do and do not return to their home
region after having received their degree. It then assumes that the
propensity to come back to a rural home region after graduation in
an urban centre is a selective process. By analysing the influence of
personal characteristics on migration behaviour, the paper iden-
tifies the graduates who are more/less likely than average to return
to the region where they grew up.
This paper addresses these issues in the case of a rural region in
Switzerland (Jura), drawing data from a questionnaire survey
designed for that purpose. The remainder of the paper is organised
as follows: Section 2 reviews the literature on the selectiveness of
internal migration, with an emphasis on young adults with a uni-
versity degree; Section 3 presents the spatial context and discusses
the research design and methods; and Section 4 presents empirical
results on the effects of various variables on the propensity to re-
turn to the home region. The summary and conclusion of the
findings are then set forth in Section 5.
2. Theoretical discussion
2.1. A life course approach to graduate migration
Adopting the perspective of migration as a decision embedded
in the graduates’ life course implies an assumption or recognition e
that internal migration is a complex phenomenon that goes
“beyond the economics” (Fielding, 1992a) and beyond “the

http://dx.doi.org/10.1016/j.jrurstud.2014.04.009
0743-0167/Ó 2014 Elsevier Ltd. All rights reserved.
124 P. Rérat / Journal of Rural Studies 35 (2014) 123e132
parochial focus on labour-motivated graduate migration” (Sage
et al., 2013). This therefore implies that research needs to show
greater appreciation of the “non-economic” issues of migration
(Halfacree, 2004). According to this perspective, migration is not
only explained by factors related to the structure of the labour
market (differentials in wages or job opportunities), as put forward
by neoclassic and structuralist approaches (Lundholm et al., 2004).
Nor is it only explained by a comparison of the financial costs and
benefits in the short and long term, as postulated by human capital
theory. Thus the underlying hypothesis tested in this paper is that
to return or not to the home region is more than a strict question of
(mis)match between supply and demand in the labour market, but
depends also on criteria related to social ties (friends, family,
partner), a sense of belonging, living environment, etc. (Haartsen
and Thissen, 2013; Rérat, 2013b).
Our argument is that the decision to move is a choice made
under certain constraints (Rérat, 2012a, 2013a). In other words, it
depends on the migrant’s needs and preferences, within a choice
set determined by various constraints or structures. In Bourdieu’s
terms (see Rye, 2011), structures can be both objective (e.g. the jobs
in a specific field that are located within a given distance or
reachable within a reasonable time) and subjective. In the latter
case, habitus (social laws, norms and values) may mediate between
individual drivers and social structures. In other words, people have
different dispositions (according to their socialisation), and they act
strategically on the basis of these. Rye (Rye, 2011) speaks of
structured freedom to qualify the migration aspirations and de-
cisions of rural youth. Even though we are not denying the growing
role of physical, economic and social mobility, we argue that in-
dividuals, even within a small and seemingly homogeneous group,
do not all display the same migration behaviours.
The population under study (young adults from a rural area who
have graduated from university) shares some important charac-
teristics: they come from the same region, belong to the same age
group and have reached the same level of education.1 However,
beyond their common characteristics, graduates may be quite
different in terms of socio-economic background, marital status,
field of study, etc. All these characteristics related to the biography
of individuals are likely to influence their migration behaviour (see,
for example, King and Shuttleworth, 1995; Belfield and Morris,
1999; Corcoran et al., 2010; Faggian et al., 2006). This shows the
importance of adopting a life course perspective in our research.
The life course approach is a way of structuring a complex set of
events that include decisions about occupational, marital and
housing careers (Mulder and Clark, 2002). It stresses the need to
take into account three kinds of interdependence (Heinz et al.,
2009, 16e17): (1) an interdependence of the past, the present and
the future, and thus a path dependence of the life course; (2) an
interdependence between the different spheres of action that
constitute the multi-dimensionality of the life course (family, ed-
ucation work, leisure, etc.) and (3) a multi-level interdependence
between individual action and political, economic, social and cul-
tural contexts, “since life course patterns are embedded in macro-
social structures and cultural beliefs and guided by market oppor-
tunities, institutions and social networks” (Heinz et al., 2009).
In this paper, the analysis of graduates’ migration behaviour has
been placed in the light of a triple biography (Courgeau, 1985),
which we have named the socio-familial trajectory, the migration
1
University graduates represent a limited e although rising e proportion of
young adults accounting to about 15% in Switzerland (SFSO, 2012). This small
proportion is explained by the fact that the Swiss education system places
importance on other forms of education (apprenticeships and professional training)
to a greater extent than many other European countries.
Fig. 1. Graduates’ triple trajectories.
trajectory and the professional trajectory (Fig. 1). These trajectories
encompass various dimensions that may influence the decision to
return or not to the home region. In other words, the hypothesis of
the life course perspective is that the graduates’ behaviour (or in-
tentions) in one life domain (e.g. marital status, career, etc.) has an
impact on their behaviour in other domains (in this case,
migration).
The objective of this paper is to determine which characteristics
are discriminant in the graduates’ propensity to return to their
home region once their university studies are completed.2 It is
important to highlight here that the relationships between the
actors’ characteristics and their migratory practices are not deter-
ministic, but probabilistic, and that the nature of causality in the
social world is a matter of chance rather than of destiny (Rye, 2011).
The links between the dependent variable (migration behaviour)
and the independent variables (relevant characteristics related to
their life course) may be of various natures. As stated in the liter-
ature review below, the link may be causal but also show an as-
sociation effect, as migration is usually part of a wider life project.
The following sections review the various dimensions of the
three trajectories according to the literature on the migration of
young graduates and highly skilled people.
2.2. Socio-familial trajectory
A graduate’s socio-familial trajectory includes elements such as
gender, life course position (whether he/she lives in couple and/or
has children), his/her partner’s characteristics (region of origin and
level of education) and his/her family background (parents’ socio-
economic status).
2.2.1. Gender
Recognition of the influence of gender varies among studies on
young graduates’ internal migration. For some scholars, gender
does not play a significant role (Belfield and Morris, 1999), whereas
others find a higher propensity to mobility among women
(Venhorst et al., 2010; Faggian et al., 2006), as well as greater as-
pirations and ambitions (Bjarnason and Thorkindsson, 2006). The
impact of gender may be indirect, and may affect other phenom-
ena: access to tertiary education, local employment structure or
construction and experience of rurality (Corbett, 2007).
2
It therefore does not aim to address migration motivations (which may change
over a graduate’s life course, as factors other than finding work may grow in
importance) or the decision making process.
 P. Rérat / Journal of Rural Studies 35 (2014) 123e132
Gender is one of the psychological and social factors affecting
the choice of both a field of study and an employment sector
(Chazal and Guimond, 2003). This may lead to the over-
representation of a gender in certain employment sectors, and the
overrepresentation of these sectors in the economic structure of a
peripheral region could induce a differentiated orientation of
migration flows between males and females. Corbett (2007), for
example, speaks of “gendered structure of opportunity”. Corcoran
et al. (2010, 210) observe such a phenomenon in Australia, where
the high percentage of female graduates in remote areas results
from the combination of two separate effects: firstly, there are more
female graduates to start with, and secondly, graduates in rural
areas are working in sectors that have traditionally been more
female-dominated, such as education and health.
2.2.2. Life course position
A person’s life course can be defined as a sequence of events
forming the pattern of his/her life from birth to death. As far as
migration and mobility are concerned, the most important transi-
tions on a personal level are leaving the parental home, forma-
tion(s) of couple(s), separation(s), birth of child(ren), and later on
retirement (Boyle et al., 1998). Life course position influences
housing needs, most notably due to changes in the household size
and structure, and also aspirations related to the residential
context. The decision to settle in cities or to move to the suburbs, for
example, is closely tied with life course position: non-family
households (persons living alone, flat sharers, childless couples)
mostly settle in cities, whereas households with children mostly
move to the suburbs (Rérat, 2012b). Evidences of an identical
phenomenon but on the interregional scale have, for example, been
found in Ireland, where families are overrepresented in the
migratory flows to rural areas (Stockdale and Catney, 2012).
2.2.3. Partner’s characteristics
For people living in a couple, the decision to migrate is made not
by just one person, but by a household. The influence of partners on
migration behaviours has been addressed in various ways. The
notion of “linked lives”, for example, has been used to underline the
fact that a move is to be understood as a joint decision or
compromise between both partners (Huinink, 2009). Other studies
have addressed the trade-offs within couples when both partners
pursue a professional career (Green, 1995).
Two characteristics related to partners appear to be of impor-
tance according to the literature on internal migration: the part-
ner’s region of origin and his/her level of education. Firstly, if the
partner comes from the same region (and thus both members of
the couple know the region, and have family there as well as
friends), the return propensity is greater (Labrianidis and Vogiatzis,
2012). Secondly, if both partners are pursuing a professional career,
then the decision to migrate implies the need to take into account
both of these careers. This raises the question of the impact of the
partner’s level of education on migration in peripheral regions
since they are characterized by a limited labour market. This point
is even more important given that an educational homogamy is
observed, i.e. couples in which both partners share the same level of
education (Suter et al., 2009).
2.2.4. Socio-economic background
Since the classic work by Bourdieu and Passeron (1977), many
studies have highlighted the contribution of the educational system
to the reproduction of social inequalities. This has been interpreted
as a consequence not only of economic factors, but also of the
habitus, a system of unconscious representations that encompasses
social laws, norms and values and that leads social behaviour.
125
Of interest here, therefore, are not inequalities in accessing
higher education, but the impact of socio-economic background on
migration behaviour. Several studies have noticed that graduates
whose father or mother occupies a managerial position are char-
acterised by a higher level of mobility (Belfield and Morris, 1999)
and a lower propensity to return to a peripheral region (Jamieson,
2000; Rye, 2011).
The concept of habitus has been used to interpret diverging
migration behaviours in three non-exclusive ways. Firstly, Jamieson
(2000) observes that the socialisation to mobility (both social and
spatial) increases with socio-economic background. Secondly, Rye
(2011) shows in the case of Norway that the social environment
intervenes in the evaluation of rurality, and that this evaluation
implies specific residential preferences and migration behaviours.
More precisely, young adults from the working class see the rural
setting and their region of origin more positively, and tend to return
more than average. Finally, the habitus also influences students’
chosen field of study and, subsequently, their choice of employ-
ment sector. The mechanism in this case is similar to the one
identified in respect to gender, given the under/overrepresentation
of some economic sectors in the labour market of rural areas.
2.3. Migration trajectory
A migration decision forms a part of a series of events in the life
course of an individual. It is therefore likely to be influenced by the
migration trajectory, that is to say, the moves and places of resi-
dence prior to the current situation. Therefore, this research has
taken into account not only the region of origin of graduates, but
also that of their parents, as well as the location of the university
that each graduate attended.
2.3.1. Region of origin of graduates and their parents
The migration experiences of graduates, and also of their par-
ents, play an important role in their subsequent behaviour
(Corcoran et al., 2010; Faggian et al., 2006). Residential practices
must therefore be studied within the context of the biographies or
histories in which they take place and the spaces that have been
involved. Gotman (1999) distinguishes two kinds of space in
childhood, the first of which is the reference space. Reference
spaces are the places where the ancestors lived: the living space of
his/her grandparents, the birthplace of his/her parents, etc. Such
spaces may form part of an individual’s historical memory without
necessarily being experienced, and enable him/her to inscribe
himself/herself in the fatherly or motherly line. The second kind of
space is the founding space. Founding spaces are the places of the
living memory, the places of residential familiarisation and social-
isation during childhood and adolescence.
Having lived most of one’s childhood and adolescence in one
particular region (founding space), or even in the region where
one’s parents grew up (reference space) can induce a higher sense
of belonging and a more developed social network within that
region, and may consequently influence the propensity to return.
Thissen et al. (2010), for example, find that young adults in rural
regions in Belgium and the Netherlands are less likely to out-
migrate when their parents were also born in these regions. On
the other hand, a family history marked by migration would imply
that young adults are more socialised to migration (Jamieson,
2000), and would therefore be less likely to return to their home
region.
2.3.2. Location of the university
Moving to university constitutes a first step in a long process of
detachment from networks of primary social relations such as
family and childhood friends. Some authors consider therefore that
126 P. Rérat / Journal of Rural Studies 35 (2014) 123e132

migration for higher education purposes is not only a move across
physical space, but is just as much a journey in social space (Rye,
2011). In addition, “migrating to learn” implies a process of
“learning to migrate” (Li et al., 1996). In the Swiss case for example,
graduates who studied in their home canton are far less likely to
migrate than the ones who had to move to attend higher education
(Schmidlin, 2007).
Given the distance between universities and their home canton,
almost all students from Jura live in the region where they study.
This migration does not necessarily imply real independence, but
rather an intermediary situation of residential semi-autonomy.
Multi-local practices are observed, between a dwelling occupied
during the week and returns to the family home on weekends or
during the holidays.3
The location of the university, or more precisely the distance to
the home region, could have an impact on migration behaviour.
This could be explained by the fact that the further the family home
is from the student’s place of study, the less easy it is to commute
back on a weekly basis. This could be of importance, as the fre-
quency of returns may have an influence on social ties (family,
friends, social life, etc.) in the home region on the one hand, and on
the development of a new social network in the destination region
on the other.
2.4. Professional trajectory
For almost all graduates, the life course transition studied in this
paper is a move into the professional world. Since regions differ in
their economic structure, labour market and job opportunities, the
region in which graduates settle after university will likely depend
on their choice of career, which in turn is likely to depend on what
they studied. Thus a graduate’s university course (field of study,
kind of degree, grants) and field of work are likely to influence his/
her migration behaviour and propensity to return to a rural region
characterised by a segmented or truncated labour market (Gordon,
1995).
2.4.1. Kind of degree
It is generally said that urban hierarchy matches with the hier-
archy of qualifications and fields of specialisation required in the
labour market, and that spatial mobility increases with the level of
education (Caro, 2006; Baron and Perret, 2006). More specifically,
and herein lies the brain drain hypothesis, the propensity to leave a
peripheral region (or not to return there) increases with the level of
qualification (Corbett, 2007; Thissen et al., 2010; Corcoran et al.,
2010). This could be observed among university graduates ac-
cording to their degree (bachelor’s, master’s or doctoral degree).
2.4.2. Field of study and field of work
The propensity to migrate varies according to the graduates’
field of study and of work.4 In the case of Australia, Corcoran et al.
(2010) show that graduates in teaching and health are more likely
to settle outside the main urban centres. This probability increases
further with the degree of periphericity of a region. In Scotland,
Stockdale (2006) highlights the importance of the public sector and
of self-employment, the latter of which is primarily a strategy used
to cope with a limited labour market (‘survival self-employment’),
and rarely a process aiming to create firms that provide other jobs.
3
Unlike some other contexts, such as the UK, access to higher education does not
go together with a tradition of termly or annual migration towards universities.
4
In strictly neo-classic and structuralist perspectives, only the fields of study and
activity would matter as they express the degree of matching between a graduate
and a regional labour market.
In other words, both fields of study and fields of work define the
range of possibilities available in terms of job opportunities.
Consequently, both have a significant influence on migration flows.
2.4.3. Grants
Does having received a grant influence migration behaviour?
This question has, among others, been the subject of concern in
countries characterised by the negative migration balance of their
highly qualified citizens. A study of graduates from several coun-
tries in Africa, Asia and Latin America has shown that grant re-
cipients have a greater propensity to return to their home country
(Glaser, 1978), as these graduates feel indebted when the grant
comes from their home region (and not the destination country).
The question remains open as to whether this also occurs on an
interregional scale.
2.4.4. Year of graduation
According to the literature, the time span since graduation may
play a double role on the propensity to return. When the length of
time spent outside the region of origin gets longer, the social ties
that are kept there may weaken, whereas the opposite happens in
the destination region. In addition, a residential, or locational,
inertia may grow over time and make a further move (back to the
home region) less likely (Halfacree and Rivera, 2012).
Another mechanism has also been highlighted. In the case of
Scotland (Fielding, 1992a,b; Findlay et al., 2009), many migrations
take place towards London and Southeast England at the beginning
of the career (expressions such as “escalator region” or “elevator
region” have been used to describe this phenomenon). In this case,
migration is mainly motivated by the desire to secure a job and to
gain professional experience. Once a certain level of social mobility
in their work career is achieved, many Scots then return to their
home region in the middle of their professional life (more than 75%
return before they are 45). Even though our research does not cover
such a long period, it will enable us to measure the impact on
migration behaviour of the length of stay outside the region of
origin.
All of the variables discussed above are addressed in this paper
in the case of a rural Swiss region. The following section presents
the case study and the methodology.
3. Case study and methodology
3.1. Spatial context
This paper addresses the migration behaviours of young adults
coming from Canton Jura, a French-speaking region located in the
north-west of Switzerland, after graduation (all universities are
located outside the region) (Fig. 2).
Canton Jura shares the usual characteristics of a peripheral and
rural region, given the spatial context of Switzerland.5 In 2010, it
had a total of 70,000 inhabitants (ranked 20th out of 26 cantons in
terms of size). The demographic growth is less dynamic than in the
rest of the country (population increase of 2.0% in Jura versus 9.2%
in the whole country between 2000 and 2010; rank 22nd). There is
an overrepresentation of low skilled workers and of jobs in the
agricultural and industrial sectors, and consequently an under-
representation of highly skilled workers and jobs in the service
sectors. The per capita income in Jura accounts for only 70% of the
Swiss value (rank 26th).
5
It is important to note that what is meant by peripheral and rural is highly
context-dependent and varies greatly between countries.
 P. Rérat / Journal of Rural Studies 35 (2014) 123e132
Fig. 2. Graduates’ current place of residence.
Switzerland has been characterised over the last four decades by
rapid metropolitan growth (mainly in Zurich, Geneva-Lausanne
and Basel), from both a demographic and an economic point of
view (Rérat, 2012a). While this has not been the case in Canton Jura,
it is not far from Basel, Switzerland’s third largest city; however,
since Basel is German-speaking, the language barrier prevents Jura
from fully benefiting from its proximity to the third Swiss city.
3.2. Identification of the target population
This paper is taken from a wider research project, whose aim
was to investigate the various migration patterns and behaviours of
young graduates from Jura based on a questionnaire survey and in-
depth interviews.6 The target population for this study has been
defined as all young people from Canton Jura (that is to say, whose
family lived in the region when they left high school), who grad-
uated from university between 2000 and 2010.7 Almost all of them
had to move to attend university, meaning that they had a first
experience of migration on their own, of living in an urban resi-
dential context and of building social ties outside their home
region.
In order to identify subjects for the study, a survey could not be
sent to a random selection of graduates, as no address lists exist.
Therefore, the snowball sampling method was chosen instead.
Sixty of the researcher’s acquaintances, who belonged to the target
6
A dozen in-depth interviews were carried out in order to deepen the under-
standing of the quantitative results with qualitative material.
7 9
The time between graduation and the survey differs considerably between
graduates (from six months to ten years). This may have an impact on some fea-
tures (e.g. older graduates may be more likely to live in a couple than younger
ones). However, this does not explain why a certain group of graduates (e.g. a
certain kind of household) is overrepresented in the home region when the year of
graduation is controlled.
127
population, were contacted by email, presenting the research ob-
jectives and requesting the email addresses of relevant graduates.
When a new address was received, the same message was sent,
creating a multiplier effect. An email was also sent to all members
of the cantonal administration, and a press conference was
organised a few days before Christmas, a period during which most
graduates visit their family.
This approach was found to be very successful: 550 emails were
sent to the researcher and 1280 individuals were identified. A total
of 924 people filled in the online questionnaire (response rate of
72%), of which 498 were university graduates; these formed the
study sample addressed in this paper.8 The sample represents an
estimated 40% of the number of university graduates coming from
Canton Jura and who obtained their degree between 2000 and
2010.9
More important than the proportion of the target population
that received and completed the questionnaire is the issue of rep-
resentivity. Biases may occur, as not all members of the target
population have equal probability of being reached by the snowball
sampling method. As there is little official data available, only a few
comparisons can be made between the target population and the
sample. They show very few noticeable differences (except the
overrepresentation of graduates from the University of Neuchatel,
which is to be expected since the researcher studied and works
8
The remaining 426, which are not addressed in this paper, were graduates from
“universities of applied science” (another form of higher education more focused on
practical aspects) and graduates participating in further full-time education.
The official statistics do not take into account the number of graduates but the
number of degrees, which may lead to double counting (most graduates have more
than one degree). The 40% quoted is calculated on the basis of people with a
master’s degree (433 in the sample; 1079 in total). The number of bachelor’s degree
has not been taken into account here as 90% of these graduates go on with a
master’s degree.
128 P. Rérat / Journal of Rural Studies 35 (2014) 123e132

Table 1
Migration behaviour according to graduates’ socio-familial trajectories.

Variable Modality Number of % of graduates % of graduates Statistical test

graduates living in Jura living elsewhere (Chi-square)

Total e 498 40.0 60.0 e

Sex Female 250 41.6 58.4 n.s.
Male 248 38.3 61.7
Kind of household Childless couples 193 32.6 67.4 p < .001
Households with children 127 58.3 41.7
Persons living alone 104 34.6 65.4
Flat shares 49 12.2 87.8
Living with parents 20 90.0 10.0
Children No 366 33.6 66.4 p < .001
Yes 132 57.6 42.3
Partner’s region of origin Jura 176 63.1 36.9 p < .001
Other 138 20.0 80.0
Partner’s level of education Low 50 70.0 30.0 p < .001
Medium 54 46.3 53.7
High 201 35.3 64.7
Father’s level of education Low 239 45.2 54.8 p < .001
Medium 90 45.6 54.4
High 160 28.1 71.9
Mother’s level of education Low 294 43.5 56.5 p < .05
Medium 128 35.9 64.1
High 58 27.6 72.4

there and thus had more contacts there). However, these differ-
ences do not affect the reliability and representivity of the sample,
as they relate to features that neither increase nor decrease the
propensity to return to one’s home region.
According to the survey, the proportion of graduates from the
University of Neuchatel returning to Jura (41.5%) is not statistically
different from the whole population. Indeed, 40% of all young
people from Canton Jura (199 of the 498) who graduated from
university between 2000 and 2010 have returned to live there. The
others have mainly settled in the urban centres in the French-
speaking part of Switzerland (Fig. 2), while about 7% live in a
different linguistic region and another 7% live abroad. This paper
aims to compare the characteristics of graduates who have
returned to their home region with the ones who have not in order
to identify the features that influence migration behaviours.
A limitation of the dependent variable defined here is its
dichotomous nature (the fact that graduates lived either in Jura or
elsewhere when the survey was carried out). It is important to
acknowledge that migration patterns may be more complex, as
shown in the UK by Sage et al. (2013). For example, they may
involve numerous moves, which may be either temporary or long
term. The choice to focus on the current place of residence (and not
to take into account the other moves that may have happened
between graduation and the survey) seems justified, however,
given the size of the sample (the numbers would be too small if
split up), the small size of the country and of its French-speaking
part (long-distance commuting may replace internal migration;
see Rérat, 2013b) and the fact that entering the labour market has
been rather easy in the years 2000 due to the economic conditions
in Switzerland (meaning that the parental home is not often
needed as a safety nest on the medium term).
4. Empirical results and discussions
A comparison of graduates by place of residence has been made
in several steps. The first analyses are bivariate, and compare the
propensity to return in the light of specific features of graduates’
socio-familial trajectories (Table 1), migration trajectories (Table 2)
and professional trajectories (Table 3). The number of graduates
representing each modality of the independent variables is shown

Table 2
Migration behaviour according to graduates’ migration trajectories.

Variables Modalities Number of % of graduates % of graduates Statistical test

graduates living in Jura living elsewhere (Chi-square)

Total e 498 40.0 60.0 e

Graduate’s region of origin Jura 479 40.5 59.5 n.s.
Other 19 26.3 73.7
Father’s region of origin Jura 381 43.8 56.2 p < .001
Other 110 25.5 74.5
Mother’s region of origin Jura 378 43.7 56.3 p < .01
Other 118 28.0 72.0
University Neuchâtel 183 41.5 58.5 p < .01
Lausanne 81 39.5 60.5
Swiss Federal Institute of Technology of 49 26.5 73.5
Lausanne
Fribourg 47 51.1 48.9
Geneva 41 36.6 63.4
Swiss Federal Institute of Technology of 20 45.0 55.0
Zurich
Bern 15 86.7 13.3
Basel 10 40.0 60.0
Other Swiss universities 12 25.0 75.0
Abroad 16 25.0 75.0
 P. Rérat / Journal of Rural Studies 35 (2014) 123e132 129

Table 3
Migration behaviour according to graduates’ professional trajectories.

Variable Modalities Number of % of graduates % of graduates Statistitical test

graduates living in Jura living elsewhere (Chi-square)

Total e 498 40.0 60.0 e

Field of study Social and human sciences 133 44.4 55.6 n.s.
Exact and natural sciences 106 30.2 69.8
Humanities 92 43.5 56.5
Law 64 40.6 59.4
Economic science and business administration 45 40.0 60.0
Technical sciences 32 50.0 50.0
Medical sciences 19 42.1 57.9
Degree Bachelor 41 65.9 34.1 p < .001
Master 454 37.7 62.3
Scholarship Yes 284 44.4 55.6 p < .05
No 214 34.1 65.9
Employment status Public sector 271 43.2 56.8 p.<.1
Private sector 149 30.9 69.1
Non-employed 34 41.2 58.8
Interns 23 51.5 43.5
Self-employed 20 40.0 60.0
Business sector Teaching 96 68.8 31.3 p < .001
Public administration 71 46.5 53.5
Health and social services 58 34.5 65.5
Higher education 58 1.7 98.3
Media and culture 33 33.3 66.7
Architecture, engineering, environment 32 50.0 50.0
Other services (IT, trade, etc.) 27 25.9 74.1
NGOs 24 41.7 58.3
Industry 24 20.8 79.2
Banks and insurance companies 23 30.4 69.6
Lawyers, notaries and trust companies 18 50.0 50.0

in the tables. The total number of respondents may differ between
the variables because of non-responses and the fact that some
graduates may not be concerned (e.g. having a partner).
The chi-square test was applied to cross tabulations to check
whether any significant differences exist. This shows whether cell
counts in the cross table deviate from their expected values if re-
turn migrants and out-migrants had the same profile. The test does
not indicate, however, the direction of the relationship, and nor
does it indicate whether the result hinges on one modality or more.
A further e and important e drawback of the chi-square test is
that it disregards conditional effects. In other words, a variable may
be insignificantly related to the dependent variable when taken
individually, but when corrected for other mediating effects, it may
well turn out significantly. For this reason, a logistic regression was
also carried out in order to measure the influence of each variable
all other things being held constant.
4.1. Socio-familial trajectory
A graduate’s life course position (measured by the kind of
households) has a strong influence on his/her propensity to return
to Jura. Two groups of graduates are more numerous in Jura: those
who live with their parents while they wait to either enter the la-
bour market or gain residential independence (90.0%)10, and those
who live in a couple and have children (58.3%). Thus in terms of life
course, settling in Jura complies with two logics: a temporary re-
turn to the family home, or a return rooted in a couple’s plans to
start a family. While 60% of graduates with one or more children
live in Jura, less than a third of graduates without children have
settled in this region. Thus it seems that return migration appears
less attractive for non-family households and that the decision to
start a family occurs later among graduates living in urban areas.
There is no difference between men and women in their return
propensity. This shows that on the whole there is no gendered
structure of opportunity in the local labour market. More precisely
it seems that there is a balance between the fields of work where
the majority of graduates are female (health and teaching) or male
(engineering, industry but also long distance commuting).
Partners’ characteristics have a clear impact on the migration
behaviour of graduates living in couples.11 If the partner is also from
Jura, a return is more likely (63.1% versus 20% if the partner spent
his/her childhood and adolescence elsewhere). His/her level of
education also plays a substantial role: the higher it is, the lower
the likelihood of settling in Jura. These results show the importance
of social ties and attachment (when both partners come from Jura)
and the limiting role of the labour market (when both partners
have attended tertiary education).
The propensity to return varies according to a graduate’s social
background, measured by the level of education reached by his/her
parents. Graduates return more often when their parents have not
achieved a university degree. This may be explained based on the
overrepresentation of certain occupations among graduates with
lower middle-class backgrounds, which correspond more to the
local labour market (teaching, for example12). More detailed
empirical data would be necessary to test the other hypotheses put
forward in the literature, such as varying socialisation to social and
spatial mobility, and a varying construction and representation of
rurality.

10
This group is quantitatively not important (20 graduates in 498), and its
behaviour is quite logical given that parents live in Canton Jura unless they have
moved since their children started their study.
11
In the questionnaire, living in a couple was defined as sharing a flat.
12
Graduates whose father attended tertiary education account for 32.6% of all
graduates. The percentage is of 21.3% among graduates working as teachers, which
is the lower proportion among all employment sectors (listed in Table 3).
130 P. Rérat / Journal of Rural Studies 35 (2014) 123e132
4.2. Migration trajectory
When Canton Jura is the region of origin13 of graduates’ par-
ents (the so-called reference space), the graduates’ propensity to
return is more marked. This could be explained by the fact that for
these graduates, the region of origin engenders a more developed
sense of belonging and a larger social network due to the presence
of the extended family, which then has an effect on migration
behaviour. An identical relationship is observed for the region of
origin of graduates themselves (namely their founding space).
However, given the limited number of those who spent most of
their first 18 years outside of Jura, the link is not statistically
significant.
Migratory patterns vary to some extent according to the uni-
versity attended.14 The case of Bern, however, needs to be put into
perspective: until the year 2000, this German-speaking university
offered a course in French specifically for students wanting to teach
in secondary schools in Canton Jura and the French-speaking re-
gions of Bern. This explains the frequent return of graduates from
this university.
Students who studied abroad or in certain Swiss non-French-
speaking universities (Lucerne, Lugano, St. Gallen or Zurich) were
less likely to return, but it is worth noting that these universities
attract only a small proportion of students from Jura in the first
place (6%). The differences between the other universities do not
enable us to distinguish a net trend regarding the role played by
distance between the cities where the universities are located and
Canton Jura.15 In other words, the distance does not significantly
affect the propensity to return. Universities located on average
90 min by train from Jura’s capital, Delémont, are characterised by a
greatly varying proportion of graduates going back to Jura: 51.1% for
the University of Fribourg, 45.0% for the Federal Polytechnic School
of Zurich, 38.1% for the University of Lausanne and 26.5% for the
Swiss Federal Institute of Technology of Lausanne. The difference
between the latter two, which are located next to each other, re-
veals the influence of other parameters, such as specialisation in
the student’s field of study (and their compatibility with the
regional labour market).
4.3. Professional trajectory
Overall, no significant difference is found between fields of
study. Only two groups of graduates differ from the overall pro-
portion of 40% of returns: only 30% of graduates in exact and nat-
ural sciences (maths, biology, chemistry, etc.) have settled in their
home region, while the figure is 50% for graduates in technical
sciences (engineering, architecture, agronomy, etc.). The broad
definitions that had to be used for statistical purposes may, how-
ever, hide some deeper differences between specialisations.
Type of degree, however, is found to influence migration
behaviour: the higher the qualification, the less likely a graduate is
to return. Graduates with only a bachelor’s degree e who represent
10% of the total e return much more often (65.9%) than graduates
with a master’s (37.7%).16
13
In this research, we defined the region of origin as the region where someone
spent most of his/her time until the age of 18.
14
When a graduate had attended more than one university, the one attended
most recently was taken into account.
15
Train journeys between Delémont (the capital of Jura) and the cities hosting
universities amount to 30 min (Basel), 53 min (Neuchatel), 58 min (Bern), 1h33
(Lausanne), 1h34 (Fribourg), 1h37 (Zurich), 1h42 (Lucerne), 2h04 (Geneva), 2h52
(St. Gallen) and 4h22 (Lugano).
16
The graduates who have achieved both a master’s and a doctoral title since the
year 2000 have returned in an even smaller proportion (7 in 37, or 18.9%).
The relationship between scholarship and migration is also
tangible, in the sense that recipients of scholarships have returned
more than average. However, the influence of a grant is less
important than the social background in which graduates grew up,
which raises the question of the mechanism behind this observa-
tion. Face-to-face interviews carried out among graduates show
that a sense of being indebted to the home region may exist, but
does not really act as an incentive to return; however, graduates
may for a few years keep Jura as their official place of residence (and
thus pay taxes there) in Canton Jura for a few years while working
and living mainly in another region. This way, they continue to pay
taxes in Jura, thus paying back part of the investment.17
People working in the public sector also return to Jura a little bit
more often than average (43.2%), but employees in the private
sector are less likely to settle in their home region (31.3%). Self-
employed people show migration behaviour reflecting exactly
that of the overall graduate population (40.0%).
Taking into account the employment sectors specifies migration
patterns in regard to the local economic structure (although it has
to be noted that graduates do not only rely on the cantonal labour
market as 17.4% of graduates living in Jura work in other cantons).
Teaching is the career path that enables the most to return: seven
teachers out of ten returned to their home region. Two further
groups are close to equilibrium, with as many out-migrants as re-
turn migrants: the category of activities related to architecture,
engineering and environment, and that of lawyers, notaries and
trust companies.
There is no correlation between the year of completion of the
degree and the propensity to return to Jura (R2 ¼ .002; n.s.). The
lowest proportion of returning graduates is recorded in 2008
(28.0%), and the highest the following year (44.6%). Such a variation
can be explained by the small size of the sample when divided by
year of graduation. In other words, there is no difference between
graduates who have several years of professional experience and
those who have recently left university. This overall trend may,
however, hide in- and out-migration flows that cancel each other
out. More information about the successive moves over the whole
life course would also be needed to reject definitely the hypothesis
of the effect of time.
4.4. Logistic regression
Logistic regression was carried out in order to consider all of the
independent variables simultaneously and in so doing to summa-
rise the information and to determine the influence of each of them
on the dependent variable, which consists in living in Jura (value 1)
or elsewhere (value 0).
The logistic regression model omits all cases with missing values
for any of the independent variables, and thus a total of 447 re-
spondents were included in the model. The influence of each in-
dependent variable was then measured by odds ratio e referred to
as exp(B) e which represents the ratio of odds of an event (e.g.
returning to Jura) occurring in one group compared to another (a
reference group). If the odd-ratio is higher (or lower) than one, the
odds of returning to Jura is more (or less) likely in the group under
study than in the reference group.
In order to carry out this analysis, some preliminary work had to
be done. Some of the variables were redefined in order to avoid
overly small numbers or redundancies. Typologies were created for
variables such as partners’ characteristics and employment sector.
Other variables were then eliminated by the model because they
17
Cantons pay a certain amount of money to the cantons hosting universities
according to the number of students from the region.
 P. Rérat / Journal of Rural Studies 35 (2014) 123e132
were not discriminant. This included variables that had already been
identified as having no impact on migration patterns: sex, year of
graduation, region of origin and field of study. For the last three
however, the small size of the sample has to be taken into account
and makes it difficult to definitely reject the hypothesis of an impact
on migration. Additional variables such as the type of degree and the
university attended were removed because their influence was not
statistically significant when the whole set of variables was taken
into account. This is explained for the type of degree by the fact that
the majority of university graduates holding only a bachelor degree
work as teachers (which is seen in the influence of the field of work).
For the university attended, specialisation in the student’s field of
study and its compatibility with the regional labour market seem to
be more important than the distance with Jura.
The variables contained in the model highlight the importance
of life course position (kind of household and having children or
not), family background (father’s region of origin and level of ed-
ucation) and employment sector (Table 4).18 Life course position
was found to exert a decisive influence on migration behaviour. For
example, having children increases the likelihood of settling in Jura
by 207.6% (in other words, the likelihood of returning is multiplied
by 3.076). This shows that the decision to migrate often takes place
in the context of a life project (e.g. starting a family). An even higher
value is obtained by persons living with a partner from Jura who
does not have a degree (þ257% when compared with singles). The
propensity to migrate is identical, however, between singles and
those graduates who live with a partner who comes from Jura and
does have a university degree.
A partner who has mostly lived in another region significantly
decreases the likelihood of returning to Jura, by 85%. This may be
due to the fact that the lack of familiarity with an area and the
absence of an extended social network do not create positive rep-
resentations and motivations to move there. In such cases, the
partner’s level of education has little effect (while for couples
where both partners come from Jura, the propensity to return is
lower when the partner has a university degree as well). This may
imply that job opportunities are not the only explanation for a
graduate’s decision not to return to the home canton, but that other
elements also influence the decision (the location of the partner’s
social ties and his/her residential aspirations).
Analysis of the results confirms that the family background in
which graduates grew up has an important influence on migration. All
other things being equal graduates whose fathers spent their youth in
Jura return more often than others (þ182.5%); this could be explained
by the fact that the parents’ region of origin induces a stronger
attachment to the canton (due to the presence of the wider family, for
example). Graduates whose fathers did not attend a higher education
institution are also more likely to return (þ75.6%). According to the
literature discussed earlier, parents’ level of education may be asso-
ciated with a certain habitus, according to their social class, notably in
the choice of field of study and employment sector as well as in terms
of residential aspirations and construction of rurality. Further in-
depth interviews would be needed to test these hypotheses.
Finally, it is worth highlighting the importance of the sector in
which graduates are employed. Two sectors in particular stand out:
teaching greatly facilitates a return (þ320% in comparison with
graduates without any paid occupation), whereas working in
higher education makes it almost impossible (96.5%). Differences
are not significant, however, between employees in the public
18
The reported model is restricted to the variables whose effect on migration
behaviour is significant. Given the problem of the sample size for some modalities
with a very limited number of observations, it has been decided to remove the
other variables.
131
sector (outside teaching), in non-governmental organisations, in
the private sector (including self-employment) and in the health
sector and social services.19 Other tests have been carried out
(notably with the variables previously presented in the bivariate
analyses), but the results were identical, that is to say a significant
impact for teaching and higher education but not of the other
employment sectors.
It might seem surprising that variables related to the economic
structure do not stand out more clearly in the analysis, given the
restricted size of the local labour market in Jura. An explanation
could partly reside in the definition of the categories, which are
quite heterogeneous. For example, working in the industrial sector
or in services can refer to very diverse situations in terms of edu-
cation, required skills and career perspectives. Despite this reser-
vation, the analysis shows a central finding: when controlling for a
set of variables, the life course position and the family background
are crucial factors in migration behaviour.
5. Conclusion
In a context characterised by wider access to higher education
and in which human capital is seen as a crucial factor in regional
development, the migration behaviour of young university gradu-
ates has become of utmost importance. This paper has addressed
the migration flows and counter flows of graduates in the case of a
rural region in Switzerland (Jura) and compared the profile of the
graduates who returned to their home region (40%) with that of the
ones who did not (60%). Based on a tailor-made questionnaire, it
analysed the propensity to return to the home region according to
various variables reflecting the biography of graduates.
The results highlight that young graduates’ migrations are se-
lective, although the seeming homogeneity of this population
group (same age range, same level of education, same home region)
means that it is important to take into account both dominant flows
and counter flows. Analysis of the results indicates regularities and
suggests that the propensity to return differs according to several
characteristics referring to graduates’ socio-familial, migration and
professional trajectories.
The propensity to return to Jura does not vary according to labour
market-related variables only. Even though the definitions of the
fields of study and employment sectors are quite broad and may
consequently hide some differences in the propensity to return, it is
interesting to observe the central role played by variables such as life
course position and family background. These results give a first
indication that graduates’ migration motivations are related to a
diversity of logics and are more complex than has been implied by
traditional (economic) interpretations, which have considered in-
ternal migration only as the result of a spatial mismatch in the labour
market (Fielding, 1992a; Haartsen and Thissen, 2013; Halfacree,
2004; Lundholm et al., 2004; Rérat, 2013a; 2013b; Sage et al., 2013).
The decision to migrate appears to be embedded in the life
course of an individual and to depend on his/her partner (level of
education, home region), whether or not they have children, and
their family background (parents’ socio-economic status and
migration history). The behaviour associated with some life do-
mains (e.g. the choice of a partner, plans to start a family, etc.) has a
clear impact on the propensity to return to the home region, even
when their separate impact is controlled. Migration appears by the
same token to be a choice that may be constrained by structures
that are both objective (labour market and job opportunities) and
19
This category gathers civil servants, employees and self-employed people,
which is the reason why it has been kept.
132 P. Rérat / Journal of Rural Studies 35 (2014) 123e132

Table 4 Fielding, A., 1992a. Migration and culture. In: Champion, A., Fielding, A. (Eds.),
Logistic regression of the graduates’ propensity to return to the home region. Migration Processes and Patterns, Research Progress and Prospects, vol. 1.
Belhaven Press, London, pp. 201e212.
exp(B) Sig. Fielding, A., 1992b. Migration and social mobility: south east England as an esca-
lator region. Reg. Stud. 26, 1e15.
Children
Findlay, A., Mason, C., Houston, D., McCollum, D., Harrison, R., 2009. Escalators,
No Ref.
elevators and travelators: the occupational mobility of migrants to South-East
Yes 3.076 *** England. J. Ethn. Migr. Stud. 35, 861e879.
Family status Glaser, W., 1978. The Brain Drain: Emigration & Return. Pergamon Press, Oxford.
Does not live in couple Ref. Gordon, I., 1995. Migration in a segmented labour market. Trans. Inst. Br. Geogr. 20,
Partner from Jura, without HE degree 3.569 *** 139e155.
Partner from Jura, with HE degree .942 n.s. Gotman, A., 1999. Géographies familiales, petites migrations et générations. In:
Partner not from Jura, without HE degree .163 *** Bonvalet, C, Gotman, A., Grafmeyer, Y. (Eds.), La famille et ses proches: l’amé-
Partner not from Jura, with HE degree .147 *** nagement des territoires. INED/Presses universitaires de France, Paris.
Father’s level of education Green, A.E., 1995. The geography of dual career households: a research agenda and
Higher education degree Ref. selected evidence from secondary data sources for Britain. Int. J. Popul. Geogr. 1,
Other 1.756 ** 29e50.
Father’s region of origin Halfacree, K., 2004. A Utopian imagination in Migration’s Terra incognita?
Acknowledging the non-economic worlds of migration decision-making. Popul.
Outside Jura Ref.
Space Place 10, 239e253.
Jura 2.825 ***
Halfacree, K.H., Rivera, M.J., 2012. Moving to the countryside. and staying: lives
Employment sector beyond representations. Sociol. Rural. 52, 92e114.
Not employed Ref. Haartsen, T., Thissen, F., 2013. The successefailure dichotomy revisited: young
Public sector 1.392 n.s. adults’ motives to return to their rural home region. Children’s Geogr. http://
Private sector 1.166 n.s. dx.doi.org/10.1080/14733285.2013.850848.
Teaching 4.200 ** Heinz, W.R., Huinink, J., Weymann, A. (Eds.), 2009. The Life Course Reader. In-
Higher education .035 *** dividuals and Societies Across Time. Campus, Frankfurt.
Health and social work 1.252 n.s. Huinink, J., 2009. Linked lives, families, and intergenerational relations. In:
NGOs 2.366 n.s. Heinz, W.R., Weymann, A., Huinik, J. (Eds.), The Life course reader: Individuals
Constant .694 n.s. and Societies across Time. Campus, Frankfurt am Main, pp. 303e310.
Jamieson, L., 2000. Migration, place and class: youth in rural area. Sociol. Rev. 48,
Note: n ¼ 447, R2 ¼ .33 (Cox & Snell), .44 (Nagelkerke). Model c2 (8) ¼ 176.01**. 203e223.
n.s. ¼ non significant; *p < .1; **p < .05; ***p < .01. King, R., Shuttleworth, I., 1995. The emigration and employment of Irish graduates:
the Export of high-quality labour from the periphery of Europe. Eur. Urban Reg.
Stud. 2, 21e40.
Labrianidis, L., Vogiatzis, N., 2012. Highly skilled migration: what differentiates the
subjective (socialisation to spatial and social mobility, and to “brains” who are drained from those who return in the case of Greece? Popul.
rurality, etc.) (Rye, 2011). Space Place (online first).
On the whole, the results show the relevance and the importance Li, F.L.N., Findlay, A., Jowett, A.J., Skeldon, R., 1996. Migrating to learn and learning to
migrate: a study of the experiences and intentions of International student
of adopting a life course perspective when looking at migration
migrants. Int. J. Popul. Geogr. 2, 51e67.
(Courgeau, 1985; Mulder and Clark, 2002). Researchers should take Lundholm, E., Garvill, J., Malmberg, G., Westin, K., 2004. Forced or free movers? the
into account the three interdependences mentioned earlier (Heinz motives, voluntariness and selectivity of interregional migration in the Nordic
countries. Popul. Space Place 10, 59e72.
et al., 2009): the interdependence between the past and the pre-
Mulder, C.H., Clark, W.A.V., 2002. Leaving home for college and gaining indepen-
sent (thus considering, for example, the family’s migration history dence. Environ. Plan. 34, 981e999.
and the current place of residence); the interdependence between Ravenstein, E.G., 1876. The Birthplace of the People and the Laws of Migration.
different spheres of action (family, migration, education, work) and Geogr. Mag. 3.
Rérat, P., 2012a. Gentrifiers and their choice of housing: characteristics of the house-
the interdependence between individual actions and structures holds living in new developments in Swiss cities. Environ. Plan. 44, 221e236.
(both objective and subjective). These observations open new di- Rérat, P., 2012b. The new demographic growth of cities: the case of reurbanisation
rections of research in migration studies that go beyond the case of in Switzerland. Urban Stud. 49, 1107e1125.
Rérat, P., 2014. Highly qualified rural youth: why do young graduates return to their
graduates coming from rural regions and could be adopted in Home region? Children’s Geogr. 12 (1), 70e86.
research into other migration phenomena as well. Rérat, P., 2013b. Après le diplôme. Les parcours migratoires au sortir des hautes
écoles. Editions Alphil e Presses universitaires suisses, Neuchâtel.
Rye, J.F., 2011. Youth migration, rurality and class: a Bourdieusian approach. Eur.
References Urban Reg. Stud. 18, 170e183.
Sage, J., Evandrou, M., Kalkingham, J., 2013. Onwards or homewards? Complex
Baron, M., Perret, C., 2006. Bacheliers, étudiants et jeunes diplômés: quels systèmes graduate migration pathways, well-being, and the “Parental safety net”. Popul.
migratoires régionaux? Esp. Géogr. 1, 44e62. Space Place. http://dx.doi.org/10.1002/psp.1793.
Belfield, C., Morris, Z., 1999. Regional migration to and from higher education in- Schmidlin, S., 2007. Regionale Abwanderung von jungen Hochqualifizierten in der
stitutions: scale, determinants and outcomes. High. Educ. Q. 53, 240e263. Schweiz: Emprische Analyse der Hochschulabsolventenjahrgänge 1998 bis
Bjarnason, T., Thorkindsson, T., 2006. Should I stay or should I go? Migration ex- 2004. Office fédéral de la statistique, Neuchâtel.
pectations among youth in Icelandic fishing and farming communities. J. Rural SFSO, 2012. Statistical Yearbook of Switzerland 2012. NZZ librio & Swiss Federal
Stud. 22, 290e300. Statistical Office, Zurich & Neuchatel.
Bourdieu, P., Passeron, J.-C., 1977. Reproduction in Education, Society and Culture. Smith, D.P., Rérat, P., Sage, J., 2014. Youth migration and spaces of education. Chil-
Sage, London. dren’s Geogr. 12 (1), 1e8.
Boyle, P., Halfacree, K., Robinson, V., 1998. Exploring Contemporary Migration. Smith, D.P., Sage, J., 2014. The regional migration of young adults in England and
Longman, Essex. Wales (2002e08): a ‘conveyor-belt’ of population redistribution? Children’s
Caro, P., 2006. La dimension spatiale des systèmes formation-emploi. Esp. Géogr. 35, Geogr. 12 (1), 102e117.
223e240. Stockdale, Aileen, 2006. Migration: Pre-requisite for Rural Economic Regeneration?
Chazal, S., Guimond, S., 2003. La théorie de la dominance sociale et les choix Journal of Rural Studies 22, 354e366.
d’orientation scolaire et de rôles sociaux des filles et des garçons. L’orientation Stockdale, A., Catney, G., 2012. A life course perspective on urban-rural migration:
sc. prof. 32, 595e616. the importance of the local context. Popul. Space Place (online first).
Corbett, M., 2007. All kinds of potential: women and out-migration in an Atlantic Suter, C., Perrenoud, S., Lévy, R., Kuhn, U., Joye, D., Gazareth, P., 2009. Swiss Social
Canadian coastal community. J. Rural Stud. 23, 430e442. Report 2008. Switzerland Measured and Compared. Seismo, Zurich.
Corcoran, J., Faggian, A., McCann, P., 2010. Human capital in remote and rural Thissen, F., Droogleveer Fortuiln, J., Strijker, D., Haartsen, T., 2010. Migration in-
Australia: the role of graduate migration. Growth Change 41, 192e229. tentions of rural youth in the Westoek, Flanders, Belgium and in the Veenko-
Courgeau, D., 1985. Interaction between spatial mobility, family and career life loniën, The Netherlands. J. Rural Stud. 26, 428e436.
cycle: a French survey. Eur. Sociol. Rev. 1, 139e162. Venhorst, V., Van Dijk, J., van Wissen, L.J.G., 2010. Do the best graduates leave the
Faggian, A., McCann, P., Sheppard, S., 2006. An analysis of ethnic differences in UK peripheral areas of The Netherlands. Tijdschr. Econ. Soc. Geogr. 101, 521e537.
graduate migration behaviour. Ann. Reg. Sci. 40, 461e471.