EXPLANATIONS OF POLITICAL EURO-SCEPTICISM AT THE INDIVIDUAL, REGIONAL AND NATIONAL LEVELS

The reduction of sovereignty of the nation-state to decide on certain policy domains is one of the central elements of the recent European Constitution. However, there is widespread resistance to such a reduction of national sovereignty, an issue that has seldom been the focus of previous studies, with the exception of, for example, Dalton and Eichenberg (1998), De Winter and Swyngedouw (1999) and Hooghe (2003). Recently, Vössing (2005) showed the relevance of nationality for political euro-scepticism (i.e., political preferences to decide on policies at the national level rather than at the European Union (EU) level, in the period 1990–1994).¹ Luedtke (2005) stressed the importance of national identity in decisions over immigration policies. Lubbers and Scheepers (2005) provided evidence that political euro-scepticism is distinct empirically from instrumental euro-scepticism (i.e., considering the EU to be a bad thing for the national country). Moreover, political euro-scepticism turns out to be far more widespread than instrumental euro-scepticism, which has already been widely researched (Anderson 1998; Anderson and Reichert 1996; Deflem and Pampel 1996; Eichenberg and Dalton 1993; Hooghe and Marks 2005, Gabel 1998a,b; Mahler et al. 2000; Marsh 1999).

Given this state of the art, we will focus on the macro- and micro-level explanations of political euro-scepticism, building on studies by Díez Medrano (2003) and Gabel (1998a), borrowing explanations for instrumental euro-scepticism. Most prior research focused merely on economic and political theories. We will add rival theories and hypotheses on nationalism as proposed by Duchesne and Frognier (1995), De Master and Le Roy (2000), McLaren (2001), Díez Medrano (2003), Luedtke (2005) and Hooghe and Marks (2005). By elaborating upon these rival or complementary theories, and by deriving contradictory or complementary hypotheses, we will be able to test these hypotheses more systematically and more rigorously than have earlier studies. The overview of previous research provides a somewhat fragmented picture of macro-level explanations, which we will re-organise and test systematically for political euro-scepticism. We will perform our research not only at the individual and country levels, as is done by most research, but we will also theorise about and test regional-level effects of euro-scepticism, as has been proposed by Mahler et al. (2000) and Díez Medrano (2003). By using multilevel techniques (Snijders & Bosker 1999), we think we can raise the research to a higher level. With the emphasis on macro-level explanations based on nationalist drives, we test for the first time to what extent macro-level variation in euro-scepticism is associated with nationalist cultural expressions. Our research question then reads: To what extent is political euro-scepticism explained at the individual-, regional- and country levels? More specifically, we will answer the question of what extent nationalist characteristics drive political euro-scepticism, in addition to political and economic characteristics.

Turning to theories on euro-scepticism, we propose following Gabel (1998a) by making a distinction between utilitarian, or economic, explanations and political explanations. Where the former emphasise the outcome of a cost-benefit evaluation to be crucial for people's attitude towards the EU, the latter emphasises the importance of political interest, knowledge and trust. Here we add the importance of the nation-state, as proposed by Duchesne and Frognier (1995), De Master and Le Roy (2000), Carey (2002), Christin and Trechsel (2002) and Hooghe and Marks (2005), referring to nationalist motivations driving euro-scepticism. These three perspectives will be elaborated at the individual and contextual (regional and national) levels.

The body of literature about nationalist sentiments related to ‘keeping the EU out’ is growing. De Master and Le Roy (2000: 419) were among the first to emphasise the attention to motivations of euro-scepticism related to ‘preservations of national integrity or fear of foreign influences’. They claim that foreigners could be regarded as different from the national in-group and hence perceived to threaten the integrity of a nation (this point has been emphasised by De Vreese and Boomgaarden (2005) as well). This idea may be substantiated theoretically by (derivations from) social identity theory (Tajfel and Turner 1979; Tajfel 1981; Brown 1995). This theory proposes that individuals have the fundamental need to perceive their in-group as superior to many out-groups (i.e., in-group bias). Subsequently, they apply favourable characteristics to themselves they perceive among members of the in-group via a mental process labeled ‘social identification’, and they value out-groups negatively via mechanisms of ‘social contra-identification’ (e.g., Adorno et al. 1982 [1950]; Levine and Campbell 1972; Brewer 1986; Coenders et al. 2004). Hence, to the extent that we may assume that in-group bias translates into attributing high value to (national) traditions, we may infer that these national sentiments drive resistance to policies directed toward ‘integrated nation-states’, which consequently may be considered to be at the expense of national sovereignty (De Master and Le Roy 2000: 425) (i.e., political euro-scepticism).

Similarly, immigration and the presence of foreigners, legitimised by EU regulations (Schengen), may be considered a threat to national traditions, irrespective of whether the foreigners are from within the EU or not (Sniderman et al. 2000), in turn inducing political euro-scepticism. De Master and Le Roy showed the relationship between xenophobia and instrumental euro-scepticism, however, only with bivariate analyses. They urged researchers to test the relation using multivariate analysis. Hooghe and Marks (2005) have done so fruitfully with respect to attitudes on European integration, while Luedtke (2005) has used it for explaining support for immigration policies and De Vreese and Boomgaarden (2005) for voting in referenda about the EU. Following this research, we set out to test to what extent it holds true that people who attach high value to (national) traditions or who are opposed to immigrants from European countries or who perceive a threat from immigrants in general support political euro-scepticism more strongly.

Coenders et al. (2004) have shown that nationalist attitudes are particularly widespread among low-educated people, working-class people and people with lower incomes. Previous research (Quillian 1995; Kunovich 2004) has, moreover, extensively shown that anti-out-group attitudes are strongly prevalent particularly among the lower strata – that is, those who have to compete with immigrants (or fear they have to) over scarce resources within similar social strata (Kiehl and Werner 1999). On the assumption that nationalist and anti-out-group attitudes drive political euro-scepticism, we hypothesise that euro-scepticism strongly prevails among people with lower education, manual workers and people with a low income, as they attach more value to national traditions, are more strongly opposed to immigrants from European countries and perceive a stronger threat from immigrants. Moreover, religious people, elderly people and people living in the countryside were found to be more nationalistic (Coenders et al. 2004). These categories of people adhere to local or national traditions and consistently resist cosmopolitan spheres, possibly inducing political euro-scepticism. We expect therefore that religious people, elderly people and people living in the countryside will subscribe strongly to political euro-scepticism as well as they are more likely to attach importance to (local or national) traditions.

Previous studies have not taken into account characteristics at the regional or national level referring to nationalism to explain variations in euro-scepticism, as has been done for economic and political characteristics. Following social identity theory (Tajfel 1981), we expect that in countries where people emphasise their national identity more strongly, there is stronger opposition to the EU. We expect that in countries where people are widely exposed to domestic or national cultural expressions rather than to international expressions, national identity is stronger and hence people are more inclined to political euro-scepticism. Similarly, we assume that national (non-religious) celebrations (e.g., mostly being independence days, remembrance days or royal celebration days) strengthen the national identity and hence increase political euro-scepticism.

At the regional level, we suggest that nationalism or regionalism are expressed either by the region's relation to Brussels or to that of the nation-state. Diéz Medrano (2003) has shown that regions that are further away from Brussels are more likely to be indifferent to what happens there. Hence, we expect that geographical distance might induce political euro-scepticism. Even though for many people distances are easier to bridge both due to the information society and cheap international flights, the saliency of what happens more nearby will be higher than what happens far away. Moreover, some regions have troubled relations with the nation-state to which they belong. Some of those regions have some autonomy, while others are striving for independence. In these regions, people are less likely to support the national state and are expected to be less politically euro-sceptic.

Building on ethnic competition theory (Olzak 1992; Quillian 1995), previous studies have revealed relationships between the presence of out-groups, enforcing processes of contra-identification, inducing exclusionism (Scheepers et al. 2002) or extreme right-wing voting behaviour (Lubbers et al. 2002). Consequently, we expect that higher numbers of foreigners produce stronger social contra-identification and a consistently stronger political euro-scepticism.

According to Gabel (1998a), the utilitarian approach has been most successful in explaining whether people believe that EU membership is good or bad. People experience different costs and benefits from membership and are therefore expected to differ in their attitudes toward the EU (Gabel and Palmer 1995). The general proposition is that in circumstances in which people are expected to gain more – in economic terms – from the EU, they subscribe less to political euro-scepticism.

Higher educated and higher occupational categories have been assumed to profit more from free movement of people and goods (Ultee 1989); Eichenberg and Dalton (1993) posited this as the ‘human capital hypothesis’. As these privileged categories have better opportunities for applying their talents in an international setting than lower educated people, they are expected to be more benevolent toward European integration (Gabel and Palmer 1995) and are expected to subscribe less to political euro-scepticism. Similarly, internationalisation would be profitable for higher income groups as they can ‘exploit the greater investment opportunities provided by the more open financial markets’ (Frieden 1991; see Gabel 1998a: 337). Conversely, this ‘capitalist hypothesis’ (Gabel and Palmer 1995) predicts that lower income groups adhere more strongly to political euro-scepticism. We would like to add that we expect that, in particular, socio-cultural specialists (mostly people working in the public sector like teachers and social scientists) demonstrate less political euro-scepticism as internationalisation is unlikely to harm their positions, while Europe creates opportunities for them as explained by the human capital hypothesis.

Some of the European countries are net contributors, while others profit strongly. As a result of the Structural Funds, differences in the extent that nations profit from EU membership exist not only between countries, but between regions as well. Consequently, we expect net-contributing countries as well as regions with smaller funds gained from the EU, to subscribe more strongly to political euro-scepticism. This relation has been corroborated empirically by Marsh (1999), Mahler et al. (2000) and Gabel and Palmer (1995), whereas Eichenberg and Dalton (1993) found no effect. With regard to political euro-scepticism, De Winter and Swyngedouw (1999) found hardly any effect from EU budgetary returns.

Other conditions under which gains from the EU are larger are the percentage of total trade with the EU and a positive trade balance (‘mercantilist hypothesis’: Gabel and Palmer 1995) as well as under a low inflation rate (Mahler et al. 2000). Eichenberg and Dalton (1993) claim that, among other characteristics, inflation taps economic hardship, contrasting with the EU's promise on increased prosperity Europe-wide. The EU would be blamed in times of recession (Marsh 1999). Comparing inflation rates between EU countries should inform us whether rising prices explain why people are more likely to favour political euro-scepticism in some countries. Finally, we expect that regions profiting more strongly from European tourists will be more likely to evaluate the EU positively and therefore to subscribe less to political euro-scepticism.

The general proposition following from political science tradition is that people who are less politically informed will subscribe more strongly to euro-scepticism. This ‘cognitive mobilisation’ explanation is derived from Inglehart's idea that well-developed skills are needed to understand what the EU is about (Inglehart 1970). Moreover, any knowledge (message-independent) would increase support for European integration (Gabel 1998a). Following this notion, we would expect the higher educated to be less politically euro-sceptical as they are expected to be more interested in politics. The same proposition applies to differences in euro-scepticism between men and women and between age groups as previous studies have shown that men and the elderly are more interested in politics (Gabel 1998a).

As the EU is described by some as an elite project that ordinary people know little about, it is expected that people adopt the views on European integration that the political parties actually take (Gabel 1998a). All governments in this study supported EU enlargement in 2002–2003. An individual's positive evaluation of the national government and the economic performance of the country, then, should go hand-in-hand with lower levels of euro-scepticism, as has previously been shown by Anderson (1998). Anderson argued that the effect of government dissatisfaction should controlled for the effect of political distrust in general: People, who are in general distrustful, are expected to subscribe strongly to political euro-scepticism, or even more strongly because of the ‘longstanding debate centering around the democratic deficit of the EU’ (Anderson 1998: 576).

At the contextual level, we follow previous research proposing that EU membership duration has induced more knowledge of the EU and has thus decreased political euro-scepticism (Anderson and Reichart 1996; De Winter and Swyngedouw 1999; Marsh 1999). Again, some countries may gain more from membership than others, according to the security principle, as the argument of ‘No more war in Europe ever again’ has been central to further European cooperation. De Winter and Swyngedouw (1999) proposed to test the effect of the population size of the country, assuming that small countries would perceive stronger threat from large countries and therefore would be less inclined to political euro-scepticism. Both Gabel and Palmer (1995) and Díez Medrano (2003) translated the security principle into the percentage of deaths during the Second World War. They showed that more deaths in a country increased approval of further European integration. We will test if this holds also for political euro-scepticism. Moreover, we will include more countries than before – particularly Poland, where the suffering was particularly hard. Finally, building on the security principle, we expect that regions at the outer EU borders, as compared to regions at intra-EU borders, experience larger threats from foreign non-EU powers and are therefore less inclined to political euro-scepticism. Previously, Diéz Medrano (2003) and Gabel and Anderson (2002) took this explanation of intra-EU borders into account, though they interpreted it as larger familiarity and more inter-regional funding, respectively.

Recently, Lubbers and Scheepers (2005) have shown that methodological artefacts (like the percentages of missing values) contribute to explanations of political euro-scepticism, though without including other contextual explanations. In this contribution we propose to account for regional and national patterns of missing values as well. In Figure 1, we present a heuristic model containing a system of explanations derived from theories on nationalistic, economic and political theories. We include social background characteristics relevant to political euro-scepticism as well as intermediates like socio-political attitudes. Moreover, we took national and regional contextual characteristics into account. We suggest that this overall model is more systematic in terms of different rival theories at different levels for explaining political euro-scepticism and is thus more appropriate for testing such rival hypotheses more rigorously than previous research in order to find decisive determinants for political euro-scepticism.

Data were derived from the European Social Survey (ESS 2004a). An extensive data description is available (ESS 2004b; Jowell et al. 2003). Detailed information on country-specific stratification of samples can be found in the country reports as provided by the ESS team in the ESS Documentation Report 2002/2003 (http://www.europeansocialsurvey.org). We included the 15 EU ‘old Member States’ as well as four of the new ones: Poland, the Czech Republic, Hungary and Slovenia. Data from Switzerland and Norway were also included, even though these countries are not members of the EU. Furthermore, we distinguished between East and West Germany due to the large differences between these regions, both currently and historically.

The regional level consists mostly of the geographical units defined by Eurostat as NUTS2-level territories. However, for Belgium, France, Germany and the United Kingdom, the higher level (NUTS1) territories were used. For the Netherlands, we reduced the 40 NUTS3 units to 12 NUTS2 units. As for Denmark, Ireland and Slovenia, a geographical breakdown hardly exists at the NUTS2 level, so here the NUTS3 level is used. For the Czech Republic, information is available on the 14 Kraje – the Czech NUTS3 level.

The measure of political euro-scepticism has been taken into account in the Eurobarometer surveys as well. People are asked to indicate at what level they believe various policies should be mainly decided. A difference from the Eurobarometer surveys is, however, that people are given the possibility to answer ‘at an international level’ and ‘at a regional level’ as well as the ‘European level’ and ‘national level’. For the present purpose, we follow previous research based on the Eurobarometers, distinguishing only a national decision level versus a European decision level (De Winter and Swyngedouw 1999; Hooghe 2003; Lubbers and Scheepers 2005).²

The policy items refer to ‘protection of the environment’, ‘fighting crime’, ‘immigration’, ‘defence’, ‘social welfare’, ‘aid to developing countries’, ‘agriculture’ and ‘interest rates’. In accordance with previous results, the international level is preferred concerning ‘international’, trans-border issues like protection of the environment and fighting crime (Niedermayer 1995; Hooghe 2003). Social welfare is preferred by a large majority at the national domain. Immigration policies are more undecided, which was also found for the Eurobarometer data (Lubbers and Scheepers 2005). We created an index by summing the policy domains on which people preferred to have intra-national decisions rather than international decisions, resulting in a maximum score of eight. The more people prefer policies to be decided at the national rather than the international (EU–) level, the more they value national sovereignty over the EU (or the international) level and, consequently, the higher the level of political euro-scepticism. This measurement is normally distributed. Having the highest score of eight and hence preferring the nation-state to decide on all policy domains is reported by 5% of the European population. At the other extreme (i.e., the score zero), 8% always prefers the international level. Where people had one or two missing values on of the eight items presented, we substituted them with the score of one if the policy was preferred to be decided at the national level by a majority of the respective nation's population. This reduced the percentage of missing values to 5.4%, but hardly altered the mean score in political euro-scepticism in a country.

As can be seen in Table 1, there is a large variation across countries both in the average score on political euro-scepticism and in the percentages of missing values. In Ireland and the United Kingdom, the nation is most often favoured as the level where policies should be decided. The nation is least often favoured in Belgium, the Netherlands and West Germany. Since in a previous study variation in missing values turned out to affect regional and national variation in euro-scepticism (Lubbers and Scheepers 2005), we include aggregated measures of this at the national and regional levels

To measure educational attainment, we used information on the highest educational level of the respondent in years. In order to assign a numerical value for the respondents who were still studying at the time of survey, we took their study length at the time of the interview. The social class measurement follows the EGP classification (Erikson et al. 1983), but we additionally distinguished cultural specialists from technical specialists as proposed by Güveli et al. (2005). Monthly net household income was measured using a standard number of categories with standard ranges. To compare incomes between countries, the mean income was set to one for each country separately. Urbanisation was measured by means of five categories, as judged by the respondent. We used information on religious attendance, which we categorised into ‘never’, ‘rarely’, ‘once a month’ and ‘once or more a week’. For age, we included a quadratic term as well.

We used numerous contextual measurements as referred to in the hypotheses. We will outline these measurements briefly. Detailed information can be derived from the authors upon request.

We found various measures to tap nationalistic or regionalist expressions at the contextual level. First, we measured to what extent national cultural expressions are dominant compared to foreign (mostly American) influences, by percentages of domestic films and music consumption (UNESCO 2000). As both measures correlate strongly (r=0.53), we summarised the two into one single measure of domestic popular culture. European countries differ also with respect to the extent to which they use their own language in television. We see television dubbing as an expression of national culture (Hassanpour 2004; Koolstra et al. 2002; Luyken 1991). Finally, we included a measurement on the number of public holidays that were not religious-based in order to take into account national celebrations (EU 2004).

With figures on non-EU nationals, we follow the publication of the EUMC by Coenders et al. (2005; Eurostat 2003). Distance to Brussels is calculated as the straight line between Brussels and the capital or main city of a region as derived from the regional information site from the European Commission (EC 2005). We follow Díez Medrano (2003), taking the natural logarithm of this measurement. All regions that have gained some autonomy from the national state, or have some support for separatist parties, are coded an ‘autonomous region’.

EU budgetary balance is taken for 2002 from the European Commission report on allocation of EU expenditures (EC 2003) and is based on the so-called ‘UK rebate definition’. At the regional level, we calculated the yearly allocated funds per 100 capita in the period 2002–2006, which were taken from European Structural Fund publications from the European Commission (2004a). For the new Member States, the figures for 2004–2006 are taken from European Commission (2004b). We calculated the region's agricultural funds by taking the number of agricultural workers as a share of the total agricultural workers in a country and divided the country's total agricultural funds by this ratio (EC 2004a; Eurostat 2002). Percentage of trade with other countries of the EU as a share of total trade and the trade balance are taken from Eurostat (2005). Inflation of prices has been derived from the IMF (2003) and reflects the 2002 situation. At the regional level, we included a measurement of the number of foreign tourists in hotels and campsites overnight in 2002 (Eurostat 2004), in millions. The Spanish Balearic Islands were most often frequented (42 million visitors), while the least tourists found their way into Polish Opolskie.

Years of membership are calculated as related to the situation of 2002. With regard to small countries, we followed the definition of less than 28 seats in the European Parliament (EP 2004). Regions on borders were either coded as situated at an ‘intra-EU border’ or at an ‘outer-EU border’. If both cases applied, we chose to code it as an ‘outer-EU border’. For calculating war deaths, we followed Gabel and Palmer (1995), but with a difference: we also included the deaths incurred in the Spanish Wars, the Greek Civilian War and the Russian-Finnish War. For the Eastern European states as well as former East Germany, we also included the victims of the communist regimes. As a source of deaths during wars, we used Matthew White's website Twentieth Century Atlas (White 2005). This website combines sources also used by Gabel and Palmer (1995), but adds other sources as well. To calculate the war deaths per 1000 capita, we used the documentation source of the Mannheim Centre for European Social Research (Rothenbacher 2002), which provides data on the population of the respective countries for 1939.

For the attitudes towards out-groups, we disentangled a measurement of the attitude towards European immigrants and the attitude towards immigrants in general. The former taps intra-EU migration and has gained relevance during the enlargement of the EU. The latter is a measurement without specification of the origin of the immigrants. Factor analysis showed that these two measurements were empirically distinct. The attitude toward European immigrants was measured by two items asking to what extent people think immigrants from either poorer or richer European countries should be allowed to enter the country. As the two items correlated at 0.65, we took the items together as one measurement of attitude toward European immigrants. To obtain a measurement that evaluates immigrants in general, we computed a scale of perceived threat from immigrants, combining six items regarding the extent to which people believe immigrants pose a threat to economy and culture. The six items turned out to form a reliable scale (Cronbach's alpha = 0.84). People were also asked to what extent they believe it is important to follow traditions and customs. This single item was used to measure attachment to traditions. Political distrust was measured with four items, which were transformed into items with similar scale lengths, running from 0 to 10, where 10 means ‘no trust’. Reliability analyses provided satisfactory statistics (Cronbach's alpha = 0.76). We computed one scale of political distrust by taking the mean of the scores on the four items. Dissatisfaction with the government was measured straightforwardly on a 10-point scale. Similarly, people were asked to evaluate the state of the country concerning the economy. Political interest was measured straightforwardly, asking people to what extent they were interested in politics on a 4-point scale running from ‘not at all’ to ‘very interested’.

To test our hypotheses, we used multilevel modeling (Snijders and Bosker 1999), nesting individuals in regions, and regions in countries – with the computer program MlWin. We started out with a so-called ‘null-model’, estimating only the extent of variation in political euro-scepticism between individuals, regions and countries. Model 1 of Table 2 shows that variation in political euro-scepticism exists at all three levels. The individual level accounts for 86% of the variance, the country level for about 11.5%, and the regional level about 2.5%. These figures are in line with previous figures based on Eurobarometer data (Lubbers and Scheepers 2005).

In the second model, individual characteristics are included. Higher educated people are less inclined to political euro-scepticism, corroborating the human capital hypothesis, consistent with previous research. With respect to class, we find that manual workers and their supervisors are much more politically euro-sceptic than the higher technical professionals, also in support of the human capital hypothesis. However, the routine non-manual workers and the self-employed do not differ significantly from the higher technical professionals. Within the higher class positions, we find a difference between higher socio-cultural professionals and higher technical professionals, with the former being much less euro-sceptical. No other social class category subscribes less to political euro-scepticism. In contrast to previous findings on support from farmers for European integration (e.g., subscribing that ‘membership is a good thing’), farmers are more likely to be politically euro-sceptical as compared to the higher technical professionals.

A lower level of political euro-scepticism is found among men and the middle aged, which we derive from the negative coefficient of age and the positive coefficient for the quadratic term of age. The capitalist hypothesis is also corroborated with these data. The higher income groups are less likely to favour political euro-scepticism than the lowest income groups. We find that people in the countryside are more politically euro-sceptical than city dwellers, just as regular church-goers are more politically euro-sceptical compared to people who never attend church. Inclusion of individual-level characteristics decreases the variance parameters of the national and regional levels, implying the existence of composition effects. As countries and regions differ in, for example, the level of educational attainment, this explains why in some countries and regions the level of political euro-scepticism is higher. The explained variance at the individual level is quite low with 3.1% – as is often found in research on euro-scepticism.

In Table 3 we include country and region characteristics. At the country level, the degrees of freedom are restricted; therefore we included the context characteristics block by block. We then tested model 3 in which only those contextual characteristics are included that turned out to be significant in one of the models. We find that in countries with more migrants and where television dubbing is common practice, people subscribe less to political euro-scepticism, which refutes our expectations. The effect of domestic popular culture is non-significant. Although the effect of numbers of non-religious national celebrations is in the expected direction, the effect is not significant. At the regional level, a strong effect is found regarding distance to Brussels: the further a region is from Brussels, the more euro-sceptical are the people in that region.

Concerning economic country characteristics, we only corroborate the hypothesis that in countries with higher percentages of trade with other EU countries, people are less politically euro-sceptical. At the regional level, the finding that the larger the agricultural returns, the more people favour political euro-scepticism refuted the formulated hypothesis. The effect from the number of tourists is as expected: in regions with more foreign tourists, people are less politically sceptical. With respect to political characteristics, we find that sustained membership is accompanied by less political euro-scepticism. Although people in small countries and in countries with more war deaths seem less sceptical towards European decision making, none of these effects reaches significance. At the regional level, we cannot corroborate the hypothesis that people in bordering regions (either intra- or outer-EU) are less politically euro-sceptical. Regional differences in missing values turn out to affect the level of political euro-scepticism. When a larger share of the respondents in a region has a missing value on our dependent variable, the extent of political euro-scepticism is lower than the country's average. This finding suggests that in regions where more people provide an answer, the likelihood is larger that people are sceptical.

In model 4, presented in Table 4, we include the seven intermediate characteristics we hypothesised upon, with three of them being highly significant predictors of political euro-scepticism. People who object to EU immigrants entering their country subscribe strongly to political euro-scepticism. The more people perceive a threat from immigrants, the more they adhere to political euro-scepticism. Third, we find that the stronger people distrust politicians in general, the stronger they are politically euro-sceptical. There is a small effect from political interest (at p<0.10 level), the direction of which is consistent with previous research – namely, the stronger the interest, the lower the levels of euro-scepticism. These socio-political attitudes explain to a large extent why the lower educated and manual workers are more supportive of political euro-scepticism and, to a lesser extent, why this holds for lower income categories and men. We derive this from comparing the parameters of these characteristics from model 2 in Table 2 and model 4 in Table 4. Comparing effects of religion and urbanisation, we find that the effects have not decreased due to the inclusion of the socio-political characteristics, implying that there should be other explanations.

European citizens face a European constitution proposing to transfer many policy domains from the nation-state to Brussels, thereby reducing national sovereignty. Although a majority of citizens subscribe to common policies on topics of environment protection and the fight against drugs, people show much scepticism in the socio-welfare policy domain. In this contribution we focused on explanations for this political euro-scepticism in 21 European countries. From a theoretical perspective, we argued that the traditional explanations of instrumental euro-scepticism from economic and political theory should be complemented with theories on nationalism and anti-out-group attitudes. We formulated hypotheses to explain variation in political euro-scepticism at the individual, regional and national levels. With the ESS data – that we enriched with regional- and national-level data – we were able to test our hypotheses as rigorously as possible to provide robust empirical insights.

With regard to differences at the individual level – the level at which most variance exists – we found support for the human capital and capitalist hypotheses, as was previously found for instrumental euro-scepticism. People with higher education and higher income are less politically euro-sceptical. Moreover, manual workers turned out to be more sceptical than higher technical specialists. The least sceptical, though, turned out to be the socio-cultural specialists. It seems that this category of ‘teachers, doctors and social scientists’ has little to fear from further European integration and predominantly perceives opportunities. Remarkably, we found that farmers were about the most sceptical regarding transferring policies to Brussels, despite the large subsidies this category receives from the EU.

We could not corroborate the hypothesis that it is, in particular, the attachment to traditions that drives euro-sceptics. We found, rather, that euro-scepticism prevails among people who fear European immigrants and who perceive a threat from immigrants, which was also proposed by De Master and Le Roy (2000). These attitudes are related particularly strongly to political euro-scepticism, as we derived from social identity theory. Since we know from previous research that national sentiments are generally rather strongly related to out-group attitudes, we suggest that attachment to traditions is not associated with political euro-scepticism due to the inclusion of anti-out-group attitudes: we may conclude that it is resistance to immigrants that fosters political euro-scepticism more strongly than attachment to traditions.

Country and regional differences in political euro-scepticism were explained by characteristics derived from all three theoretical perspectives, though many determinants in previous research included were found to be non-significant. Membership duration decreases political euro-scepticism. Moreover, in countries that are more dependent on EU countries for their trade, people are less politically euro-sceptic. Surprisingly, in countries with larger proportions of non-Western emigrants, people are less euro-sceptic, as is the case in countries where television dubbing is the common practice, contrary to theoretical arguments that these circumstances would trigger nationalist sentiments and increase political euro-scepticism. Alternatively, we formulate that in countries with television dubbing, people are more aware of their national identity and therefore fear losing it to a lower degree.

At the regional level, people living further from Brussels were found to be more politically euro-sceptic. Moreover, in regions where more foreign tourists are received annually, people are less euro-sceptic. Finally, in regions estimated to be receiving more agricultural subsidies, people were found to be more sceptical. This seems likely to be interpreted by localism theory, as those are the European rural regions. It is, however, remarkable that neither at the national nor the regional levels do financial burdens or benefits play a role in explaining political euro-scepticism.

We have shown that political euro-scepticism is associated particularly strongly to fears about European immigrants and losses of wealth and traditions due to the inflow of new immigrants. The EU not only has to explain to what extent Europe deals with money and bureaucracy (to meet political distrust), and what Europe is about (to inform people who know little about the EU), but above all to understand people's fears of losing national identity through the inflow of EU immigrants.

This contribution has been presented and discussed at the WAPOR 2005 conference in Cannes, and at the ECSR Conference 2005 in Paris. Moreover, it was also discussed in a seminar at the University of Trento. Thanks go to the Department of Sociology of the University of Trento, Italy, making working on this contribution possible. Moreover, the project is financed by the Netherlands Organisation for Scientific Research.

Marcel Lubbers is assistant professor at the Department of Sociology at the Radboud University Nijmegen, the Netherlands.

Peer Scheepers is Professor of Methods in Social Sciences at the Radboud University Nijmegen, the Netherlands.