Feeling European in a globalised world and the role of mobility, networks, and consumption: A comparative approach to British exceptionalism Open Access

The rise of populism and nationalism has raised with great urgency concerns about which factors lead citizens of European nations to feel European. These issues reached new intensity with the success of the Brexit campaign in the 2016 British European referendum. On the face of it, the British vote to leave could be interpreted as a revival of nationalism, associated with a populist backlash driven by those who have not benefited from the ‘European project’ and may not feel part of it (Goodwin & Heath, 2016; Hobolt, 2016). Our paper responds to a need to explore further European identity (or the lack of it) by re-assessing the role of cross-border practices. Following the lead of transactional theorist Karl Deutsch, and the recent emphasis by Delhey, Deutschmann, Graf, and Richter (2014) and Delhey, Deutschmann, and Cirlanaru (2015) on assessing the relevance of the national, European, and global reference frame for people’s everyday activities and interactions, we analyse the EUCROSS survey on the nature of cross-border practices in Denmark, Germany, Italy, Romania, Spain, and the UK in 2012. Our aim is to reflect on the complex ways that mobility histories, social networks, and cultural consumption might be associated with people’s attachment to Europe and to the world as a whole. We underline that cultural consumption practices have often been ignored so far, and we are breaking new ground by considering how they may play a pivotal role in the formation of a European identity.

Another distinctive contribution of our paper is disentangling how wider processes of globalisation might affect both national and European identifications (on which see Delhey et al., 2014; Duchesne et al., 2010), and to broaden our analysis to include a wider range of practices and identities associated with globalisation and Europeanisation. At a very general level, there may be some convergence between European and global transnational practices, linked to certain core rationalistic cultural assumptions (Meyer, 2009). However, Europeanisation and globalisation may refer to very different – even potentially clashing – processes. The colloquial term of ‘Fortress Europe’, for instance, points to the way that an idea of Europe can be mobilised in opposition to global migration flows (Delanty, 2002). Self-identification with Europe might thus rely on different dimensions from those which relate to globalisation (Bruter, 2005; Delhey et al., 2014). Having said this, research has shown that people’s identification with Europe need not be mutually exclusive compared to other identifications, such as national and local ones (Díez Medrano & Gutiérrez, 2001; Risse, 2003; Rother & Nebe, 2009).

This article has two main aims: first, to explore the relations between national, European, and global self-identification, and, second, to explore how these supranational feelings are associated with networks, mobility, and cultural consumption practices both at a European level and at a national level. Using a more sophisticated analysis of transactionalism than has often hitherto been the case, we will be able to better understand the specificity of the British case, which should not simply be seen as a nationalist reaction so much as a complex interweaving of global identities and internal divisions. We show that the British are not, in any obvious ways, more nationalist than other Europeans, but they feel much less part of a larger entity than their own country. Furthermore, when they do express supranational feelings, their historic global connections, notably including those of the empire, can lead to a more singular kind of cosmopolitan, and possibly anti-European, attitude.

Debates about cosmopolitan identifications are strongly influenced by transactionalist theory. The latter was most notably developed in the work of Karl Deutsch (1954, 1969), which sees the development of a wide range of cross-border exchanges and interactions as generating broader transnational attachments. Because the European Union (EU) has emphasised the need for European integration, much substantive research in this vein has been focused on the European case. This has led to a focus on the significance of cross-border practices for the development of a European identity (Kuhn, 2015; Recchi, 2015; Recchi & Favell, 2009). A particularly striking recent example is Fligstein’s Euroclash (2008), which draws attention to the way that a European ‘field’ with its own rules for the game has developed as a result of transnational practices and of European-oriented professionals and managers. It also shows how less-advantaged groups in European nations feel marginal or even excluded from this European project, and thus how this might generate a potential backlash.

To date, most research in the transactionalist vein has therefore focused on the intra-European dimensions and showed that people who move across the EU are more likely to feel European than the rest of the population (see the Pioneur project¹ led by Recchi & Favell, 2009). Similarly, the EUMARR project² coordinated by Díez Medrano focuses on binational marriage between EU citizens and argues that it equally has a positive link with European identification (Schroedter, Rössel, & Datler, 2015). Mau, Mewes, and Zimmermann (2008) also show the links between stays abroad and regular and private transnational relations and cosmopolitan attitudes. Social networks and physical mobility thus seem to foster a European and global identification.

Given these arguments, amidst the evident proliferation of cross-European interactions in recent decades, the apparent backlash against the EU comes as something of a surprise. Part of the problem here is that this research has not addressed the precise mechanisms by which contacts can have these effects. Is there a difference if one is socialised early within a foreign context? And, if yes, is there a difference of impact when this context is European and when it is not? Perhaps also, the practices of virtual mobility (Canzler, Kaufmann, & Kesselring, 2008; Urry, 2007) have distinctive links with European and global identifications? Furthermore, there is the possibility, explored by Delhey and his associates (2014), that ‘European closure’ means that some people with ties to other parts of the world may feel a lack of sympathy with Europe. This is the argument of Flemmen and Savage (2017), who claim that political support for Brexit came not simply from the ‘excluded white working class’ (as Fligstein might predict), but also a privileged group of ‘white imperial racists’ who are highly internationalised, but to the Anglophone nations of the former British empire.

This line of argument can be extended to emphasise that it is the symbolic, as much as the practical and experiential components which are central to developing European or global identifications. Savage, Bagnall, and Longhurst (2005) argue that younger generations in northern English cities who have openly espoused a cosmopolitan identity might actually have a rather limited range of actual practices which extend outside Britain (see also Savage, Wright, & Gayo-Cal, 2010). Instead, their cosmopolitanism marks a symbolic resistance to what are seen as old-fashioned nationalist idioms rather than being based on extensive knowledge or experience of other nations. Furthermore, Favell (2008) and Andreotti, Le Galès, and Moreno Fuentes (2013) show that those middle-class European citizens who actually live and/or work outside their home nations come to realise the difficulties of genuine European identification.

Currently, there is little research examining how consumption practices might play a role in these transnational identifications. Although Calhoun (2002) perceives these as superficial forms of cosmopolitan engagement, it has been argued that they can have considerable power (Kendall, Woodward, & Skrbis, 2009; Woodward, Skrbis, & Bean, 2008). Drawing on neo-institutionalism emphasising an embedding ‘world culture’ (Meyer, 2010; Meyer, Boli, Thomas, & Ramirez, 1997), Lizardo (2005) shows that feelings of world citizenship (Boli & Thomas, 1997) are associated with omnivorous cultural consumption (that is, eclectic patterns of consumption crossing the boundaries between high and low culture – see Peterson & Kern, 1996; Peterson & Simkus, 1992). However, there is no significant relationship with European identification. Following the logic of Delhey et al. (2014), self-identification with Europe could operate following a nationalist logic and might not be associated with inclusive and open global patterns of consumption. In a similar vein to that of Lizardo, Rössel and Schroedter (2015) found that self-identification with the world was more relevant to the understanding of a wide cultural knowledge and taste than self-identification with Europe.³

Similarly, Bennett et al. (2009) contend that British cultural engagement with Europe is possibly declining as interests in European ‘high culture’ also fade among well-educated younger generations. They demonstrate that whereas older middle-class Britons show high levels of engagement with European artists, musicians, and writers, younger generations are more drawn to Anglophone reference points, especially those emanating from the United States and (related) mass culture (see also Savage et al., 2010). Similar arguments are made by Meuleman and Savage (2013) in their discussion of the Dutch cultural field.

By considering the potential significance of consumption practices, and in recognising the difference between European and global transnationalism, this article seeks to offer a new empirical contribution to the debates about the role of cross-border practices on self-identities in the European space and at the national level. This analysis will enable us to assess whether and to what extent Britain is different from other countries in the continent. We will show that an anti-European attitude does not necessarily imply the absence of supranational identification. This is the crucial issue that we take up in this paper, which we operationalise through three more specific questions.

1. Are individuals’ identifications to Europe different from a wider cosmopolitan identification to the world?

2. To what extent are supranational self-identifications associated with networks, consumption, and mobility practices and traditional sociodemographic variables?

3. What kind of comparative differences within Europe are revealed by this focus on actual practices and consumption patterns? Is the UK different from other countries in this respect?

Our sample comes from the EUCROSS⁴ survey, a wide-ranging survey of physical and virtual mobility and practices of cultural consumption as reported by nationals of six European countries (Denmark, Germany, Italy, Romania, Spain, and the UK). The EUCROSS fieldwork was conducted between June 2012 and January 2013. Computer-assisted telephone interviews (CATI) were carried out by trained interviewers, all of whom were native speakers. Sample selection was based on random digit dialling. In total, 6016 respondents were interviewed by phone in 2012 – that is, around 1000 per country. The cooperation rate varied between 9% in the UK and 38% in Romania (see Pötzschke, Ciornei, & Apaydin vom Hau, 2014). This is a low rate for the UK; however, our sample has been randomly drawn and is large enough to reliably illustrate interesting patterns. We do not have any reason to believe that there are systematic biases in the data collection for the UK and the low rate is likely to be related to contextual factors such as a high number of marketing phone calls in this country.⁵ As has been established, response rates (related to the cooperation rates) are not necessarily a key indicator of survey data quality (Holbrook, Krosnick, & Pfent, 2008).

The core analysis of this article refers to a multinomial logistic regression performed at the European level first, and then at the country level, examining the effect of different variables on whether respondents feel that they are citizens of the world but not European, European but not citizens of the world, or both European and citizens of the world.

The concept of European identity, which is explored in this paper, has been flexibly used in the literature so that some have started to call it ‘loose and baggy’ (Favell, 2005). Taking this into account, we have adopted here a rather encompassing definition of European identity as an active, conscious, and also emotional identification to Europe, which also corresponds to one that is frequently used in quantitative research. In comparison with Bruter’s approach (2005), our aim is not to explore the multidimensionality of European identity but to capture more broadly the extent to which people reflexively identify with Europe as a whole. This is inspired by Duchesne and colleagues’ definition of ‘European identity as a psycho-sociological or socio-political process of citizens’ attachment to the European space or to the political community designed by integration’ (Duchesne, 2010, p. 7; translated by Recchi, 2014, p. 119). We will nevertheless prefer to use the terms ‘identification’ or ‘self-identification’ with Europe (Belot, 2010) since this also prevents us from reifying a process that is, at the least, in gestation.

European and global identifications are obtained by combining two variables, previously recoded as dichotomous (with values 1, 2, 3 being recoded as 0 and values 4 and 5 as 1).

1. ‘I feel European’: please tell me how much you agree or disagree with the following statement?

2. ‘I feel citizen of the world’ (scale from 1 ‘strongly disagree’ to 5 ‘strongly agree’)

For our independent variables, we draw on questions measuring physical and virtual mobility, social networks and cultural cosmopolitan consumption and competence on Europeans’ feelings of being European and/or citizens of the world. In order to be able to disentangle a specific Europeanisation of identity, we have discriminated wherever possible between transnational practices between different European (EU) nations, and in other parts of the world, in order to allow us to differentiate specifically European processes from more global ones. Table 1 lists the independent variables associated with mobility and network practices.

The cross-border practices inserted in the model include ‘mundane’ aspects of life such as eating foreign cuisines as well as a measure of linguistic competence which can be treated as significant for cosmopolitan practices (see Table 2). We not only assess respondents’ knowledge of other languages, but also how they use it in their everyday life. We also include a measure for buying goods from sellers or providers located abroad: this might represent an additional step towards Europeanisation or globalisation, depending on the nationality of the sellers. It requires more effort than simply going to the nearby restaurant or buying music online, as people have to search for what they want. Finally, we asked people whether they were very familiar with one or several countries besides their country of residence, that is, that they knew well enough to feel comfortable in. They will enable us to assess people’s level of spatial competence.

These cross-border practices are all contextualised by sociodemographic variables such as age, gender, educational level, and the Standard International Socio-Economic Index of Occupational Status (ISEI), developed by Ganzeboom, De Graaf, and Treiman (1992). Before discussing the results of the regressions, we first present some descriptive data on national differences in terms of supranational self-identifications.

Table 3 cross-tabulates these European and world identifications, broken down by the six national samples. It indicates that although 46% of our sample has both European and world identities, there is a sizeable group – 30% – who have one, but not the other, identity. It also shows that the British responses are distinctive in two ways. Firstly, Britons are much more likely to see themselves as neither citizens of the world, nor of Europe, than any other nation. However, with 40%, there is no simple nationalist majority even in the UK. Secondly, Britons are more likely to see themselves as citizens of the world – but not of Europe – than any other nation – 23%. This suggests that the Brexit vote may not have been only a simple nationalist reaction, and that the British might have greater non-European global ties. Outlining the framing importance of history and culture over economics and geopolitics to account for national divergences in terms of support for the EU, Díez Medrano (2003) characterises the UK as exhibiting pride in feeling British and by the belief in British cultural distinctiveness compared to the (European) continent, which is inscribed historically, notably through the role of the British Empire. By contrast, in Denmark and Germany, over a quarter of respondents feel European but not citizens of the world. A greater support for the EU in Germany may be interpreted as a wish to re-build trust after its role in the Second World War and it is translated into a ‘European approach’ in school textbooks appearing as ‘a lesson to be learned from the destructive nationalism of the German past’ (Soysal, Bertilotti, & Mannitz, 2005, p. 23). In the UK, although they may have changed over time, recent curriculum analyses have shown a focus on Britishness in education, which has related to multicultural values, and on more global than European dimensions, when a supranational level is discussed (Faas, 2011). Besides, negative perceptions of other European groups have been fostered by tabloids, toying with images of invasions or floods of migrants (Fox, Morosanu, & Szilassy, 2012), and have most likely influenced people’s opinions of the EU, increased the separation between the British island and the ‘Continent’ and eventually contributed to the Brexit win.

We should also emphasise that having European or global identities is not in a zero-sum relationship with identifying with one’s nation. Table 4 reports the mean scores of national feelings, broken down by the strength of respondents’ orientations to European and world identities. For instance, in the UK, those who feel European and citizens of the world score on average 4.36 on the scale ‘Feeling citizen of their country of residence’ (where 5 means they strongly feel as such), while those who feel neither European nor citizens of the world have a lower sense of national belonging, with an average score of 4.25. It shows that in every case – including Britain then – it is those who identify with Europe and the world who also report high levels of national identity.

We use a multinomial logistic regression⁶ to assess the relations between supranational self-identifications, on the one hand, and sociodemographics, social network, and virtual and concrete mobility practices, cosmopolitan and European consumption practices, on the other hand. No causal patterns will be assumed here, given the nature of our analysis. The analysis focuses on the whole sample of European citizens.

If we follow the variables’ order of appearance in Table 5, we can see that as expected age has a positive (but small) impact on forms of supranational identity that include a European dimension (odds ratios higher than 1). Older people are more likely to feel European more generally than neither European nor global. In contrast, younger people are more likely to feel citizens of the world, but not European, compared to feeling none of these. This supports the argument by Savage and colleagues (2010) that younger people are less European oriented, and more attracted by a more general global cosmopolitanism. In addition, women are more likely than men to feel both European and global, and they are not predisposed to distinguish between these two identities.

Despite the emphasis on socio-economic and status differences in shaping European identities (notably Fligstein, 2008), in our analysis ISEI’s role is muted. It has a negative impact on feeling a citizen of the world (and not feeling European), but this needs to be interpreted with caution (its contribution on average is not significant at a 0.05 α-level over the five imputations). A closer examination of the average for each form of identity shows that feeling European only, and feeling both European and a citizen of the world, are associated with higher ISEI compared to feeling neither or ‘simply’ a citizen of the world. Yet the differences are rather small, which may explain the lack of significance to be found.

With regard to the level of education, we observe that those with a tertiary education degree have in general higher odds to identify as a world citizen but not European (than to feel neither) compared to people with a higher secondary diploma or to those who have an in-between lower and higher secondary diploma. It seems that the well-educated are especially drawn to global rather than specifically European identities. In addition, there is no significant difference between those who have achieved lower secondary education or less and those with tertiary education. If education has a role, it is certainly not a linear or straightforward one.

It is not surprising that being active in associations oriented towards other countries or cultures turns out to be strongly associated with supranational identities. This kind of associational social capital evidently tends towards a more outgoing set of identifications, both European and global.

Not being born in the country of residence has no impact on people’s identification. Since our sample is composed of national citizens, we only use this variable as an imperfect proxy for migration status and it seems that we cannot establish a link between it and people’s sense of belonging.

In terms of concrete and virtual mobility practices, none of the variables linked to an experience of life abroad before and after 18 is significant. And generally, one’s actual history of mobility practices has relatively little significance whether for Europe or for the world as a whole. Actual experience of migration does not itself appear very significant in shaping transnational identities in Europe, which is an arresting finding for addressing one vein of transactionalist theory that focuses on mobility practices.

By contrast, it makes a significant difference whether trips are to the EU or outside it. The more one travels outside the EU, the more respondents feel citizens of the world – and not European – and the more one travels in the EU, the more they consider themselves European and not citizens of the world. This is a striking demonstration of the need to differentiate between globalisation and Europeanisation.

Virtual communications are negatively associated with feeling European and not citizen of the world. The more virtual contacts people have, the less European they feel.

Furthermore, our model shows that having friends in and outside Europe increases one’s likelihood of identifying as a global citizen, whether alongside a self-identification as European or otherwise. Having a non-native partner clearly makes a big difference. The odds of those who have a partner from another part of the EU to feel European (including a global identity or not) compared to feeling neither European nor citizens of the world are at least twice those who do not have a partner from another country of the EU. Having a partner from outside the EU substantially decreases the odds of feeling both European and citizen of the world.

In addition, other variables, such as the number of countries inside and outside the EU respondents are familiar with, and music tastes, continue to be associated with a European identity – though not always positively. For instance, the more countries outside the EU people are familiar with, the lower the odds are for them to feel European but not citizens of the world, compared to having no supranational identity. However, being familiar with a large number of EU countries increases the odds of feeling both European and a citizen of the world. In addition, listening to world music increases the odds of developing supranational forms of identity. It is interesting to note that a taste for world music is positively linked to all three forms of supranational identity. World music benefits from enjoying a ‘universal’ appeal which crosses geographical boundaries. A taste for traditional European music is also associated with a European and global identity, just as is a taste for traditional music from the country of residence. A preference for European cuisine has no significant link with supranational identity, partly because it depends too much on the national cuisine of the country of residence.

Still in terms of consumption practices, following sport at an international level is positively associated with feeling both European and citizen of the world. Purchasing in the EU is positively related to European self-identifications (with or without a global one), while purchasing outside the EU is negatively related to them.

Other variables related to cosmopolitan competence are significantly related to different forms of supranational identifications. The more languages one knows, the more likely one is to feel European (but not a citizen of the world). Watching TV in another language seems to have a similar effect.

In general, our findings suggest that consumption and network practices are significant in extending transactional analyses, and that, in comparison, transnational mobility practices seem less important in differentiating European from world identifications. Actual life experience in other European nations does not necessarily make much of a difference to feeling European. It is trips to Europe, rather than living in other Europe nations, which are important. Our model seems to suggest that experiencing European diversity may reinforce less a European sense of belonging than enjoying bits of what one might think of as European culture. The coefficient patterns indicate that it would be too hasty to confuse European identification with global identification; they underscore the role of practices anchored in the European space, especially that of consumption and network ones.

Now that we have unravelled general findings for the European space, we turn to consider how they can vary at a national level.⁷ We ran identical models separately for Denmark, Germany, Italy, and the UK, as they cover different parts of Europe (North, Centre, South, Europe continental, and the UK). We omitted the other two nations since half the Romanians and Spaniards surveyed felt they were both European and citizens of the world (Table 1). In addition, for these two countries, the model may not be as reliable, because some variables have very low frequencies. The model is most explanatory for the UK, with a Nagelkerke R² ranging from 0.376 to 0.381 across the five imputed datasets, and for Germany, with a Nagelkerke R² ranging from 0.261 to 0.269. For Italy (Nagelkerke R² from 0.213 to 0.214), but particularly for Denmark (Nagelkerke R² from 0.132 to 0.143), the model is less powerful, with a lower number of significantly contributing variables. In what follows, results of each national regression will be presented graphically for key selected independent variables. We will focus on variables that have a significant contribution for one or several countries, with special attention to the British case. Although we concentrate on the graphs showing predicted probabilities, odds ratios, p values, and confidence intervals will be provided to further illustrate the role of these variables, and especially their level of significance for the national model ( Appendix shows the full model with significant odds ratios for each country).

Figure 1 presents the relationships between the predicted probabilities of supranational feelings and network practices for each selected country. In each country, people with a non-EU partner have higher probabilities of feeling citizens of the world and not European, and also lower probabilities of feeling both European and global. This latter pattern is somewhat more pronounced in the UK, where the odds ratio for British people with a non-European partner to feel both European and global rather than neither is significant and shows a negative effect (odds ratio = 0.126, p value < 0.001, 95%CI: 0.042–0.376). Note that, in the UK, having a partner from a non-EU country leads to somewhat higher probabilities of having no supranational feelings, but the difference between the two is small and should be interpreted with caution. Figure 1.

View largeDownload slide

View largeDownload slide

Close modal

View largeDownload slide

View largeDownload slide

Close modal

Similarly, having friends from outside the EU leads to a higher increase in the probability of feeling a citizen of the world rather than European in the UK compared to the other countries (odds ratio = 1.565, p value < 0.001, 95%CI: 1.320–1.856). This is a clear and direct trend in every case, where those with friends from at least six different countries have lower probabilities of no supranational identification. With regard to being a member of an association oriented towards other cultures, the pattern is similar across countries: yet the difference between members and non-members in terms of their global and European feelings is slightly greater in the UK (odds ratio = 3.630, p value < 0.001, 95%CI: 1.874–7.032).

In Figure 3, the variables ‘purchase in the EU’, ‘number of languages learnt’, and ‘watching television in another language’ reveal some British specificity. First, although all countries follow the same trend, the association between the number of languages learnt and the probability of having supranational feelings is visually pronounced in Britain. This is especially visible for the probabilities of feeling European and not a citizen of the world, and feeling neither (UK odds ratio of feeling European and not citizen of the world compared to feeling neither for that variable: 1.489, p value < 0.01, 95%CI: 1.148–1.932). In Germany, the odds of feeling both European and global compared to feeling neither also significantly increase with the number of languages learnt (1.327; p value < 0.05, 95%CI: 1.054–1.670). Purchasing in the EU has also a significant contribution to the model in both countries (according to the likelihood ratio test). The graph shows that the UK follows a similar trend to the other countries for feeling a citizen of the world and not European, or feeling the opposite. It seems, for instance, that people who purchase in the EU have higher probabilities of feeling European and not a citizen of the world (UK odds ratios comparing with feeling neither: 3.021, p value < 0.01, 95%CI: 1.512–6.038). The graphs also point to a distinctive pattern for the UK, especially for the probabilities of feeling both European and a citizen of the world. With regard to watching television in another language, Figure 3 reveals that, quite distinctively for the UK, it is mainly those who watch at least once a week but not every day who are most likely to feel most European and citizens of the world.⁸Figure 3.

View largeDownload slide

View largeDownload slide

Close modal

View largeDownload slide

View largeDownload slide

Close modal

Finally, in terms of the influence of sociodemographic variables, the UK is the country for which the level of education makes a significant contribution. Those with a university degree are more likely to have supranational feelings than to have none in comparison with those with a lower educational degree. Similarly, ISEI has a significant negative contribution only in the UK (differentiating those with a European and global self-identity compared to those who have none). Although the graphs show a tendency to feel European and a citizen of the world, or only European, associated with middle or older ages, and an inclination to feel only a citizen of the world with younger ages, the association between age and supranational feelings is only significant in Germany and Italy, and for the odds ratios of feeling European without or without a global self-identity.

We saw that our model explains more variation in the UK, where contacts with a foreign environment have an especially important role to play. The graphs demonstrate that Britain follows similar trends to other countries’ and, hence, is in a way not that different; yet, Britain’s position is particular for two reasons. First, a greater number of variables have a significant influence in the model, showing that network, consumption, and sociodemographic variables induce a greater variation among British citizens in terms of supranational feelings, compared to the other countries investigated. Importantly, cultural consumption practices are important, especially for feeling both European and a citizen of the world. The significance of consumption variables, such as those related to music or television watching, tends to corroborate the idea that people can be more cosmopolitan and European if they can feel part of a larger community which shares some cultural communality. Second, geographical boundaries seem to matter more for the links between practices or tastes and supranational identities in the UK than in the other countries. European-centred activities, such as purchasing in the EU or listening to traditional European music, are associated more directly with pro-European feelings than in other nations. Similarly, there are greater differentiations between those with and without relationships with parts of the world outside the EU (through, for instance, having friends or making trips outside the EU) among Britons than for other Europeans. Britain arguably sees a clearer differentiation between the forces of globalisation and Europeanisation.

This clearly outlines strong internal divisions about how people feel towards Europe and towards the world, which are engrained in historical and cultural dynamics and tensions specific to the UK, as already briefly discussed. As Díez Medrano showed, ‘there is less agreement about how to characterise the country in the United Kingdom’ (2003, p. 59).

It may therefore be erroneous to see Brexit voters as simple nationalists. In fact, the crucial issue seems to be the bifurcation between two different kinds of global networks, one looking to the former empire, the other to Europe. Indeed, compared to other research which indicates that younger Britons were more likely to vote in favour of staying in the EU (Goodwin & Heath, 2016), our data do not show a significant relationship with age in the British case when it comes to assess self-identification. Hence, wanting to stay in the EU may be justified by other – maybe more pragmatic – reasons than feeling European. There is, however, a very strong effect of educational qualifications for all kinds of transnational identities.

In addition, our model suggests that the socio-geographical position of the countries is also important. Research on linguistic capital has, for instance, suggested that the size of a country (defined as the size of the population) could motivate an attachment to Europe that could not be reduced to mobility, network, and consumption practices, if one looks at Italy but most especially at Denmark (Gerhards, 2014). Yet, Denmark is similar to Germany with almost a quarter of its population feeling European only. As a small nation, Denmark is one of the most globalised countries according to the 2015 KOF index (seventh)¹⁰ but its apparent openness has its limits, observable through its reluctance to further open the EU’s frontiers (Dosenrode, 2007, p. 186) and a certain Euroscepticism. As Wivel (2014, no page) explains, Denmark holds a dual position towards European integration, complaining but still loyal, partly because, as a small country, the European project is a way for Denmark to ‘maximise national interests by moderating the agendas of the great powers and viewing international institutions as the most effective way to do so’. Therefore, any pragmatic European identity Danes might have does not prevent them from feeling strongly about their own nation too (see Table 2). The lack of differentiation (or absence of impact of most independent variables) between Danes in our model might come from the fact that Denmark is a ‘highly affluent, yet highly cohesive and relatively homogenous society. Yet it combines high levels of social transnationalism with high levels of national loyalty’ (Favell, Reimer, & Solgaard Jensen, 2014, p. 161).

In comparison, Italy looks less anchored in the local and more oriented towards the world. European integration might be perceived in more global terms in this country. European identification is associated with global belonging, with more than half of its population feeling both European and citizens of the world. This global identification seems to be a long-standing orientation (Kohli, 2000) and may have a link with ‘Italy’s many diasporas’ (Gabaccia, 2003) and the movements of worldwide emigration and immigration that have characterised the history of the country.

In this paper, we have elaborated a revised transactionalist approach to transnational practices. We have shown how it is important to distinguish between transactions with European partners and with other parts of the world, and also to recognise how practices are different in their significance for wider identifications. We thus showed in detail how national identities can be consistent with being part of a wider transnational sense of belonging. Most respondents in the six nations which we studied reported high levels of identification with Europe and/or the world as a whole. Only in the UK do less than half of respondents not identify with Europe, and even here, only a minority has neither a European nor a world identity. The British case, we argue, reflects the ongoing power of extra-European ties (especially those associated with imperial Anglophone connections), rather than being a simple protest of those ‘left behind’ by globalisation.

Second, we demonstrated how there are interesting contrasts in the way these supranational feelings are associated with networks, mobility, and cultural consumption practices both at a European level and at a national level. Our findings demonstrate the importance of networks and consumption practices, such as familiarity with other countries, partners from abroad, musical taste, or purchasing abroad. In contrast, concrete physical mobility experience, and especially life experience abroad, is less significant in shaping identities than anticipated. Indeed, actual experience of spending longer time abroad, especially in the EU, seems of less importance in affecting one’s identification both at the European and national levels. Possibly, these practices are less significant for Europeans (especially for the EU-15) as they tend to take the right of freedom of movement for granted. The importance of physical mobility experience is likely to be more significant for those for whom frontiers within and around Europe are still very much in place, especially non-EU migrants (Teney, Hanquinet, & Bürkin, 2016).

As the literature has long established (e.g. Díez Medrano, 2010; Hooghe & Marks, 2004) and as we have already suggested ourselves, the links between identity and support for European integration are not straightforwardly related. Yet, our analysis has helped us to question the validity of some hypotheses regarding the Brexit vote, such as the nationalist backlash. With our model being most explanatory for the UK, it is evident that the British case cannot be reduced to a straightforward nationalist current, since the supranational self-identifications of the British are very complex, linked to different patterns of mobility, network, and lifestyle.

In conclusion, we have extended the analysis of transnational practices to show how intra- and extra-European interactions tend towards complex patterns of identification which can hold global and European identities apart and are not necessarily reduced to a simple dichotomy between affluent Europeans and nationalist ‘left behind’. This is very clear in the British case. By recognising how European identifications are not necessarily part of a larger process of global identification, we are much better able to understand the nature and gravity of current challenges.

Thanks to comments by Adrian Favell, Simon Hix, Aaron Reeves and Céline Teney for their critically constructive comments as well as all members of the EUCROSS team for their support (especially Ettore Rechi and Juan Díez Medrano). This publication uses data from the EUCROSS project which was funded as part of the EC’s 7th Framework Programme (grant agreement: 266767).

No potential conflict of interest was reported by the authors.

Model in Denmark, Germany, Italy, and the UK (odds ratios).