Regional identity and support for integration: An EU-wide comparison of parochialists,inclusive regionalist,and pseudo-exclusivists

Abstract

The rise of “new regionalism” is one of the most salient features of the post-Cold War international order. Despite the resurgence of regionalism in Europe, little consensus exists on how regional identity impacts public opinion toward the European Union. To remedy this problem, this study examines the impact of three types of individual-level regional identity on support for integration: parochialism (exclusive regionalism), inclusive regionalism, and pseudo-exclusive regionalism. Contrary to scholarly expectations, the multilevel analysis reveals that inclusive regionalists are as equally Eurosceptic as parochial regionalists. In general, regional identity depresses support for integration unless it is expressly combined with a supranational identity. This finding holds true even in minority nations, where respondents are, on the whole, less Euro-friendly.

Keywords

European Union public opinion exclusive and inclusive identities minority nations regional identity support for integration

The recent secession referendums in Scotland and Catalonia underscore that regionalism is a legitimate political force in the European Union (EU) that warrants closer investigation. The central assumption in the literature (Jolly, 2007; Keating, 1998; Risse, 2010) is that regionalists are supportive of European integration, particularly in regions like Scotland and Catalonia, where regional identity is constructed along “civic” rather than “ethnic” lines, making elites and citizens more progressive, forward thinking, and Euro-friendly (Keating, 1998). However, this assumption was formulated in the heyday of the “Europe of the Regions” idea and may no longer reflect the positions of contemporary regionalists, who have increasingly become more Eurosceptic (Elias, 2008). The same literature assumes that the attitudes of voters are aligned with those of parties. Citizens who self-identify as regionalists are construed as Europhiles. This belief is predicated on the notion that pro-EU elite attitudes filter down to constituents. This assertion has gone untested in the literature, but has important political implications at a time when the EU seeks to legitimize its existence, given the Eurozone and migrant crises.

The debates surrounding Scottish and Catalan independence highlight the potential discord between the attitudes of political elites and their constituents. In both regions, elites have created a narrative where subnational identity is coupled with more inclusive attitudes, greater demands for sovereignty, and integration support. The reality about the attitudes of Scottish and Catalan citizens is murkier. Scottish citizens have become increasingly Eurosceptic since the early 2000s, and a vote for independence from the UK does not automatically translate into support for greater integration (Curtice, 2016). Catalan Euroscepticism is less pronounced than in Scotland, although there is evidence that Catalans with an exclusively subnational identity are less likely to support integration than those with a dual Catalan–Spanish identity (Brigevich, 2012; Llamazares and Reinares, 1999). While we have yet to see results from more recent surveys, one anticipates that the EU’s support of Spanish over Catalan interests will produce greater Euroscepticism in Catalan citizens.

While both referendums failed to result in secession, it has become clear that a significant proportion of regional citizens are not able to reconcile their regional and national identities. What is not clear is the way that these citizens perceive their region’s future in the context of the EU. What is the relationship between regional identity and support for integration? How is this relationship mediated by exclusive and inclusive identities at the individual level? Is there evidence that regions with a distinctive identity (i.e. minority nations), like Scotland and Catalonia, are less Eurosceptic than those without a distinctive identity?

EU public opinion studies highlight that the way individuals reconcile their territorial identities conditions support for integration (Carey, 2002; Citrin and Sides, 2004; Hooghe and Marks, 2005; McLaren, 2006). Individuals who combine their national and European identities (inclusive identitarians) are more likely to support integration than those who resist incorporating a European dimension into their conception of self (exclusive identitarians). While several studies explore the impact of exclusive/inclusive regional identity on integration support, there is no scholarly consensus on the nature of that impact. The problem is twofold. First, these studies do not consistently define or operationalize exclusive regionalism. Second, the dependent variable differs across studies and only partially captures integration support.

This article aims to overcome these challenges by employing a multilevel model using the 2008 European Values Study (EVS). Drawing on Inglehart’s (1977) continuum of parochialism–cosmopolitanism, I operationalize exclusive and inclusive regionalism in the three dominant ways identified in the literature: parochial, inclusive, and pseudo-exclusive.

I construct a dependent variable that combines items related to fear about the EU, support for further enlargement, and confidence in EU institutions. Furthermore, I control for how EU support varies across non-distinctive regions, distinctive regions, and minority nations.

The findings challenge the mainstream assumption that inclusive regionalists are primed to support integration and that parochial regionalism is on the decline in Europe. Parochial regionalists are the most prevalent of the three identity types, and while they are the most Eurosceptic (both amidst the three regional identity categories and when compared to non-regionalists), there is no significant difference between them and the inclusive regionalists. The pseudo-exclusive regionalists are significantly more supportive of integration, although this group is small, even in minority nations where this identity type is presumed more prevalent. The results refute the notion that respondents from minority nations and distinctive regions are more favorably predisposed to the EU and unequivocally show that regional identity, unless explicitly paired with a supranational identity, dampens integration support.

Defining regional identity

Regional identity is a type of collective identity and is a key element in the construction of regions as social and political spaces and systems of action (Keating, 1998; Paasi, 1986). However, it is difficult to define precisely what this identity consists of, or how it affects collective action and politics, because the concept of region is open to contestation. A region may be a territorial delineation of similar topographical characteristics, a group of people interested in their own well-being and autonomy, and/or an area of political and administrative activity (Karolewski, 2007: 11). This lack of agreement on what constitutes a region presents a challenge to EU scholars striving to unpack the effects of regional identity.

Regional identity does not exist in a political vacuum but relies on political and social entrepreneurs to give it context and meaning (Karolewski, 2007; Keating, 1998). Regional identities are based on “imagined communities” (Anderson, 1991) that require citizens to relate to people whom they only know from secondhand experiences. To make these “imagined communities” meaningful spaces for collective action, regional elites engage in a process of region building that defines the character of that region and prescribes the norms of in-group/out-group relations. The extent to which regional identity is compatible with the idea of an integrated Europe relies on whether regional elites choose to frame integration as a beneficial process for their community.

Another challenge is that the salience of regional identity varies across individuals and territorial units. Individuals have many social identities, which fluctuate in importance depending on context (Citrin and Sides, 2004). In this sense, the objective identity of a region (one that is rooted in cultural, political, or economic distinctiveness) is not the same as an individual’s subjective regional identity, or the sense of attachment she feels to her region (Paasi, 1986). Attempts to explain the effects of regional identity must account for this distinction. For some individuals, the subjective identity may take precedence over any objective vision of their regional community, and may condition support for integration in a contradictory manner than the one endorsed by regional elites.

Individual-level identity and support for integration

Territorial identities are collective identities, which are defined as the psychological links between individuals and the social groups to which they belong (Tajfel, 1981, 1982). A strong collective identity fosters feelings of mutual obligation among group members, impelling attachment and loyalty to the in-group, sometimes at the expense of the out-groups.

Brewer (1999: 188) posits that social identity is derived from two opposing human needs: inclusion and differentiation. The former motivates individuals to immerse themselves in social groups, while the latter drives the need to feel distinct from the group. The equilibrium one reaches in balancing these two needs varies across individuals. For some, the need for inclusion eclipses the need for differentiation, leading these individuals to seek out membership in broad, all-encompassing social groups (nation, Europe). For others, the need for differentiation surpasses that of inclusion, and these individuals draw sharper boundaries around their lower-order identity groups (local or regional communities).

Application of social identity theory in the EU context focuses on the distinction between exclusive and inclusive identities. Exclusive identitarians feel attached to only one territorial level, while inclusive identitarians express attachment to multiple levels. Individuals with exclusive attachment to the nation are less likely to support integration than inclusive identitarians (Citrin and Sides, 2004; Hooghe and Marks, 2005). While the focus in the literature remains largely on national identity, numerous studies have explored the impact of exclusive regionalism on integration support.

There are two broad ways to conceive of exclusive regionalism. The first is to conceptualize it as a relationship between regional and national identities. Here, an exclusive regionalist has a strong regional attachment but a negligible national attachment. An inclusive regionalist combines her regional and national attachments in varying but significant proportions. This operationalization of exclusive regionalism frequently employs the Moreno Question¹ (Moreno, 1999), and is typically used to analyze the potentially conflicting identities in minority nations (see Brigevich, 2012; Llamazares and Reinares, 1999).

The second way is to evaluate the compatibility of a respondent’s regional identity with the local, national, and European levels (Berg, 2007; Brigevich, 2012). Specifically, Inglehart (1977) relies on a battery of questions that ask respondents which territorial level they are attached to first and second, and shows that strong local and regional attachments are mutually reinforcing, as are strong European and global attachments. This leads him to conceptualize territorial attachments as falling somewhere on a parochial–cosmopolitan continuum, with those declaring only lower-order attachments (town, region) at one end and those with higher-order attachments (Europe, world) at the other. Inglehart explains that the shift from parochialism to cosmopolitanism occurs when an individual becomes more urbanized, literate, and politically engaged with a community larger than her village or tribe. He shows that cosmopolitans have broader horizons and are more likely to exhibit values that predispose them to support European integration.

The focus of this analysis is on the Inglehart measure, which most closely comports with the expectations laid out in social identity theory about in-group/out-group boundaries. It is also the most applicable measure for the set of all EU countries, given that they span the range of unitary states on the one hand (e.g., Greece, Hungary, Poland) to highly federalized states on the other hand (e.g., Germany, Belgium, Austria).

Toward a typology of exclusive and inclusive regionalism: Hypotheses

Within the literature, it is possible to identify three types of regional identity that correspond to the exclusive/inclusive identities paradigm. The first of these is a parochial regionalism, which manifests itself in attachment to the subnational level at the expense of higher-order levels (i.e. exclusive regionalism). Mid-20th-century theories of modernization and political development treat this regional particularism as a “revolt against modernity” (Lipset, 1985) that stands in stark opposition to the dual processes of modernization and Europeanization (Inglehart, 1977). The second is a modern regionalism that is future oriented and inclusive and is expressed in simultaneous territorial attachment to regional and national levels (Keating, 1998). Third, regional attachment can also be expressed as cosmopolitan localism, where societal interests are aimed at developing both a sense of local community and participating in the international context (Moreno, 1999, 2006). This regional pseudo-exclusivism pairs antagonism toward the national level with strong attachment to the supranational level (Brigevich, 2012; Risse, 2010).

Parochial regionalists (exclusive regionalists)

Parochial regionalism “generally involves the traditional sense of the region, as a place, a culture, and a way of life rather than the modern sense of the region as an actor in the political life of the nation and in Europe” (Keating, 1998: 89). Parochialists are largely depoliticized and conservative in orientation, as well as less geographically mobile and older. They are also more likely to have lower levels of cognitive mobilization and less likely to hold post-materialist values (Inglehart, 1977). These attributes are associated with higher levels of Euroscepticism (Fligstein, 2008; McLaren, 2006). Parochialists are “relatively uninterested in regional autonomy, seeing the region in purely cultural or topographic perspective, and preferring traditional mechanisms of representation in central and local government, including partisan and clientelistic networks” (Keating, 1998: 89). This manifestation of regionalism, as a type of cultural identity, helps explain why there is a large proportion of regional identitarians in countries that lack strongly institutionalized political or administrative regions (Antonsich, 2010; Brigevich, 2016; Marks, 1999).

This type of identity is associated with parochial practices, which are characterized by the refusal to cooperate or trade with outsiders because this lowers the returns to members of parochial networks (Bowles and Gintis, 2004: 5). With the growing heterogeneity and multiplicity of identities in an ever-enlarging EU, parochial regionalists are the least likely to see their preferences addressed or realized at the European level and are more likely to protect their parochial networks against EU encroachment (Brigevich, 2012; Karolewski, 2007). Hence, parochialists are less inclined to develop a European attachment or to support European integration. European integration is a modern project and therefore hostile to their quest to preserve the traditional way of life. Parochialists are more likely to view European integration as both a cultural threat that dilutes their regional identity and an economic threat that redistributes resources from the in-group to the out-group.

H1: Parochial regionalists are less likely to support integration.

Inclusive regionalists (regional–national identitarians)

While parochialists find it difficult to reconcile their lower-order and higher-order identities, drawing a more stringent line around their subnational in-group, inclusive regionalists exhibit territorial identities that are more fluid and complementary. For these individuals, territorial identity is not perceived as a zero-sum game, where one necessarily trades off between two territorial levels to attain a sense of well-being and achieve greater control over one’s life (Lawler, 1992). Rather, inclusive regionalists reconcile their regional and national identities simultaneously while deriving positive feelings from each identity type. Inclusive regionalists “share their institutional loyalties at both levels of political legitimacy without any apparent fracture between them,” (Moreno, 2006: 2).

Territorial identities tend to be mutually amplifying: attachment to one level provokes strong identification with additional levels (Berg, 2007; Marks, 1999). While this is clearly not the case for parochialists, for inclusive regionalists, strong attachment to region and nation should breed identification with the European level. Given that identifying with Europe and support for integration are strongly correlated (Brigevich, 2016; Citrin and Sides, 2004; Hooghe and Marks, 2005), inclusive regionalists should have a greater proclivity to evaluate the EU positively than exclusive identitarians.

H2: Inclusive regionalists are more likely to support integration than parochial regionalists.

Pseudo-exclusive regionalists (regional–supranational identitarians)

Finally, the literature identifies a specific subtype of regionalism that is most closely tied to minority nationalism. In regions with a distinct cultural or political history (i.e. Basque Country, Catalonia, Scotland, Northern Ireland), regional identity goes hand in hand with a European identity, while regional and national identities are perceived as incompatible (Brigevich, 2012; Olsson, 2007; Risse, 2010: 71). This is attributed to the decades-long subjugation of historical nationalities to the center, where regional customs and languages were perceived as threatening to the homogenizing policies of the modern nation-state. The result is persistent animosity between the periphery and the center, as well as a resurgence of regional identity. The rejection of the national level for these types of regionalists does not translate into hostility toward other higher-order levels. Instead, the EU is perceived as a champion of regional interests and a means to circumvent the state.

To apply the social identity theory paradigm, for this type of individual, the regional level satisfies the need for differentiation, while the European level satisfies the need for inclusion. As these individuals mix lower- and higher-order attachments, they have more fluid identity boundaries, and should be less likely to fear the diversity of the EU. Furthermore, pseudo-exclusive regionalists should be more likely to support integration than inclusive regionalists, given that one of their territorial attachments (European) is expressly cosmopolitan and correlates highly with EU support.

H3: Pseudo-exclusive regionalists are more likely to support integration than inclusive regionalists.

Regions as distinctive spaces

Regional context itself conditions pro-EU sentiment. Because regional identities are socially constructed and continuously redefined by political entrepreneurs, support for integration is contingent on whether regional elites see utility in EU membership to carry out their political agenda. Two types of regions stand to gain most from integration and should exhibit higher support: regions with a distinctive culture but without nationalist aspirations and minority nations, or communities that constitute a numerical minority within the central state and possess ethnic, linguistic, or religious characteristics differing from the rest of the population (Olsson, 2007). The “nationalism” component implies a doctrine of self-determination, although one that has been modified in recent times to accommodate the changing function of the state in the newly integrated Europe (Lynch, 1996).

Distinctive regional identity can influence support for integration in several ways. Multilevel governance in the EU has created new structural opportunities for regional actors to acquire greater competencies vis-à-vis the state (Hooghe and Marks, 2001; Jeffery, 2000; Marks and Wilson, 2000). Regions can influence EU policy by opening regional offices in Brussels and lobbying EU institutions to achieve their agenda. The EU offers a political space where states can be challenged and circumvented (Keating and Hooghe, 2001). The Committee of the Regions gives regions some consultative power, making it difficult for European decision makers to ignore regional concerns. Support for integration should be higher in distinctive regions, as their elites have much to gain by participating in the integration process.

H4: Support for European integration is higher in distinctive regions than in nondistinctive regions.

As a politicized expression of regional distinctiveness, minority nations should be the most supportive of the EU due to economic, cultural, and geopolitical reasons. The single market reduces economic penalties for regional political autonomy by affording regions opportunities to move beyond the national economic sphere in search of new trade partners (Jolly, 2007; Marks and Wilson, 2000). The EU’s policies of championing minority rights and language protection help strengthen minority nationalist identity (Laitin, 2001). Additionally, the EU’s open borders have removed barriers for regional communication and coordination, allowing minority nations to form alliances across state lines. This pro-EU sentiment in the regional elites should then filter down to the region’s citizens.

H5: Support for integration is highest in minority nations.

However, some scholars highlight current Euroscepticism amongst minority nationalist parties because access to the EU decision-making process is more limited than previously anticipated (Elias, 2008; Hepburn, 2007). Jeffery (2000) points out that present theories of the EU-regions’ relationship overstate regional capacities to reach beyond the central state and ignore the fact that much of the region’s renewed influence is carried out through the central state.

The coding criteria for minority nations and distinctive regions are derived from Olsson (2007) and Fitjar (2010). All 63 regions encompass populations that are either ethno-culturally, linguistically, or religiously distinct from the rest of the country, or some combination of all three. The difference between the two categories is that minority nations have at least one political party that advocates for greater autonomy for the region with the intent of securing greater protection for its regional minority, or outright secession for the region. While some distinctive regions also have regional parties, these parties do not challenge the integrity of state power. Table A in the Online appendix lists the 40 distinctive regions and 23 minority nations in this analysis.

Operationalizing regional identity

To replicate Inglehart’s (1977) measure of parochialism-cosmopolitanism, I use two EVS 2008 questions that ask respondents: “Which of these geographical groups would you say you belong to first of all? And secondly?”Respondents choose from a list of options that includes their locality, region, country, Europe, and the world. These two answers are combined to form eight different identity types.

Table 1 shows the distribution of respondents across the eight identity categories, while Table B in the Online appendix provides information on how the categories were constructed.

Table 1.

Distribution of respondents across identity categories.

	Full sample		Minority Nations		Distinctive regions
	N	%	N	%	N	%
Regionalists
Parochial regionalist	8705	23.94	935	27.04	1106	20.15
Inclusive regionalist	5695	15.66	648	18.74	904	16.47
Pseudo-exclusive regionalist	852	2.34	154	4.45	123	2.24
Non-regionalists
Local–national identitarian	13,290	36.56	968	27.99	1982	36.10
Local–supranational identitarian	1623	4.46	184	5.32	299	5.45
National–supranational identitarian	4400	12.1	365	10.56	789	14.37
Cosmopolitan	679	1.87	89	2.57	132	2.40
Ambiguous identitarian	543	3.06	115	3.33	155	2.82
Total	35,787	100	3,458	100	5490	100

Source: EVS (2008).

The first group, the parochial regionalists, is made up of individuals with only subnational (local/regional) attachments. They are the second largest group in the sample (24.32%). The second group are the inclusive regionalists, who combine regional and national identities. This group are smaller than the parochialists, although they are the third most prevalent group in the analysis (15.66%). The third, and final, regionalist group are the pseudo-exclusivists, who combine regional and supranational attachments. The small size of this group (2.34%) suggests that subnational and supranational identities are not so easily combined.

The rest of the categories comprises the non-regionalists: The fourth group are the local–national identitarians, who are the most prevalent in the sample (36.56%) and should exhibit middling levels of integration support, given that they have both lower-order and middle-order identities. The fifth group are the local–supranational identitarians (4.46%) and the sixth are the national–supranational identitarians (12.1%). These two groups should have higher levels of support, given that one of their identities is higher-order. The seventh and eighth groups are the smallest in the sample: the cosmopolitans (1.87%) and the ambiguous identitarians (3.06%). As cosmopolitans identify with the broadest territorial categories, they are assumed to be the most inclusive and the most supportive of European integration. The ambiguous identitarians are respondents who answered “don’t know” to either of the two questions, or both. Because they appear confused about some elements of their territorial identity, I expect that they are not extraordinarily inclusionary or exclusionary, and should have average levels of support for integration.

Table 1 also shows the distribution of respondents from the identity categories across minority nations and distinctive regions. Surprisingly, there is a similar distribution of respondents from each category across the whole sample and the two region types. One expects the proportion of all three types of regionalists to be higher in minority nations and distinctive regions, where regional identity is more salient. While that is marginally true for minority nations, the difference in proportion is small. However, the data are aggregated at the European level, and further lines of inquiry might explore the extent to which these distributions vary by country and by region.

Data and variables

The individual-level data come from the EVS 2008 survey. Although other data sets, i.e. the 2015 Eurobarometer, the 2013 International Social Survey Programme (ISSP), include questions about territorial identity and EU support, the EVS is the most appropriate for this study. It is the only recent cross-national data set, to my knowledge, that includes the Inglehart-style measure of territorial identity, thereby allowing me to replicate his analysis of parochialism versus cosmopolitanism.² Second, the EVS has a comprehensive list of EU-related questions, which are used to construct the dependent variable.

One potential complication with the EVS is that the data are 10 years old; the next wave is scheduled for release in 2019. It is probable that European attitudes have changed significantly in the past 10 years. This problem is less acute for the independent variables of interest. Multiple studies show that territorial identities remain largely fixed over time (Antonsich, 2010; Brigevich, 2016; Risse, 2010). Using 1991–2007 Eurobarometer data, Antonsich (2010) demonstrates that the proportion of respondents feeling attached to the local, regional, and national levels consistently hovers between 85% and 92%, while attachment to Europe is more volatile, ranging between 40% in 2002 and 67% in 2007. A cursory analysis of Eurobarometer 2015 data reveals that the older patterns hold.

The problem is more acute when considering the shift in attitudes toward integration. Armingeon and Guthmann (2014) show that public trust in national institutions declined significantly with the onset of the economic crisis, and this decline is particularly salient in countries where austerity policies were imposed by external actors. Armingeon and Guthman attribute this decline to increasing citizen awareness of the limited capacity for national governments to pursue sovereign economic policy in the face of supranational institutions, which chips away at the sense of democratic legitimacy imparted onto national governments. Armingeon et al. (2016) also show that this loss of legitimacy translates into erosion of support for democratic institutions at both the national and European level, especially in deficit countries in the Eurozone periphery. Democratic satisfaction with the national and European levels drops significantly 2008 onward, marking a period of externally imposed internal devaluation policies. The EVS data were collected between 2008 and 2010, when the economic crisis was in full swing, and should reflect the decline in EU support found in Armingeon’s work. However, the data do not account for the more recent migrant crisis nor Brexit, which further exacerbate support for integration. As such, the results presented here are likely a more conservative estimate of heightened Euroscepticism that will be reflected in future data.

Twenty-four EU member countries are represented in the analysis. Cyprus, Luxembourg, and Malta do not have a significant regional tier and are excluded, as is Croatia, which joined the EU after 2008. The sample consists of 36,355 individuals, averaging about 1500 respondents from each member state.

There is a significant amount of missing values on the dependent variable and several control variables: income, the proxy argument, the unemployed dummy, and cognitive mobilization. Hence, analyses were carried out twice: once by employing list-wise deletion, and once using Stata’s multiple imputation technique. Multiple imputation (m = 20) recouped 6913 missing observations, for a sample total of 28,659. These two options did not yield substantially different results, with the exception of increasing the significance level of the unemployed dummy and strengthening the interaction effects in Model 3. As such, I present the imputed models here and the original models in Table D of the Online appendix.

The data are collected at the individual level, but these individuals are nested within regions and countries with a distinctive context that may affect individual-level attributes differently. Hence, I use a multilevel model that allows me to account for variation in integration support at the individual, regional,³ and national levels without assuming a single level of analysis (Rabe-Hesketh and Skrondal, 2005).

Dependent variable: EU Support Scale

Support for integration encompasses a variety of attitudes towards economic, political, and cultural policies. I follow Vasilopoulou’s (2011) definition that highlights four distinct components of integration support: (a) support for economic policies such as the monetary union and the single market; (b) support for enlargement; (c) support for embracing European multiculturalism; and (d) general support for the EU and the transfer of power and competencies to the EU level.

I construct an EU Support Scale that combines a respondent’s answers to seven EU-related questions that tap into Vasilopoulou’s definition. The first five ask respondents how much they fear the building of the EU because of (a) loss of social security; (b) loss of national identity and culture; (c) our country paying more and more into the EU; (d) loss of power in the world for our country; (e) loss of jobs in the country. The response option ranges from 1 = very afraid to 10 = not afraid at all. The sixth question asks respondents how much confidence they have in the EU, with 1 = none at all and 4 = a great deal. The seventh question asks respondents whether they feel that enlargement has gone too far = 1 or whether it should go further = 10. Because one of the questions is on a different scale, I standardize each variable and take the average of the seven variables to form the EU Support Scale. The scale ranges from a minimum value of −1.36 to a maximum value of 1.81. Higher values indicate more support for integration.

To assess how well these seven questions form a single scale measuring the same concept, I look at their Cronbach’s Alpha coefficient. The coefficient is 0.82, indicating that the questions do a generally good job of capturing one concept. Figure A in the Online appendix displays the frequency distribution of the EU Support Scale and shows that the scale largely conforms to a normal distribution. Given this and the continuous nature of the scale, linear regression is implemented in the analysis.

Territorial identity variables

Each of the eight identity categories is operationalized as a binary variable. National–supranational identitarians, the literature’s equivalent of inclusive nationalists, are used as the baseline category. To assess the impact of regional distinctiveness, I include a dummy variable for being a distinctive region resident and a minority nation resident.

Control variables

An alternative explanation is that individuals use proxies to form opinions about integration (Anderson, 1998). Most citizens have poor working knowledge of the EU. Instead, they use information about something they know well (their own governments) to formulate opinions about something they know less well (integration). Previous research shows that respondents who feel satisfied with their country are more likely to support EU institutions and their country’s participation in them (Anderson, 1998; Gabel, 1998). Hence, I include a measure for how much confidence a respondent has in her country’s government (proxy argument).

Utilitarian accounts of public opinion find that individuals who are more likely to benefit economically from integration are more supportive of the process (Fligstein, 2008; Gabel, 1998; McLaren, 2006). The literature identifies the traditional winners of integration as wealthier, better educated, younger, and male. This analysis controls for income, levels of education, age, and gender. Individuals who have encountered extensive unemployment may be more likely to associate that unemployment with the effects of economic integration, and are expected to support the EU less. Furthermore, the cognitive mobilization hypothesis (Gabel, 1998; Inglehart, 1977) posits that individuals who are more interested in politics and more in tune to the news are more familiar with the concept of integration and are less likely to fear it. Table C in the Online appendix shows the descriptive statistics for the control variables.

Finally, I include a number of country-level variables, as national economic performance drives public opinion toward integration (Armingeon and Ceka, 2013). I control for whether the country is a net recipient or donor of EU funds. I operationalize recipient/donor status by looking at the percentage of a country’s gross domestic product (GDP) that is contributed to the EU’s budget, with negative values indicating donor status. I expect that respondents in recipient countries are more supportive of integration given they are the beneficiaries of a fiscal transfer. Next, I control for the health of a county’s economy by creating variables that measure a country’s deficit, debt, GDP growth, unemployment rate, and level of social spending. Poor economic performance should translate into greater Euroscepticism. I use data from the Eurostat website for the year 2007.

Empirical analysis

I first confirm that there is significant variation in support across the individual, regional, and national levels. If there is negligible variation at the regional and national levels, then regular regression is more appropriate than a multilevel model. I conduct an analysis of variance (ANOVA) on the EU Support Scale. The variance components can be used to partition the variance across the three levels. In this way, we can get a better understanding of the relative importance of each level in the analysis (Snijders and Bosker, 1999). The analysis reveals that 22% of the variance is at the national level and 17% is at the regional; we can proceed to analyzing the various models using the multilevel modeling technique.

The results of the multilevel analysis are shown in Table 2, and support two of the five hypotheses.

Table 2.

Determinants of integration support.

Predictors	Model 1	Model 2	Model 3	Model 4
Fixed effects
Constant	.184*** (.050)	.157 (.452)	.167 (.452)	.153 (452)
Identity variables
Parochialist	−.279*** (.013)	−.224*** (.013)	−.235*** (.014)	−.216*** (.014)
Inclusive regionalist	−.243*** (.014)	−.209*** (.014)	−.216*** (.014)	−.207*** (.015)
Pseudo-exclusive identitarian	.043* (.026)	.050** (.025)	.043 (.027)	.066*** (.027)
Local–national identitarian	−.225*** (.012)	−.188*** (.012)	−.198*** (.012)	−.181*** (.013)
Local–supranational identitarian	−.005 (.020)	.017 (.019)	.013 (.020)	.047** (.021)
National–supranational identitarian	omitted	omitted	omitted	omitted
Cosmopolitan	.221*** (.028)	.222*** (.027)	.211*** (.029)	.242*** (.030)
Ambiguous identitarian	−.246*** (.028)	−.181*** (.027)	−.171*** (.029)	−.197*** (.029)
Minority nation resident (MNR)		−.021 (.046)	−.102* (.057)	−.021 (.046)
Distinctive region resident (DRR)		−.032 (.033)	−.032 (.032)	.004 (.040)
Interaction effects
MNR × Parochialist			.112*** (.043)
MNR × Inclusive regionalist			.074* (.045)
MNR × Pseudo-exclusive regionalist			.082 (.067)
MNR × Loc.–Nat. Ident.			.110*** (.043)
MNR × Loc.–Supra. Ident.			.057 (.061)
MNR × Nat.–Supra. Ident.			omitted
MNR × Cosmopolitan			.105 (.080)
MNR × Ambiguous Ident.			−.060 (.085)
DRR × Parochialist				−.046 (.035)
DRR × Inclusive regionalist				−.005 (.036)
DRR × Pseudo-exclusive regionalist				−.098 (.070)
DRR × Loc.–Nat. Ident.				−.039 (.031)
DRR × Loc.–Supra. Ident.				−.169*** (.050)
DRR × Nat.–Supra. Ident.				omitted
DRR × Cosmopolitan				−.104* (.067)
DRR × Ambiguous Ident.				.146* (.080)
Individual-level controls
Proxy argument		.169*** (.005)	.169*** (.005)	.168*** (.005)
Cognitive mobilization		.020*** (.003)	.020*** (.003)	.020*** (.003)
Education		.067*** (.003)	.067*** (.003)	.067*** (.003)
Income		.016*** (.002)	.016*** (.002)	.016*** (.002)
Age		−.002*** (.000)	−.002*** (.000)	−.002*** (.000)
Male dummy		.054*** (.007)	.054*** (.007)	.054*** (.007)
Unemployed dummy		−.026*** (.010)	−.026*** (.010)	−.026*** (.010)
Country-level variables
Recipient of European Union funds		−.067 (.070)	−.067 (.070)	−.067 (.070)
Deficit		.001 (.022)	.001 (.022)	.001 (.022)
Debt ratio		.003 (.003)	.003 (.003)	.003 (.003)
GDP growth		−.015 (.028)	−.015 (.028)	−.015 (.028)
Unemployment rate		.028 (.028)	.028 (.028)	.028 (.028)
Welfare spending		−.043** (.018)	−.043** (.018)	−.043** (.018)
Variance components
Country level	.231 (.035)	.206 (.032)	.206 (.032)	.206 (.032)
Regional level	.169 (.009)	.162 (.009)	.162 (.009)	.162 (.009)
Individual level	.630 (.003)	.604 (.003)	.604 (.003)	.604 (.003)
N	28,659	28,659	28,659	28,659

Note: GDP: gross domestic product.

***

p < .01. **p < .05. *p < .1 Multilevel linear regression coefficients with standard errors in parentheses.

Model 1 includes only the territorial identity variables, while Model 2 introduces the controls. The identity variables perform consistently across the two models and the magnitude of their coefficients is only marginally reduced with the introduction of the controls, indicating that that the identity variables exert an independent effect on integration support. In the case of pseudo-exclusive regionalists, including the controls increases the predictive power of the variable.

Model 2 reveals that parochial regionalists have the lowest levels of support for integration, confirming H1. Surprisingly, the inclusive regionalists have a negative coefficient (−.209) that is only marginally smaller than that of the parochialists (−.224). Means testing reveals that there is no difference between parochialists and inclusive regionalists, leading me to reject H2. The pseudo-exclusive regionalists are more supportive of integration than national–supranational identitarians, thereby confirming H3.

The results clearly indicate that respondents with a primary or secondary attachment to Europe are more EU-friendly. Pseudo-exclusive regionalists, local–supranational identitarians, and national–supranational identitarians have commensurate levels of integration support, while the cosmopolitans are the most supportive. Ambiguous identitarians have a more negative view of integration, although they are by no means the most Eurosceptic.

Figure 1 graphs the predicted values of integration support across the identity categories, with all control variables held at their means. The figure shows that a strong regional identity without a strong supranational identity dampens integration support. This makes it difficult to assert that regionalism has shifted from the “old” exclusionary type to a “new” inclusionary form. However, a strong regional identity is not the only culprit here: national identity also exerts a Eurosceptic effect.

Both the minority nation and distinctive regional resident dummy coefficients fail to reach statistical significance, leading me to refute H4 and 5. In fact, the coefficients for both dummy variables are negative, hinting that respondents from distinctive regions may be more Eurosceptic. However, it is possible that the impact of the identity variables differs significantly within minority nations and distinctive regions. It is plausible that parochialists in minority nations are less Eurosceptic than their non-minority nation counterparts, or that pseudo-exclusive regionalists are more Europhile in minority nations, as the literature predicts. Therefore, I include interaction terms between the identity categories and being a minority nation resident in Model 3 and a distinctive region resident in Model 4. In Model 3, the minority nations dummy reaches significance at the 0.1 level and is negative. Furthermore, there is a statistically significant difference between parochialists and local–national identitarians in minority nations versus nonminority nations. The inclusive regionalist coefficient is significant at the 0.1 level only. In Model 4, there is a significant difference between local–supranational identitarians in distinctive regions versus non-distinctive regions and minority nations. The cosmopolitan and ambiguous identitarian coefficients reach significance at the 0.1 level.

Because interaction terms are difficult to interpret, I graph the predicted values of integration support across the identity categories in minority nations versus nonminority nations (Figure 2) and in distinctive regions versus other regions (Figure 3).

Three patterns are discernable in Figure 2. First, parochialists in minority nations are more supportive of integration than those in nonminority nations and they are less Eurosceptic than their inclusive regionalist counterparts. In this specific instance, having only a lower-order identity does not translate into the most exclusionary attitudes, as social identity theory predicts. That status is reserved for the ambiguous identitarians. Second, and counter to expectations, inclusive regionalists and pseudo-exclusive regionalists are more Eurosceptic in minority nations, although the pseudo-exclusive regionalists are the second most supportive group, following the cosmopolitans. Third, a supranational component to identity does not increase integration support in the same measure as in nonminority nations. Pseudo-exclusive identitarians, local–supranational identitarians, and national–supranational identitarians exhibit less integration support in minority nations than in nonminority nations. One explanation is that individuals with a European identity component might be more critical of European institutions in minority nations specifically. Feeling attached to a territorial level does not automatically translate into political support for that level (Brigevich, 2016). In minority nations, where the political stakes for regional policy are higher, it is possible that the public has grown frustrated with the inability of regional actors to extract meaningful concessions from the EU, as Hepburn (2007) and Elias (2008) suggest.

The results in Figure 3 largely mirror those in the previous two figures. However, it is clear that respondents from distinctive regions with a non-politicized identity are less supportive of integration than their counterparts in the other two region types, with the exception of two categories: the national–supranational identitarians and the ambiguous identitarians. The pseudo-exclusive regionalists are almost half as likely to support integration in distinctive regions, the cosmopolitans are a quarter less likely, and local–supranational identitarians are three-quarters less likely. While there are no current explanations in the literature for this, I wager this is an extension of the phenomenon observed in minority nations. However, in distinctive regions without an autonomist political party to advocate for their agenda, the inability to have the region’s voice heard at the supranational level is more acute, and thereby produces more Euroscepticism.

Table 3 reports the percent change in the dependent variable when moving from the baseline category (national–supranational identitarian) to one of the other identity categories, while holding the control variables at their means. Overall, a change in identity type has a weak to moderate effect on integration support. In non-distinctive regions, parochial and inclusive regionalism decrease integration support by roughly 7%, while a pseudo-exclusive regional identity increases it by 1.5%. In minority nations, parochialism and an inclusive regional identity dampen integration support by roughly 4%, whereas a pseudo-exclusive regional identity boosts it by 4%. In distinctive, non-politicized regions, parochialists and inclusive regionalists are 8% and 7%, respectively, less likely to support integration than national–supranational identitarians, while pseudo-exclusive regionalists are 4% more likely.

Table 3.

Marginal effects of predictors on integration support, percent change.

Predictors	Full sample	Minority nation	Distinct. region
Parochial regionalist	−7.07	−3.85	−8.27
Inclusive regionalist	−6.61	−4.45	−6.73
Pseudo-exclusive identitarian	1.58	3.95	−1.01
Local–national identitarian	−5.94	−2.75	−6.95
Local–supranational Ident.	0.54	2.21	−3.66
Cosmopolitan	7.01	9.98	4.36
Ambiguous identitarian	−5.72	−7.29	−1.61
Proxy argument	15.98
Cognitive mobilization	2.53
Education	12.76
Income	5.53
Age	−4.74
Male dummy	1.71
Unemployed dummy	−0.85

Note: For identity variables, percent change is calculated for moving from the baseline category of national–supranational identitarian.

Turning to the controls, all individual-level variables reach statistical significance. Before interpreting the results, it is important to note that public opinion analyses which evaluate the impact of one type of political behavior (i.e. confidence in national institutions) on a second similar type of political behavior (i.e. confidence in the EU) potentially suffer from endogeneity problems. Foster and Friedan (2017: 520) show that individuals who have trust in their national government have nine times greater odds of trusting the EU compared to those who do not trust the national government. Removing the national trust variable from their EU trust model improves the predictive power of some of Foster and Friedan’s other variables. While space constraints prevent me from evaluating endogeneity issues here, future research will wish to keep this problem in mind.

Table 3 reports the marginal effects of moving from the minimum value of each variable to the maximum. As suggested by Foster and Friedan’s work, the proxy argument has the greatest impact on integration support of all the variables. Individuals who have a great deal of confidence in their national government are 16% more supportive of integration than those with no confidence. Respondents who have completed a tertiary level of education are 13% more supportive of integration than those with only a primary level. Individuals with higher incomes are more likely to support integration, as are respondents who follow the news daily, as opposed to not at all. Furthermore, younger individuals and males exhibit higher levels of support, while the unemployed are slightly less supportive. Country-level variables perform poorly in the model. Only the amount of social spending has an impact on integration support. Countries that spend more on welfare programs are less likely to support integration. One possible explanation is that the country-level variables are constructed from 2007 data, taken one year before the onset of the economic crisis. As Armingeon et al. (2016) show, EU support drops after 2007.

Conclusion

The analysis presented here contributes to the discourse on regionalism and public support for European integration in several ways. First, it evaluates the scholarly convention that Europe has moved toward a more inclusionary type of regionalism that bolsters support for integration. On the contrary, the findings indicate that this attitudinal shift has yet to happen. Second, this study extends the inclusive/exclusive national identity framework to the regional level and shows that there are three types of regionalists. Parochial (exclusive) regionalists are the least supportive of integration. However, quite surprisingly, inclusive regionalists are as Eurosceptic as parochialists. The most Europhile group are the pseudo-exclusive regionalists, who combine their regional identity with a supranational identity. Third, I show that the type of region has an impact on integration support. On the whole, living in a minority nation or a distinctive region decreases integration support for most of the identity types outlined in this article. These finding suggest that regionalists may have grown increasingly frustrated with the inability of their regional institutions to carry out their political agenda at the European level.

This analysis aggregates minority nations and distinctive regions into two categories and does not distinguish between different types of minority nationalism or regional distinctiveness (linguistic, religious, etc.). The literature has documented that certain minority nations (i.e. Catalonia, Scotland) are more Euro-friendly than others (i.e. Flanders, Basque Country). Future works should focus on unpacking the differences between civic identity-based versus ethnic identity-based regional movements. Analyses of minority nationalist attitudes toward the EU are likely to become even more significant at a time when the integrity of state boundaries is challenged by secessionist referendums, like those in Scotland and Catalonia. Future works may also wish to look at why regionalists are more Eurosceptic. Is their Euroscepticism motivated by economic, cultural, or political considerations? Finally, are they taking political cues from regional elites, or are they informing the public discourse on integration?

Supplemental Material

EUP-17-1344.R2 replication - Supplemental material for Regional identity and support for integration: An EU-wide comparison of parochialists, inclusive regionalist, and pseudo-exclusivists

Supplemental material, EUP-17-1344.R2 replication for Regional identity and support for integration: An EU-wide comparison of parochialists, inclusive regionalist, and pseudo-exclusivists by Anna Brigevich in European Union Politics

Supplemental Material

Supplemental material for Regional identity and support for integration: An EU-wide comparison of parochialists, inclusive regionalist, and pseudo-exclusivists

Supplemental material for Regional identity and support for integration: An EU-wide comparison of parochialists, inclusive regionalist, and pseudo-exclusivists by Anna Brigevich in European Union Politics

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Anderson

(1991) Imagined Communities: Reflections on the Origin and Spread of Nationalism, New York: Verso.

Anderson

(1998) When in doubt, use proxies: Attitudes toward domestic politics and support for European integration. Comparative Political Studies 31(5): 569–601.

Antonsich M (2010) Exploring the correspondence between regional forms of governance and regional identity: The Case of Western Europe. European Urban and Regional Studies 17(3): 261–276.

Armingeon

Ceka

(2013) The loss of trust in the European Union during the great recession since 2007: The role of heuristics from the national political system. European Union Politics 15(82): 82–107.

Armingeon

Guthmann

(2014) Democracy in crisis? The declining support for national democracy in European countries, 2007–2011. European Journal of Political Research 53(3): 423–442.

Armingeon

Guthmann

Weisstanner

(2016) How the Euro divides the union: The effect of economic adjustment on support for democracy in Europe. Socio-Economic Review 14(1): 1–26.

Berg

(2007) Multi-Level Europeans: The Influence of Territorial Attachments on Political Trust and Welfare Attitudes, Göteborg, Sweden: Göteborg Studies in Politics, p. 104.

Bowles

Gintis

(2004) Persistent parochialism: Trust and exclusion in ethnic networks. Journal of Economic Behavior and Organization 55(1): 1–23.

Brewer

(1999) Multiple identities and identity transition: Implications for Hong Kong. International Journal of Intercultural Relations 22(2): 187–197.

10.

Brigevich

(2012) Peeling back the layers: Territorial identity and EU support in Spain. Regional and Federal Studies 22(2): 205–230.

11.

Brigevich

. Why don’t you like us?: The impact of exclusionary attitudes on European identity construction and support for integration. In: Paper presented at the 2014 Midwest Political Science Association annual conference. 2014. Chicago, 3–6 April.

12.

Brigevich

(2016) Regional identity and support for Europe: Distinguishing between cultural and civic social identity in France. Regional & Federal Studies 24(4): 475–507.

13.

Carey

(2002) Undivided loyalties is national identity an obstacle to European integration. European Union Politics 3(4): 387–413.

14.

Citrin

Sides

(2004) More than nationals: How identity choice matters in the new Europe. In: Herrmann

Brewer

Risse

(eds) Transnational Identities: Becoming European in the EU, Lanham: Rowman and Littlefield, pp. 161–185.

15.

Curtice J (2016) ‘Scottish social attitudes: From Indyref1 to Indyref2? The state of nationalism in Scotland. ScotCen Social Research. Available online: http://www.ssa.natcen.ac.uk/media/38910/ssa16-2fr8m-1ndyref-2-1ndyr8f-tw0-two.pdf.

16.

Elias

(2008) From Euro-enthusiasm to Euro-scepticism? A re-evaluation of minority nationalist party attitudes towards European integration. Regional and Federal Studies 18(5): 557–581.

17.

European Values Study (2008). EVS 2008 Integrated Dataset, 4th wave. [Data file and codebook.] Retrieved from http://www.europeanvaluesstudy.eu/page/survey-2008.html.

18.

Fitjar RD (2010). Explaining variation in sub-state regional identities in Western Europe. European Journal of Political Research 49(4): 522–544.

19.

Fligstein

(2008) Euroclash: The EU, European Identity, and the Future of Europe, New York: Oxford University Press.

20.

Foster

Friedan

(2017) Crisis of trust: Socio-economic determinants of Europeans’ confidence in government. European Union Politics 18(4): 511–535.

21.

Gabel

(1998) Public support for European integration: An empirical test of five theories. Journal of Politics 60(2): 333–354.

22.

Hepburn

(2007) The New Politics of Autonomy. Territorial Strategies and the Uses of European Integration by Political Parties in Scotland, Bavaria and Sardinia, 1979–2005, Florence: Cadmus Online Publications (EUI).

23.

Hooghe

Marks

(2001) Multilevel Governance and European Integration, Baltimore: Rowman and Littlefield.

24.

Hooghe

Marks

(2005) Calculation, community, and cues: Public opinion on European integration. European Union Politics 6(4): 419–443.

25.

Hooghe

Marks

Schakel

(2010) The Rise of Regional Authority: A Comparative Study of 42 Democracies, NY: Routledge.

26.

Inglehart

(1977) The Silent Revolution: Changing Values and Political Styles Among Western Publics, Princeton: Princeton University Press.

27.

Jeffery

(2000) Sub-national mobilization and European integration: Does it make any difference? Journal of Common Market Studies 38(1): 1–23.

28.

Jolly

(2007) The Europhile fringe? Regionalist party support for integration. European Union Politics 8(1): 109–130.

29.

Karolewski

(2007) Regionalism, nationalism, and European integration. In: Karolewski

Suszycki

(eds) Nationalism and European Integration: The Need for New Theoretical and Empirical Insights, New York: Continuum, pp. 9–32.

30.

Keating

(1998) The New Regionalism in Western Europe: Territorial Restructuring and Political Change, Northampton, MA: Edward Elgar Publishing.

31.

Keating

Hooghe

(2001) Bypassing the nation-state? Regions and the EU policy process. In: Richardson

(ed.) European Union. Power and Policy-Making, New York: Routledge, pp. 239–256.

32.

Laitin

(2001) National identities in the emerging European state. In: Keating

McGarry

(eds) Minority Nationalism and the Changing International Order, New York: Oxford University Press, pp. 54–113.

33.

Lawler

(1992) Affective attachments to nested groups: A choice-process theory. American Sociological Review 57(3): 327–339.

34.

Lipset S (1985). Consensus and Conflict: Essays in Political Science, volume 1. New Brunswick, NJ: Transaction Publishers: 253–294.

35.

Llamazares

Reinares

(1999) Identificaciones territoriales, ciudadanía europea y opinión pública Española. In: Llamazares

Reinares

(eds) Aspectos politicos y sociales de la integración europea, Valencia: Tirant Lo Blanch, pp. 179–200.

36.

Lynch

(1996) Minority Nationalism and European Integration, Cardiff: University of Wales Press.

37.

Marks

(1999) Territorial identities in the European Union. In: Anderson

(ed.) Regional Integration and Democracy: Expanding on the European Experience, Lanham, MD: Rowman and Littlefield, pp. 69–91.

38.

Marks

Wilson

(2000) The past in the present: A cleavage theory of party response to European integration. British Journal of Political Science 30(3): 433–459.

39.

McLaren

(2006) Identity, Interests, and Attitudes to European Integration, Houndsmills: Palgrave Macmillan.

40.

Moreno

(1999) Local and global: Mesogovernments and territorial identities. Nationalism and Ethnic Politics 5(3/4): 61–75.

41.

Moreno

(2006) Scotland, Catalonia, Europeanization and the “Moreno question”. Scottish Affairs 54: 1–21.

42.

Olsson

(2007) Regional minority nationalist attitudes toward European integration. In: Karolewski

Suszycki

(eds) Nationalism and European Integration: The Need for New Theoretical and Empirical Insights, NY: Continuum, pp. 52–66.

43.

Paasi

(1986) The institutionalisation of regions: A theoretical framework for understanding the emergence of regions and the constitution of regional identity. Fennia 164(1): 105–146.

44.

Rabe-Hesketh

Skrondal

(2005) Multilevel and Longitudinal Modeling Using Stata, College Station, TX: Stata Press.

45.

Risse

(2010) We the European Peoples? Identity, Public Sphere, and European Democracy, Ithaca, NY: Cornell University Press.

46.

Snijders

TAB

Bosker

(1999) Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modeling, London: SAGE Publications.

47.

Tajfel

(1981) Human Groups and Social Categories, Cambridge: Cambridge University Press.

48.

Tajfel

(1982) Social Identity and Intergroup Relations, Cambridge: Cambridge University Press.

49.

Vasilopoulou

(2011) European integration and the radical right: Three patterns of opposition. Government and Opposition 46(2): 223–244.