http://orcid.org/0000-0002-4758-4213
![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
ABSTRACT
Today, populism has gradually become one of the most talked about, most studied phenomena, both within and beyond academia. Most studies of populism focus on its conceptualisation, operationalisation, measurement or its outcomes. However, adding to the growing empirical analysis of populism, we propose to study populism as a regional-level phenomenon and explain regional patterns of variation in the populist demand. To do so, we develop a series of theoretical arguments from, which we subsequently test empirically. Specifically, we argue that higher levels of regional populism demand are associated with (i) economic hardship, (ii) strong institutional autonomy, (iii) strong territorial identity, and (iv) greater distance to elites. We construct a populist index for 143 regions across nine countries and combine this with a unique and rich regional database. While we find that populism holds distinct regional patterns and there is support for classic predictors like economic hardship, we are also able to provide some unique insights into the regional foundations of populism, most notably the predictive power of regional identity and the distance to national elites.
Acknowledgements
Christian H. Schimpf acknowledges the support by the University of Mannheim’s Graduate School of Economic and Social Sciences funded by the German Research Foundation. We presented a first draft of this manuscript at the 'French-German Research Colloquium on Nationalism, Populism and Regionalism in a Europe in Crisis' at the Internationales Zentrum für Kultur- und Technikforschung (IZKT) in Stuttgart. We thank Felix Heidenreich for this opportunity and the colloquium participants for their feedback. We are grateful to various members of Team Populism (https://populism.byu.edu) for their comments on earlier versions. All remaining errors are our own.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes on contributors
Steven M. Van Hauwaert ([email protected]) is the principal investigator of the Global Public Opinions Project (GPOP). His primary research interests are in the fields of comparative political behaviour and public opinion, as well as populism and political extremism. His most recent academic contributions have appeared in Acta Politica, Comparative European Politics, Electoral Studies, the European Journal of Political Research, European Societies and the Journal of European Integration. He is also an associate editor of the ECPR journal Political Research Exchange (PRX).
Christian H. Schimpf ([email protected]) is a Ph.D. candidate in political science at the Graduate School of Economics and Social Sciences (University of Mannheim). He is also a data processing specialist with the Comparative Study of Electoral Systems (CSES) Secretariat at the GESIS Leibniz Institute for the Social Sciences, Mannheim. His research interests include comparative political behaviour, populism, and public opinion. His research has been published, among others, in Political Studies and Local Government Studies.
Régis Dandoy ([email protected]) is a researcher at the Center for Local Politics at the Ghent University and at the Waseda Institute for Advanced Studies at the Waseda University and guest professor at the FLACSO-Ecuador. His main research interests deal with Belgian politics, comparative federalism, local and regional elections, e-voting and party manifestos.
ORCID
Steven M. Van Hauwaert http://orcid.org/0000-0002-4758-4213
Christian H. Schimpf http://orcid.org/0000-0002-4049-7758
Régis Dandoy http://orcid.org/0000-0003-3593-0024
Notes
1 The ‘thin-centred’ part indicates that populism provides some "individually shaped coat hangers" other ideologies can use to build on Freeden (Citation1997: 5), while it simultaneously promotes some degree of abstractness that proves rather beneficial for cross-contextual comparisons (cf. Sartori Citation1970).
2 Some of these studies are even more specific and argue that a populist demand translates into ‘populist attitudes'. This would then result in an individual-level analysis of the populist demand. While we recognise this is undoubtedly a relevant field of study, neither the concept of ‘populist attitudes', nor the individual-level analysis of populism falls within the scope of our study.
3 From a theoretical perspective, one could also study demand-side populism from a local (Pollock et al. Citation2015) or state perspective (cf. Akkerman et al. Citation2014). However, politics at the regional level, esp. regional elections, are increasingly important (Thorlakson Citation2016) and can even shape national politics (cf. Schakel Citation2013). To this extent, regions provide meaningful contexts that may shape peoples’ views. Furthermore, as the world becomes more connected, the state - as the dominant political entity - has become less critical (Agnew Citation2013). At the same time, regions closely intertwined with the state, but their relation to the state has undergone significant changes in the wake of these developments as evident by the growing number of regionalist and separatist movements in Europe (Applegate Citation1999). Thus, regions do constitute meaningful entities to study in particular after observing that certain populist parties, like the Italian Lega Nord and the Vlaams Belang, began as [and to some extent still are] regionalist parties (cf. McDonnell Citation2006).
4 For each country, demographic quotas were set before fieldwork in order to obtain nationally representative samples. YouGov initially fielded panel samples for the UK, Germany, France and Sweden, with the remaining countries (Poland, Italy, Spain, Greece, Switzerland) being fielded shortly after that by country-specific panel providers. For these nationally representative samples, YouGov draws a sub-sample of the panels that is representative in terms of age (adults), gender, social class and type of newspaper (upmarket, mid-market, red-top, no newspaper), and invites this sub-sample to complete a survey. For more detailed information regarding this active sampling strategy, we refer to Twyman (Citation2008) and the YouGov website. For an overview of the regions by country, including the number of individuals per region, see Table A1 in the Online Appendix.
5 We recode some of the regions since these are not fully identical between our datasets. Most notably, we reduce the number of regions in Sweden to a limited number of five broader regions. For this particular reason, we check the robustness of our results by excluding Sweden from all our analyses; yet, our results remain largely the same. For more information on the regional division, we refer to Table A1 in the Online Appendix.
6 For the results of a general confirmatory factor analysis of our populist items, we refer to Table A3 in the Online Appendix. A measurement invariance test (equal means) suggests our populist items have a similar meaning across countries (Prob>chi2 = 0.000; RMSEA = 0.087; CFI = 0.950). A significant likelihood ratio test between the variant and invariant models returns a p-value < 0.001 and confirms that populism has the same model form in our nine countries under analysis. For country-by-country descriptive statistics, we refer to Table A5 in the Online Appendix. For a more detailed analysis of these eight items, using item response theory, see Van Hauwaert et al. (Citation2018).
7 For the exact wordings and descriptive statistics, see Table A2 in the Online Appendix.
8 We recognise that CFA scores generally do not meet the aggregation criteria for necessary and sufficient conditions that constitute populism (cf. Wuttke et al. Citation2017). The set of items available to us, however, were not build to account for the multidimensionality that populism entails (e.g. Castanho Silva et al. Citation2018). Hence, we proceed here using the CFA scores.
9 Long-term unemployment (more than 12 months unemployment) is preferred to the total unemployment figures as it better reflects the structural economic situation of the region and is not influenced by short-term conjuncture and seasonal fluctuations.
10 The authors have collected scores for Poland, Switzerland and the UK regions.
11 While our indicator (based on the proportion of the total population with a country of birth different from the current country of residence) does not take into account migrants from the second and third generations and considers national citizens born abroad as part of this share of population, we believe it to be a good proxy to assess the variation of immigrants in the structure of the regional population.
12 For descriptive statistics of all covariates, see Table A4 in the Online Appendix. For descriptive statistics of our dependent variable by country, see Table A5 in the Online Appendix.
13 Because we only focus on the regional level, no hierarchical structure permit for more advanced modelling techniques. This is particularly the case because our data set only includes nine countries, thereby rendering multi-level modelling highly problematic (cf. Stegmueller Citation2013). However, we do take into account the possibility that there may be systematic country differences in our analyses by also fitting a model with country fixed effects (cf. Table A7 in the Online Appendix). Both models provide us with similar substantive results, thereby indicating the robustness of our results. To further test the robustness of our results, we also provide an analysis with individual-level correlates, using a fixed-effects model. In this model, we test how, on the individual level, respondents’ populism demand varies by regional factors identified to relate to general populism demand. For the results of this model, we refer to Table A8 in the Online Appendix. Furthermore, we provide bootstrapped statistics (cf. Table A6 in the Online Appendix) from 1,000 randomly drawn samples to take into account potential outliers.
14 Since some studies indicate that immigration is only associated to (right-wing) populism in the case of economic hardship (e.g. Arzheimer Citation2009; Golder Citation2003), we also tested for the interaction between immigration and economic indicators, but substantive results remained identical.