In the current discourse it is frequently stated that in the course of European integration and globalisation we witness more intensified and more frequent transactions spanning across national borders. These assumptions relate not only to transactions in the economic sphere or to forms of political co-operation, but to the individual lifeworlds as well. Yet concerning the latter dimension, research into the patterns and dynamics of interpersonal interaction, relationships, and forms of mobility across national borders is scarce. This paper is a contribution towards filling this research lacuna. It addresses the question of the extent to which horizontal Europeanisation, understood as different forms of individual cross-border activities within the European Union, depends on characteristics at the country level. From a comparative perspective we will analyse the transnational mobility and cross-border networks of European citizens within a European context. Empirically, we refer to data from the Eurobarometer 65.1 (2006), which allows us to answer the question how certain contextual variables, such as internationalisation, modernisation, and characteristics such as the respective country's geography, affect people's participation in cross-border activities. Based on a sample of 25 European countries, we will demonstrate that geographic characteristics drive social transnationalism less than internationalisation, Europeanisation, and modernisation, which contribute to a proliferation of individual transnational activities across the European Union.
Introduction
In recent decades the rise of world markets, technological revolutions in transportation and communication, and the flood of images and messages from new media have brought what was once distant and inaccessible within easy reach of the individual (Held et al.1999; Robertson 1992; Rosenau 2003). The territorial and social demarcations that characterised nation-states are rapidly disappearing; this is reflected not only in new forms of governance and economic globalisation, but in transnational social integration, mobility, affiliations and networks (Mau and Büttner 2010). Social relationships and interactions have expanded across physical space, leading to lives that are more transnational and mobile than ever. As a result of these processes scholarly interest in new forms of cross-border activities and interactions, often labelled ‘social transnationalism’ (Mau 2010), has grown strongly.
European integration as a process of regional integration and supra-national institution-building can be considered the most advanced attempt to transform the traditional nation-state. This process involves not only economic and political integration and intensified exchanges, but is likely to affect social networks and forms of mobility of individuals as well. In understanding Europe as a space for intensified cross-border interaction, this article is linked to recent writings on the sociology of Europeanisation which are interested in the impact of European integration on peoples’ lives (Díez Medrano 2008, 2011; Favell 2008; Fligstein 2008; Favell and Guiraudon 2009; Gerhards 2010a; Mau and Verwiebe 2010; Roose 2010; Andreotti and Le Galès 2011; Favell and Recchi 2011; Kuhn 2011). Key factors for a possible increase in cross-border connectivity at the individual level are the elimination of barriers to the movement of people, especially in the Schengen area, the liberalisation of the transportation and telecommunication sectors, which lowers the costs of travelling and communicating across borders, the creation of a common market, and the implementation of mobility and exchange schemes such as Erasmus. It can be assumed that Europe, more than many other parts of the world, is a space of dense cross-border interaction and intense cross-border networking, paradigmatically called ‘horizontal Europeanisation’ (Beck and Grande 2007; Mau and Verwiebe 2010). However, the growth of transnational interconnectedness in an increasingly integrating Europe is often taken for granted in political science theories of European integration. Research on the extent and degree of transnational exchanges and contacts between Europeans is still rare, and data on horizontal cross-border connections and movements is quite limited. Nonetheless, aside from ‘top-down’ political processes, transnational interactions in everyday life possess a quality of their own, and it is necessary to have a closer look at this horizontal dimension of European integration.
In this paper we will approach the issue of horizontal Europeanisation from the individual perspective. For that reason we will make use of survey data from the Eurobarometer (EB 65.1, 2006). This survey covering all the European Union member states records respondents’ interactions with people from other nationalities as well as cross-border mobility. The central aim of this article, however, is not to describe different levels of transnational involvement of the ‘average citizen’, but to explain differences between individuals living in different European societies. Thus, our primary task is to explain which specific context factors located at the national level drive (or hinder) the proliferation of cross-border activities at the individual level. We will focus on factors such as key geographic conditions (population and territory), modernisation (economic wealth and human development), and internationalisation (Europeanisation, globalisation, and migration). As it can be assumed that differences between countries also result from country-specific compositions of the national populations, we will additionally control for a set of individual-level variables. Therefore, we will use proxies for social status such as education and occupational class which are assumed to impact on the differences in the degree of ‘transnationality’ as well as socio-demographic characteristics such as age and gender (Mau and Mewes 2009).
Social transnationalism and horizontal Europeanisation
As we are interested in the expanding spaces of mobility and interaction of individuals, we will now turn to the concept of transnationalism, which provides the broad frame of this study (Sklair 2001; Faist 2004; Pries 2007; Vertovec 2009). The adjective transnational was first used in political science to designate the rise of new supra-governmental or inter-governmental forms of regulation and the growing significance of political, social and economic players beyond the framework of the nation-state (Keohane and Nye 1973). The concept was also adopted early on by the fields of economics and organisational research, which faced the problem that many of their central actors and corporations could no longer be described as national but had rather become multinational. In sociology most studies using the term ‘transnationalism’ focus on specific groups, areas or transnational activities, such as diaspora groups and ethnic minorities (Cohen 1997), migrants1 (Grasmuck and Pessar 1991; Basch et al.1994; Pries 1996), transnational social movements (Cohen and Rai 2000; Tarrow 2005), business networks (Yeung 1998), transnationally organised crime (Williams and Vlassis 2001), globalised forms of purchasing in the area of household services and social care (Anderson 2000; Ehrenreich and Hochschild 2003), or the transnational capitalist class (Sklair 1991, 2001).
In a broad sociological understanding transnationalism refers to the relations, networks, and practices across nation-state borders (Mau 2010). Ulrich Beck (2000: 32) states: ‘“Transnational” implies that forms of life and actions emerge whose inner logic comes from the inventiveness with which people create and maintain social lifeworlds and action contexts where distance is not a factor’. According to de Swaan (1995), the objective of a sociology of transnationalism is the analysis of transnational relations, i.e., those relationships which people maintain across national borders, either directly or through the intermediation of various organisations. In this view it is less international relationships, but rather the relations between the inhabitants of various nations which are of interest. By the same token Roudometof (2005: 119) suggests: ‘To go beyond the conventional understanding of transnationalism as a facet of international migration, it is necessary to conceptualize transnational interactions as taking place among people and institutions in two or more separate “containers” or nation-states’.
From this perspective, the context of European integration is particularly interesting as the EU strives to create an ‘ever closer union among the peoples of Europe’ (emphasised in the Treaties of Rome and Maastricht). In sociology, the term ‘horizontal Europeanisation’ denotes a variety of cross-border interactions between European countries in terms of communication, the exchange of ideas and meanings, collective mobilisation across borders as well as cross border mobility and networks. The latter, in particular, is rather close to the whole branch of studies on ‘social transnationalism’ (Mau 2010) which is interested in interactions between people and movement of people across borders. Taking this perspective, the focus on social relations and types of cross-border connectivity between people and social groups moves to the fore. It has been highlighted that processes of ‘horizontal Europeanisation’ in terms of contacts, interactions, and social relationships between different European countries as well as various forms of pan-European mobility, are at least as important for European integration as processes of vertical integration, meaning the establishment of supra-national institutions (Beck and Grande 2007; Favell 2007; Favell and Guiraudon 2009; Mau and Büttner 2010; Mau and Verwiebe 2010). Ever since Durkheim's (1960 [1893]) seminal work The Division of Labour in Society, the link between growing societal interdependence and solidarity has been one of the core theorems of sociological analysis. Durkheim assumed that intensified exchange and involvement engender social and emotional bonds between the interacting strangers. Adopting Durkheim's approach, it stands to reason that the more interwoven the European member-states are, the more likely bonds across national borders are to develop. In fact, Durkheim himself indicated that such a possibility might come about for European nations: ‘(…) [A]mong European peoples there is a tendency to form by spontaneous movement a European society which has, at present, some idea of itself and the beginning of organisation. If the formation of a single human society is forever impossible, a fact which has not been proved, at least the formation of continually larger societies brings us vaguely near the goal’. (Durkheim 1960 [1893]: 405f.).
Along this line Karl W. Deutsch's (1953, 1972) ‘transactionalism’ defines integration as dependent on the density and intensity of the exchange of information, capital, goods, and people (see Fligstein 2008). He emphasises that economic and political transactions between countries also bring about more close-knit exchanges between individuals. In turn, according to this theory, through the intensification of cross-border exchange, more and more people become aware of the advantages of this exchange and also participate in transnational interaction. As a result, social distance can be overcome and prejudices dispelled. Certain aspects of these processes of bottom-up Europeanisation, however, cannot be deduced from the political process of integration, since social lifeworlds have a momentum of their own. In addition, it is an open question as to whether all forms of horizontal Europeanisation necessarily serve to push forward European integration. Favell (2005: 1115), for example, argues:
Political scientists think of voting and ‘revealed preferences’, of course, but ‘being European’ nowadays is as much likely to be about this, as it is about shopping across borders, buying property abroad, handling a common currency, looking for work in a foreign city, taking holidays in new countries, buying cheap airline tickets, planning international rail travel, joining cross-national associations – and a thousand other actions facilitated by the European free movement accords.
This means that the creation of a European social space is not only about the emergence of a new political centre and shifting political loyalties, but also about intensified cross-border transactions, mobility and networks which are horizontal in nature (Favell 2008; Fligstein 2008). However, compared to the multiplicity of studies dealing with processes of ‘vertical integration’, ‘horizontal Europeanisation’ in terms of networks and mobility of people has been under-researched. Our article addresses this research lacuna by focusing on cross-border mobility and networks.
Research questions and hypotheses
As a broad point of departure, we are interested in explaining the different degrees of transnational activities in and across various European Union member states (horizontal Europeanisation). As a more general point, we assume that Europe is a relatively integrated social space, with a high degree of supra-national coordination and integration, a dense network of communication and transportation infrastructure, a common market, and deinstitutionalised borders as in the Schengen zone. These characteristics ought to be good preconditions for opening the ‘nation-state container’ and facilitating cross-border exchanges between European citizens. Fligstein (2008) even makes the point that the ‘success of integration’ highly depends on the opportunities of Europeans to travel and to engage in interaction. In his view, the degree of the transnationalisation of Europeans determines the future of the European polity. It has been shown, for example, that interacting with people from other European countries as well as travelling to other EU countries is positively associated with European identification and support for European integration (Fligstein 2008, chapter 6; Kuhn 2011). On this basis, it has been suggested that transnational experiences within the European context may also strengthen the attachment to the European project.
However, we also know that the process of integration does not take place in a uniform way; rather, it appears to be fragmented across and within different societies. Fligstein distinguishes different social groupings with different opportunities of access to a ‘European society’. However, while Fligstein focuses on group-specific opportunity structures and transnational activities, he pays less attention to the differences in the level of horizontal Europeanisation between countries. Here is where our study begins. We focus on explaining differences in the level of transnational activities between European member states. In our view the emphasis on context factors allows the view on determinants and dynamics of European social transnationalism to be broadened. Based on the literature, we view geography, the level of modernisation, and the degree of societal internationalisation as major factors determining the opportunity structure.
Size of territory and population
It can be assumed that population size impacts on social transnationalism as the likelihood of interacting with citizen from other countries decreases with population size and vice versa. With a few hundred million fellow members it is more likely that the networks of individuals are absorbed internally, whereas the members of small societies are more prone to establish external contacts. Moreover, we assume not only that population size might matter, but also that territory size should: As a rule of thumb it can be assumed that the size of the territory country is negatively correlated with the number and frequency of outside contact (Geser 1992; Gerhards 2010b). The most striking argument in this context is that the circumference of a geometrical area grows in degressive relation to its inner area (Geser 1992). Thus, it can be assumed that a small nation has, in relation to its size, proportionally more interaction with its neighbours than a larger country (Gerhards 2010a: 173). Moreover, living in a small state also indicates a comparably high chance of living close to the national borders, thereby increasing the opportunities of transnational interaction.2 Finally, small-sized states’ internal markets offer only limited possibilities for internal trade. As a consequence, small countries are forced to open up to external markets in order to maintain or to develop high(er) standards of living (Eisenstadt 1985). Thus, the labour force in these countries is expected to be more internationalised than their counterparts inhabiting bigger countries. In sum, we assume that the individual probability of engaging in different forms of social transnationalism is dependent on the population size and the territory size of a given country, that is, people living in countries that are comparably small in size and have a relatively small number of inhabitants are more prone to engage in transnational activities within the EU than others (Hypothesis H1).
Modernisation
Modernisation usually refers to a multidimensional social process that brings about economic growth as well as social and political mobilisation, and transforms the political order through democratisation and bureaucratisation. Modernisation theory is a theory used to summarise the modern transformations of the social order as well as of social life and attempts to identify factors which explain or measure ‘social progress’. While the first transition was the one from traditional to industrial societies, the second marks the transition from industrial to post-industrial societies (Bell 1973). Mass production and bureaucratic organisations of industrial society are being replaced by a greater role of knowledge-intensive service sector employment and education. This also goes along with changes in core social values. Inglehart's (1990, 1997) theory, for example, claims that once a society has embarked on modernisation, fundamental changes in the belief systems of the citizens take place. More ‘modernised’ countries pioneer the development of postmodern values which emphasise self-expression instead of deference to authority and are tolerant of other groups, even regarding exotic things and cultural diversity as stimulating and interesting, not threatening. According to the United Nations Development Program, the concept of human development refers to the process of widening the options of persons, giving them greater opportunities for education, health care, income, employment, etc. This concept can easily be applied to our research question. With a higher level of modernisation one would expect more opportunities to travel and to make contacts abroad (for a similar relationship with transnational linguistic capital, see Gerhards 2010a: 175ff). We already know that, at the individual level, income and education are good predictors for the degree of individual transnationality (Mau and Mewes 2009), as income gives people more resources for mobility, while education increases people's cognitive competences to deal with different settings and people. The same link should hold at the country level, with countries scoring high on modernisation also being more ‘transnationalised’. Therefore, we maintain that the higher the level of modernisation in a given country, the more likely do its inhabitants engage in different forms of transnational activities within the EU (Hypothesis H2).
Internationalisation
Karl W. Deutsch's ‘transactionalism’ (1957, 1968) has emphasised that political integration leads to intensified communication and exchange between different nations. According to Deutsch, it is of political importance that integration does not only take place at the economic level, but also at the level of personal contacts and interactions (Delhey 2004; Mau 2010). Deutsch describes the dynamics of this process as a cybernetic model – a self-propelled process – within which economic and political transactions trigger individual-level interactions. With the intensification of border-crossing exchanges at the political and economic level, more and more people would start to become aware of the advantages of these flows and start to participate in transnational interaction. Though one can justifiably question the optimistic perspective of such a prediction, we can derive the hypothesis that integration processes (such as the creation of a common market, etc.) should increasingly encourage more individual cross-border transactions (Fligstein 2008). Also, social transnationalism should be more pronounced the more globalised a country is, because globalisation may work as an amplifier of cross-border interactions at the individual level (Kuhn 2011). Thus, there should be a positive association between political and economic European integration and horizontal Europeanisation at the individual level. When it comes to the factors driving transnational activities, we draw on a broad literature and expect two other context-factors to be important, namely globalisation (Robertson 1992; Held et al.1999; Castells 2000; Rosenau 2003) and migration (e.g., Basch et al.1994; Castles and Miller 2003; Vertovec 2003). More precisely, we assume that both of these processes engender transnational orientations and practices. Migration into a country makes meaningful as well as casual interaction between natives and foreigners more likely. With regard to globalisation we assume that in order to maintain the border-crossing flow of goods, capital, and money, there likewise must be a labour force that is transnationally or globally active (Sassen 1991). Thus, we assume that countries with a fairly high level of globalisation should have also a more transnationally engaged population. In sum, we assume that internationalisation is a decisive factor for transnational involvement at the micro-level. Our related hypothesis reads as follows: The higher the level of internationalisation of a given country (Europeanisation. globalisation, migration), the more likely do its citizens take part in different forms of transnational activities within the EU (Hypothesis H3).
Data, variables, and method
We will test our hypotheses using data derived from the Eurobarometer 65.1 (in the following: EB 65.1). This survey was conducted in 2006, providing information on different forms of social transnationalism by people who were at least 15 years old and who had finished their education at the time of the interview. After listwise deletion of cases, our sample contains information on respondents from the following 25 European countries: Austria (n=914), Belgium (n=928), Cyprus (n=461), the Czech Republic (n=949), Denmark (n=812), Estonia (n=890), Finland (n=887), France (n=930), Germany (n=1410), Greece (n=910), Hungary (n=946), Ireland (n=872), Italy (n=883), Latvia (n=884), Lithuania (n=863), Luxembourg (n=456), Malta (n=460), the Netherlands (n=963), Poland (n=899), Portugal (n=834), Slovenia (n=848), Slovakia (n=976), Spain (n=863), Sweden (n=912), and the United Kingdom (n=1211). In total, our sample consists of 21,961 individuals.
Response variables
The EB 65.1 contains two items which provide us with information about Europeans’ transnational activities within the EU context (horizontal Europeanisation). Firstly, item QA5 192 (‘In the last 12 months, have you visited another EU country’) taps into the dimension of cross-border mobility to other EU countries. Secondly, item QA5 194 (‘In the last 12 months, have you socialised with people from another EU country’) refers to the establishment of transnational networks within the EU. The response categories were (1) ‘yes, on several occasions’, (2) ‘yes, once or twice’, and (3) ‘no’.
Explanatory variables
As the primary scope of our study deals with the question of how specific characteristics of countries influence different modes of social transnationalism within Europe, we start by discussing our explanatory variables at the higher level. Nevertheless, taking into account that variation between countries might not only be due to the way they are geographically, politically and economically shaped, but also due to differences in regard to the composition of their populations, we will also discuss important control variables at the individual level.
Macro variables
At the country level we employ indicators for the three different dimensions we already discussed before, namely:
- 1.
Geography
- 2.
Modernisation
- 3.
Internationalisation
Each of these three dimensions is operationalised on the basis of at least two macro-level indicators. Starting with the geographic dimension, we take into account two different characteristics of the countries in our sample.3 Firstly, we check for the effect of the total population size, asking whether the numbers of people living in the European societies affect the probability of taking part in processes of horizontal Europeanisation. Secondly, we monitor territory size (land area in km2), assuming that individuals living in larger countries are less likely to engage in cross-border interaction than people from comparably small countries. In order to check whether it is only very large and very populous countries whose citizens are unlikely to get involved in transnational activities within the EU, we also check for the effects of the logarithm of population and territory sizes.
As regards the measurement of the dimension of modernisation we refer to two indicators. Firstly, we refer to the Human Development Index (HDI).4 As opposed to purely economic measures, the HDI is a composite index capturing life expectancy (at birth), education (measured by the adult literacy rate and the combined primary, secondary, and tertiary gross enrolment ratio) and the standard of living (measured by the natural logarithm of the gross domestic product per capita at purchasing power parity). The values of the HDI range from 0 to 1, with a 1 indicating the highest possible level of human development. Secondly, we check the purely economic measure of GDP per capita when examining socioeconomic aspects determining the level of social transnationalism.
Finally, we examine the role of internationalisation by employing three different indicators, thereby assuming that all of the indicators affect the probability of engaging in social transnationalism positively. Firstly, we monitor the effect of European integration, operationalised by the duration of EU membership.5 Here, we assume that the probability of participating in transnational activities increases with the years of membership. The underlying assumption is that the European integration process works as an opportunity structure for cross-border interactions, although, networks and the infrastructure of horizontal Europeanisation need time to develop. We know that the level of connectedness in terms of transport and communication differs between old and new member states (Mau and Verwiebe 2010: 304ff.). Secondly, we control for globalisation by referring to the KOF index of economic globalisation (Dreher et al.2008). This index is composed of manifold information on border-crossing economic flows, given that it relies on information about trade (percent of GDP), foreign direct investment (flows and stocks), portfolio investment (percent of GDP), hidden import barriers, mean tariff rates, taxes on international trade (percent of current revenue) and capital account restrictions. The index ranges, theoretically, from 0 to 100, with 100 indicating the highest possible level of economic globalisation.6 We expect that the more globalised countries also score higher in terms of horizontal Europeanisation. Moreover, we want to investigate whether migration at the macro-level impacts on social transnationalism at the individual level. Therefore, we employ the value ‘share of foreign-born population’, expecting that individuals from countries with high levels of migration are more inclined to actively take part in the process of EU-cross-border activities at the micro-level.7
Individual level control variables
As we assume the between-country variation to be partly dependent on different national compositions of the social structure, we control for important socio-demographic characteristics of our respondents.
2.4.1. Education and occupational class: From a study that deals with the relationship between class and border-crossing activities in the German population (Mau and Mewes 2009), we know that it is the group with high levels of education that might be regarded as the pioneers of social transnationalism, as their ‘social networks frequently stretch across national borders and they are significantly more active in terms of transnational mobility’ (Mau and Mewes 2009: 180). Also, education might be regarded as an indicator for ‘transnational competence’ (Koehn and Rosenau 2002), which refers to the ability to understand other codes, conventions, attitudes, and modes of behaviour as well as to communicate and cooperate across cultural boundaries. Similarly, we expect social class to either drive or prevent participation in social transnationalism (cf. Mau and Mewes 2009). Overall, we assume that it should be non-manual workers and managers rather than blue-collar workers and lower classes that are involved in the forms of horizontal Europeanisation in which we are interested. Nevertheless, we acknowledge that in some member states, especially in those that joined the Union relatively recently, it might be predominantly labour migration that drives cross-border interaction within the European Union. In this latter perspective, we would expect high levels of cross-border interactions especially among working-class members. Yet, taking into consideration that levels of labour migration within the European Union are still relatively low, we consider this last effect to be rather weak.8 Following Knutsen's (2006) approach to developing a class scheme on the basis of Eurobarometer data, we employ a very rough proxy for class membership in the sense of Erikson and Goldthorpe (1992), based on the items V633 and V634 in the EB65.1. In the cases where categorisation of the current occupational group failed we used the respondents’ previous occupations as an indicator for class membership. We distinguish between the following occupational groups: employers, self-employed in the primary industries (such as farmers, fishermen, etc.), higher-level non-manuals, other non-manuals (we merged the categories ‘medium-level non-manuals’ and ‘low-level non manuals’), workers, students, unemployed, and people who never had a paid job. Thus, in line with Fligstein (2008) we expect that the probability of becoming a transnational European increases with education and occupational status: ‘It is the educated, professionals, managers, and other white-collar workers who have the opportunity to travel, speak second languages, and interact with people like themselves in different countries’ (ibid.: 123).
2.4.2. Age: We follow Edmunds and Turner (2005: 572), who assume that we are currently witnessing the formation of a new global generation that ‘both shares its information and ideas across borders and acts with global impact’. We therefore assume that the younger the people are, the more prone they are to participate in different forms of social transnationalism. Also, it may be assumed that ‘[o]lder people will be less adventurous than younger people, and less likely to have learned other languages, or to hold favourable views of their neighbours; moreover, they will probably remember who was on which side in World War II. They will be less likely to want to associate with or have curiosity about people from neighbouring countries’ (Fligstein 2008: 127). Here, age is employed in years.
2.4.3. Nationality and migratory background: With regard to transnational activities migrants, compared to the ‘native’ population, may be regarded as an exceptional group. (Vertovec 2009). As a matter of fact, most of the research on transnationalism clearly has its roots in the field of migration studies (e.g., Basch et al.1994; Portes et al.1999). These studies focus on the observation that more and more migrants seem to establish lifestyles in which they (and even the next or third generation) develop strong transnational ties to more than one home country, thereby blurring the congruence of social space and geographic space. This could also be a relevant factor when looking at cross-border interactions within the EU. Thus, we monitor the migratory background of the respondents in two ways: on the one hand, we control for nationality (native/foreign), on the other hand we take into consideration the parental background (father and/or mother born in country/abroad).
2.4.4. Gender: While the feminisation of migration is part of the intra-European migration process (Favell 2008), a study with an empirical focus on the German population has shown that men tend to have (slightly) more transnational contacts and to have lived more often abroad compared to women (Mau 2010). Furthermore, it is known that women are still more occupied with domestic responsibilities than men, regardless of whether they are economically active or not (Turner and Grieco 2000). This might reduce women's chances of making contacts with people from different contexts than the household, neighbourhood and family. From research on mobility and transportation, we also know that women's trips are confined within a smaller geographical area than those of men (Jones et al.1983). Hence, we expect men to be more involved in social transnationalism than women, though the differences are not expected to be striking.
2.4.5. Degree of Urbanisation: Against the background of the ‘Global City’ debate, it is suggested that bigger cities act as central nodes in the world economy, bringing together people from various countries and cultures. Although not every country in our sample possesses a ‘Global City’ in the strict sense of Saskia Sassen (1991), we assume that people who are living in large towns are comparatively more involved in social transnationalism within the European context than people in less densely populated areas (that is, rural areas and small towns).
Statistical method
To disentangle contextual and individual-level effects regarding our two response variables, we used multi-level modelling techniques. Keeping in mind that the two response variables discussed above are ordinal-scaled, we decided to conduct multilevel ordered logistic regression models (Snijders and Bosker 1999: 229–33; Hox 2010: 141–50). This kind of multilevel model is formulated as a threshold model, i.e., the outcome of the response variable is conceived as the result of an underlying non-observed continuous variable (Snijders and Bosker 1999: 223). With regard to the interpretation, the threshold parameters are of secondary importance, given that they reflect the marginal probabilities of the outcome categories only (therefore, we omitted them from our tables in the following section). The parameters of the model can be interpreted in principle just as in the hierarchical linear model (ibid.: 231).
In our analysis, we focus on a test of our initial hypotheses, concentrating on the question whether certain macro variables impact the likelihood of participating in transnational activities. In this context, we want to recall that in logistic multi-level models the individual-level variance is always fixed, namely to π2/3 (≈3.29) (Snijders and Bosker 1999: 224). Nevertheless, controlling for relevant individual-level explanatory variables is deemed to be important against the background of our research question, given that a certain degree of variance between countries always tends to be due to the fact that countries differ in regard to their socio-demographic structure. Thus, uncontrolled country-level differences in the social composition may easily lead to effects at the higher level of analysis.
All of the multilevel models were calculated with MLWIN 2.20, using 2nd order Penalized Quasi-Likelihood estimation (PQL), which is regarded as superior compared to 1st order Maximum Quasi-Likelihood (MQL), given that it produces more reliable level-2 estimates and more precise standard errors (Hox 2010: 119ff.). Moreover, a sandwich estimator was used to obtain robust standard errors, which are considered to lower the influence of statistical outliers and to be more reliable in cases in which the data violate the normal distribution assumption (Maas and Hox 2004).
Results
Taking a first glance at the descriptives (Table 1), we observe remarkable variation between countries. Luxembourg, a very small and centrally located country, especially turns out to be an outlier when it comes to participation in the two observed forms of social transnationalism. More than 80 percent of Luxemburg respondents crossed their national borders during the last 12 months at least once. Similarly, 80 percent of respondents from this country socialised with EU foreigners within this particular time period. The Netherlands rank second with 74 percent having made a trip in the year before the interview, while 65 percent socialised with people from other EU countries. Also the Scandinavian countries score relatively high on horizontal Europeanisation. Though some of them joined the EU relatively recently (Sweden and Finland only in 1995) there is a long history of cooperation and allowing free travel between Scandinavian countries. For example, the Nordic Passport Union, established in the 1950s, allows Scandinavian citizens to travel and reside in other Nordic countries without a passport or a residence permit. Passport checks for aliens at internal Nordic borders were already removed in 1958. From the data, we also see that several of the former EU candidate countries (those countries that joined the EU in 2004) exhibit relatively low levels of social transnationalism (with Slovenia and Slovakia being an exception). In Estonia, Hungary, Latvia, Lithuania, Malta, and Poland people with transnational experience are clearly outnumbered by those without transnational activities during the 12 months prior to the interview. Also, respondents from Greece, Portugal, and Spain turn out to be less transnationalised than average.
| . | Visited another EU country . | Socialized with people from another EU country . | ||
|---|---|---|---|---|
| Country . | Once or twice . | On several occasions . | Once or twice . | On several occasions . |
| Austria | 32.5 | 31.4 | 23.7 | 22.0 |
| Belgium | 29.8 | 31.0 | 13.4 | 33.5 |
| Cyprus (Republic) | 32.8 | 3.0 | 15.2 | 34.5 |
| Czech Republic | 28.0 | 18.0 | 15.4 | 16.6 |
| Denmark | 27.6 | 39.4 | 11.1 | 33.4 |
| Estonia | 15.4 | 9.8 | 16.6 | 28.0 |
| Finland | 35.6 | 13.2 | 28.5 | 24.5 |
| France | 17.5 | 9.9 | 13.8 | 23.3 |
| Germany | 28.3 | 18.7 | 21.2 | 34.8 |
| Greece | 7.0 | 2.4 | 11.6 | 14.9 |
| Hungary | 11.1 | 10.0 | 8.4 | 7.2 |
| Ireland | 35.2 | 11.2 | 27.5 | 21.2 |
| Italy | 24.3 | 4.2 | 26.3 | 9.7 |
| Latvia | 14.4 | 4.0 | 14.8 | 20.5 |
| Lithuania | 10.3 | 5.3 | 14.4 | 17.1 |
| Luxembourg | 23.0 | 60.1 | 10.5 | 71.9 |
| Malta | 17.8 | 6.1 | 16.3 | 28.0 |
| The Netherlands | 35.0 | 39.6 | 20.0 | 54.9 |
| Poland | 12.5 | 5.9 | 11.9 | 15.1 |
| Portugal | 14.0 | 5.5 | 10.2 | 8.4 |
| Slovenia | 23.6 | 26.9 | 14.7 | 23.0 |
| Slovakia | 24.0 | 16.0 | 23.8 | 21.3 |
| Spain | 12.6 | 4.2 | 12.2 | 10.8 |
| Sweden | 39.6 | 17.8 | 17.7 | 37.0 |
| United Kingdom | 32.4 | 12.7 | 30.1 | 22.5 |
| . | Visited another EU country . | Socialized with people from another EU country . | ||
|---|---|---|---|---|
| Country . | Once or twice . | On several occasions . | Once or twice . | On several occasions . |
| Austria | 32.5 | 31.4 | 23.7 | 22.0 |
| Belgium | 29.8 | 31.0 | 13.4 | 33.5 |
| Cyprus (Republic) | 32.8 | 3.0 | 15.2 | 34.5 |
| Czech Republic | 28.0 | 18.0 | 15.4 | 16.6 |
| Denmark | 27.6 | 39.4 | 11.1 | 33.4 |
| Estonia | 15.4 | 9.8 | 16.6 | 28.0 |
| Finland | 35.6 | 13.2 | 28.5 | 24.5 |
| France | 17.5 | 9.9 | 13.8 | 23.3 |
| Germany | 28.3 | 18.7 | 21.2 | 34.8 |
| Greece | 7.0 | 2.4 | 11.6 | 14.9 |
| Hungary | 11.1 | 10.0 | 8.4 | 7.2 |
| Ireland | 35.2 | 11.2 | 27.5 | 21.2 |
| Italy | 24.3 | 4.2 | 26.3 | 9.7 |
| Latvia | 14.4 | 4.0 | 14.8 | 20.5 |
| Lithuania | 10.3 | 5.3 | 14.4 | 17.1 |
| Luxembourg | 23.0 | 60.1 | 10.5 | 71.9 |
| Malta | 17.8 | 6.1 | 16.3 | 28.0 |
| The Netherlands | 35.0 | 39.6 | 20.0 | 54.9 |
| Poland | 12.5 | 5.9 | 11.9 | 15.1 |
| Portugal | 14.0 | 5.5 | 10.2 | 8.4 |
| Slovenia | 23.6 | 26.9 | 14.7 | 23.0 |
| Slovakia | 24.0 | 16.0 | 23.8 | 21.3 |
| Spain | 12.6 | 4.2 | 12.2 | 10.8 |
| Sweden | 39.6 | 17.8 | 17.7 | 37.0 |
| United Kingdom | 32.4 | 12.7 | 30.1 | 22.5 |
Source: Eurobarometer 65.1 (2006), n=21,961.
Let us now turn to the results of our multivariate analyses. Here we want to stress again that we estimated ordinal multilevel models with regard to two different response variables. As we assume that the two items are related to the same latent dimension of ‘social transnationalism’, we do not discuss the results in separate sections. Our initial models (Table 2, models 0a and 0b) allow us to check how much of the total variance is due to differences between countries. Following Snijders and Bosker's (1999: 231) approach towards computing the intraclass correlation coefficient (in the following: ICC) in multilevel models for ordinal response data,9 we firstly observe that 18 percent of the variance regarding the dependent variable ‘visiting other EU countries’ (in the following ‘trips’) is due to differences between countries. Secondly, in the case of the response variable ‘socializing with people from other EU countries’ (in the following: ‘networks’), the ICC is 0.11, indicating that 11 percent of the total variance is located at the higher level, whereas the residual 89 percent is due to differences between individuals.
| . | Dependent variable ‘visit’ . | Dependent variable ‘socialize’ . | ||||||
|---|---|---|---|---|---|---|---|---|
| . | Model 0a . | Model Ia . | Model 0b . | Model Ib . | ||||
| . | Coeff. . | Sig. . | Coeff. . | Sig. . | Coeff. . | Sig. . | Coeff. . | Sig. . |
| Migratory background (1 = yes) | 0.177 (0.047) | * | 0.491 (0.068) | * | ||||
| Foreign nationality (1 = yes) | 0.570 (0.107) | * | 0.795 (0.169) | * | ||||
| Female (1 = yes) | −0.274 (0.029) | * | −0.245 (0.035) | * | ||||
| Age | −0.012 (0.001) | * | −0.015 (0.002) | * | ||||
| Education | 0.120 (0.006) | * | 0.145 (0.011) | * | ||||
| Community size | ||||||||
| Rural area or village (Ref.) | Ref. | Ref. | ||||||
| Small/Middle town | 0.049 (0.033) | 0.054 (0.044) | ||||||
| Large town | 0.080 (0.037) | 0.179 (0.057) | * | |||||
| Social class | ||||||||
| Employers | 0.783 (0.054) | * | 0.776 (0.082) | * | ||||
| Higher-level non-manuals | 0.815 (0.062) | * | 0.879 (0.057) | * | ||||
| Other non-manuals | 0.479 (0.037) | * | 0.405 (0.040) | * | ||||
| Self-employed in the primary industries (farmers. Fishermen etc.) | −0.058 (1.103) | −0.280 (0.146) | ||||||
| Workers (Ref.) | Ref. | Ref. | ||||||
| Unemployed | −0.095 (0.063) | 0.089 (0.070) | ||||||
| Never had paid work | −0.009 (0.090) | 0.033 (0.078) | ||||||
| Variance (country-level) | 0.741 (0.209) | 0.647 (0.185) | 0.420 (0.120) | 0.371 (0.106) | ||||
| Notes: Ref., Reference category. Unstandardised coefficients are stated. Standard errors appear in parentheses. Sig., *p≤0.05. | ||||||||
| Source: Eurobarometer 65.1 (2006). ni=21,961; nj=25. | ||||||||
| . | Dependent variable ‘visit’ . | Dependent variable ‘socialize’ . | ||||||
|---|---|---|---|---|---|---|---|---|
| . | Model 0a . | Model Ia . | Model 0b . | Model Ib . | ||||
| . | Coeff. . | Sig. . | Coeff. . | Sig. . | Coeff. . | Sig. . | Coeff. . | Sig. . |
| Migratory background (1 = yes) | 0.177 (0.047) | * | 0.491 (0.068) | * | ||||
| Foreign nationality (1 = yes) | 0.570 (0.107) | * | 0.795 (0.169) | * | ||||
| Female (1 = yes) | −0.274 (0.029) | * | −0.245 (0.035) | * | ||||
| Age | −0.012 (0.001) | * | −0.015 (0.002) | * | ||||
| Education | 0.120 (0.006) | * | 0.145 (0.011) | * | ||||
| Community size | ||||||||
| Rural area or village (Ref.) | Ref. | Ref. | ||||||
| Small/Middle town | 0.049 (0.033) | 0.054 (0.044) | ||||||
| Large town | 0.080 (0.037) | 0.179 (0.057) | * | |||||
| Social class | ||||||||
| Employers | 0.783 (0.054) | * | 0.776 (0.082) | * | ||||
| Higher-level non-manuals | 0.815 (0.062) | * | 0.879 (0.057) | * | ||||
| Other non-manuals | 0.479 (0.037) | * | 0.405 (0.040) | * | ||||
| Self-employed in the primary industries (farmers. Fishermen etc.) | −0.058 (1.103) | −0.280 (0.146) | ||||||
| Workers (Ref.) | Ref. | Ref. | ||||||
| Unemployed | −0.095 (0.063) | 0.089 (0.070) | ||||||
| Never had paid work | −0.009 (0.090) | 0.033 (0.078) | ||||||
| Variance (country-level) | 0.741 (0.209) | 0.647 (0.185) | 0.420 (0.120) | 0.371 (0.106) | ||||
| Notes: Ref., Reference category. Unstandardised coefficients are stated. Standard errors appear in parentheses. Sig., *p≤0.05. | ||||||||
| Source: Eurobarometer 65.1 (2006). ni=21,961; nj=25. | ||||||||
Next (Table 2, models Ia and Ib), we introduced our set of individual-level control variables. The results tend to confirm our initial assumption that both of our dependent variables tap into the same dimension, given that with one exception (community size, as indicated by the question of living in a rural area, a town or a bigger city), the effects of the level-1 variables turn out to be similar for the two different response variables. This allows us to summarise the main results in regard to the likelihood of participating in social transnationalism. Across the countries of the European Union, the probability of participating in horizontal Europeanisation clearly increases with education, confirming similar results from a study conducted in Germany (Mau and Mewes 2009). Likewise, we observe an effect of social class, indicating that employers and non-manuals – compared to workers – have higher odds of taking part in cross-border interaction.
In relation to the factor ‘age’, we observe that the older people get, the less likely they are to make trips abroad and to socialise transnationally. Moreover, given the negative effect of gender, women's lifeworlds still seem to be more strongly anchored in the national arenas than men's ones. Unsurprisingly, people with a migratory background and/or a foreign nationality are more likely to be involved in social transnationalism in the European context than ‘indigenous’ citizens of the countries in question. This result is perfectly in line with those from studies into migrants’ transnationalism (e.g. Al-Ali and Koser 2002; Conradson and Latham 2005; Faist 2004).
Shifting our attention to the change in the level-2 variance that the introduction of the control variables has brought about, we see in regard to both cases of horizontal Europeanisation that a considerable part of the level-2 variance (about 13 percent in the case of transnational trips and about 12 percent with regard to the ‘networks’-response variable) can be explained by so-called ‘composition effects’. This implies that only a part of the between-country variance is due to the fact that the European societies in our sample are quite differently composed when it comes to the factors age, sex, education, urbanisation, migration, and class membership. The major part of the between-country variance remains unexplained, urging us even more to check for specific features of the countries in our sample.
Therefore, we introduce our macro-level indicators of interest. As already pointed out, we distinguished three different dimensions which are assumed to drive European transnationalism, namely the countries’ geographies and populations, their degrees of modernisation and their degree of internationalisation, including a measure for European integration. Each of these dimensions is measured by at least two macro-level indicators. Tables 3 and 4 summarise the results from 18 separate multilevel analyses (nine models for each of the response variables), in which we controlled for one macro-level indicator at a time (with an otherwise identical set of those individual-level control variables that were already portrayed in the models Ia and Ib).
| Model . | Dimension . | Coeff. (SE) . | Sig . | Variance (Country level) (SE) . | Explained variance (country level) . |
|---|---|---|---|---|---|
| Geographic dimension | |||||
| IIa | Population size | −0.005 (0.006) | 0.629 (0.179) | 15.1% | |
| IIIa | Population size (log) | −0.096 (0.115) | 0.627 (0.179) | 15.4% | |
| IVa | Territory size | −0.001 (0.001) | 0.591 (0.169) | 20.2% | |
| Va | Territory size (log) | −0.153 (0.115) | 0.588 (0.168) | 20.6% | |
| Modernization | |||||
| VIa | Human Development Index/100 | 0.145 (0.034) | * | 0.424 (0.121) | 42.8% |
| VIIa | GDP per capita (in 1000 US$) | 0.005 (0.001) | * | 0.314 (0.090) | 57.6% |
| Internationalization | |||||
| VIIIa | Duration of EU-membership (years) | 0.019 (0.008) | * | 0.539 (0.154) | 27.3% |
| IXa | Economic globalization (KOF index) | 0.067 (0.027) | * | 0.535 (0.153) | 27.8% |
| Xa | Foreign-born population (percent of population) | 0.052 (0.017) | * | 0.514 (0.147) | 30.6% |
| Notes: Each of the nine models contains the set of individual-level variables used in model Ib and the country-level depicted in the respective row of the table. log, natural logarithm of the respective variable. Sig., *p≤0.05. Unstandardised coefficients are stated, standard errors appear in parentheses. | |||||
| Source: Eurobarometer 65.1, ni=21,961; nj=25. | |||||
| Model . | Dimension . | Coeff. (SE) . | Sig . | Variance (Country level) (SE) . | Explained variance (country level) . |
|---|---|---|---|---|---|
| Geographic dimension | |||||
| IIa | Population size | −0.005 (0.006) | 0.629 (0.179) | 15.1% | |
| IIIa | Population size (log) | −0.096 (0.115) | 0.627 (0.179) | 15.4% | |
| IVa | Territory size | −0.001 (0.001) | 0.591 (0.169) | 20.2% | |
| Va | Territory size (log) | −0.153 (0.115) | 0.588 (0.168) | 20.6% | |
| Modernization | |||||
| VIa | Human Development Index/100 | 0.145 (0.034) | * | 0.424 (0.121) | 42.8% |
| VIIa | GDP per capita (in 1000 US$) | 0.005 (0.001) | * | 0.314 (0.090) | 57.6% |
| Internationalization | |||||
| VIIIa | Duration of EU-membership (years) | 0.019 (0.008) | * | 0.539 (0.154) | 27.3% |
| IXa | Economic globalization (KOF index) | 0.067 (0.027) | * | 0.535 (0.153) | 27.8% |
| Xa | Foreign-born population (percent of population) | 0.052 (0.017) | * | 0.514 (0.147) | 30.6% |
| Notes: Each of the nine models contains the set of individual-level variables used in model Ib and the country-level depicted in the respective row of the table. log, natural logarithm of the respective variable. Sig., *p≤0.05. Unstandardised coefficients are stated, standard errors appear in parentheses. | |||||
| Source: Eurobarometer 65.1, ni=21,961; nj=25. | |||||
| Model . | Dimension Geographical dimension . | Coeff. (SE) . | Sig. . | Variance (Country Level) . | Explained variance (country-level) . |
|---|---|---|---|---|---|
| IIb | Population size | −0.001 (0.005) | 0.370 (0.106) | 11.9% | |
| IIIb | Population size (log) | −0.139 (0.080) | 0.341 (0.098) | 18.9% | |
| IVb | Territory size | −0.001 (0.001) | 0.356 (0.102) | 15.2% | |
| Vb | Territory size (log) | −0.170 (0.067) | * | 0.305 (0.088) | 27.4% |
| Modernization | |||||
| VIb | Human Development Index/100 | 0.081 (0.033) | * | 0.309 (0.089) | 26.4% |
| VIIb | GDP per capita (in 1000 US$) | 0.004 (0.001) | * | 0.215 (0.062) | 48.8% |
| Internationalization | |||||
| VIIIb | Duration of EU-membership (years) | 0.013 (0.006) | * | 0.324 (0.093) | 22.9% |
| IXb | Economic globalization (KOF index) | 0.045 (0.019) | * | 0.306 (0.088) | 27.1% |
| Xb | Foreign-born population (percent of population) | 0.053 (0.014) | * | 0.265 (0.076) | 36.9% |
| Notes: Each of the nine models contains the set of individual-level variables used in model Ib and the country-level depicted in the respective row of the table. log, natural logarithm of the respective variable. Sig., *p≤0.05. Unstandardised coefficients are stated, standard errors appear in parentheses. | |||||
| Source: Eurobarometer 65.1, ni=21,961; nj=25. | |||||
| Model . | Dimension Geographical dimension . | Coeff. (SE) . | Sig. . | Variance (Country Level) . | Explained variance (country-level) . |
|---|---|---|---|---|---|
| IIb | Population size | −0.001 (0.005) | 0.370 (0.106) | 11.9% | |
| IIIb | Population size (log) | −0.139 (0.080) | 0.341 (0.098) | 18.9% | |
| IVb | Territory size | −0.001 (0.001) | 0.356 (0.102) | 15.2% | |
| Vb | Territory size (log) | −0.170 (0.067) | * | 0.305 (0.088) | 27.4% |
| Modernization | |||||
| VIb | Human Development Index/100 | 0.081 (0.033) | * | 0.309 (0.089) | 26.4% |
| VIIb | GDP per capita (in 1000 US$) | 0.004 (0.001) | * | 0.215 (0.062) | 48.8% |
| Internationalization | |||||
| VIIIb | Duration of EU-membership (years) | 0.013 (0.006) | * | 0.324 (0.093) | 22.9% |
| IXb | Economic globalization (KOF index) | 0.045 (0.019) | * | 0.306 (0.088) | 27.1% |
| Xb | Foreign-born population (percent of population) | 0.053 (0.014) | * | 0.265 (0.076) | 36.9% |
| Notes: Each of the nine models contains the set of individual-level variables used in model Ib and the country-level depicted in the respective row of the table. log, natural logarithm of the respective variable. Sig., *p≤0.05. Unstandardised coefficients are stated, standard errors appear in parentheses. | |||||
| Source: Eurobarometer 65.1, ni=21,961; nj=25. | |||||
Starting with the discussion of the statistical importance of the geographic dimension, we observe that neither the absolute size of the population nor the logarithm of the size play any role in the proliferation of transnational activities in the European context. This applies both to the ‘mobility-model’ (response variable 1) and the ‘network-model’ (response variable 2). Yet, the results differ when it comes to the effect of the territory size. Whereas the absolute territory size does not affect the probability of participating in any of the forms of horizontal Europeanisation, the logarithm of this variable alters the likelihood of establishing transnational networks negatively. This means that, compared to bigger countries, it is especially the geographically very small countries whose inhabitants display a higher probability of engaging in interaction with other people in the EU.
Next, we checked whether our modernisation indicators influence the probability of making trips abroad and of socialising across borders within the EU. Here, our results clearly corroborate our initial assumptions: both the Human Development Index and GDP per capita affect the likelihood of engaging in the two dimensions of social transnationalism positively. The reduction of total variance brought about through these two indicators is remarkable. This holds especially true in the case of GDP per capita: Just by controlling for this particular county characteristic, about half of the residual between-country variance can be explained. This applies both to the ‘mobility model’ (VIIa) and to the ‘network model’ (VIIb). Thus, it may be concluded that modernisation is a relatively strong determinant of social transnationalism within the EU.
Finally, it was examined whether the processes of internationalisation and supra-nationalisation at the higher levels go along with processes of transnationalisation at the level of individuals. More precisely, we investigated whether EU membership, globalisation and migration trigger the proliferation of transnational lifestyles among the Europeans. Again, our results are unambiguous: Both in the case of the ‘mobility model’ and the ‘networks model’, the propensity towards social transnationalism increases with the length of EU membership, the level of economic globalisation, and the proportion of the foreign-born population within a society.
Concluding our discussion of the empirical results, Table 5 summarises the findings with regard to higher-level effects on the two response variables. Despite the finding that the logarithm of territory size reveals a negative impact on the likelihood of establishing transnational social contacts within the EU, we observe no further effect of the countries’ geographic characteristics. Therefore, our results lead us rather to reject hypothesis H1, according to which people living in small countries are more likely to engage in social transnationalism. In contrast, our data clearly corroborate our hypotheses H2 and H3, according to which the processes of internationalisation and supra-nationalisation as well as modernisation affect social transnationalism within the EU positively.
| Characteristic . | Visit another EU country . | Socialize with people from another EU country . |
|---|---|---|
| Population size | 0 | 0 |
| Population size (log) | 0 | 0 |
| Territory size | 0 | 0 |
| Territory size (log) | 0 | – |
| Human Development Index | + | + |
| GDP per capita | + | + |
| EU membership (years) | + | + |
| Economic globalization (KOF index) | + | + |
| Foreign-born population | + | + |
| Notes: log, natural logarithm of the respective variable. 0, no effect; +, significant positive effect; –, significant negative effect (p≤0.05). | ||
| Source: Eurobarometer 65.1 (2006); ni=21,961; nj=25. | ||
| Characteristic . | Visit another EU country . | Socialize with people from another EU country . |
|---|---|---|
| Population size | 0 | 0 |
| Population size (log) | 0 | 0 |
| Territory size | 0 | 0 |
| Territory size (log) | 0 | – |
| Human Development Index | + | + |
| GDP per capita | + | + |
| EU membership (years) | + | + |
| Economic globalization (KOF index) | + | + |
| Foreign-born population | + | + |
| Notes: log, natural logarithm of the respective variable. 0, no effect; +, significant positive effect; –, significant negative effect (p≤0.05). | ||
| Source: Eurobarometer 65.1 (2006); ni=21,961; nj=25. | ||
Conclusion
At a very general level, transnationalism refers to multiple ties and interactions linking people or institutions across the borders of nation-states (Vertovec 2009). Taking a comparative European perspective, we took stock of two specific manifestations of social transnationalism, namely undertaking trips to other EU countries and socialising with other EU citizens. Though this is only a limited set of variables, both can be understood as forms of horizontal Europeanisation. More precisely, we examined whether there are certain country-level characteristics at play when it comes to the proliferation of transnational connectedness of individuals across European countries. On the basis of comparative European data (Eurobarometer 65.1), it could be shown that there are indeed specific macro-level factors that determine migrants’ as well as non-migrants’ participation practices of horizontal Europeanisation. It is one of the findings that the likelihood of making trips to other EU countries or of establishing friendships with people from other member states is closely associated with major social transformations of our time, namely modernisation, European integration and globalisation. Moreover, it could be established that the likelihood of engaging in the establishment of transnational networks is ceteris paribus higher in very small EU countries than on average. In contrast, territory size does not play any role when it comes to trips into other countries. Here a further differentiation between holiday trips and those for other purposes would be highly desirable. Unfortunately, we lack the data to go further in that direction. In contrast to our initial working hypothesis, the population size of a given country does not seem to play any role within the process of social transnationalism.
Against the background of the Europeanisation debate, we suggest that there are huge differences between the EU member states in the intensity of taking part in cross-border interaction and the possible building of a European society. Moreover, our results clearly show that social transnationalism is by no means evenly distributed across Europe. Obviously, it is especially the highly educated upper and middle classes that might be regarded as pioneers of social transnationalism, whereas the European working classes remain relatively strongly attached to the social spaces defined by national borders. These different levels may well become relevant for the Europeanisation process if we take into account that frequent contacts between people from different countries tend to promote supra-national values and identification (Fligstein 2008; Mau et al.2008a, b; Kuhn 2011). Thus, much will depend on whether the ongoing process of European integration brings about new forms of transnational activities and connectedness which also involve less transnationalised parts of the populations.
However, given that our data are cross-sectional in nature, we cannot make clear-cut predictions about the future. For example: is convergence between the European member states contributing to an increase or a decrease of transnational activities? Based on our results, we would argue that labour migration might decrease, while certain forms of mobility (tourism, etc.) and socialising among the Europeans may increase. One pertinent finding is that more prosperous nations tend to produce relatively high levels of horizontal Europeanisation, whereas those lagging behind exhibit a lower degree of transnational activities and an above average share of locally bound people. However, we should bear in mind that cross-border activities are not only structured by economic or political factors, but that they may develop independently of these driving forces through other primarily social means. Yet, together with others (Fligstein 2008) we share the idea that more opportunities for people to interact with other Europeans shape the way people feel about themselves as European citizens, the ‘other’ Europeans and about the project of European integration as a whole.
Footnotes
Migration research – with all its merits – dominates sociological transnationalisation research, even to the extent that the definition of transnationalisation is only applied to migrants. For example, Basch et al. (1994: 7) write ‘We define “transnationalisation” as the process by which immigrants forge and sustain multi-stranded relations that link together their societies of origin and settlement’.
We are aware of the fact that these rules should not apply very well to insular countries.
Population figures and information on territorial characteristics were obtained from the CIA World Factbook, which collects basic information on the entire world of 193 nation-states that are recognised by the UN at the time of writing this article. https://www.cia.gov/library/publications/the-world-factbook/ (accessed December 12, 2009).
The values of the HDI-Index and more detailed information can be queried online: http://hdr.undp.org/en/humandev/ (accessed December 17, 2009).
The duration of EU membership (LEU) was calculated through the formula: LEU=2006 – year of joining the EU.
The KOF data for 2006 can be queried online: http://globalization.kof.ethz.ch/ (accessed December 17, 2009).
Data on the shares of the foreign-born population in our country-sample are derived from the Migration Information Source and refer to EUROSTAT information from 2005: http://migrationinformation.org/charts/pop-table2-jun06.cfm (accessed September 29, 2010).
In view of the fact that the majority of immigrants in Europe are from outside Europe, our hypothesis concerning the relationship between class and social transnationalism would be less clear-cut if the border-crossing interactions were not confined to the realm of the European Union only but to the world as a whole.
In terms of calculating the ICC in ordinal multilevel analysis, we refer to the threshold model for the unobserved underlying variable (Snijders and Bosker 1999: 231). In this way, the residual intraclass correlation coefficient can be estimated as follows:
ICC = residual level-2 variance(Model 0)/(residual level-2 variance(Model 0) + residual level-1 variance(Model 0)). Here, it has to be taken into account that the ‘logistic distribution for the level-one residual implies a variance of π2/3 (≈3.29)’ (Snijders and Bosker 1999: 224).
References
Steffen Mau is professor of Political Sociology and Comparative Social Research and Vice Dean of the Bremen International Graduate School of Social Sciences (BIGSSS). His recent publications include: European Societies. Mapping Structure and Change (together with Roland Verwiebe), 2010; Social Transnationalism, 2010; Welfare States: Construction, Deconstruction, Reconstruction (3 volumes, edited together with Stephan Leibfried), 2008; Justice Legitimacy and the Welfare State (edited together with Benjamin Veghte), 2007. Forthcoming in 2012 is the volume Liberal States and the Freedom of Movement: Selective Borders, Unequal Mobility (with Heike Brabandt, Lena Laube and Christof Roos)
Jan Mewes is a COFAS Marie Curie postdoctoral fellow at Umeå University (Sweden), Department of Sociology. He has worked with the Chair of Political Sociology and Comparative Social Research at the University of Bremen. His research focuses on migration, social capital, and the welfare state. He is author of ‘Cosmopolitan Attitudes Through Transnational Social Practices?' (together with Steffen Mau and Ann Zimmermann, Global Networks, 2008) and of ‘Unraveling Working Class Welfare Chauvinism' (together with Steffen Mau, forthcoming in 2012, in Stefan Svallfors (ed.): Contested Welfare States: Welfare Attitudes in Europe and Beyond)