Jump to ContentJump to Main Navigation
The Europe of ElitesA Study into the Europeanness of Europe's Political and Economic Elites$

Heinrich Best, György Lengyel, and Luca Verzichelli

Print publication date: 2012

Print ISBN-13: 9780199602315

Published to Oxford Scholarship Online: May 2012

DOI: 10.1093/acprof:oso/9780199602315.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use. date: 23 November 2020

Elite foundations of European integration: a causal analysis

Elite foundations of European integration: a causal analysis

Chapter:
(p.208) 10 Elite foundations of European integration: a causal analysis
Source:
The Europe of Elites
Author(s):

Heinrich Best

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199602315.003.0010

Abstract and Keywords

This chapter examines the causes of variation in the Europeanness of European economic and political elites. It shows that attachment to Europe (emotion), the wish to strengthen European integration (cognition), and to strive for a common foreign policy (conation) are loosely coupled and vary strongly between European countries, between elites and the general population, and between economic and political elites. This challenges the thesis that European integration is based on a pan-European consensus within elites and between elites and non-elites. Results of analyses show weak and fragmented evidence for the impact of elites’ situs, status, previous biographical experience, and religious affiliation on their Europeanness, whereas a consistent impact comes from the level of Europeanness of other elite groups, indicating a strong effect of mutual cue-taking and peer-pressure at national level. The impact of these and other findings on the future of the process of European integration is discussed.

Keywords:   Europeanness, structural functionalism, theory of cognitive mobilization, cultural capital, post-functionalism, political extremism, multilevel governance, Eurelitism, European unification, religion, institutional networks, responsiveness, European Commission, institutional networks, population gap, elite theory, national identity

The study into the ‘Europe of Elites’ that has formed the basis of this book departed from the assumption that European integration can be conceptualized as a process of elite integration leading to an accord between national elites over their enduring cooperation and peaceful competition in the frame of pan-European institutions. According to this theoretical approach a comprehensive process of elite integration provides the normative and secures the structural basis for the establishment and operation of a European system of multilevel governance. Indeed, results presented in this book furnish evidence for the emergence of a ‘Eurelitism’ that is characterized by a stronger attachment to Europe, a stronger support for the process of European integration, and a stronger willingness to transfer substantial elements of national sovereignty to the European level than is found in the general population (see Chapters 1 and 8). Results also show a broad and strong consensus among national elites in their view that membership in the EU has benefited their countries. Among economic elites this assessment is nearly unanimous, and among their political counterparts there is an average agreement level of 94 per cent; in the general population, however, about a quarter of all respondents, on average, disagree (see Chapter 2 and Figure 10.4).

Other findings though have challenged the consensus thesis and provided a much more differentiated picture of partial consent and division over important facets of Europeanness between and within European national elites (see Chapter 2). Whereas Eurosceptic or Europhobic positions are rarely expressed in all dimensions of Europeanness, discordant and inconsistent answer patterns are the rule. We maintain that analyses of European integration as an elite process have to take this contradictory plurality in the expressions of Europeanness into account because Europe-related elite behaviour (p.209) and in particular elites’ policy preferences will reflect these patterns. It is pivotal for our understanding of Eurelitism, therefore, to identify and explain the causes of variation in the Europeanness of European economic and political elites.

This chapter pursues this agenda and provides a comprehensive causal analysis of the Europeanness of national political elites (see Chapter 1). Comprehensive means here that we will test causal models, which include a wide array of factors expected to explain the variance in the degree (favourable or unfavourable) of elites’ Europeanness. Some of these causal factors have already been introduced and discussed in earlier chapters of this book, such as contextual influences emanating from national polities and societies (see Chapter 7) or the effects of elites’ embeddedness in institutional and individual networks of supranational cooperation (Chapters 4 and 7). Others, like intra-elite cue taking or cognitive mobilization, will be introduced here.

10.1 Europeanness: the Explanandum

‘Europeanness’ will be measured according to the theoretical concepts and indicators introduced in the introductory chapter of this book. There, we identified a cognitive, an emotive, and a conative dimension of Europeanness emanating from elites’ attitudes towards European integration, their attachment to Europe, and their preparedness to accept a transfer of control over sovereignty rights to a supranational level in several key policy areas. We have further substantiated how these facets of Europeanness may be rooted in deeper mental layers of attitude formation. In order to examine the applicability of our conceptual suggestions, in particular regarding the concept of Europeanness, a short review of the dependent variables used in this chapter is appropriate. The concept of Europeanness is represented here by three items that were used identically in all versions of the elites and general population questionnaires, each capturing one of the three dimensions of Europeanness (see Chapter 12 Appendix).

  • The emotive dimension: ‘People feel different degrees of attachment to their town or village, to their region, to their country, and to Europe. What about you? Are you very attached, somewhat attached, not very attached, or not at all attached to…’ (attachment to Europe is reported here).

  • The cognitive-evaluative dimension: ‘Some say European unification should be strengthened. Others say it has already gone too far. What is your opinion?’ Answer categories range from ‘0’ meaning unification ‘has already gone too far’ to ‘10’ meaning ‘should be strengthened’.

  • (p.210) The projective or conative dimension: ‘Thinking about the European Union over the next ten years, can you tell me whether you are in favour or against the following: ‘A single foreign policy towards outside countries’. Answer categories range from ‘1’ meaning ‘strongly in favour’ to ‘4’ meaning ‘strongly against’.

The three items chosen to represent Europeanness also capture the three different subject areas addressed in the IntUne questionnaire, i.e. identity (here: attachment to Europe), representation (here: strengthening of unification), and scope of governance (here: common EU foreign policy), and therefore form a link between the construct of Europeanness used in this book and the general concepts used in the theoretical framework of the IntUne project (see Chapter 1; Cotta and Isernia 2009). For the purpose of this chapter, foreign policy was selected as a pivotal policy competence from a set of other policy areas covered in the IntUne questionnaire because it refers to the core competence of a sovereign state and its status as an independent subject of international law.

The three items representing the construct of Europeanness are significantly correlated, indicating that there is some communality between emotive, cognitive-evaluative, and projective-conative orientations towards Europe. Generally, connections are stronger within the elites than within the population. Within the elites, stronger connections prevail between the conative and cognitive dimensions than between these and the emotive dimension of Europeanness. Within the population, correlations between emotive and cognitive dimensions are stronger. Nevertheless, the correlations are far from being deterministic in all samples (the Pearson coefficients vary between 0.139 and 0.356; see Table 10.1), meaning that the three dimensions capture different and distinguishable facets of the construct. Earlier chapters of this book, in particular the contribution by Cotta and Russo, have further ascertained and defined the multidimensionality of elites’ attitudes towards Europe.

Contrary to expectations of an integrated ‘Eurelite’, we see massive differences between countries in all three dimensions of Europeanness and between all sub-samples targeted by the IntUne surveys (see Figures 10.110.3). Among political elites, attachment to Europe varies between 66 per cent ‘very attached’ in Poland and 10 per cent ‘very attached’ in the United Kingdom. In the management of large companies, banks, and employers’ organizations, extreme values of strong attachment to Europe vary between 79 per cent (France) and 10 per cent (United Kingdom), and between 46 per cent (Hungary) and 10 per cent (United Kingdom) in the general population.

Strong support (answer categories 8–10) for a strengthening of European unification are expressed by 67 per cent of Spanish and 59 per cent of Greek political elites, but only by 13 per cent of their Estonian colleagues, who are (p.211)

Table 10.1. Correlations between dimensions of Europeanness (Pearson’s r)

Attachment to integration

Attachment to single foreign policy

Integration with single foreign policy

Political elites

0.332***

0.251***

0.356***

Economic elites

0.243***

0.139***

0.273***

General Population

0.227***

0.194***

0.204***

Significance levels (two tailed): *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

preceded by only 15 per cent strongly pro-European British politicians. Among economic elites we find Estonia and United Kingdom again at the bottom of the ranking (0 [!] and 5 per cent, respectively) which is now topped by Austria (69 per cent) and Belgium (61 per cent). In the general population the range between top and bottom is smaller, but again we find the United Kingdom and Estonia (13 and 15 per cent respectively) at the bottom of the ranking, and Italy and Portugal (46 and 45 per cent respectively) at the top.

The project of a single foreign policy of the EU is strongly supported by 90 per cent of the Italian and 84 per cent of the Greek political elite, but only by 28 per cent of their Czech and 12 per cent of their British colleagues. Among the economic elites, the respective positions are occupied by Italy (90 per cent) and France (72 per cent) at the top of the ranking and the United Kingdom

Elite foundations of European integration: a causal analysis

Figure 10.1. Dimensions of Europeanness––attachment to Europe (% very attached)

Note: wording of question: see Appendix, item id01d.

(p.212)
Elite foundations of European integration: a causal analysis

Figure 10.2. Dimensions of Europeanness––unification should be strengthened (% strongly in favour)

Note: wording of question: see Appendix, item rp08.

Elite foundations of European integration: a causal analysis

Figure 10.3. Dimensions of Europeanness––single foreign policy (% strongly in favour)

Note: wording of question: see Appendix, item sg03_3.

(p.213) (0 per cent) and the Czech Republic (38 per cent) at the bottom. In the general population, the distance between extremes is again smaller with Bulgaria (42 per cent) and Greece (39 per cent) occupying the top ranks and Denmark (18 per cent) and the United Kingdom (14 per cent) being at the bottom.

Results of these country rankings challenge the consensus thesis of European unification and integration. In terms of their emotive, cognitive-evaluative, and conative-projective orientations towards Europe, European political and economic elites display huge variations at national level. In the general population the spread between countries is smaller but still sizeable. Only among Italian political and economic elites can the project of European integration count on a majority of strong supporters in all dimensions of Europeanness. In other countries, like the United Kingdom and Estonia (here with the exception of support for a common foreign policy) we found only small minorities of respondents with strong pro-European orientations. Although there is a tendency of Southern European political elites to be more pro-European and for elites from CEE-countries to show more Euroscepticism, the positioning of most countries varies strongly between the different rankings.

In some countries, elites seem to build elements of Europeanness into their very specific concepts of national identity––e.g. Poland as the defender of ‘true’ European values––or into elite strategies to promote national sovereignty––e.g. Europe as a shield against Russian attempts on Estonian independence. The national use and reinterpretation of European topoi may explain why elements of Europeanness sometimes appear in seemingly implausible and even contradictory combinations. Europe as the shell and support for a national revival and redefinition seems to be a particularly attractive solution for elites of the new member states of Central and Eastern Europe that had been force fed with internationalist ideologies under Soviet hegemony (Best 2009). Elites in these countries are naturally reluctant to embark on a new experience of internationalism and submission to a distant centre under European auspices. It was a challenging task of this book to give answers to the question of what factors underlay these between-countries differences and whether there are regional patterns recognizable in this European diversity (see, in particular, Chapter 7).

An important limiting factor of European elites’ Europeanness is the differences found between the elites and the general population in our survey. This elites–masses gap is seen in all the European countries included in both the elites and the general population surveys, and in all facets of Europeanness. With regard to both economic and political elites, we see an elite–masses differential giving economic and political elites an advantage in Europeanness over the general population (Hooghe 2003). Only in a very small minority of elites–masses comparisons does the general population have a lead over the elite in terms of Europeanness. In view of these results we can justifiably speak (p.214) of a ‘Europe of Elites’. The elite–masses differential is particularly large when it comes to the evaluation of the present state of European integration. The actual population is by far more sceptical towards a strengthening of European unification and the cession of rights of sovereignty to European institutions (here: the transfer of foreign policy competences from the member states to the EU) than elites. In view of these results, the failure to introduce a European constitution or a new constitution-like Treaty of the Union does not come as a surprise. The elite–masses gap limits the Europeanness of Europe’s political elites in that it restricts their options to broaden the European Union if this requires a referendum. As well as creating a strong temptation to enter into anti-integrationist populism, it also decreases the attractiveness of the European Union for economic elites, because a ‘Europe of citizens’ would probably be protectionist and restricted by the preservation of rights of national sovereignty, thus being somewhat of an impediment to economic freedom.

The limiting effect of the general population’s Euroscepticism on elites’ Europeanness must be, however, fairly inconsistent, because the elite–population gap varies considerably between countries. It comes close to zero in the United Kingdom and attains particularly high values for both economic and political elites in Belgium and France. CEE-countries show on average a somewhat smaller elite–population margin than their Western and Southern neighbours. The question of whether the elite–masses gap is indeed a causal factor influencing the Europeanness of individual members of the elite is a question that cannot, however, be decided on the basis of comparisons between aggregate data. To investigate the causes of elites’ Europeanness we need to embark on confirmatory data analyses of individual elite data. This is investigated in more depth in the following paragraphs of this chapter.

10.2 Exploring the Causes of Elites’ Europeanness

The observations reported in the previous paragraph require us to test causal models for each indicator separately and, because we assume that the formation of political and economic elites’ attitudes towards Europe and European integration follows a different logic, to differentiate between these two elite groups (see also Chapter 1). This latter decision is based mainly on the consideration that, for political elites, divergences over Europe are a subject of public controversy consequential in their competition for power. Economic elites, on the other hand, are expected to relate to Europe mainly as a market offering a set of opportunities to be utilized in their competition for financial gain.

For causal analyses of European political and economic elites’ Europeanness, the national sub-samples of each elite sector are merged separately to form comprehensive samples of ‘European’ economic and political elites. Weighting (p.215) procedures are not applied, so that all political and economic elites are included in the data analysis with the same weighting. A Multiple Regression Analysis is used to test causal models of European elites’ Europeanness. The three dimensions or facets of Europeanness (emotive; cognitive-evaluative; projective-conative) are tested separately for each elite group resulting in six Multiple Regression Models.

We derive the theoretical propositions regarding the determinants of elites’ Europeanness from general theoretical explanations of public support for European integration. This ‘universalistic’ approach is used as the starting point for empirical scrutiny, because we assume that elites are rooted in and closely related to their domestic societies and polities. The question as to what extent this is true in the given case, and in particular whether it has an impact on elites’ Europeanness, is central to this chapter, with the answer expected to provide important insights into the nature of Eurelitism. The specific status of elites is also considered in our explanatory models, in particular by including elite–masses differentials and elite–elite cues in the models.

In designing the causal or explanatory models for the investigation of Europeanness, we start by drawing on the theoretical propositions of structural-functionalism which have dominated—mostly implicitly––academic discourse about the foundations of European unification. Although functional integration theory has obvious shortcomings––as outlined in the introductory chapter––because of its teleological and harmonistic biases, the general framework of structural-functionalism may be considered as a source and guideline for the formulation of hypotheses and the choice of indicators. In its general form, this theoretical approach imputes ‘as rigorously as possible, to each feature, custom, or practice, its effect on the functioning of a supposedly stable, cohesive system’ (Bourricaud 1981: 94). Our short outline of an explanans for European elites’ Europeanness, which focused on the self-interests, belief systems, and network capital of European elites as a basis for their Europeanness, departed from a structural-functionalist position (see Chapters 1 and 4). In more general terms, it explained elites’ Europeanness by the advantages elites derive from the political and economic consequences of integration and the allegiances they show to supranational foci of identity. It also maintained that the integration and unification of Europe is an elite process emanating from the social structure of elite systems, i.e. from their interests, norms, traditions, interactions, and institutions.

In reformulating this general proposition into empirically testable statements, we start with the general hypothesis that European elites’ Europeanness will be determined by their level of embeddedness in transnational networks (i.e. their European ‘contact capital’), the scope of their biographical experiences (temporal and territorial), their cultural capital, and their religious and political belief systems (Fligstein 2008). However, this general hypothesis, (p.216) which holds that individual status, situs, and socialization experiences define their relation to the European context and thereby their level of Europeanness, has to be broken down further into partial hypotheses for empirical examination. First, we assume that where general political belief systems are linked to statist and nationalistic ideologies they will impact on elites’ Europeanness, because both are difficult to reconcile with European integration and the opening of national markets (Hooghe 2007; Schlesinger 2007). In this study, political ideology is measured by respondents’ positioning on an 11-point left–right self-rating scale.

We also assume that religious belief systems and denominational affiliations will have an impact on elites’ Europeanness, because Christianity formed a defining element of Europe (‘Christian occident’) against the Islamic world (Nelson, Guth, and Fraser 2001). On the other hand, there are some Protestant churches strongly attached to established national states (like in Great Britain and Denmark), whereas orthodoxy has its roots in the Eastern Mediterranean and a long history of hostile relations with the ‘Latin’ church of Western and Southern Europe (Best 2009). Both religious affiliations should therefore further Eurosceptic positions. In the models, religious belief systems are represented by the religious and denominational affiliations of elite members.

Another set of hypotheses claims that ‘macro-contexts’ (De Winter 2003), such as gender, territory, and age, will have an impact on elites’ Europeanness. We expect that the younger elites are, and the wider their territorial scope of biographical experience, the stronger their Europeanness will be, because younger elites are more likely to have been socialized in a nation-transcending area of communication. In particular, extended stays abroad should further cosmopolitan orientations and the readiness to attach effective allegiances to supranational institutions (Inglehart and Robier 1978; Fuss, Garcia-Albacete, and Rodriguez-Monter 2004). With regard to the effect of gender, female elites are expected to show higher levels of Europeanness, because the European Union champions policies of gender equality. In the case of female economic elites, however, European economic integration may reduce their level of Europeanness due to negative effects on their career chances stemming from the male domination of internationally operating concerns (Nelson and Guth 2000). The territorial scope of biographical experience will be represented by the elites’ record of extended stays in other European countries.

We also assume that elites’ social and cultural capital is associated with their level of Europeanness. This hypothesis is based on Inglehart’s theory of cognitive mobilization, which claims that increases in education and access to information encourage citizens to develop a more cosmopolitan outlook that benefits support for European integration (Inglehart 1971). Higher levels of education are also expected to have a ‘functional’ link to Europeanness, because well-educated individuals are assumed to have the cognitive ability (p.217) and social competences with which to interact in the complex institutional and cultural settings of an integrated Europe. In the models, therefore, cultural capital is introduced by level of education. The social capital hypothesis assumes that the deeper the involvement of elites in institutional networks at European level, the stronger will be their Europeanness. The link is constituted by the increased subjective and objective value of transnational contact capital in the case of deepening of European integration. Involvement in institutional networks is represented by elites’ self-reported contacts with organizations and authorities at the European level.

Structuralist and functionalist theories, which see Europeanness as a result of the perceived benefits and welfare derived from the political and economical consequences of integration, have been recently challenged by ‘post-functionalist theories’ that focus instead on the range and nature of identity-forming collective historical experiences (Hooghe and Marks 2008). The core of this approach is the affective and perceptive bases for the allegiances of individuals to supranational institutions and collectivities. It is expected here that these allegiances are, to some extent, independent of the locus of an individual in the social structure. A ‘post-functionalist’ perspective, therefore, requires us to consider in the explanans attitudinal data referring to identities, subjective evaluations, and trust. The extension of our research agenda into the realm of cognitive, normative, and affective frames will be achieved by including three attitudinal variables in the multiple regression models:

  • Attachment to one’s own country. Here we expect a trade-off between emotive ties to one’s own country and attachment to Europe, because identification with different in-groups is considered to be a zero-sum phenomenon (Carey 2002). There will also be a negative impact of a strong attachment to one’s own country on the approval of further European integration and a cession of national sovereignty to the European level, because ‘the stronger the bond that an individual feels towards the nation, the less likely that individual will approve of measures that decrease national influence over economics and politics. The growth in scope of the European Union in the realm of economics, politics and culture, which have previously been under the sole control of the nation state impinges on this view of the nation’ (Carey 2002: 391). Strong attachment to one’s own nation should therefore be negatively connected to attachment to Europe. Strong attachment to one’s own nation should also reduce the acceptance of multilevel governance (deeper integration) and the transfer of authority for foreign policy to the European level.

  • No such trade-off is expected with regard to trust in the institutions of the European Union (here: the EU Commission). Research into political (p.218) support has identified belief in the integrity and performance of its core institutions as an important factor in forming an allegiance to representative democracy (Newton 2007). We expect the same mechanism to be working in relation to EU institutions, whereby trust in the EU Commission will be positively linked to elites’ Europeanness, i.e. to a high attachment to Europe, to an acceptance of multilevel governance, and to support for a transfer of authority to the European level.

  • A similar pro-integrationist effect will result from the assessment that one’s country has benefited from membership in the European Union. Perception of the utility of EU membership for own country will, therefore, increase elites’ Europeanness in all three dimensions of the concept. We have, however, to be aware that, due to the very high overall level of agreement among elites that their countries have benefited from EU membership, the power of this variable to explain variation of Europeanness will be limited.

A third set of factors influencing the Europeanness of political and economic elites is considered to be specific to power holders and decision makers. Here we are referring in particular to the cues regarding European issues that national elites receive from their domestic environments. First, however, we need to look at the relationship between elites and non-elites, which is––particularly for political elites––a factor constraining and directing their involvement in supranational institution building and European policy making. In abstract terms, the complex intertwining between the various levels of the European system of governance has been described recently as a ‘compliance-legitimacy relationship between the union and its member-states’ which is ‘constrained by the basic compliance–legitimacy relationship between member-governments and their constituents’ (Scharpf 2009b: 173). In more simplistic terms, it means that voters hold politicians accountable for their decisions regarding the European Union, which is translated by the politicians into a considerable electoral hazard. The theory of ‘constraining dissensus’ has recently put these elites–masses interactions at the centre of the argument (Hooghe and Marks 2008). We expect, therefore, to see some responsiveness on behalf of political elites in accordance with the preferences of their domestic populations regarding integration policies and allegiances towards the European Union. As these relationships concern constituents and their representatives in particular, we do not expect to see the same pattern in the relationship between economic elites and the domestic populations in their home countries (see Chapter 8). We do expect, however, to find indications for mutual cue-taking between economic and political elites. One reason is the expectation that EU-related interests and perceptions of economic and political elites are linked, which is (p.219) grounded in the fact that the EU is a political union that has evolved from an economic community and still has at its core a common market. The historical background and the actual operational requirements of running the EU imply, and to some extent require, an understanding between political and economic elites about their status in the integration process and their allegiances towards Europe. For this reason, we include the national averages of elites’ and their respective general populations’ agreement concerning the three indicators of Europeanness as contextual variables in the models.

With the inclusion of contextual variables, which provide averages of mass and elite attitudes and orientations for the countries covered in this study, we are taking the multilevel character of our data into consideration (Steenbergen and Jones 2002). Nevertheless, we have to be aware that this specification of our model is no proper multilevel analysis and that Beta estimates and the calculation of significances in Ordinary Least Squares (OLS) regressions can be seriously biased by the ‘nested’ structure of the data and the resulting distortion of standard errors. The ‘methodological concern of heteroskedasticity when pooling data from cross-sections as varied as the member countries of the EU’ (Carey 2002: 397) is complemented by our interest in the overall impact of national diversity on the variation of elites’ Europeanness. The question to be addressed here is whether and to what extent Eurelitism is basically a national phenomenon, reflecting national agendas and conditions of action.

In sum, we examine seven hypotheses that are assumed to impact on the emotive, cognitive-evaluative, and projective-conative dimensions of Europeanness. These are: elites’ position in the transnational social structure and institutional networks of Europe; the range of their historical experiences; their cultural capital; cues from peers and masses in the elites’ national environments; attitudes towards their own nation; the level of trust they confer on the EU Commission; and the perceived benefits of EU membership for their country. These are examined with ‘Attachment to Europe’, ‘Strengthening of European integration’, and ‘Support for a single European foreign policy’ as dependent variables. Regression models will be tested in two steps, with attitudinal variables being included in the second step. This procedure will allow us to ascertain the overall effects (direct and indirect) of the structural, contextual, and attitudinal variables on the dependent variables.

Results of OLS-regression analyses are presented in six separate tables (see Tables 10.210.7) for both elite groups (political and economic) and three indicators of Europeanness (attachment to Europe, attitude towards unification, and attitude towards a single foreign policy). Each table contains five differently specified models: two single-level models, one including and the other excluding attitudinal variables. It also presents three variants of the multilevel model: one ‘empty’ model, which considers only the constant (p.220)

Table 10.2. Multiple regression model for attachment to Europe––political elites

Attachment to Europe (1 = none at all; 4 = high)

political elites

Independent variables

Single-level models

Multilevel models

1

2

3

Beta

Beta

Empty model

Without context variables Beta

With context variables Beta

Self-allocation on Left–right scale (collapsed: 0-centre 5-extreme)

−0.093**

−0.010

−0.012

−0.011

Religion (Ref. cat.: none)

Orthodox

−0.028

−0.038

−0.031

−0.012

Catholic

−0.002

−0.028

−0.029

−0.036

Protestant

−0.022

−0.037

−0.067

−0.045

Other

−0.048

−0.040

−0.041

−0.036

Gender (Ref. cat.: male)

0.074*

0.066*

0.054*

0.060*

Age (mean centred)

0.063*

0.001

0.018

0.010

Education

0.059*

−0.002

0.024

0.014

Contact frequency to EU actors and institutions (1 = no contacts last year; 5 = at least once a week)

0.148***

0.112***

0.095***

0.105***

Context variables

Duration of EU Membership

−0.030

0.037

0.042

Elite consensus

0.279***

0.258***

0.275***

Elite–mass responsiveness

0.037

0.053

0.050

Attachment to own country (1 = low; 4 = high)

0.237***

0.241***

0.238***

Trust in European institutions (0 = none; 10 = high)

0.227***

0.215***

0.222***

Has your country benefited from EU membership? (0 = no; 1 = yes)

0.203***

0.203***

0.203***

Constant/intercept

−0.219

−2.087***

3.208***

0.767***

−2.290***

Adjusted R²/Maddala R²

0.116***

0.286***

0.082***

0.285***

0.300***

Intra-class correlation (ICC)

14.4%

11.9%

2%

−2Log‐likelihood

2166.210

1917.474

1896.963

Log-likelihood-ratio testa

85.589***

248.736***

20.511***

N

1034

995

995

995

995

Significance levels: *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

a -2LL single level empty model = 2251,799

and shows how much of the variance in the dependent variables can be at most explained by the aggregate level (here: country); one model that only contains individual-level independent variables; and a comprehensive multilevel model, which also includes context variables that measure country differences at the aggregate level. The discussion of results will go through clusters of independent variables sequentially, starting with ideological self-placement and religious affiliations. (p.221)

Table 10.3. Multiple regression model for attachment to Europe––economic elites

Attachment to Europe (1 = none at all; 4 = high) economic elites

Independent Variables

Single-level models

Multilevel models

1

2

3

Beta

Beta

Empty model

Without context variables Beta

With context variables Beta

Self-allocation on left–right scale (collapsed: 0-centre 5-extreme)

−0.141**

−0.160***

−0.155***

−0.153***

Religion (Ref. cat.: none)

Orthodox

0.007

−0.004

−0.066

−0.011

Catholic

0.054

0.061

0.050

0.036

Protestant

0.011

0.017

0.061

0.053

Other

0.012

−0.003

−0.006

−0.006

Gender (Ref. cat.: male)

0.049

0.052

0.050

0.054

Age (mean centred)

0.113*

0.079

0.077*

0.072

Education

−0.009

−0.012

−0.022

−0.018

Contact frequency to EU actors and institutions (1 = no contacts last year; 5 = at least once a week)

0,045

0.037

0.038

0.041

Context variables

Duration of EU membership

0.010

0.040

0.060

Elite consensus

0.186**

0.120*

0.122a

Elite–mass responsiveness

−0.017

0.004

0.009

Attachment to own country (1 = low; 4 = high)

0.293***

0.317***

0.310***

Trust in European institutions (0 = none; 10 = high)

0.167***

0.194***

0.191***

Has your country benefited from EU membership? (0 = no; 1 = yes)

0.003

0.010

0.008

Constant/intercept

1.552*

0.516

3.217***

1.669***

0.322

Adjusted R²/Maddala R²

0.074***

0.171***

0.037

0.200***

0.206***

Intra-class correlation (ICC)

9.1%

6.5%

5.0%

−2Log‐likelihood

1068.917

975.993

971.932

Log-likelihood-ratio testa

18.617***

92.924***

4.061*

N

524

499

499

499

499

Significance levels: *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

a -2LL single level empty model = 1087,534

10.3 Empirical Findings

The expected effect of political ideology is overwhelmingly supported by empirical scrutiny. With the exception of economic elites’ approval of a strengthening of European unification, political ideology has a significant effect in five out of the six regression models (Tables 10.210.7). This effect, (p.222)

Table 10.4. Multiple regression model for attitude towards unification––political elites

Attitude towards Unification (0 = against; 10 = in favour) political elites

Independent variables

Single-level models

Multilevel models

1

2

3

Beta

Beta

Empty model

Without context variables Beta

With context variables Beta

Self-allocation on left–right Scale (collapsed: 0-centre 5-extreme)

−0.070*

−0.016

−0.011

−0.016

Religion (Ref. cat.: none)

Orthodox

0.005

0.016

−0.059

0.016

Catholic

−0.106**

−0.089**

−0.072*

−0.089*

Protestant

−0.022

−0.026

−0.056

−0.026

Other

0.035

0.030

0.018

0.031

Gender (Ref. cat.: male)

0.040

0.030

0.023

0.030

Age (mean centred)

0.039

0.011

0.016

0.011

Education

0.032

−0.059

−0.051

−0.059

Contact frequency to EU actors and institutions (1 = no contacts last year; 5 = at least once a week)

0.098**

0.081**

0.079**

0.081*

Context variables

Duration of EU membership

0.048

0.059

0.059

Elite consensus

0.254***

0.240***

0.240***

Elite–mass responsiveness

0.108**

0.110**

0.110**

Attachment to own country (1 = low; 4 = high)

−0.031

−0.031

−0.031

Trust in European institutions (0 = none; 10 = high)

0.218***

0.217***

0.218***

Has your country benefited from EU membership? (0 = no; 1 = yes)

0.234***

0.234***

0.234***

Constant/intercept

−1.278

−2.960**

6.731***

3.862***

−2.956**

Adjusted R²/Maddala R²

0.129***

0.268***

0.100***

0.252***

0.280***

Intra-class correlation (ICC)

14.6%

14.4%

〈0.01%

−2Log‐likelihood

4278.003

4101.532

4065.482

Log-likelihood-ratio testa

99.753***

176.47***

36.05***

N

990

951

951

951

951

Significance levels: *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

a -2LL single level empty model = 4377,756

however, only appears after ‘folding’ the left–right scale and its transformation into an extremist–moderate scale. By a folding procedure, scale-positions left and right of the mid-point of the scale are collapsed and (ignoring signs) added, so that moderate positions receive low and extremist positions receive high values. The resulting Beta values show that political extremism, notwithstanding its position on the right or on the left, reduces elites’ attachment to (p.223)

Table 10.5. Multiple regression model for attitude towards unification––economic elites

Attitude towards unification (0 = against; 10 = in favour) economic elites

Independent variables

Single-level models

Multilevel models

1

2

3

Beta

Beta

Empty model

Without context variables Beta

With context variables Beta

Self-allocation on left–right scale (collapsed: 0-centre 5-extreme)

−0.064

−0.051

−0.054

−0.050

Religion (Ref. cat.: none)

Orthodox

−0.019

−0.005

0.056

−0.001

Catholic

0.051

0.059

0.027

0.044

Protestant

−0.066

−0.074

−0.085

−0.079

Other

−0.010

−0.018

−0.015

−0.018

Gender (Ref. cat.: male)

−0.100*

−0.086*

−0.093*

−0.087*

Age (mean centred)

0.012

0.025

0.014

0.017

Education

−0.037

−0.033

−0.037

−0.034

Contact frequency to EU actors and institutions (1 = no contacts last year; 5 = at least once a week)

0.021

0.009

0.028

0.015

Context variables

Duration of EU membership

0.011

0.010

0.023

Elite consensus

0.352***

0.353***

0.349***

Elite–mass responsiveness

−0.046

−0.070

−0.067

Attachment to own country (1 = low; 4 = high)

0.062

0.061

0.065

Trust in European institutions (0 = none; 10 = high)

0.104*

0.132**

0.112**

Has your country benefited from EU membership? (0 = no; 1 = yes)

0.001

0.017

0.007

Constant/intercept

2.983*

2.003

6.751***

5.796***

1.837

Adjusted R²/Maddala R²

0.144***

0.142***

0.094***

0.138***

0.170***

Intra-class correlation (ICC)

15.2%

13.8%

1.7%

−2Log‐likelihood

2033.860

2009.286

1990.801

Log-likelihood-ratio testa

48.83***

24.574***

18.485***

N

519

493

493

493

493

Significance levels: *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

a -2LL single level empty model = 2082,690

Europe, their acceptance of European integration, and their willingness to support a single European foreign policy in the future. The relationship between political extremism and Europeanness is, however, as the Beta values also show, rather weak and unstable. Only in the case of economic elites’ attachment to Europe does the effect of political ideology ‘survive’ the significance threshold after the inclusion of attitudinal variables in the models. We (p.224)

Table 10.6. Multiple regression model for attitude towards a single foreign policy––political elites

Attitude towards a single foreign policy (1 = against; 5 = in favour) political elites

Independent variables

Single-level models

Multilevel models

1

2

3

Beta

Beta

Empty model

Without context variables Beta

With context variables Beta

Self-allocation on left–right Scale (collapsed: 0-centre 5-extreme)

−0.074*

−0.019

−0.019

−0.019

Religion (Ref. cat.: none)

Orthodox

−0.006

0.012

0.022

0.009

Catholic

−0.044

−0.022

−0.007

−0.018

Protestant

−0.015

−0.020

−0.042

−0.021

Other

0.014

0.009

0.011

0.009

Gender (ref. cat.: male)

−0.020

−0.030

−0.013

−0.025

Age (mean centred)

0.052

0.048

0.036

0.043

Education

0.080**

0.009

−0.002

0.006

Contact frequency to EU actors and institutions (1 = no contacts last year; 5 = at least once a week)

0.010

−0.018

−0.016

−0.017

Context variables

Duration of EU membership

0.051

0.048

0.048

Elite consensus

0.287***

0.263***

0.264***

Elite–mass responsiveness

−0.024

−0.024

−0.023

Attachment to own country (1 = low; 4 = high)

−0.072*

−0.075*

−0.077*

Trust in European institutions (0 = none; 10 = high)

0.131***

0.142***

0.133***

Has your country benefited from EU membership? (0 = no; 1 = yes)

0.274***

0.272***

0.273***

Constant/intercept

0.562

0.257

4.244***

3.177***

0.257

Adjusted R²/Maddala R²

0.101***

0.216***

0.074***

0.206***

0.229***

Intra-class correlation (ICC)

13.4%

11.5%

0.6%

−2Log‐likelihood

2915.464

2762.176

2732.959

Log-likelihood-ratio testa

76.212***

153.288***

29.217***

N

1031

993

993

993

993

Significance levels: *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

a -2LL single level empty model = 2991,676

attribute this instability (which is particularly reflected in the volatility of Beta values after inclusion of attitudinal variables) to the problematic validity of the left–right scale and to the heterogeneity of extremist political camps regarding their Europeanness. The initial hypothesis holds, however, for economic elites’ attachment to Europe, where we see a strong and stable effect of political extremism, in that extremism diminishes their attachment to (p.225)

Table 10.7. Multiple regression model for attitude towards a single foreign policy––economic elites

Attitude towards a single foreign policy (1 = against; 5 = in favour) economic elites

Independent variables

Single-level models

Multilevel models

1

2

3

Beta

Beta

Empty model

Without context variables Beta

With context variables Beta

Self-allocation on left–right Scale (collapsed: 0-centre 5-extreme)

−0.110**

−0.079

−0.063

−0.079

Religion (Ref. cat.: none)

Orthodox

0.081

0.082

0.108

0.081

Catholic

0.017

−0.002

−0.027

−0.002

Protestant

0.031

0.004

−0.043

0.004

Other

−0.067

−0.078

−0.069

−0.078

Gender (Ref. cat.: male)

−0.037

−0.008

−0.007

−0.007

Age (mean centred)

−0.027

−0.037

−0.033

−0.038

Education

−0.062

−0.058

−0.059

−0.058

Contact frequency to EU actors and institutions (1 = no contacts last year; 5 = at least once a week)

−0.053

−0.073

−0.073

−0.072

Context variables

Duration of EU membership

−0.019

0.003

0.003

Elite consensus

0.304***

0.263***

0.262***

Elite–mass responsiveness

0.019

0.003

0.003

Attachment to own country (1 = low; 4 = high)

0.070

0.074

0.070

Trust in European institutions (0 = none; 10 = high)

0.123**

0.142***

0.123**

Has your country benefited from EU membership? (0 = no; 1 = yes)

0.031

0.030

0.031

Constant/Intercept

1.445

1.102

4.360***

4.147***

1.115

Adjusted R²/Maddala R²

0.093***

0.102**

0.033

0.079***

0.129***

Intra-class correlation (ICC)

13.5%

7.6%

〈0.01%

−2Log‐likelihood

1342.130

1318.113

1290.469

Log-likelihood-ratio testa

16.826***

24.017***

27.644***

N

525

496

496

496

496

Significance levels: *** sig. 〈 0.001; ** sig. 〈 0.01; * sig. 〈 0.05

a -2LL single level empty model = 1358,956

Europe. We assume that this stability and consistency can be ascribed to the fact that political extremism for economic elites means predominantly right-wing extremism.

On the other hand we found none of the expected effects of religious belief systems. Nevertheless, a weak but stable negative effect of Catholic affiliation among political elites on their attitudes towards a deepening of European (p.226)

Elite foundations of European integration: a causal analysis

Figure 10.4. Country has benefited from EU membership (% benefited)

Note: wording of question: see Appendix, item ev2.

integration suggests that Catholic politicians consider an integrated Europe as a threat to their religious values rather than as an institutional frame for the Christian occident. Conflicts with the Catholic Church over the enforcement of equality laws, and the rejection of claims to include religious references in the draft of the European constitution, may have reduced the enthusiasm of devout Catholic politicians concerning European unification.

Hypotheses assuming that the Europeanness of elites is influenced by their positioning in the social structure are either refuted, weakly supported, or work in an unexpected direction. Surprisingly, the territorial scope of previous European experiences through migration for the purpose of study had no significant impact at all and was therefore omitted from the models. Age has a significant and positive impact on both economic and political elites’ attachment to Europe, meaning that older elites show a stronger attachment to Europe. This finding runs contrary to the direction of the age effect expected according to the theory of cognitive mobilization. It seems that a wider range of historical experience of older elites has a positive effect on their attachment to Europe. The stereotype of young, cosmopolitan, and pro-European elites is, it seems, exactly that. Significant effects for age, however, are only found in the attachment to Europe models and disappear when we control for attitudinal variables.

The effect of gender is more stable but inconsistent in its direction: while, as expected, female political elites show a stronger attachment to Europe than their male colleagues, female economic elites tend to show a distinct (p.227) aversion to a further deepening of European unification. We attribute this to the potential threat posed by expanding Western companies to female-held management positions in Central and East European (CEE) countries. This interpretation conforms to other findings that the gender gap in EU support can be predominantly attributed to the positioning of women in the labour market (Nelson and Guth 2000). It is also compatible with our finding that female politicians show stronger attachment to Europe than their male colleagues, because there is no international competition in legislative recruitment markets (Best 2007).

The hypothesis that higher levels of education enhance elites’ Europeanness is partly confirmed. Education has significant and (although fairly weak) positive effects, however, only for political elites and only in relation to their attachment to Europe and their support for a single European foreign policy. The question remains open whether this finding supports Inglehart’s cognitive mobilization theory, because attachment was introduced here as an emotive concept, whereas we see no effect of level of education on the cognitive facet of Europeanness.

Stronger and more consistent support is given for the impact of network capital. We find significant and positive effects of contacts to institutions and authorities at EU level on political elites’ attachment to Europe and on their support for integrationist positions. Although no significant impact of European contact capital on any of the indicators of Europeanness is found for economic elites, results show that the involvement of national politicians in institutional networks and arenas at the European level increases their support for European integration and unification. The overall effect of European contact capital is stable after controlling for attitudes and the impact of national context; it is, however, rather weak.

The hypothesis that responsiveness to cues from reference groups and the general population influences elites’ levels of Europeanness is strongly and consistently confirmed with regard to cues coming from other elite groups. In all three dimensions of Europeanness, and in all variants of the models, we find orientation of members of the political elite influenced by the average national level of Europeanness among economic elites and vice versa. The consistency, strength, and symmetry of elite–elite responsiveness at the national level is a strong indication of the existence of integrated elite systems at that level, which are probably connected by intensive inner flows of communication and by peer pressure, thereby rectifying orientation and behaviour. This interpretation is the most plausible explanation for the somewhat bewildering finding that individual levels of Europeanness in one elite group are heavily influenced by the aggregate levels of Europeanness in the other. This suggests that systemic interconnection results in a mutual adaptation of elements of Europeanness between elite groups.

(p.228) Cues from the masses are only significant for political elites and only in the case of their attitude towards integration. This result conforms to our expectations and fits plausibly into a comprehensive model of interconnectedness. At the national level, we find a ‘half-circle of responsiveness’ that connects political elites and masses, and both elite groups reciprocally. This half-circle of responsiveness absorbs other contextual or regional factors of elites’ and masses’ Europeanness, such as the length of their country’s EU membership, which has no significant effect in any of the six models due to suppressor effects of elite–elite reciprocity and elite–masses responsiveness. If the average levels of masses’ and elites’ Europeanness are removed from the models, however, we see the expected positive effects of length of EU membership on elites’ Europeanness. The most important realization emanating from these findings is that we find here a structure of elite–elite and elite–masses’ relations that actually caves in and perpetuates national differences.

At this point of the analysis, and with corrected R2 varying between 0.14 and 0.07, the overall explanatory contribution of the independent variables is rather low. The Europeanness of European elites is only weakly anchored in their religious affiliations, political belief systems, and their locus in the social structure. The strongest impact on Europeanness comes from significant others, particularly from cues taken from fellow elites in other sectors and ––to a lesser degree and only for political elites––from the population at large. Structural-functionalist approaches are obviously only of limited value when it comes to explaining European economic and political elites’ Europeanness. This holds true, even if we include aggregate data about national contexts in our analyses. However, a marked improvement of the explanatory power of multiple regression models, more than doubling R2 values in the case of political elites, is achieved when attitudinal variables are included. The model with attachment to Europe now attains corr R2 = 0.29 for political elites (after corr R2 = 0.12) and corr R2 = 0.17 (after corr R2 = 0.07) for economic elites. However, the same effect cannot be seen for economic elites regarding the other two dependent variables.

If we look at the impact of specific attitudes we can see that an attachment to one’s own country has a significant effect in three out of the six models. However, the original hypothesis is only confirmed for political elites’ acceptance of transferring authority for foreign policy to the EU level. Here we find a weak but stable effect that results in a reduced agreement to such a transfer if attachment to one’s own country is high. The next result came as a surprise: attachment to one’s own country has a strong and consistent, but reverse effect (to what was expected) on attachment to Europe. Both are positively linked in both economic and political elites. In this case, in-group identification is not a zero-sum phenomenon, but it is mutually reinforcing in multilevel European settings. From these findings, we can refute the concept of a (p.229) ‘terminal community’, which claims there is a highest––typically national––political unit to which people feel they owe allegiance (Deutsch 1966). On the other hand, we do find support for the thesis that ‘European allegiance originates in national allegiance and that European integration depends on a primary allegiance to the nation-state and a secondary or derived allegiance to the EU’ (Van Kersbergen 2000: 19). There is, however, a caveat in that this conclusion only holds true for the affective facet of Europeanness, whereas we found no effect for the cognitive-evaluative side of the concept and a reverse effect on its conative-projective dimension. Politicians, who are closely attached to their own country, are reluctant to agree to a transfer of control over foreign policy from national to European authority. This inconsistency should warn us not to have too harmonistic ideas about the impact of multilevel allegiances on the process of European integration.

The original hypothesis concerning the effect of elites’ trust in the European commission is convincingly confirmed. This has the expected reinforcing effect in all six models and is the most consistent of all three attitudinal background variables. The belief in good (i.e. trustworthy) governance of European institutions obviously reinforces European economic and political elites’ Europeanness. This implies, however, that elites’ Europeanness is conditional on their perception of the trustworthiness of European institutions, which might become shaky ground with the fading of permissive consensus, as discussed in Chapter 1. Similar problems may arise with the perceived benefits to one’s own country of EU membership. This has the expected reinforcing effect on political elites’ Europeanness in all three models, but in none of the models for economic elites. It would be premature, however, to conclude from this result that the Europeanness of economic elites is not driven by perceptions regarding the benefits of their own country’s EU membership, because the non-relation can be easily explained by the extreme skewness of the independent variable in the case of this group: only 2.5 per cent of the economic elite respondents did not consider EU membership to be beneficial for their country, so that there is simply not enough variance in this variable to add any explanatory power to the model. Perceived benefits of EU membership are not only one of the strongest instances of pan-European elite integration; they are also an important explanatory factor for elites’ Europeanness (see Figure 10.4).

10.4 The Impact of National Elites

One of the central issues addressed in this chapter and the focus of our theoretical consideration is the thesis that European integration can be conceptualized and interpreted as a process of elite integration, which involves (p.230) national or domestic elites as the main actors in the building and operation of European institutions. We have also suggested that this process should be beyond normative and structural integration––i.e. the convergence of elites in terms of their make-up and outlooks––and entail ‘systemic integration’, i.e. an assimilation of the social and mental mechanisms that shape these elites. Only when this precondition is fulfilled can we rightfully speak of ‘Eurelitism’ (see Chapter 1); when it is not, we should continue to treat national elites as separate social groupings and avoid merging national samples into one European meta-sample. This problem can be approached technically in different ways, such as, for example, by the introduction of national units as dummy variables. A more appropriate approach, however, is to apply a multilevel analysis that considers explicitly the ‘nested’ or multilevel structure of data in specifying OLS-regression models. This raises the question of whether the consideration of the multilevel structure of our data in the causal models actually changes the fundamental patterns of variable relationships and thereby the interpretation of the models. If this is the case, multilevel analysis would mean treating national elites as unconnected collectivities and dropping the concept of Eurelitism.

Results of log-likelihood ratio tests and intra-class correlations do indeed show that multilevel modelling considering country as level specification does significantly improve the fit of the six models. We have, therefore, to be aware that we are dealing with aggregates of national elites. The other side of the coin, however, is that the adaptation of a multilevel design does not dramatically change the general pattern of OLS-regression analysis results. None of the Beta values changes signs or significance levels. In addition, the improvement of model fit achieved by multilevel modelling is modest, because we have to take into consideration that a multilevel element had already been introduced into our data when we included context variables like elite–consensus and elite–masses responsiveness. The most dramatic change in the models, and one that had a major impact on the signs and significances of the Beta values, was the introduction of attitudinal variables, which also raised R2 values to levels only slightly below those attained in the comprehensive multilevel models. The conclusion is that our elite samples have a hybrid character, in that they are displaying features indicating the effectiveness of ‘systemic integration’ at the European level, as well as the persistent significance of national arenas. Overall, however, ‘systemic integration’ seems to dominate. This is particularly the case among economic elites, where the need to respond to national constituencies is missing.

(p.231) 10.5 Conclusion

Our analyses of European political and economic elites’ attitudes towards Europe have re- or rather de constructed their Europeanness as a loosely coupled configuration of emotive, cognitive-evaluative, and conative-projective dimensions that are combined in nation-specific patterns (see Table 10.1; also Chapters 2, 4, and 6). The generative logic behind these national patterns shows that European integration and unification cannot be based on a uniform pan-European consensus, or even on a general majority of strong supporters with regard to any of the three dimensions of Europeanness. So far, the Europe of elites is a rather ephemerons construct, being based on a wide and sometimes contradictory diversity of value concerns, interests and assessments. On the other hand, this variety ensures that Europhobia in all of the three dimensions is a relatively rare configuration. The only case in our sample of countries where there is a consistent pattern of Europhobic elite orientations is the UK, which acquires an outlier status in this respect. Overall, results converge after all in the highly contradictory realization that the pan-European communality in European elites’ Europeanness results from its national diversity.

A second fundamental result of our study is the ubiquity of an elites–masses gap in terms of Europeanness. With very few exceptions, elites in general are more Europhile than the general public in all territorial and substantial aspects of our aggregate-level analyses, and economic elites in particular tend to display higher levels of Europeanness than their political counterparts. The elite–masses differential is particularly distinct with regard to the cognitive-evaluative and the projective dimension of Europeanness, i.e. when it comes to the question of whether it was right to have ceded national sovereignty to European institutions and authorities and whether this should be expanded in the future. European citizens are less prepared than European elites to accept such a cession now or in the future. We interpret the elite–masses gap as an indication of conflict over citizens’ rights, in that European citizens are reluctant to accept a Europe of elites where distant authorities cannot be sanctioned at the ballot box, or at least to a lesser degree than national governments, and only indirectly. It can be also shown that the general population is more sceptical about the benefits of EU membership for their countries than both elite groups.

Causal analyses of factors determining European political and economic elites’ Europeanness have shown a highly diverse picture that converges, however, in two main results: evidence for an impact of elites’ situs, status, previous biographical experience, or religious affiliation is fragmented, contradictory, weak, or non-existent. This means that earlier benefits received by (p.232) some elite groups from policies of the European Union, generational change, migration, and education, have no or only limited impact on European elites’ Europeanness. Political ideology, however, generates a more consistent impact: both radicals of the right and the left tend to be anti-European, although probably for different reasons. Nevertheless, although relatively consistent and statistically significant, the effect of political ideology is not very strong and disappears almost completely after the inclusion of attitudinal variables. Another, more consistent finding pointing in the direction of an integrative effect of an involvement in multilevel governance is that participation in pan-European networks increases political elites’ Europeanness.

An influential explanans of European elites’ Europeanness is the average national level of other elite groups’ Europeanness. We have interpreted this result as an effect of mutual cue-taking and peer pressure between elite groups at the national level. Together with the impact of the average national level of Europeanness in the general population on political elites’ Europeanness, we have here an indication of the existence of a half-circle of responsiveness which links elites and masses at the national level. It is obvious that these links, which operate at a national level, countervail the process of transforming national elites into a fully integrated Eurelite.

The inclusion of the three attitudinal variables in the multiple regression models confirmed the expected influence of ‘Trust in the European Commission’ and the perception of ‘Benefits of EU membership for one’s own country’ in all three models, and for both elite groups. Trust in, and perceived performance of, EU multilevel governance strengthens attachment to Europe, as well as increasing respondents’ acceptance of a stronger European integration and a transfer of foreign policy competences from the national to the European level. In sum: the perception of good European governance strengthens Europeanness and vice versa. The result for ‘Attachment to one’s own country’ came as a surprise, contradicting our expectations: there is no trade-off, but rather a strong convergence between emotive ties to the national and European focus of identity, indicating a mutual reinforcement of these ties. We also find that a strong attachment to one’s own country has no negative impact on consent to further European integration and to a transfer of authority from the national to the European level. Only in the case of the acceptance of a single EU foreign policy, and even then only for political elites, do we see the expected ‘trade-off’.

The task of integrating these findings into a comprehensive theory of regional integration might be not as challenging as it first appears. If respondents view European integration as being beneficial for their own country, then there is no contradiction between an attachment to one’s own country and an attachment to Europe, or the approval of deeper European unification. If attitudes towards a deeper unification of Europe and a transfer of authority (p.233) from the national to the European level depend mainly on the evaluation of multilevel governance, we may also turn to attitudinal data in order to explain the elite–masses differential: elites have a much more positive opinion about the success of European governance for their countries than the general population. Whereas only 2 per cent of the respondents in the economic elite samples and 6 per cent in the political elite samples say that their country has on balance not benefited from membership in the European Union, 27 per cent of the general population samples agree with this statement. This leads to the final conclusion that the process of European integration is not so much driven by deeply rooted cognitive or normative concepts of national or regional identities, or by the emergence of ever denser and wider networks of transnational cooperation, as by the daily demonstration of good multilevel governance. If this diagnosis is true and elites, like the general public, base their Europeanness on the utility and trustworthiness of European institutions, the foundations of European integration are much weaker than those of consolidated national states, which can count on the solidarity and attachment of their citizens, whatever the performance of their governments. The European Union will have quite a long way to go before a majority of its inhabitants, including its elites, proclaim: ‘My Europe, right or wrong!’, and we can conclude that it is still based on the maxim which initiated the process of integration: ‘S’unir ou périr!’