The strength of attachment: regionalism, nationalism and vote choice

AbstractRadical right parties have grown in popularity recently, leaning heavily on state-level nationalism, anti-immigrant sentiment and the promise of a homogenous nation state. There has also recently been increased devolution of power to subnational communities, who historically have resisted homogenisation and infringements on their autonomy. This implies a tension between the interests of subnational units and those of radical right parties. Using data from the 2016 and 2018 European Social Surveys and a new measure of regional identity strength constructed from the Regionalist Parties Dataset (Massetti and Schakel Citation2016) for 10 European countries, this study shows that people living in regions with strong legacies of regionalism are less attached to their national state. Second, the article demonstrates regional identity strength is negatively associated with voting for radical right parties due to their opposition to radical right parties’ exclusive state-level nationalism. Lastly, this relationship can eclipse the effect of immigration attitudes on vote choice as radical right parties’ rhetoric around immigration generally focuses on the importance of cultural homogeneity. Evidence supporting these hypotheses is found using both cross-national data and case evidence from Italy. This study underscores the importance of examining the role older social cleavages play in structuring party politics in the transnational era.Keywords: Party politicscleavage theoryregional identitymultilevel governanceradical right parties AcknowledgementsWe would like to give special thanks to Gary Marks, Liesbet Hooghe, Christina Zuber, Sean T. Norton, David Attewell, John D. Stephens, Evelyne Huber, and the three anonymous reviewers for their extremely helpful comments on various drafts of this manuscript. Previous versions of this paper were presented at the Comparative Politics working group at the University of North Carolina at Chapel Hill, where we received valuable feedback from our colleagues at the department.Disclosure statementNo potential conflict of interest was reported by the authors.Notes1 Massetti and Schakel's (Citation2016) database also includes three ‘controversial cases’ which are parties that have some association with a national party but which acts sufficiently independently as to be classified as a regionalist party for the purposes of their dataset. These are the CSU in the German region of Bavaria, the UPN in the Spanish region of Navarre, and the UUP in the UK region of Northern Ireland.2 This mirrors the strategies of nation-builders during the era of state consolidation, who, as noted previously, frequently co-opted symbols of regional identity and subsumed them into a broader, constructed national identity (Giordano and Roller Citation2001; Pasquier Citation2015). Similarly, modern radical right parties sometimes make claims that those peripheral cultures are in fact ‘purer’ expressions of a broader, unifying statewide identity.3 Analyses including Belgium may be found in Online Appendix Tables 10 and 11. Results are unchanged.4 Austria, Finland, France, Germany, the Netherlands, Norway, Spain, Sweden, Switzerland and the United Kingdom. Spain is only present in 2018 due to no radical right votes being reported in 2016. Sweden is only present in 2016 due to incompatible region codes in 2018.5 As discussed, we also exclude Belgium from our main analyses. See Online Appendix Tables 10 and 11 for analyses including Belgium. Results are unchanged.6 See Appendix Table A1 of Massetti and Schakel (Citation2016) for the full list of regionalist parties and coding decisions.7 A strength of this measure is that it smooths out the inherent volatility in electoral results over such a long time frame. One potential concern would be if there were instances in which regionalist party vote share was extremely strong early on and then dropped precipitously to zero (or near zero) over time. This would indicate that perhaps the divide was no longer salient. However, this is not the case for any of the regions in our analysis. Although the nature of politics means that there is of course some amount of volatility in regionalist party vote share across elections, it is actually relatively stable over time (i.e. fluctuations are relatively evenly dispersed over time). See Online Appendix Table 6 for a) the cumulative average vote share, b) the (over-time) standard deviation in vote share, and c) the type of electoral system at the regional level, for every region in the analysis.8 NUTS (Nomenclature of Territorial Units for Statistics) levels are cross-nationally standardized statistical units (based on population), where level 1 units are generally the largest subnational level in a given state, and level 3 units are the smallest non-local/municipal level, often corresponding to ‘counties’. Different NUTS levels may correspond to theoretically and/or practically ‘useful’ territorial units/regions in different countries (e.g. NUTS level 1 corresponds to German Länder while NUTS level 2 corresponds to Austrian Länder). In addition, the ESS and other surveys may use different NUTS levels in different countries. We choose NUTS levels in our analysis based on those in the ESS, so that we may match individual respondents to their regions.9 Again, see Online Appendix Table 6 for more detailed descriptive statistics about this variable by region.10 These are the only available variables of interest that have sufficient temporal and regional coverage to be included in our analyses.11 A regionalist faction still exists within the party, however, which could lead to possible complications or splits going forward (Albertazzi et al. Citation2018).Additional informationNotes on contributorsKaitlin AlperKaitlin Alper is a postdoctoral researcher at the Danish Centre for Welfare Studies at the University of Southern Denmark. Her research focuses on comparative welfare state politics, the politics of multilevel governance and decentralisation, and comparative political economy, especially in Western Europe. Her work has appeared in Social Forces. [kalp@sam.sdu.dk]Caroline Marie LancasterCaroline Marie Lancaster is a research methodologist at NORC at the University of Chicago, where she focuses on psychometrics, program evaluation, and survey research. She has published in journals such as British Journal of Political Science and European Journal of Political Research on topics relating to political attitudes, immigration, and the radical right. [lancaster-caroline@norc.org]