Competition for talent: retaining graduates in the Euregio Meuse-Rhine

ABSTRACT

Graduates are considered a convenient source of human capital in today’s knowledge-based economy. It is therefore crucial to understand what drives their mobility intentions to retain larger numbers of graduates. This is particularly true for peripheral regions, which need to compete with economic centres that are assumed to be more attractive. This paper adds a euregional perspective to the existing literature on graduate migration by investigating whether or not students intend to stay in the Euregio Meuse-Rhine (EMR) after graduation. It takes into account the role of hard and soft locational factors, social factors as well as individual characteristics in shaping future graduates’ mobility preferences. Using survey data from 2015 from five higher education institutions in the EMR, this paper finds that mobility intentions are determined by students’ perceptions of the quality of life, openness and career opportunities in the euroregion. In addition, distance to the partner and other social ties such as family and friends influence migration intentions.

KEYWORDS:

• Euroregion
• prospective graduates
• highly skilled migration
• human capital
• talent

Introduction

Universities play a substantial role in delivering human capital to regions (Glaeser, Kolko, & Saiz, 2001), which is why graduates are regarded as the ideal highly skilled individual to retain. Graduates not only share their new-acquired knowledge in the labour market, but can also contribute to the regional economy by residing in the region, by enjoying local consumption goods and by socially participating in the society. Multiple factors contribute to graduates’ decision to choose a certain residency (Faggian & McCann, 2009; Suter & Jandl, 2008; Venhorst & Cörvers, 2017). While hard factors such as labour market opportunities often show to be crucial in location decisions, soft locational factors such as the availability of amenities and social factors such as distance to friends and family receive increasing attention in migration theory (Florida, 2002, 2003; Musterd, Bontje, & Rouwendal, 2016; Sleutjes, 2013).

Founded in 1976, the Euregio Meuse-Rhine (EMR) is among the oldest euroregions (Perkmann, 2003). It consists of five sub-regions covering three European countries: the Southern part of the Dutch Province of Limburg, the German Zweckverband Region Aachen, the German-speaking community of Belgium and the Belgian provinces of Limburg and Liège (Euregio Maas-Rhein, n.d.). In its current strategy, EMR2020, the Euregio Meuse-Rhine Foundation (2013) recognizes the need to further integrate the euregional labour market and to prevent brain drain. There is a web of knowledge-leading educational institutions within the EMR. Graduates and the networks between universities, industries and governments are a potential source to innovation and to regional development (Harris, 1997). The almost 100,000 students studying in the EMR (Stadt Aachen, n.d.) translate into an enormous potential of human capital. Hence, the integration of graduates into the euregional labour market is one way to reach the EMR’s goals of a balanced labour market and the prevention of brain drain. The sub-regions of the EMR differ in labour supply and demand which, in search of employment, contributes to a larger pool of available jobs matching one’ s education. The EMR therefore potentially offers a larger economic functional search area (Centraal Bureau voor de Statistiek [CBS], 2015, 2017; Hensen, de Vries, & Cörvers, 2009; Planbureau voor de Leefomgeving [PBL], 2015).

This paper aims at answering the research question: What determines the preferences of prospective graduates to remain living in the Euregio Meuse-Rhine after graduation? Special attention is given to whether hard locational, soft locational or social factors are the main drivers motivating their preferences. The societal relevance of arriving at a better understanding of why graduates migrate can help make the EMR more attractive for graduates and facilitate their integration into the euregional labour market to foster economic growth and increase the region’s competitiveness (Plöger & Weck, 2014). The sub-regions of the EMR show a mixed picture of demographic development. For example, the Belgian part of the EMR shows a remarkable and positive population development, whereas the Southern part of the Dutch Province of Limburg experiences negative population growth (CBS, 2017). Retaining graduates can mitigate negative population developments, which pose serious challenges to some parts of the EMR (Elzerman & Bontje, 2015).

This paper uniquely extends the academic literature by adding a euregional dimension. Research on euroregions has experienced a recent upsurge. There is a substantial interest in cross-border cooperation and the factors conditioning its challenges and success (Medeiros, 2011; PBL, 2015; Perkmann, 2003). To the best of our knowledge this is the first study that covers the mobility preferences of prospective university graduates in a euregional perspective rather than studying it in a system of bordered national areas. We find that mobility intentions are determined by individuals’ perceptions of the quality of life, the openness (i.e. ethnic diversity, tolerance and ease of making contacts with locals) and the career opportunities in the EMR. In addition, distance to the partner and other social ties influence prospective graduates’ migration preferences.

Literature review and theoretic background

The determinants of human migration have been subject to research since the late nineteenth century. ‘Conventional wisdom holds that migration is driven by geographical differences in income, employment and other opportunities’ (Castles, De Haas, & Miller, 2014, p. 25). The move towards a knowledge-based economy results in an increasing global competition for human capital and encouraged many scholars to examine the determinants of highly skilled migration. The existing literature includes research on different groups of highly skilled migrants (Mahroum, 2000) from and within different countries (Venhorst, 2013). Since higher education graduates are considered a central source of human capital, many scholars have analysed the determinants of graduate migration in different European countries (Coniglio & Prota, 2008; Faggian & McCann, 2009; Haapanen & Tervo, 2012; King & Shuttleworth, 1995; Sykes, 2012; Van Wissen, Van Dijk, & Venhorst, 2011; Venhorst, Van Dijk, & Van Wissen, 2010; Venhorst, 2013) and the US (Hansen, Ban, & Huggins, 2003). Despite the vast amount of literature on graduate migration, no specific theory explaining migration of recent or prospective graduates specifically has been developed.

Based on recent research, determinants of migration can be clustered in four factors: hard locational factors, soft locational factors, social factors and individual characteristics. The four factors should not be seen as mutually exclusive, but rather as complementary (Castles et al., 2014; Massey et al., 1993). Migration is in this sense caused by the interplay of economic, social, cultural and political factors.

Hard locational factors

Hard locational factors are the ‘traditional economic aspects’ (Sleutjes, 2013, p. 13). They are considered the main determinants of migration in traditional migration theories such as the functionalist neoclassical theory and historical-structuralist theories. While the neoclassical theory suggests that migration results from differences in labour supply and demand across regions and that the decision to migrate is based rationally on comparing the costs and benefits of migration (Arango, 2000; Massey et al., 1993; Todaro, 1969), the historical-structuralist approach holds that individuals are not free to choose to migrate due to structural limitations (Castles et al., 2014; De Haas, 2010). Both schools are criticized for not capturing the complexity of different factors determining migration sufficiently and for characterizing individuals as rather passive (Arango, 2000; Castles et al., 2014; De Haas, 2014). Focusing on graduate migration, the availability of jobs is usually considered the most important hard locational factor (Van Wissen et al., 2011). Economic factors furthermore have been found to play the most important role in retaining individuals in shrinking regions in Portugal (Guimaraes, Nunes, Barreira, & Panagopoulos, 2016). Not only facts regarding the economic situation, but also individual perceptions regarding labour market opportunities can influence mobility intentions (Davies, 2008; Pethe, Bontje, & Pelzer, 2009).

A second hard locational factor playing a role in the choice of residence is the available transportation network because of accessibility of the workplace as well as low cost of the transportation (Lawton, Murphy, & Redmond, 2013).

Thirdly, language is seen as a hard locational factor influencing migration decisions. Being proficient in the official language of the host country can be seen as a prerequisite for entering the labour force. This might be less true in work places where the lingua franca is English, for instance, in academia, international firms and organizations. Language skills are not only a prerequisite for work, but also for social life. Accordingly, individuals are more likely to move to a country or stay in a country if they speak the official language (Adserà & Pytliková, 2015; King & Shuttleworth, 1995).

Soft locational factors

In recent years, the focus of scholars shifted from hard to soft locational factors. Soft locational factors refer to the quality of a place, for instance the quality of life, the living environment and the availability of amenities including ‘lifestyle considerations such as geography, climate, leisure time activities, recreational and cultural opportunities’ (Hansen et al., 2003, p. 141). A prominent advocate of the role of amenities is Richard Florida who established the theory of the creative class. He argues that in order for a place to be attractive for the creative class, it requires the presence of the so-called ‘3 Ts’ (Florida, 2003, p.10): technology, talent and tolerance. The theory of the creative class provides a popular explanation for migration and affects policies regarding economic development world-wide including EU policies (European Commission, 2009; Nathan, 2015).

However, the theory is often criticized for not matching reality (Nathan, 2015) and for being based on US data only (Musterd & Gritsai, 2013). To address the latter, the Accommodating Creative Knowledge (ACRE) project tested the theory in 13 European cities. The project found that Florida’s theory holds for a very limited number of European cities (Musterd & Gritsai, 2013). Furthermore, it shows ‘that job opportunities and personal net­works [ … ], and not amenities, cultural environment, openness, diversity and tolerance, are decisive for attracting skilled workers’ (Musterd & Gritsai, 2013, p. 354). The latter are found to play a larger role in retaining human capital in European cities (Musterd & Gritsai, 2013). These results underline that the role of soft locational factors is highly disputed. While some scholars argue that they play an important part in the choice of residence (Florida, 2003; Hansen et al., 2003), others contradict this view (Lawton et al., 2013; Musterd & Gritsai, 2013) or see soft locational factors as secondary factors in case hard locational factors of two places are very similar (Sleutjes, 2013).

Many soft factors arguably have an effect on graduates’ migration decisions. One prominent factor is the general living environment, which refers to housing and the natural environment. This is closely linked to the aesthetic appeal of a region, referring to urban and environmental attractiveness of a place. Urban attractiveness depends on the architecture and the number of historic buildings in a region, while environmental attractiveness concerns the proximity to nature (Malens & Van Woerkens, 2005). Beyond this, the way of life and the availability of cultural and social activities, for instance, the existence of leisure and sport facilities, museums, restaurants, cafes and bars, matter in determining graduate migration choices. The role of these factors in migration decisions is often attributed to certain individual characteristics such as age. Younger migrants are expected to see a higher value in these aspects (Florida, 2002). Some scholars even argue that the future of a place depends on its attractiveness (Glaeser et al., 2001). One should note that whether or not an individual perceives a place as attractive depends on subjective evaluations and different lifestyles (Castles et al., 2014; Davies, 2008; Pethe et al., 2009; Servillo, Atkinson, & Russo, 2012 ; Sleutjes, 2013).

Social factors

Besides hard and soft locational factors, social factors are important in determining graduates’ migration behaviour (Pethe et al., 2009). Sykes (2012) finds that being close to one’s family, friends and personal relationships plays an important role in graduates’ decisions to migrate. This assumption is confirmed by Sleutjes (2014, p. 13) who states that ‘the presence of family members or friends, [or] following a partner, [ … ] are important reasons for choosing a place of residence’. Personal links ‘connect people with certain places, including places [ … ]where they were born, where they have friends, where their family is living and where they studied’ (Musterd & Gritsai, 2013, p. 348).

Individual characteristics

Individual characteristics also determine the decision to migrate (Lee, 1966). Several studies conclude that with increasing age mobility decreases. While younger individuals are more likely to move in search of their first job or a partner, older individuals are more likely to settle down because of family and accept more permanent jobs (Faggian, McCann, & Sheppard, 2007; Van Wissen et al., 2011).

Moreover, female graduates are assumed to be more mobile than male graduates because they migrate for employment possibilities as a result of possible gender discrimination (Coniglio & Prota, 2008; Faggian et al., 2007; Venhorst et al., 2010).

In addition, research shows that university graduates are more mobile than graduates from universities of applied science and that, the field of study influences graduate mobility as it determines the extent to which jobs are available in the region. For instance, graduates from the field of economics are arguably more flexible than graduates of the field of healthcare (Sauermann, De Grip, & Fouarge, 2010; Van Wissen et al., 2011; Venhorst et al., 2010).

Given our research population, higher education graduates, one has to distinguish between foreign and domestic graduates and between students studying in their home region and those studying away from home. Foreign graduates and students studying away from home are more likely to migrate after graduation (Haapanen & Tervo, 2012; Venhorst et al., 2010). One explanation is that ‘subsequent migration is highly correlated with previous migration behavior’ (Faggian et al., 2007, p. 537). Moreover, having work experience abroad increases the propensity to move (King & Shuttleworth, 1995). Differences in mobility behaviour are related to the concept of ‘home preference’ (De Haas, 2011, p. 21): ‘most people, given the choice, prefer to stay at home’ (De Haas, 2014, p. 25). Otherwise, migration rates would be much higher, taking into consideration the world-wide economic inequalities. This assumption underlines the role of human agency, which has been downplayed by traditional migration theory (De Haas, 2014). De Haas (2014, p. 33) defines ‘human mobility as peoples’ capability to choose where to live’. It is important to stress that individuals also have the option not to move. This choice is called ‘voluntary immobility’ (De Haas, 2014, p. 26). Another concept which is closely related to that of the ‘home preference’ is that of ‘regional familiarity’, which can be acquired by residing and studying in a certain region (Venhorst, 2013). Venhorst (2013, p. 118) argues that ‘regional familiarity [ … ] appears to play an important role in work and/or residential destination choice’ of recent graduates.

Hypotheses

The main aim of this study is to identify the major determinants of prospective graduates’ migration intentions in a euregional context. Based on the reviewed literature, along the four factors distinguished above, we formulate the following hypotheses. The ordering does thereby not have implications for the relative importance of the individual hypotheses. This is because migration decisions are usually the result of a complex set of interrelated factors, which may differ from case to case.

Hard locational factors

1. A positive perception of career opportunities and the transport networks within the EMR increases the likelihood that prospective graduates intend to stay in the EMR after finishing education.

Soft locational factors

1. A positive perception of the quality of life in the EMR increases the likelihood that prospective graduates intend to stay in the EMR.

2. A positive perception of the openness in the EMR increases the likelihood that prospective graduates intend to stay in the EMR.

Social factors

1. With increasing distance to relatives, friends and partner, the likelihood that prospective graduates intend to stay in the EMR decreases.

Individual characteristics

1. We expect mobility intention to decrease with increasing age, to be higher for female graduates, to increase with an increasing level of education and to be higher for university graduates than for graduates from universities of applied sciences.

2. We hypothesize that graduates who were born in the EMR are more likely to stay and that mobility increases with previous migration experience.

Data and methodology

This study uses survey data from 2015 collected at five higher education institutions in the EMR to learn more about mobility intentions of prospective graduates.¹ The data collected between June and October 2015 includes information on 3328 individuals (RWTH Aachen = 1.926, FH Aachen = 409, UHasselt = 307, UM = 346, Zuyd Hogeschool = 340). This study targets prospective graduates, that is to say students who are in the final phase of their studies assuming that they are more likely to have thought about their future residency compared to students at the beginning of their studies. This restricts our sample to 1211 individuals (RWTH Aachen = 694, FH Aachen = 115, UHasselt = 88, UM = 117, Zuyd Hogeschool = 197).

Variables

The dependent variable investigated in this study has three possible outcomes, indicating if students intend (1) to stay in the EMR after graduation, (2) leave the EMR after finishing higher education or (3) are uncertain about their future residency.

The independent variables used to explain the mobility intentions of graduates are divided into four groups: hard locational factors, soft locational factors, social factors and individual characteristics.

Hard locational factors

We include four independent variables related to respondents’ perceptions or subjective evaluations of hard locational factors. Two binary variables indicate if respondents have a positive view of the career opportunities and the transport network in the EMR. A third binary variable specifies if respondents regard proficiency in the official language important when choosing their residency. Furthermore, English language proficiency is used as a binary variable.

Soft locational factors

Additionally, we account for three different independent variables referring to respondents’ perception of soft locational factors. Respondents were asked to indicate their view on soft locational factors in the region where they study, namely the living environment, cultural and social activities, aesthetic appeal of the region, way of life, ethnic diversity, tolerance and ease of making contacts with locals. Based on a factor analysis, these variables were clustered into two variables. The first, which is referred to as the ‘view on the quality of life’ refers to the living environment, cultural and social activities, the aesthetic appeal of the region and the way of life. The second combines ethnic diversity, tolerance and ease of making contacts with locals and is called the ‘view on the openness’. Both are binary variables indicating if respondents have a positive view. Moreover, we consider if respondents find the quality of life important when choosing the place of residence. Based on a factor analysis, the binary variable combines the importance of the living environment, cultural and social activities, the aesthetic appeal of the region and the way of life.

Social factors

Two independent variables refer to the role of social factors in graduates’ migration decision. First, respondents were asked to indicate how important friends and family are when choosing their residency. Based on a factor analysis the answers from both questions were merged into one binary variable called ‘social ties’ indicating if respondents find social ties important in choosing their place of residence. Second, the survey contains information on the relationship status of the respondents and where their partner lives. This information has been merged into one variable called ‘distance to partner’. There are four possible categories, namely the reference category ‘living together with the partner’, ‘partner living in the same city or region’, ‘partner living in the same or another country’ and ‘no partner’.

Individual characteristics

Individual aspects relate to gender; age; place of birth and residence, distinguishing between places inside and outside the EMR. In addition, we use individual data related to education. We consider if respondents study at a university or a university of applied sciences. Moreover, we take into account if respondents pursue a Bachelor’s degree, a Master’s degree, or a PhD, whereby Bachelor’s students are used as the reference group. Additionally, we distinguish between the best 25 per cent of students and the rest, based on their current grade point average. Furthermore, fields of study are taken into consideration. We distinguish between social sciences; business and economics; law; science, maths and computing; engineering, manufacturing and construction; health and welfare; as well as behavioural science and life science. A binary variable specifies if respondents plan to continue studying or do a PhD or if they have other plans. Three independent variables are linked to the previous migration of the respondents: First, a continuous variable indicates how often the respondents changed their place of residence since their 16th birthday. Second, a binary variable shows if respondents moved for their studies. Third, a categorical variable specifies if respondents have working experience abroad, have study experience abroad or if they have other experiences abroad.

Methodology

The analysis is divided into two parts. To start, we distinguish between stayers, leavers and uncertain respondents, exploring differences between the three groups based on a descriptive analysis of individual characteristics factors, the importance of different locational factors and the respondents’ perceptions of these factors in the EMR. In this preliminary analysis, we use chi-squared and one-way analysis of variance (ANOVA) tests to see whether there are differences between the combination of all three outcomes stay, leave and being uncertain based on these factors (see Table 1). In the second part, we estimate a multinomial logistic regression model using a stepwise approach. We distinguish between four different groups of variables throughout the models: (1) individual characteristics, (2) hard locational factors, (3) soft locational factors and (4) social factors.

Table 1. Descriptive statistics of variables of interest.

		Leavers	Stayers	Uncertain	All respondents
Background information
Mean age**		24.74	24.95	25.20	24.98
Female***		46.08%	53.70%	41.87%	46.41%
Born in EMR***		18.25%	62.96%	27.64%	34.02%
Living in EMR***		71.39%	98.15%	84.96%	84.06%
Moved for studies***		83.80%	52.78%	78.46%	73.33%
Mean previous migration***		2.08	1.42	1.76	1.72
Partner***		69.11%	78.09%	58.74%	67.30%
University***		75.70%	65.43%	78.86%	74.24%
Degree***	BA	39.75%	49.38%	34.76%	40.30%
	MA	53.92%	45.06%	56.91%	52.77%
	PhD	6.33%	5.56%	8.33%	6.94%
Top 25%*		30.08%	22.61%	26.72%	26.73%
Important when choosing residency
Social ties***		66.58%	83.33%	76.02%	74.90%
Quality of life**		91.14%	85.19%	87.20%	87.94%
Openness**		68.35%	58.95%	67.28%	65.40%
(International) Transport		54.68%	47.22%	51.22%	51.28%
Career opportunities**		84.05%	81.17%	88.62%	85.14%
Language proficiency***		57.72%	72.22%	59.55%	62.35%
Good view of factor in EMR
Quality of life***		75.95%	92.90%	88.01%	85.38%
Openness***		70.13%	84.88%	81.71%	78.78%
(International) Transport**		61.27%	70.06%	66.06%	65.57%
Career opportunities**		34.18%	44.14%	39.02%	38.81%
N		395	324	492	1211

*p < .1; **p < .05; ***p < .01.

Note: Chi²- tests were applied for binary variables and one-way analysis of variance (ANOVA) tests and a subsequent Bartlett’s tests for continuous variables.

Results

Descriptive analysis

Examining the migration intentions of prospective graduates studying in the EMR descriptively, we find that the share of respondents intending to stay is the smallest (26.75 per cent, N = 324) compared to the share of those intending to leave (32.62 per cent, N = 395) and the share of those who are still uncertain (40.63 per cent, N = 492). This shows that even towards the end of their studies a majority has not made up their mind about their plans after graduation. The most important reason for respondents’ migration intentions are work (50.73 per cent), the living environment (14.64 per cent) and the partner (17.15 per cent). About 35.19 per cent of potential stayers intend to move internally. In most cases, they intend to move within the sub-region where they currently live. The same is found for the uncertain respondents. If they stay they are more likely to remain living in the same sub-region. Hence, there is very little movement between the individual sub-regions within the EMR. In addition, the largest share of respondents who intend to move within the EMR plan to work or pursue a PhD after graduation (77.19 per cent). A possible explanation for their internal movement is that they will be able to afford better housing once they earn money.

Table 1 presents descriptive statistics of some variables of interest. Regarding individual characteristics, the average age is the only variable, which does not show statistically significant differences between the three groups, which might be related to our selected group of young people. Comparing stayers, leavers and uncertain individuals, we find in line with earlier research that among stayers there are larger shares of respondents who have been born in the EMR (62.96 per cent), who are living inside the EMR while studying (98.15 per cent), who have a partner (78.09 per cent) and who are pursuing a Bachelor’s degree (49.38 per cent). In contrast, larger shares of leavers moved before for studying (83.80 per cent) and belong to the top 25 per cent of students (30.08 per cent). On average, leavers have more previous migration experience (2.08 times). Among the uncertain group we find the largest share studying at university level (75.70 per cent), pursuing a Master’s degree (56.91 per cent) or a doctorate (8.33per cent).

Additionally, Table 1 presents the shares of respondents, who rate various locational factors important when choosing their residency. A relative larger share of stayers considers social ties (83.33 per cent) and proficiency in the official language (72.22 per cent) as important. A relative larger share of leavers consider the quality of life (91.14 per cent), openness (68.35 per cent) and transport network important. A relative larger share of uncertain respondents finds career opportunities (88.62 per cent) important. With the exception of the importance of the transportation network, stayers, leavers and uncertain individuals differ significantly from each other.

Finally, Table 1 shows the shares of stayers, leavers and uncertain individuals, who have a positive view on different locational factors in the EMR. As expected, we find that stayers have the most positive and leavers the least positive views on all four categories: quality of life, openness, transport network and career opportunities. The difference in proportion between all three groups is statistically significant. There are no statistically significant differences between the individual sub-regions of the EMR regarding the descriptive statistics.

Regression analysis

To analyse the impact of the different determinants simultaneously, we estimated a multinomial logistic regression² on the full sample of students from all five institutions of higher education. To test if the patterns of migration intentions of graduates are the same on all sides of the border, we also run logistic regressions for each EMR sub-region separately. No notable differences are found between the individual sub-regions of the EMR. This points to the interesting fact that the same processes play a role in determining the migration preferences of prospective graduates on all sides of the border. Hence, programmes at a euregional level to retain prospective graduates to the euroregion can be recommended.

Tables 2 and 3 present the results of the multinomial logistic regression, including relative risk ratios (RRR), robust standard errors (RSE) and statistical significance of the independent variables. The baseline for the dependent variable is the intention to leave the EMR after finishing education.

Table 2. Stepwise models comparing migration intentions of stayers and leavers (reference category).

	Model 1		Model 2		Model 3		Model 4		Model 5
Independent variables	RRR	RSE	RRR	RSE	RRR	RSE	RRR	RSE	RRR	RSE
Individual characteristics
Female	1.34*	0.23	1.30	0.24	1.23	0.24	1.14	0.22	0.91	0.19
Age	1.08***	0.03	1.13***	0.04	1.13***	0.04	1.13***	0.04	1.10**	0.04
Living in the EMR	13.03***	5.77	15.79***	7.17	18.1***	8.38	18.56***	8.69	19.78***	9.54
Born in the EMR	4.43***	0.97	4.33***	1.00	4.25***	0.99	4.39***	1.03	3.89***	0.93
Not moved for studies	1.90***	0.46	1.69**	0.45	1.81**	0.49	1.74**	0.47	1.72**	0.48
Previous migration	0.95	0.07	1.00	0.08	1.04	0.08	1.06	0.08	1.09	0.09
Institution
University			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
University of applied sciences			1.43	0.36	1.46	0.37	1.58*	0.42	1.39	0.38
Degree
Bachelor’s			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Master’s			1.48*	0.35	1.57*	0.38	1.53*	0.37	1.44	0.36
PhD			0.65	0.31	0.73	0.35	0.64	0.31	0.51	0.25
Top 25% students			0.78	0.16	0.84	0.17	0.80	0.17	0.76	0.16
Job as stepping stone towards future career			1.19	0.23	1.16	0.23	1.19	0.24	1.18	0.24
Field of studies
Social Sciences			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Business & Economics			1.43	0.53	1.39	0.54	1.51	0.59	1.68	0.68
Law			2.57*	1.26	2.42*	1.24	2.84**	1.48	3.13**	1.72
Science, Maths & Computing			1.45	0.51	1.23	0.45	1.39	0.52	1.60	0.61
Engineering, Manufacturing & Construction			1.08	0.35	0.98	0.33	1.07	0.37	1.04	0.37
Health & Welfare			3.17**	1.50	2.85**	1.40	4.10***	2.05	3.97***	2.03
Behavioural Science & Life Science			1.53	0.75	1.39	0.69	1.60	0.81	1.77	0.91
Experience abroad
None			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Working			0.43***	0.14	0.49**	0.16	0.51**	0.17	0.52**	0.17
Schooling			0.74	0.18	0.76	0.18	0.72	0.18	0.73	0.19
Other			0.49**	0.16	0.54*	0.18	0.58	0.19	0.62	0.22
Continue studying			2.15***	0.45	2.19***	0.46	2.20***	0.47	2.34***	0.52
Hard locational factors
View on the career opportunities					1.66***	0.32	1.41*	0.28	1.42*	0.29
View on the transport system					1.32	0.25	1.04	0.20	0.96	0.19
Language proficiency important					1.65***	0.31	1.68***	0.33	1.69***	0.34
Fluent in English					0.93	0.20	0.94	0.21	1.05	0.24
Soft locational factors
Quality of life important							0.56**	0.15	0.53**	0.15
View on the quality of life in the EMR							3.24***	0.92	2.94***	0.86
View on the openness in the EMR							1.91***	0.45	2.15***	0.52
Social factors
Distance to partner
Living together									Ref.	Ref.
Same city/region									0.91	0.23
Same/other country									0.23***	0.08
No partner									0.50***	0.12
Social ties important									2.07***	0.47
Constant	0.00***	0.00	0.00***	0.00	0.00***	0.00	0.00***	0.00	0.00***	0.00
N	1172		1172		1172		1172		1172
McFadden’s Pseudo R^2	10.1%		12.0%		12.8%		14.6%		17.4%
AIC	2310.7		2322.9		2319.8		2285.4		2230.7
BIC	2381.6		2545.8		2583.2		2579.2		2565.1

Notes: Base outcome: leave.

*p < .10, **p < .05, ***p < .01.

Table 3. Stepwise models comparing migration intentions of leavers (reference category) and undecided.

	Model 1		Model 2		Model 3		Model 4		Model 5
Independent variables	RRR	RSE	RRR	RSE	RRR	RSE	RRR	RSE	RRR	RSE
Individual characteristics
Female	0.92	0.13	0.90	0.14	0.89	0.14	0.84	0.14	0.87	0.15
Age	1.07***	0.03	1.08***	0.03	1.09***	0.30	1.08***	0.03	1.09***	0.04
Living in the EMR	1.87***	0.34	2.04***	0.39	2.11***	0.41	2.12***	0.42	2.25***	0.46
Born in the EMR	1.49**	0.29	1.47*	0.30	1.47*	0.30	1.49*	0.31	1.41	0.30
Not moved for studies	1.11	0.23	1.09	0.25	1.11	0.26	1.09	0.26	1.05	0.25
Previous migration	0.92	0.05	0.95	0.05	0.96	0.06	0.96	0.06	0.99	0.06
Institution
University			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
University of applied sciences			1.09	0.24	1.11	0.24	1.15	0.26	1.08	0.25
Degree
Bachelor’s			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Master’s			1.51**	0.30	1.52**	0.30	1.47*	0.30	1.51**	0.32
PhD			1.18	0.46	1.19	0.48	1.06	0.42	1.16	0.47
Top 25% students			0.83	0.14	0.85	0.14	0.84	0.15	0.86	0.15
Job as stepping stone towards future career			0.98	0.16	0.95	0.15	0.98	0.16	1.00	0.17
Field of studies
Social Sciences			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Business & Economics			1.38	0.46	1.38	0.46	1.42	0.48	1.46	0.51
Law			1.91	0.84	1.88	0.84	20.7	0.95	1.99	0.96
Science, Maths & Computing			1.26	0.40	1.19	0.38	1.27	0.42	1.30	0.43
Engineering, Manufacturing & Construction			1.08	0.32	1.05	0.32	1.11	0.35	1.05	0.33
Health & Welfare			1.50	0.68	1.41	0.65	1.79	0.85	1.66	0.80
Behavioural Science & Life Science			1.80	0.78	1.72	0.75	1.87	0.83	1.91	0.85
Experience abroad
None			Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.	Ref.
Working			0.78	0.20	0.78	0.20	0.81	0.20	0.81	0.21
Schooling			0.77	0.14	0.75	0.14	0.73	0.14	0.68*	0.14
Other			0.76	0.18	0.75	0.19	0.79	0.20	0.77	0.20
Continue studying			1.38*	0.24	1.38*	0.24	1.36*	0.24	1.41*	0.26
Hard locational factors
View on the career opportunities					1.23	0.19	1.11	0.18	1.11	0.18
View on the transport system					1.22	0.18	1.03	0.16	0.99	0.16
Language proficiency important					1.08	0.16	1.07	0.16	1.03	0.16
Fluent in English					1.03	0.18	1.04	0.19	1.06	0.20
Soft locational factors
Quality of life important							0.70	0.17	0.67*	0.16
View on the quality of life in the EMR							2.09***	0.44	2.01***	0.43
View on the openness in the EMR							1.43*	0.27	1.51**	0.30
Social factors
Distance to partner
Living together									Ref.	Ref.
Same city/ region									1.28	0.30
Same/ other country									0.73	0.17
No partner									1.71***	0.34
Social ties important									1.65***	0.28
Constant	0.13***	0.09	0.07***	0.06	0.05***	0.04	0.03***	0.03	0.02***	0.02
N	1172		1172		1172		1172		1172
McFadden’s Pseudo R^2	10.1%		12.0%		12.8%		14.6%		17.4%
AIC	2310.7		2322.9		2319.8		2285.4		2230.7
BIC	2381.6		2545.8		2583.2		2579.2		2565.1

Notes: Base outcome: leave.

*p < .10, **p < .05, ***p < .01.

Model 1 includes information on the individual characteristics that are likely to influence the migration intention. The results show that respondents are more likely to stay with increasing age, if they are living in the EMR during their studies and if they were born in the EMR (p < .01). Similarly, not having moved for one’s studies is associated with an increased probability of staying in the region. As expected, these results are consistent throughout the different models for the group of stayers relative to the group of leavers. These results can be seen as evidence for the ‘home preference’, the effect of ‘regional familiarity’ and the hypothesis that people who have moved previously, are more likely to move again. People who were born in the EMR and reside there while studying are arguably more familiar with the region and are, therefore, more likely to stay than others. Other than expected, the estimations do not reveal a gender effect. A possible explanation for this is that while studying, respondents are not aware of possible gender discriminations in the labour market yet, or more optimistic, gender discrimination (or perception of discrimination) is less of a problem in the EMR.

Model 2 adds information regarding education and work experience. Other than expected, we find that students studying for a Master’s degree (p < .10) are more likely to stay compared to Bachelor students. However, this effect is only marginally significant and changes when adding additional variables in the following models. The same is true for respondents studying law (p < .10) or health and welfare (p < .05) relative to students of the social sciences. This is in line with earlier research (Venhorst et al., 2010, p. 524) showing that graduates of the field of health care are assumed to be less flexible and mobile. Respondents with experiences abroad, such as work (p < .01) or study experiences (p < .05), are less likely to remain living in the EMR after graduation. Those planning to continue studying are twice as likely to remain living in the EMR (p < .01).

Model 3 adds a set of hard locational factors. Respondents who have a good view on career opportunities and respondents who find the language proficiency important, are 65 per cent more likely to stay relative to those who have a negative view on career opportunities or do not find language skills important when choosing their place of residence (p < .01). Other than expected, the estimations do not reveal any effects of the transport system and proficiency in the English language.

In Model 4, we add soft locational factors. Our first two factors, the importance attached to the quality of life when choosing one’s residency and the view on the quality of life in the EMR, need to be interpreted together. We find that respondents who find the quality of life important in choosing their residency are generally 50 per cent less likely to stay (p < .05). At the same time, however, we find that students with a positive view on the quality of life are on average three times more likely to stay relative to the ones with a negative view (p < .01). This indicates that students who find the quality of life important and have a positive view on the EMR with respect to this aspect are most likely to stay whereas students who find the quality of life important but have a negative view are most likely to leave.³ In addition, students with a positive view on the openness in the EMR are on average three times more likely to stay relative to the ones with negative views (p < .01). This result regarding the openness is in line with the findings of the ACRE project which concludes that factors such as openness, tolerance and diversity play a role in retaining individuals in Europe (Musterd & Gritsai, 2013).

Model 5 finally adds social factors as a last group of variables. In comparison to respondents living together with their partner, those who have a partner living further away in the same or in another country are more likely to leave (p < .01). The same is true for the respondents without a partner (p < .01). In addition, we find that respondents who find social ties important in choosing their residency are two times more likely to stay relative to those who do not (p < .01). As the effect of the partner variable depends on where the partner lives, we suggest that the social ties variable is subject to similar circumstances. Looking at the shares of respondents who rate social ties important considering their place of birth and migration intention, we find that 53.53 per cent of the respondents who consider social ties important were born in the EMR and intend to stay after graduation. Similarly, 37.38 per cent were born elsewhere and intend to leave. These findings can be interpreted as an indicator that respondents prefer living near their social ties. The significance of the variables remains unchanged throughout the five different models. Yet, the RRR of the respondents holding a Master degree loses significance in the full model.

In addition, our multinomial logistics regression compares the uncertain group to the group intending to leave. Again, the results are similar across all five models specified which is why we focus on the results of the full model here.⁴ The results indicate that respondents are more likely to be uncertain with increasing age (p < .01), if they are currently living in the EMR (p < .01), if they are pursuing a Master’s degree (p < .05) and if they intend to continue studying (p < .10). Moreover, having a good view of the quality of life and the openness in the EMR is associated with an increased likelihood of being uncertain ((p < .01 and (p < .05 respectively). Similarly, respondents who are not in a relationship and who rate social ties important when choosing their residency are also more likely to be uncertain (p < .01). Finding the quality of life important when choosing one’s residency is associated with an increased probability of intending to leave. This effect is however only marginally significant (p < .01). An important difference between the models comparing leavers relative to stayers and the models comparing leavers relative to uncertain respondents is that in the latter variables referring to the field of studies, previous migration experience, the view on career opportunities and language proficiency are not found to have a statistically significant effect. Interestingly, the measurements testing the goodness of fit between the different models implies that Model 4 including the hard locational factors has less explanatory potential as compared to the models including soft and social factors. This contradicts the literature on graduate migration, which considers hard locational factors to be crucial in location decisions. The overall result, however, indicates that it is the interplay between the different factors explaining mobility intentions.

Conclusion and discussion

As the competition for talent increases, it becomes more and more important to understand what influences migration decisions of prospective graduates as a convenient source of human capital. In this paper, we analyse the determinants of students’ intentions to remain living in the EMR after graduation. The innovative aspect of our paper is that it applies a euregional perspective.

Even towards the end of their studies, the largest share of future graduates is still uncertain about their future residency. Of those who have decided already, a larger share intends to leave the EMR after finishing education. We find that migration intentions of prospective graduates depend on a variety of hard locational, soft locational, social and individual factors. These findings are in line with previous studies, which suggest that migration is caused by the interplay of various economic, social, cultural and political factors. Furthermore, distance to the partner and other social ties such as family and friends influence migration intentions of students studying in the EMR. In addition, migration intentions are influenced by individual perceptions of the quality of life, the openness and the career opportunities in the EMR. If students have a good view on these factors, they are more likely to stay in the EMR. In comparison, soft and social factors seem to have more explanatory power in the models than hard factors only. Our findings are in line with the ACRE project, which concludes that soft factors such as openness and tolerance are important factors in retaining individuals in the European context. The effect of individual characteristics, particularly the place of birth and place of residence underline the importance of the so-called ‘home preference’ and ‘regional familiarity’ in choosing one’s residency. Other factors which were expected to impact graduate migration intentions, such as the transport system, gender, the level of education and the type of higher education institution were not found to have an effect.

The findings are consistent between the individual sub-regions pointing to the interesting fact that the same processes play a role in determining the migration preferences of prospective graduates on all sides of the border. Hence, programmes at a euregional level to retain prospective graduates to the euroregion can be recommended.

The findings allow us to formulate additional policy recommendations to retain larger numbers of graduates in the EMR. First, one should encourage students to move to the EMR instead of commuting to university from outside the EMR. As our results show, respondents who do not live inside the EMR while studying are extremely unlikely to intend to live inside the EMR in the future. One suggestion is to provide attractive and affordable housing for students. In addition, universities should encourage their Bachelor’s students to continue studying at the same university. The longer they live and study in the region, the greater their ‘regional familiarity’ becomes and the more likely they are to stay after finishing education. This is particularly true for students who did not live in the EMR before studying. Generally, more of an effort should be made to tie students to the region if they come from outside the EMR. We acknowledge that this is a difficult undertaking considering that the results of this study underline the importance of individual perceptions on hard and soft locational factors in the retention of graduates. However, we suggest that one possible way to achieve this is to further eliminate border barriers and thereby improve the quality of life. Moreover, providing students with more information on career opportunities and opportunities related to the quality of life such as cultural and leisure time activities on all three sides of the borders might be a good idea to influence graduates’ perceptions of the EMR. Given that nearly 90 per cent of the students in our population indicate that the quality of life is an important aspect when deciding on the place of residence, keeping up the quality of life in the EMR on the high level currently achieved (Agit, n.d.; DGStat, n.d.; Zuidlimburg, n.d.) is crucial for retaining graduates in the EMR.

There are many projects which deal with these issues already. Unfortunately, it is beyond the scope of this paper to discuss them. Therefore, one suggestion for further research is to analyse which projects are in place and how effective they are in retaining graduates. An additional area for further research is to examine to what extent migration intentions were actually realized. This is relevant because migration does not necessarily take place directly after graduation and hence it does not reveal one’s mobility behaviour over time, which is relevant for regional development. Finally, this paper does not take into consideration the role of locational factors in regions the EMR competes with. It would be interesting to identify these competing regions and compare them to the EMR.

Acknowledgements

The authors would like to thank the anonymous referee and Frank Cörvers for the constructive feedback and comments. In addition, the authors would like to thank RWTH Aachen, FH Aachen, Universiteit Hasselt, Maastricht University and Zuyd University of Applied Science for supporting the implementation of our survey by inviting their students to participate.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The implementation of the survey, on which this project is based, was supported by the Research Centre for Education and the Labour Market – Maastricht University.

Notes

1 The survey was carried out by Maastricht University in cooperation with the five institutes of higher education. Students were approached through their student email account.

2 We tested the Independence of Irrelevant Alternatives assumption (IIA) of the multinomial logit model applying a Hausman-McFadden test. We run the full model including all three outcomes (leave/stay/uncertain) against the restricted model in which we exclude the group of respondents indicating to be uncertain whether to move or not. The results of the test show that the full model is the correct specification (Cameron & Trivedi, 2009).

3 Adding an interaction variable between the importance attached to the quality of life and the view on the quality of life in the EMR confirm these findings.

4 There is one exception: While in general respondents who were born in the EMR are more likely to be uncertain about their future residency, the variable loses its statistical significance when adding social factors in Model 5.