Scholarly article on topic 'Calculation, Community and Cues: Public Opinion on European Integration'

Calculation, Community and Cues: Public Opinion on European Integration Academic research paper on "Political Science"

Share paper
Academic journal
European Union Politics
OECD Field of science

Academic research paper on topic "Calculation, Community and Cues: Public Opinion on European Integration"

European Union Politics

Calculation, Community and Cues: Public Opinion on European Integration

Liesbet Hooghe and Gary Marks European Union Politics 2005 6: 419 DOI: 10.1177/1465116505057816

The online version of this article can be found at:

Published by: ®SAGE

Additional services and information for European Union Politics can be found at: Email Alerts: Subscriptions: Reprints: Permissions: Citations:

>> Version of Record - Nov 4, 2005 What is This?

Calculation, Community and

Public Opinion on European Integration

^ Liesbet Hooghe

University of North Carolina at Chapel Hill, USA and Vrije Universiteit Amsterdam, Nethlands

Gary Marks

University of North Carolina at Chapel Hill, USA and Vrije Universiteit Amsterdam, Netherlands


This article summarizes and extends the main lines of theorizing on public opinion on European integration. We test theories of economic calculus and communal identity in a multi-level analysis of Eurobarometer data. Both economic calculus and communal identity are influential, but the latter is stronger than the former. We theorize how the political consequences of identity are contested and shaped - that is to say, politically cued - in national contexts. The more national elites are divided, the more citizens are cued to oppose European integration, and this effect is particularly pronounced among citizens who see themselves as exclusively national. A model that synthesizes economic, identity, and cue theory explains around one-quarter of variation at the individual level and the bulk of variation at the national and party levels.

European Union Politics

DOI: 10.1177/1465116505057816 Volume 6 (4): 419-443 Copyright© 2005 SAGE Publications

London, Thousand Oaks CA, New Delhi


■ cueing

■ European integration

■ identity

■ political economy

■ public opinion

What drives citizens to support or oppose European integration? The question is as old as the European Union, and it has been the subject of some one hundred articles, yet there is no scholarly consensus on the answer. There are three main families of explanation. Most research on the topic builds on trade theory to conceptualize a calculus of economic costs and benefits. The presumption is that citizens evaluate the economic consequences of European integration for themselves and for the groups of which they are part, and that such consequences motivate their attitudes. An alternative line of explanation draws on the psychology of group membership to examine how social identities, including, above all, national identities, constrain support for European integration. These two families of theorizing have often been pitted against one another as mutually exclusive conceptualizations. But a new line of research, drawing on cognitive and social psychology, challenges this either/or thinking by examining how political cues - grounded in ideology or in elite communication - mediate the effect of economic calculation and community membership.

These approaches conceive the European Union in contrasting ways. Economic theories view the EU as a regime that facilitates economic exchange, with profound distributional consequences for individuals arising from differences in asset mobility and for countries arising from varieties of capitalism. Social identity theory conceives of the European Union as a polity overarching established territorial communities, and considers how public opinion is constrained by citizens' conceptions of their identities. Cue theory regards the European Union as an extension of domestic politics, and infers that public attitudes are therefore guided by domestic ideology and domestic political organizations.

This article has three purposes. First, we take stock of the field to convey the current state of knowledge and, hence, our point of departure. The study of public opinion on European integration is fast-moving, and it is useful to compare the explanations that are now on the table. Our second purpose is to evaluate the relative causal power of the two most compelling explanations - economic theory and identity theory - in a way that proponents of each would find reasonable. In earlier work we find that both theories bite, but that identity appears the more powerful influence (Hooghe and Marks, 2004). Our third purpose is to build on this analysis to theorize how economic calculation and identity are cued by elites. Given that the European Union is rarely foremost in citizens' minds, we need to understand how interests and identity come to bear on European integration. The resulting model explains slightly more than one-quarter of the variance at the individual level and the bulk of variance at the country and party levels.

Theorizing support for European integration Economic models

European integration has engendered new forms of competition and, hence, new inequalities (Kriesi and Lachat, 2004). In general, trade liberalization and increased factor mobility advantage those with higher levels of human capital, and hurt those with less (Anderson and Reichert, 1996; Gabel 1998a, 1998b; Inglehart, 1970). Trade liberalization increases the international substi-tutability of labor because firms are more able to shift production across borders, and this intensifies job insecurity for less-skilled workers (Rodrik, 1997). International economic openness puts pressure on welfare systems and shifts the burden of taxation from mobile factors of production to immobile factors (Huber and Stephens, 2001; Scharpf, 2000). Following Gabel (1998b), we hypothesize that respondents' general level of education picks up these mobility effects (education).

Economic internationalization affects the relative scarcity of assets in a national economy depending on prior factor endowments (Brinegar and Jolly, 2005). According to the Heckscher-Ohlin theorem, trade benefits individuals who own factors with which the national economy is relatively well endowed and hurts individuals who own factors that are relatively scarce (Mayda and Rodrik, 2002; O'Rourke and Sinnott, 2001). Hence, in the most capital-rich member states we expect unskilled workers to be Euro-skeptic and managers or professionals to be Euro-supportive, whereas in labor-rich member states we expect the reverse (manual worker*gross national income and professional*gross national income).1

Theories of public opinion derived from individual egocentric calculation have been extended in two directions. First, subjective as well as objective factors have been taken into account. Second, sociotropic evaluations concerning one's group (in this case, country) can be theorized alongside egocentric evaluations. The corresponding four lines of theorizing are represented in Figure 1.

Citizens may be sensitive to their sociotropic or collective economic circumstances (cell II in Figure 1), as well as to those that affect them individually (cell I in Figure 1). It seems reasonable to expect residents of countries that are net recipients of European Union spending to support European integration, and those in donor countries to oppose it (fiscal transfer) (Anderson and Reichert, 1996; Brinegar et al., 2004; Diez Medrano, 2003). The same logic is often at work in regional or federal states, where poorer regions champion centralization to increase redistribution whereas prosperous regions favor decentralization (Bolton and Roland, 1997).

Objective evaluation Subjective evaluation

Egocentric I Education Manual worker*GNI Professional*GNI III Personal economic prospects

Sociotropic Fiscal transfer National economic prospects

Type of capitalism

Figure 1 Economic models of public opinion on European integration.

Sociotropic preferences may be shaped by political-economic institutions (Brinegar et al., 2004; Ray, 2004). The European Union encompasses countries with contrasting systems of economic coordination: liberal, social democratic, continental/Christian democratic, and mixed (Hall and Soskice, 2001). Citizens' cost-benefit calculations concerning European integration are likely to be influenced by the type of capitalism in which they live and work. Political-economic institutions are costly to change, and hence we expect citizens in more peripheral systems - liberal and social democratic - to be Euro-skeptical (type of capitalism: liberal, social democratic, continental/Christian democratic, mixed).

Subjective economic evaluations can be expected to influence public opinion on European integration alongside objective factors (Anderson, 1998; Christin, 2005; Eichenberg and Dalton, 1993; Rohrschneider, 2002). European integration is perceived by most citizens to shape their economic welfare in a general sense. Citizens who feel confident about the economic future - personally (cell III) and for their country (cell IV) - are likely to regard European integration in a positive light, whereas those who are fearful will lean towards Euro-skepticism (personal economic prospects and national economic prospects).

The economic approach to public opinion is likely to be most valid when economic consequences are perceived with some accuracy, when they are large enough to matter, and when the choice a person makes actually affects the outcome. To the extent that these conditions are not present, attitudes may be sensitive to group identities (Chong, 2000; Elster, 1990; Sears and Funk, 1991; Young et al., 1991).


The premise of social identity theory is that 'who one is' depends on which groups one identifies with. Humans evolved a capacity for intense group

loyalty long before the development of rational faculties. These loyalties can be extremely powerful in shaping views towards political objects (Massey, 2002; Sears, 1993; Sniderman et al., 2004). The strongest territorial identities are national, and we suspect that such identities constrain preferences concerning European integration.

The European Union meshes national and European governments in a system of multi-level governance that pools sovereignty over important aspects of citizens' lives. To the extent that European integration makes it more difficult for national governments to pursue distinctly national preferences, it undermines national self-determination and blurs boundaries between distinct national communities.

European integration reinforces multiculturalism. It erodes exclusionary norms of 'us' and 'them' that are deeply rooted in the creation of European national states. Kriesi and Lachat (2004) observe that individuals who strongly identify with their national community and who support exclusionary norms tend to perceive European integration as a threat. De Vreese and Boomgaarden (2005) show that anti-immigration sentiment is associated with Euro-skepticism. Similarly, McLaren finds that '[a]ntipathy toward the EU is not just about cost/benefit calculations or about cognitive mobilization .. . but about fear of, or hostility toward, other cultures' (McLaren, 2002: 553). Not only does European integration create economic losers and winners; it provokes a sharp sense of identity loss among defenders of the nation (national attachment) and among anti-cosmopolitans (multiculturalism).

The relationship between national identity and European integration is double-edged. On the one hand, national identity and European identity may reinforce each other (Citrin and Sides, 2004; Klandermans et al., 2003). It is not unusual for citizens to have multiple identities - to feel, for example, Catalan, Spanish, and European - at one and the same time (Diez Medrano and Guttiérez, 2001; Marks, 1999; Marks and Llamazares, forthcoming). Haesly (2001) finds positive, rather than negative, associations between Welsh and European identities and between Scottish and European identities. Klandermans and his co-authors (2003) detect a cumulative pattern of identities, in which farmers who identify with Europe tend also to identify with their nation. Risse (2002) conceptualizes the relationship as akin to a marble cake in which multiple identities are meshed together. Van Kersbergen (2000) conceives of European allegiance as embedded in national allegiance. Citrin and Sides find that 'while the nation retains primacy in most people's minds, the growing sense of Europeanness implies that more people are integrating a sense of belonging to two overlapping polities' (2004: 170).

But it is also true that opposition to European integration is couched as defense of the nation against control from Brussels. Radical right political parties in France, Denmark, Italy, Belgium, and Austria tap nationalism to

reject further integration, and since 1996 such parties have formed the largest reservoir of Euro-skepticism in the EU as a whole (Hooghe et al., 2002; Taggart, 1998). Christin and Trechsel (2002) find that the stronger the national attachment and national pride of Swiss citizens, the less likely they are to support membership in the European Union. Carey (2002) shows that national attachment combined with national pride has a significant negative effect on support for European integration. Luedtke (2005) finds a strong negative association between national identity and support for EU immigration policy.

To resolve these conflicting expectations, we need to theorize how national identity can both reinforce and undermine support for European integration. Diez Medrano (2003) argues that national histories are crucial. Analyzing patterns of discourse in the UK, Spain, and Germany, Medrano finds that English Euro-skepticism is rooted in Britain's special history of empire, that West German pro-Europeanism reflects Second World War guilt, and that the Spanish tend to support European integration as proxy for modernization and democratization (Diez Medrano, 2003). A research team led by Stráth and Triandafyllidou (2003) links party programs, public opinion, educational curricula, and media within nine EU countries. These studies emphasize the stickiness of national identity within unique national contexts.

Can one generalize about the connection between national identity and public opinion? We begin with the basic distinction between exclusive and inclusive national identity, and we hypothesize that citizens who conceive of their national identity as exclusive of other territorial identities are predisposed to be considerably more Euro-skeptical than are those who conceive their national identity in inclusive terms (exclusive national identity). We know, for example, that individuals who identify themselves exclusively as Belgian or exclusively as Flemish are more likely to oppose multi-level governance than are those who identify themselves as both Belgian and Flemish (Maddens et al., 1996). We expect to find something similar at the supranational level (Hooghe and Marks, 2004).

However, the impact of identity on political attitudes is neither automatic nor uniform (Diez Medrano, 2003; Kriesi and Lachat, 2004). The connection between individuals' communal identity and their attitude toward European integration appears to be politically constructed, as we theorize in the next section.

Political cues

The premise of cue theory is that underlying values and interests need to be primed to become politically salient. An experiment examining immigrant attitudes among Dutch citizens finds that individuals who are prompted to

think about national identity are much more likely to oppose immigration than are respondents whose personal identity is primed (Sniderman et al., 2004). This has directed attention to cognitive short-cuts, contextual factors, and elite cues, each of which help a person respond to a survey question about an issue that is seldom on his or her mind (Feldman, 1988; Steenbergen and Jones, 2002; Zaller, 1992).

The cues that appear most relevant to European integration arise in member states. The European Union is part of a system of multi-level governance that encompasses domestic political arenas, and so one would expect domestic politics to shape public views on European integration. A stream of research examines how national contexts frame views on European integration. Taking off from Franklin et al.'s 'uncorking the bottle' model (1994 and 1995) and Anderson's national proxies model (1998), this research emphasizes the quality of national governance (Sánchez-Cuenca, 2000), national democratic performance (Rohrschneider, 2002), or incumbent support and political ideology (Ray, 2003a). Here we hypothesize that public opinion is constrained by political ideology, political parties, and political elites in those domestic arenas.

Political choice in European domestic politics is structured by a general left/right dimension. Previous research has found that this dimension - in the aggregate - has little bite on public opinion on European integration (left/right) (Ray, 2003a, 2003b; Van der Eijk and Franklin, 1996). But some writers have suggested that the implications of left and right for public opinion on European integration depend on a country's political-economic institutions (Brinegar et al., 2004; Brinegar and Jolly, 2005; Marks, 2004; Ray, 2004). In most countries, European integration has become a left-leaning project because it holds out the prospect of continental-wide regulation. However, citizens in social democratic societies are likely to perceive European integration as a source of regulatory competition, and hence as a constraint on market regulation. Hence in social democratic systems, we expect the Left to be opposed to European integration and the Right to be supportive (left/right*social democratic capitalism).

Literature on American public opinion suggests that citizens are cued by political elites (Druckman, 2001; Zaller, 1992: 97-117). In Europe, the most important political organizations connecting elites to the public are political parties, and we hypothesize that individuals who say that they support a particular party will tend to follow that party's position on European integration (party cue) (Steenbergen and Jones, 2002). Cues are likely to be strongest when elites conflict over an issue (Ray, 2003b; Steenbergen and Scott, 2004). Elite conflict punctures passive support for European integration -transforming the 'permissive consensus' that predominated during the EU's

first three decades into a 'constraining dissensus'. We hypothesize that the greater the divisions among political parties and national elites on European integration, the more citizens are likely to oppose the process (elite division). We follow Zaller (1992), Ray (2003b), and Steenbergen and Jones (2002) by modeling the causality as elite driven.2

Research on national political parties tells us that conflict over European integration has, in large part, become a struggle over national community values: what does it mean to be British, French, or Greek, and how does this connect to European integration? We hypothesize that citizens who see themselves as exclusively national are particularly receptive to elite warnings that European integration harbors unacceptable foreign influence. We theorize an interaction: the deeper elite division in a country, the more will exclusive national identity be harnessed against European integration (elite division*exclusive national identity). In countries where the elite is divided on the issue, exclusive national identity is likely to rear its head. In countries where the elite is squarely behind the European project, we expect national identity to lie dormant or to be positively associated with support for integration.


Table 1 summarizes 11 models of public opinion on European integration. The table lists the dependent variable used in each analysis, the method of analysis, the proportion of the variance explained, and, in italics, the most powerful independent variables. These models are, in our view, the most interesting, influential, and/or original analyses to have appeared over the past decade. They also represent the major directions in theorizing. Direct comparison of results across these models is complicated because the dependent variable varies, as do populations, time points, and methods. But some general lessons can be learned.

Most models, like the field as a whole, emphasize political-economic variables. Identities are far less prominent, though we over-sample in this respect by including McLaren's cultural threat model (2002), Carey's identity model (2002), and Diez Medrano's framing Europe model (2003).

The European Union is a moving target, and it is not surprising that analyses of public opinion have changed over time. Up to the mid-1990s and the Maastricht Treaty, the EU was essentially a means to institutionalize market integration, and analyses of public opinion reflected this. Gabel's book Interests and Integration (1998b), from which we draw the policy appraisal and national political economy models, is primarily concerned with economic costs and benefits, as is Anderson and Reichert's economic benefits model (1996). Another stream of work (not represented in Table 1) examines

cross-national variation in support in terms of aggregate economic factors (Carrubba, 1997; Eichenberg and Dalton, 1993 and 2003).

Several models engage domestic politics. Anderson's national proxies model (1998), Rohrschneider's democratic performance model (2002), and Steenbergen and Jones' party cue model (2002) draw attention to national political-institutional variables, of which party support or party cue appear especially powerful. Brinegar et al.'s types of capitalism model (2004) highlights how national political-economic institutions refract left/right positioning on European integration.

In the 1990s, elite conflict on Europe intensified, radical right parties became the Euro-skeptical vanguard, and scholars began to analyze communal identities as sources of public opinion. Carey's identity model (2002) provides evidence that regional, national, and European identities structure EU public opinion. McLaren's cultural threat model (2002) demonstrates that negative attitudes towards the EU reflect general hostility toward other cultures. Diez Medrano's model (2003) attempts to generalize how different national histories frame conceptions of national identity and Europe.

Our analysis builds on these insights. We compare the relative influence of economic calculation and communal identity, and we propose a simple, but encompassing, model that explains around one-third of the variance in public support for European integration.

Method and data

To measure support for European integration we combine three complementary elements of support: the principle of membership, the desired speed of integration, and the desired direction of future integration. The results reported below are robust across these component measures. This and other variables in our analysis are detailed in the appendix (see the appendix3 also for descriptive statistics).4

We use multi-level analysis to probe variation at the individual, party, and country level. Our presumption is that political parties and countries are irreducible political contexts that interact with individual attributes to produce political effects - in this case, support for or opposition to European integration. To the extent that individuals are clustered in parties and countries, they should not be regarded as independent units of analysis. Ignoring this biases standard errors because residuals will co-vary across the higher-level groups. By specifying predictors for clustered data across the relevant clusters, one is less likely to mis-specify parameters (Steenbergen and Jones, 2002).

Table 1 Public opinion on European integration

o ■a

Economic models

Gabel/Palmer's economic voting model (1995)

Anderson/ Reichert economic benefits model (1996)


policy-appraisal model (1998b)

Gabel's national political economy model (1998c)

Brinegar et al.'s types of capitalism model (2004)

Dependent variable

Economic calculation

Individual factors

Collective factors

Membership + unification

Occupation, income, education, proximity to border Evaluation of national economy,a national benefit


Occupation,c income, education

EU trade, budget returns


Occupation,dincome, education, proximity to border


Human capital,a relative wage, occupation EU trade

Perceived/desired speed

Type of capitalism, Structural funding

Community and identity

Cultural threat Identity

Political cues


Party/elite cues Other political cues

Other factors

Country dummies


OLS pooled time series (EU) .38


Gender, age, length of membership No

OLS over different years (EU) .04-. 10

Gender, age Yes

OLS pooled time series (EU) . 13-. 14

Political stabilityf



OLS pooled time series (EU) . 11 —. 13

Values,h left/right self-placement Party cue1

OLS (EU) .17

Notes: The strongest variables in each model are italicized.

a Retrospective evaluation. b Benefit question. c Dummy for farmer. d Dummies: farmer, professional, manual worker. e Occupation/ income interaction. f % vote parties opposing democratic capitalism. g WWII deaths. h Views on welfare state, gender equality. 1 See Steenbergen and Jones (2002).

Anderson's national Rohrschneider's Steenbergen/ Jones' McLaren's cultural Carey's identity Diez Medrano's

proxies model democracy model party cue model threat model model framing Europe model

Non-economic models (1998) (2002) (2002) (2002) (2002) (2003)

Dependent variable

Economic calculation

Individual factors

Collective factors

Community and identity

Cultural threat


Political cues


Party/elite cues Other political cues

Other factors

Country dummies Method


Evaluation of personal economy

Support for EU government

Evaluation of personal economy

Membership + desired speed

Evaluation of national Evaluation of single economy market, national economy

Party support a

System support,b government support

Not applicable OLS

(individual countries)


Perception EU representation, satisfaction with EU democracy

Not applicable

OLS; multi-level analysis

(individual countries) .23-0.40

Party cued

Opinion leadership

Multi-level analysis (EU)

Membership + benefit

Occupation, income, education, proximity to border

Perceived economic threate

Perceived cultural threatf


Occupation, income, education; evaluation of personal economy Evaluation of national economy

Perceived cultural threats

national pride,h territorial attachment'

Left/right self-placement


Opinion leadership

Ordered LOGIT (EU)

59% correct

Index of support for EMU and CFSP

Occupation, income, education

Inflation, growth, trade, objective 1 region

Perceived cultural threat

European attachment

Party supportj

Catholic country; distance to Brusselsk No

Multi-level analysis (EU)

Notes: The strongest variables in each model are italicized.

a Voted for establishment party. b Satisfaction with national democracy. c (1) EU government responsible to EU Parliament? (2) More power for EP good/bad? (3) EP more/less important role? d EU support among political parties, assigned as value to party supporters. e Minorities abuse social benefits. f Religious practices of minorities threaten way of life. g EU threatens national identity, language. h Interaction national pride, exclusive national identity. i Local, region, nation, Europe. i Government party; working class vs. bourgeois party. k Also WWII deaths, new democracy.

Table 2 The base model

Unstandardized coefficients Standard errors

Fixed effects: constant 64.697** 2.965

Variance components

Country level 110.838** 46.446

Party level 70.261** 13.013

Individual level 593.384** 9.669

-2* log likelihood 70985

Note: * p< .01 ** p< .001

We use five controls throughout. Consistent with prior work, we expect support to be greater among opinion leaders, respondents knowledgeable about European politics, men, and younger individuals. We also control for European attachment so that our measure of exclusive national identity does not tap absence of European identity. European attachment is strongly associated with support for European integration (R = .30), and its inclusion as a control variable imposes conservatism in estimating the significance and effect of identity variables of theoretical interest.


Let us begin by examining the extent to which variation in public opinion on European integration is clustered among countries and political parties. An empty ANOVA model partitions the total variance into discrete variance components. The ANOVA model, hereafter described as the base model, is shown in Table 2. The individual level accounts for 76.6% of the variance across the sample; the party level accounts for 9.1%; and the country level accounts for 14.3%.

Table 3 presents the results of the multi-level analysis. Each of the theories we discuss has some power. Models 1 and 2 confirm that citizens respond to economic stimuli. Variables that tap occupation along with personal and collective expectations are significant in both models, though they are not particularly powerful when compared with variables that tap type of capitalism (model 1) or fiscal redistribution among countries (model 2).

In model 1, mean support for European integration is more than 25 points lower (on a 100-point scale) in Scandinavian social democratic political

economies than in our reference category, the mixed political economies of France and southern Europe. Model 1 accounts for almost four-fifths of the country variance in our data, considerably more than model 2 or model 3. But what can one infer from the strength of dichotomous varieties of capitalism variables? The three dummy variables in model 1 specify groupings of countries that share distinctive social, political, and cultural features alongside particular types of capitalism. To be sure, Scandinavians tend to be far more Euro-skeptical than southern Europeans, but is this because they have a distinctive political economy, or because Scandinavians have particular identities that lead them to resist rule from Brussels?

To probe further, one must replace country names with variables. In model 2 we replace the dummies representing groups of countries with a measure of fiscal redistribution. Because most redistribution in the EU is from the richer countries of the north to the poorer countries in the south, fiscal transfer is strongly correlated (R > .30) with three of the dummy variables for country groupings. Adding the variable to model 1 creates unstable coefficients. In model 2 and under the controls exerted in subsequent models, fiscal transfer is significant and powerful.

Figure 2 illustrates this by estimating the relative effect of eight influential independent variables, including fiscal transfer. The solid boxes encompass the inter-quartile range and the whiskers indicate the range between the 5th and the 95th percentile, holding all other independent variables at their means. An individual in Germany at the 5th percentile on fiscal transfer has a mean score of 66.3 on support for European integration, whereas an individual in Greece, at the 95th percentile, has a mean score of 81.5, controlling for all other variables in our analysis. The differing length of the 95% whiskers in Figure 2 for this variable indicates that its association with support for European integration is not linear. Fiscal transfer sharply delineates four countries (Greece, Portugal, Spain, and Ireland) that receive the bulk of cohesion funding and that tend to be pro-EU.

Three variables that tap identity - exclusive national identity, multicultural-ism, and national attachment - are featured in model 3. This model is not as efficient as either model 1 or model 2 in accounting for country variance, but it is considerably better at explaining variance at the party and at the individual level. The reduction in the chi-square (-2*log likelihood) from model 2 to model 3 is 714, and model 3 costs four fewer degrees of freedom. This identity model explains 21.9% of the total variance (excluding European attachment).

The double-edged character of identity is apparent: national identity both contributes to, and diminishes, support for European integration. Attachment to one's country is positively correlated with support for European integration

■c o

o ■u


□ Inter-quartile range for independent variable in fully specified model where all other variables

are held at their means.

5-95% range for independent variable in fully specified model where all other variables are held at their means.

Figure 2 Effects of independent variables.

in bivariate analysis.5 But national identity is Janus-faced: in some circumstances it collides with European integration.

The extent to which national identity is exclusive or inclusive is decisive. A Eurobarometer question compels respondents to place either European or national identity above the other, and separates those who say they think of themselves as 'only British (or French, etc.)' from those who say they have some form of multiple identity. Estimates for exclusive national identity are negative, substantively large, and significant in the presence of any and all controls we exert. On average, and controlling for all other variables, an individual in our sample who claims an exclusive national identity scores 63.2 on our scale for support for European integration, compared with 76.1 for a person who does not. The difference, 12.9%, is indicated by the solid box in Figure 2.

Two methodological issues arise in relation to our claim that national identity shapes public opinion on European integration. The first concerns causal priority. Is national identity exogenous with respect to public support for European integration? Are we right to assume that national identity causes support, and not the reverse? Our approach is confirmed by empirical research arguing that national factors shape public attitudes on European integration, rather than the reverse (Kritzinger, 2003; Van Kersbergen, 2000). It seems plausible to place identities, especially national identities, earlier in the causal chain than support for or opposition to a particular political system, particularly one as distant to most citizens as the European Union. National identities are more deeply rooted in respondents' minds than are attitudes towards European integration, and, to the extent that one finds an association between them, it seems sensible to argue that identities are causally prior.

The second issue concerns measurement. The Eurobarometer question concerning exclusive national identity is far from perfect for our purpose. The measure we use taps national identity by asking whether respondents see themselves as exclusively national or have some form of national and European identity. We control for European attachment to diminish the influence of European identity in our results. We also find that degrees of European identity - whether respondents say they see themselves as national first or European first - have little statistical bite. Consistent with our exclusive national identity argument, the difference between respondents with exclusive national identity and any form of mixed identity is considerably greater than the differences among those with varying forms of mixed identity. Average support for European integration is 53.3 on our scale for respondents who have exclusive national identity, and varies between 75.1 and 80.4 across the remaining categories. The active agent in our analysis is, therefore, the divide between individuals with exclusive national identity and those who

Table 3 Public opinion on European integration: Calculation, community, cues

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Constant (fixed effects) 81.156** 70.290** 76.216** 74.032** 75.570** 70.550**

(2.792) (2.653) (2.760) (2.295) (1.840) (1.709)

Economic calculation (fixed effects)

Education 1.543** 1.527** 0.997* 0.970* 0.927*

(0.357) (0.357) (0.340) (0.339) (0.338)

Professional/manager*gross national income 0.119* 0.118* 0.079 0.072 0.069

(0.046) (0.046) (0.044) (0.044) (0.043)

Manual worker*gross national income -0.065* -0.065 -0.040 -0.035 -0.033

(0.029) (0.029) (0.028) (0.028) (0.028)

Personal economic prospects 2.368** 2.378** 2.257** 2.334** 2.314**

(0.649) (0.649) (0.616) (0.615) (0.615)

National economic prospects 4.018** 3.994** 3.125** 3.025** 3.115**

(0.468) (0.468) (0.446) (0.445) (0.444)

Fiscal transfer 4.633* 4.525** 3.408** 3.330**

(1.589) (1.333) (1.014) (0.864)

Capitalism: liberal

Capitalism: continental/Christian democratic Capitalism: social democratic


(4.841) -11.257* (3.809) -25.419** (4.134)

o ■u

Table 3 continued

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Community and identity (fixed effects) National attachment

Exclusive national identity


Political cues (fixed effects) Left/right

Party cue

Elite division

Elite division*exclusive national identity Left/right*social democratic capitalism

-2 *log likelihood

R2 (including European attachment) R2 (excluding European attachment)

69683 .26 .19

69695 .20 .12

-1.598** (0.420) -13.260** (0.562) 4.532** (0.310)

-1.638** (0.418) -12.954** (0.560) 4.191** (0.310)

-1.699** (0.416) -12.835** (0.559) 4.173** (0.309)

0.039 (0.138) 2.417** (0.241) -6.439** (2.191)

-1.691** (0.415)

4.153** (0.309)

-0.312 (0.144) 2.317** (0.221)

-5.136** (0.223) 1.580** (0.294)

Notes: Multi-level analysis using MLWiN. N = 7641; countries equally weighted. * p< .01, ** p< .001

attest to some mix of inclusive national and European identity rather than between those who have different degrees of mixed identity.

Model 4 combines economic and identity variables and is a large improvement over models 1 to 3. This model reveals that economic and identity variables tap different aspects of public opinion. Model 5 introduces political parties, elite divisions, and left/right ideology. Party cue is the fifth most powerful variable in Figure 2, while divisions within and across political parties and divisions within the political elite, summarized by the variable elite division, are extraordinarily influential. The inter-quartile range in Figure 2 for elite division is second only to that for exclusive national identity. A citizen in Sweden, the country with the most divided elite, scores on average 60.8 on our scale for support for European integration, whereas a citizen in Spain, with the least divided elite, scores 76.1. This 5-95% range is represented by the whiskers in Figure 2. Model 5 explains 33.5% of total variance when we strip out the effect of European attachment, and it is more powerful than any model in Table 1.6

Model 6, our final model, is in one key respect simpler than model 5. It combines the two most influential variables in model 5, elite division and exclusive national identity, into a single interactive term that provides information about the level of elite division only for individuals having exclusive national identity. Our hunch, derived from what we know about American public opinion, seems to be on the right track: divisions within a country's elite interact with exclusive national identity to shape attitudes on European integration. Model 6 includes another interaction term, left/right*social democratic capitalism, which has a significant positive coefficient consistent with our hypothesis that political-economic institutions refract ideological positioning. In social democratic systems, the Left's response to European integration is framed by its defense of welfare provisions that appear anomalous in a wider European context, whereas the political Right welcomes European norms.


The European Union is an extremely versatile institution. It is an international regime that facilitates economic exchange; it is a supranational polity that exerts political authority over its citizens; and it is part of a system of multilevel governance that encompasses national politics. In this paper we show that the motivations underlying public opinion on European integration draw on all three perspectives. Citizens take the economic consequences of market integration into account, both for themselves and for their countries. They evaluate European integration in terms of their communal identities and their

views towards foreigners and foreign cultures. Further, their attitudes are cued by their ideological placement and by elites and political parties.

A multi-level model that synthesizes these perspectives is considerably more powerful than one that does not. The model we put forward in this paper uses 12 degrees of freedom to account for 25.6% of variance at the individual level and almost all variance at the country and party levels. Comparison with prior models is not easy given variations in method and dependent variables, but, for the first time, readers can review leading models and their basic findings side by side.

Economic interests and communal identities do not speak with a single voice across the European Union, but interact with national institutions and elites. The implications of ideology for public opinion on European integration vary with the expected effect of integration on welfare states. More integration means one thing for welfare spending in a country such as the UK, which has a low level, and quite another in a country such as Sweden, where welfare spending is high. This is reflected in ideological positioning with respect to European integration.

We find that exclusive national identity provides a key to public opinion, but the extent to which exclusive national identity bites on support for European integration depends on how divided national elites are. Where elites are united on Europe, national identity and European integration tend to coexist; where they are divided, national identity produces Euro-skepticism.

National identities are formed early in life. Children as young as six or seven know full well whether they are Spanish, German, or Swedish (Druckman, 1994). Yet the political consequences of national identity are constructed in debate and conflict. We suggest that such construction takes place primarily in domestic arenas, and is cued by political elites and political parties, but we need better data, and data over time, to delve more deeply into the causal connections between elite and public attitudes.

Theories of public opinion on European integration have lagged behind first-hand description. Journalists and close observers of the public mood have for some time emphasized that national identities constrain support for European integration, yet all but a few scholarly analyses have focused on economic calculation. Our finding that identity is influential for public opinion on European integration extends research linking identity conceptions to attitudes on immigration and race (Citrin et al., 1990; De Vreese and Boomgaarden, 2005; Luedtke, 2005; Sears, 1993; Sniderman et al., 2004).

Our analysis suggests that the influence of communal identities may reach well beyond race or immigration. A policy with clear distributional consequences may still be evaluated as an identity issue. Research on trade liberalization has produced the unexplained finding that citizens with strong

national attachment tend to oppose trade liberalization both in the United States and across OECD countries. National attachment appears to be a more powerful influence than conventional economic factors, a finding that is all the more striking because it has emerged in two independent tests of economic, not identity, theories (O'Rourke and Sinnott, 2001; Mayda and Rodrik, 2002). Clearly, we have much to learn about how economic calculation and identity shape public opinion, and about how their effects vary across political contexts.

Notes -

For comments and advice we are grateful to Mark Aspinwall, Stefano Bartolini, Tanja Börzel, Gerda Falkner, Rusanna Gaber, Peter Hall, Elizabeth Gerber, Adrienne Héritier, Seth Jolly, Hans-Peter Kriesi, David Lake, Christiane Lemke, Ivan Llamazares, Catherine Netjes, Thomas Risse, Edeltraud Roller, Dieter Rucht, David Soskice, Alexander Trechsel, Anna Triandafyllidou, Bernhard Wessels, and the Steiner political science discussion group at the University of North Carolina, Chapel Hill. Earlier versions were presented at the Vrije Universiteit Amsterdam; the Social Science Research Center, Berlin; the European University Institute, Florence; the Center for European Studies at Harvard University; Humboldt University, Berlin; the University of Michigan, Ann Arbor; Princeton University; the Institute for Advanced Studies, Vienna; and the 2003 APSA Meeting, Philadelphia. We received institutional support from the Center for European Studies at the University of North Carolina, the Alexander von Humboldt Foundation, and the Social Science Research Center, Berlin. We alone are responsible for errors.

1 The literature is divided on how to test this hypothesis. Gabel (1998c) interacts occupational dummies for low- and high-skilled individuals with relative wages (or income). This operationalization has been criticized on the grounds that relative wage/income data at the individual level do not adequately capture variation in national contexts (Brinegar and Jolly, 2005). We follow Gabel by interacting occupation with income, but we use a national variable - gross national income (GNI) - to tap country variation.

2 This assumption is controversial, and has been recently challenged (Carrubba, 2001; Gabel and Scheve, 2004). It seems sensible to model the party-public interaction as conditional on the salience of an issue for the public. Recent data and research suggest that European integration has indeed become salient in some recent national elections, though it is not clear how this has varied over time (Evans, 1999; Netjes, 2004; Tilman, 2004).

3 The appendix can be found on the EUP web page.

4 Data are from Eurobarometer 54.1 (Hartung, 2002; fieldwork in Fall 2000). The data set was made available by the Mannheimer Zentrum für Umfragen, Methoden und Analysen (ZUMA). We include only respondents for whom we have values on all variables in the full model to assure commensurabil-ity across explanatory models, and we weight each country to have equal sample size. Neither of these decisions affects our results. There are no

significant differences in means and standard deviations between our restricted sample and the full sample (minus Luxembourg, and each country equally weighted; see appendix). No variables reported here shift in sign or significance across the restricted and full samples or across weighted and unweighted samples.

5 The association between national attachment and support for European integration is usually insignificant under controls, and becomes negative when we control for European attachment.

6 When we follow Gabel and Palmer (1995) and include national benefit ('Do you believe your country has benefited from European integration') on the left-hand side of the equation along with country dummies, model 5 explains 42.6% of the variance.


Anderson, Christopher J. (1998) 'When in Doubt, Use Proxies: Attitudes toward Domestic Politics and Support for European Integration', Comparative Political Studies 31: 569-601.

Anderson, Christopher J. and M. Shawn Reichert (1996) 'Economic Benefits and Support for Membership in the E.U.: A Cross-National Analysis', Journal of Public Policy 15: 231-49.

Bolton, Patrick and Gérard Roland (1997) 'The Breakup of Nations: A Political Economy Analysis', Quarterly Journal of Economics 112: 1057-89.

Brinegar, Adam and Seth Jolly (2005) 'Location, Location, Location: National Contextual Factors and Public Support for European Integration', European Union Politics 6(2): 155-80.

Brinegar, Adam, Seth Jolly and Herbert Kitschelt (2004) 'Varieties of Capitalism and Political Divides over European Integration', in Gary Marks and Marco Steenbergen (eds) European Integration and Political Conflict, pp. 62-89. Cambridge: Cambridge University Press.

Carey, Sean (2002) 'Undivided Loyalties: Is National Identity an Obstacle to European Integration?', European Union Politics 3(4): 387-413.

Carrubba, Clifford J. (1997) 'Net Financial Transfers in the European Union: Who Gets What and Why?', Journal of Politics 59: 469-96.

Carrubba, Clifford J. (2001) 'The Electoral Connection in European Union Politics', Journal of Politics 63: 141-58.

Chong, Dennis (2000) Rational Lives: Norms and Values in Politics and Society. Chicago: University of Chicago Press.

Christin, Thomas (2005) 'Economic and Political Basis of Attitudes towards the EU in Central and East European Countries in the 1990s', European Union Politics 6(1): 29-57.

Christin, Thomas and Alexander Trechsel (2002) 'Joining the EU? Explaining Public Opinion in Switzerland', European Union Politics 3(4): 415-43.

Citrin, Jack and John Sides (2004) 'Can There Be Europe without Europeans? Problems of Identity in a Multinational Community', in Richard Herrmann, Marilynn Brewer and Thomas Risse (eds) Transnational Identities: Becoming European in the EU, pp. 161-85. Lanham, MD: Rowman & Littlefield.

Citrin, Jack, Beth Reingold and Donald P. Green (1990) 'American Identity and the Politics of Ethnic Change', Journal of Politics 52: 1124-54.

Commission of the European Union (2001) Allocation of 2000 EU Operating Expenditure by Member State. European Commission, Budget Directorate, September.

De Vreese, Claes and Haj Boomgaarden (2005) 'Projecting EU Referendums: Fear of Immigration and Support for European Integration', European Union Politics 6(1): 59-82.

Diez Medrano, Juan (2003) Framing Europe: Attitudes to European Integration in Germany, Spain, and the United Kingdom. Princeton, NJ: Princeton University Press.

Diez Medrano, Juan and Paula Guttierez (2001) 'Nested Identities: National and European Identity in Spain', Ethnic and Racial Studies 24: 753-78.

Druckman, Daniel (1994) 'Nationalism, Patriotism, and Group Loyalty: A Social Psychological Perspective', Mershon International Studies Review 38: 43-68.

Druckman, James N. (2001) 'On the Limits of Framing Effects: Who Can Frame?', Journal of Politics 63: 1041-66.

Eichenberg, Richard and Russell Dalton (1993) 'Europeans and the European Community: The Dynamics of Public Support for European Integration', International Organization 47: 507-34.

Eichenberg, Richard and Russell Dalton (2003) 'Post-Maastricht Blues: The Welfare State and the Transformation of Citizen Support for European Integration, 1973-2002', mimeo, Tufts University.

Elster, Jon (1990) 'When Rationality Fails', in Karen Schweers Cook and Margaret Levi (eds) The Limits of Rationality, pp. 19-51. Chicago: University of Chicago Press.

Evans, Geoffrey (1999) 'Europe: A New Electoral Cleavage?', in Geoffrey Evans and Pippa Norris (eds) Critical Elections: British Parties and Voters in Long-Term Perspective, pp. 207-22. London: Sage.

Feldman, Stanley (1988) 'Structure and Consistency in Public Opinion: The Role of Core Beliefs and Values', American Journal of Political Science 32: 416-40.

Franklin, Mark, Michael Marsh and Lauren McLaren (1994) 'Uncorking the Bottle: Popular Opposition to European Unification in the Wake of Maastricht', Journal of Common Market Studies 32: 455-73.

Franklin, Mark N., Cees van der Eijk and Michael Marsh (1995) 'Referendum Outcomes and Trust in Government: Public Support for Europe in the Wake of Maastricht', in Jack Hayward (ed.) The Crisis of Representation in Europe, pp. 101-17. London: Frank Cass.

Gabel, Matthew (1998a) 'Public Support for European Integration: An Empirical Test of Five Theories', Journal of Politics 60: 333-54.

Gabel, Matthew (1998b) Interests and Integration: Market Liberalization, Public Opinion, and European Union. Ann Arbor: University of Michigan Press.

Gabel, Matthew (1998c) 'Economic Integration and Mass Politics: Market Liberalization and Public Attitudes in the European Union', American Journal of Political Science 42: 936-53.

Gabel, Matthew and Harvey Palmer (1995) 'Understanding Variation in Public Support for European Integration', European Journal of Political Research 27: 3-19.

Gabel, Matthew and Kenneth Scheve (2004) 'Estimating the Effect of Elite

Consensus on Public Opinion Using Instrumental Variables', mimeo, University of Kentucky.

Haesly, Richard (2001) 'Euroskeptics, Europhiles and Instrumental Europeans: European Attachment in Scotland and Wales', European Union Politics 2(1): 81-102.

Hall, Peter A. and David Soskice (eds) (2001) Varieties of Capitalism: The Institutional Foundations of Comparative Advantage. Oxford: Oxford University Press.

Hartung, Harald (2002) Eurobarometer 54.1: Building Europe and the European Union. The European Parliament, Public Safety, and Defense Policy, November-December 2000. Brussels: Commission of the European Communities.

Hooghe, Liesbet (2003) 'Europe Divided? Elites vs. Public Opinion on European Integration', European Union Politics 4(3): 281-305.

Hooghe, Liesbet and Gary Marks (2004) 'Does Identity or Economic Rationality Drive Public Opinion on European Integration?', PS: Political Science and Politics 37: 415-20.

Hooghe, Liesbet, Gary Marks and Carole Wilson (2002) 'Does Left/Right Structure Party Positions on European Integration?', Comparative Political Studies 35: 965-89.

Huber, Evelyne and John Stephens (2001) Political Choice in Global Markets: Development and Crisis of Advanced Welfare States. Chicago: Chicago University Press.

Inglehart, Ronald (1970) 'Cognitive Mobilization and European Identity', Comparative Politics 3: 45-70.

Klandermans, Bert, José Manuel Sabucedo and Mauro Rodriguez (2003) 'Inclu-siveness of Identification among Farmers in the Netherlands and Galicia', European Journal of Social Psychology 34: 279-95.

Kriesi, Hanspeter and Romain Lachat (2004) 'Globalization and the Transformation of the National Political Space: Switzerland and France Compared', paper prepared for a Workshop on the Analysis of Political Cleavages and Party Competition, Duke University, 16-17 April.

Kritzinger, Sylvia (2003) 'The Influence of the Nation-State on Individual Support for the European Union', European Union Politics 4(2): 219-42.

Luedtke, Adam (2005) 'European Integration, Public Opinion, and Immigration Policy: Testing the Impact of National Identity', European Union Politics 6(1): 83-112.

McLaren, Lauren (2002) 'Public Support for the European Union: Cost/Benefit Analysis or Perceived Cultural Threat?' Journal of Politics 64(12): 551-66.

Maddens, Bart, Roeland Beerten and Jaak Billiet (1996) Ethnocentrism and Nationalism: Towards a Contextual Approach. Leuven: Department of Sociology (ISPO).

Marks, Gary (1999) 'Territorial Identities in the European Union', in Jeffrey J. Anderson (ed.) Regional Integration and Democracy: Expanding on the European Experience, pp. 69-91. Lanham, MD: Rowman & Littlefield.

Marks, Gary (2004) 'Conclusion: European Integration and Political Conflict', in Gary Marks and Marco Steenbergen (eds) European Integration and Political Conflict, pp. 235-59. Cambridge: Cambridge University Press.

Marks, Gary and Ivan Llamazares (forthcoming) 'Multi-level Governance in Southern Europe: European Integration and Regional Mobilization', in P. Nikiforos Diamandouros, Richard Gunther and Gianfranco Pasquino (eds) The

Changing Functions of the State in the New Southern Europe. Baltimore, MD: Johns Hopkins University Press.

Massey, Douglas (2002) 'Presidential Address: A Brief History of Human Society: The Origin and Role of Emotion in Social Life', American Sociological Review 67: 1-29.

Mayda, Anna Maria and Dani Rodrik (2002) 'Why Are Some People (and Countries) More Protectionist Than Others?', mimeo, MIT.

Netjes, Catherine (2004) 'Sleeping Giant or Much Ado about Nothing? Examining the Role of Attitudes towards European Integration on National Vote Choice in Denmark', mimeo, Free University of Amsterdam.

O'Rourke, Kevin H. and Richard Sinnott (2001) 'The Determinants of Individual Trade Policy Preferences: International Survey Evidence', paper presented at the Brookings Trade Policy Forum, 10-11 May.

Ray, Leonard (2003a) 'Reconsidering the Link between Incumbent Support and Pro-EU Opinion', European Union Politics 4(3): 259-79.

Ray, Leonard (2003b) 'When Parties Matter: The Conditional Influence of Party Positions on Voter Opinions about European Integration', Journal of Politics 65: 978-94.

Ray, Leonard (2004) 'Don't Rock the Boat: Expectations, Fears, and Opposition to EU Level Policy Making', in Gary Marks and Marco Steenbergen (eds) European Integration and Political Conflict, pp. 51-61. Cambridge: Cambridge University Press.

Risse, Thomas (2002) 'Nationalism and Collective Identities. Europe versus the Nation-State?', in Paul Heywood, Erik Jones and Martin Rhodes (eds) Developments in West European Politics, pp. 77-93. London: Palgrave Macmillan.

Rodrik, Dani (1997) Has Globalization Gone Too Far? Washington DC: Institute for International Economics.

Rohrschneider, Robert (2002) 'The Democracy Deficit and Mass Support for an EU-wide Government', American Journal of Political Science 46: 463-75.

Sánchez-Cuenca, Ignacio (2000) 'The Political Bases of Support for European Integration', European Union Politics 1(2): 147-71.

Scharpf, Fritz (2000) 'The Viability of Advanced Welfare States in the International Economy: Vulnerabilities and Options', Journal of European Public Policy 7: 190-228.

Sears, David O. (1993) 'Symbolic Politics: A Socio-Psychological Theory', in Shanto Iyengar and William J. McGuire (eds) Explorations in Political Psychology, pp. 113-49. Durham, NC: Duke University Press.

Sears, David O. and Carolyn Funk (1991) 'The Role of Self-Interest in Social and Political Attitudes', Advances in Experimental Social Psychology 24: 1-91.

Sniderman, Paul M., Louk Hagendoorn and Markus Prior (2004) 'Predisposing Factors and Situational Triggers: Exclusionary Reactions to Immigrant Minorities', American Political Science Review 98: 35-49.

Spence, Jacqueline [s.d., 1997] The European Union: A View from the Top. Waver, Belgium: EOS Gallup Europe of top national decision-makers in member states.

Steenbergen, Marco and Bradford S. Jones (2002) 'Modeling Multilevel Data Structures', American Journal of Political Science 46: 218-37.

Steenbergen, Marco and David J. Scott (2004) 'Contesting Europe? The Salience of European Integration as a Party Issue', in Gary Marks and Marco

Steenbergen (eds) European Integration and Political Conflict, pp. 165-92. Cambridge: Cambridge University Press. Strath, Bo and Anna Triandafyllidou (eds) (2003) Representations of Europe and the Nation in Current and Prospective Member States: The Collective State of the Art and Historical Reports. Brussels: European Commission, Directorate-General for Research, EUR 20736. Taggart, Paul (1998) 'A Touchstone of Dissent: Euroscepticism in Contemporary Western European Party Systems', European Journal of Political Research 33: 363-88.

Tilman, Erik R. (2004) 'The European Union at the Ballot Box? European Integration and Voting Behavior in the New Member States', Comparative Political Studies 37: 590-610.

Van der Eijk, Cees and Mark N. Franklin (eds) (1996) Choosing Europe? The European Electorate and National Politics in the Face of Union. Ann Arbor: University of Michigan Press. Van Kersbergen, Kees (2000) 'Political Allegiance and European Integration',

European Journal of Political Research 37: 1-17. World Bank (2003) World Development Indicators Database. URL (consulted April

2004): Young, Jason, Cynthia J. Thomsen, Eugene Borgida, John L. Sullivan and John H. Aldrich (1991) 'When Self-Interest Makes a Difference: The Role of Construct Accessibility in Political Reasoning', Journal of Experimental Social Psychology 27: 271-96.

Zaller, John R. (1992) The Nature and Origins of Mass Opinion. Cambridge: Cambridge University Press.

About the authors

Liesbet Hooghe is Professor of Political Science, University of North Carolina at Chapel Hill, Hamilton Hall #3265, Chapel Hill NC 27599, USA, and Professor of Political Science, Chair in Multilevel Governance, Vrije Universiteit Amsterdam. Fax: +1 919 962 0432 E-mail:

Gary Marks is Burton Craige Professor of Political Science, University of North Carolina at Chapel Hill, Hamilton Hall #3265, Chapel Hill NC 27599, USA, and Professor of Political Science, Chair in Multilevel Governance, Vrije Universiteit Amsterdam. Fax: +1 919 962 0432 E-mail: