ABSTRACT
Right-wing populist (RWP) parties and attitudes are flourishing in many European countries. This is often linked to the emergence of a new cleavage that juxtaposes the victims and beneficiaries of globalisation. Using data from the 2008 European Values Study (EVS) this paper investigates how globalisation affects both (1) the overall level of RWP attitudes in 27 EU countries, and (2) the attitudinal gap between victims and beneficiaries of globalisation. The results show (1) a clear and consistent effect of social position on RWP attitudes in all countries, (2) effect of globalisation on both overall levels of RWP attitudes, as well as (3) on the attitudinal gap between victims and beneficiaries. However, the determinants of the overall level and of the gap differ, and this explains the observed regional differences in Europe regarding the distribution of RWP attitudes.
Introduction
The spectre of right-wing populism (RWP) is haunting Europe. Over the past decades, RWP parties across the European Union gained substantial electoral support. They all share certain political views, combining an anti-migrant and anti-EU agenda with a distrust of politics and institutions. Such parties can only thrive when they find fertile soil for such views among the electorate (Akkerman et al., 2014; Van Hauwaert & Van Kessel, 2018).
It is common to interpret RWP as a backlash against globalisation, reflexive modernity, post-modernity, etc. (Beck, 2008; Betz, 1994; Inglehart & Rabier, 1986; Kriesi et al., 2008). The institutions symbolising the successes of modernity are weakened by globalisation and are increasingly less able to cope with the challenges of globalisation and to respond to the needs and expectations of segments of society. Large scale transformations, including globalisation, lead to the emergence of new cleavages (Deegan-Krause, 2007; Kriesi et al., 2008). According to the literature it is the losers and winners, or victims and beneficiaries, of globalisation who oppose each other on this new cleavage, and are the victims most attracted by RWP ideas. As a thin ideology (Mudde, 2004), populism is able to unite people with different backgrounds, interests and grievances, i.e. the different groups of globalisation victims.
Our central research question is to what extent globalisation creates a new cleavage opposing victims and beneficiaries of globalisation. First, we explore whether globalisation in Europe consistently creates victims receptive to RWP ideas and beneficiaries who are relatively immune to such ideas. Second, we want to understand how variations in exposure to globalisation affect both a country’s overall level of RWP attitudes as well as the attitudinal gap between beneficiaries and victims. We explicitly acknowledge that different European regions may be differently affected by globalisation (Deegan-Krause, 2007; Enyedi, 2008; Randall, 2001), and, therefore, include Eastern European (EE) as well as Western and Northern (WNE) and Southern European (SE) countries, rather than only the WNE countries by which much of literature is inspired.
Background
For a ‘true’ societal cleavage, three conditions need to be fulfilled (Bartolini & Mair, 1990; Deegan-Krause, 2007; Emanuele et al., 2020): (1) a structural (demographic, socio-economic) condition that distinguishes groups on both sides of the cleavage; (2) a normative one that requires these groups to hold clearly opposed values and opinions; and (3) an institutional one that requires interest organisations and/or political parties organizing around the cleavage. The extent to which these conditions are fulfilled varies widely over countries and cleavages. RWP-attitudes, therefore, only reflect a new cleavage to the extent that these attitudes are embedded in specific groups in society, and likewise, when the electorates of RWP parties are concentrated in these groups.
Cleavages originating in the industrial and national-democratic revolutions of the 18th and 19th centuries shaped the political landscape in western democracies throughout most of the twentieth century (Hooghe & Marks, 2018; Lipset & Rokkan, 1967). Over the past few decades, the traditional cleavages weakened as shown by the increased electoral volatility and the emergence of new parties less linked to traditional cleavages. Recent societal developments, including post-industrialisation, post-modernism, educational expansion, welfare state expansion, working class embourgeoisement and secularisation (Azmanova, 2011; Franklin, 1992; Inglehart, 2007; Inglehart & Rabier, 1986; Kriesi, 2014; Martin, 2000) defuse traditional cleavages and reduce their salience, thus weakening the association between structural position, political attitudes and voting behaviour (Enyedi, 2008), while globalisation created the conditions for a new cleavage.
Globalisation captures a series of interrelated, complex, and multifaceted transformations and related economic, political, social, and cultural developments (Stefan, 2013) which imply the increase in scale of social life and the growing interdependency among regions and countries, and the weakening of boundaries (Gygli et al., 2019). Although globalisation is a global process, it does not manifest itself uniformly in all countries and regions. The starting point for the globalisation cleavage literature is the Western European experience. For instance, Kriesi’s landmark study focused exclusively on WNE countries (Kriesi et al., 2006, 2008). In these countries economic globalisation implied increased foreign trade and financial flows which resulted in a deindustrialisation, a shift towards a service economy and some neo-liberal reforms. Social and cultural globalisation meant here not only large-scale migration, first as guest workers and later as refugees, and increased social and cultural diversity but also a shift towards post-materialist and more individualistic value patterns. Political globalisation in these countries implied to loss of sovereignty to international organisations of which the EU is the most visible one. Other countries and regions, however, may experience other transformations and/or the transformations may yield different social and political divisions (Randall, 2001). For instance, in low- and middle-income countries globalisation stimulates further industrialisation as industries relocate from higher to lower-income countries, but also out-migration as large groups seek a better live in higher-income countries.
A cleavage requires its structural and normative bases to be congruent, i.e. distinct groups need to oppose each other holding opposite views regarding the conflict axis (Bartolini & Mair, 1990; Deegan-Krause, 2007; Emanuele et al., 2020; Lipset & Rokkan, 1967). Globalisation creates new and increased economic, political and cultural competitions, which oppose beneficiaries and victims – or ‘winners’ and ‘losers’ – of this transformation (Bergh, 2021; Bornschier, 2010; Bornschier & Kriesi, 2013; Emanuele et al., 2020; Enyedi, 2008; Grande & Kriesi, 2012; Harteveld, 2016; Hawkins et al., 2017; Hooghe & Marks, 2018; Kriesi et al., 2006; Langsæther & Stubager, 2019; Rooduijn, 2018; Van der Brug & Van Spanje, 2009). While some benefit from globalisation and improve their social position, others experience or fear a deterioration of their position, diminishing opportunities and exclusion from full participation in society, and the undermining of their old certainties and identities (Helbling & Jungkunz, 2020; Rippl & Seipel, 2018; Teney et al., 2014; van der Walt, 2020). The cost–benefit balance can vary widely. Those for whom the costs consistently outweigh the benefits form the victims of globalisation, while those who enjoy many benefits and incur few costs are the beneficiaries. In between these two extreme categories, however, there are many other categories for which the cost–benefit distribution is less skewed.
The two normative poles of the globalisation cleavage, integration vs. demarcation reflect these different experiences (Bornschier, 2010; Dolezal & Hutter, 2012; Grande & Kriesi, 2012; Helbling & Jungkunz, 2020; Kriesi, 2014; Kriesi et al., 2006; Lachat & Dolezal, 2008; Van der Brug & Van Spanje, 2009). The integration pole favours integration in the globalised world and approves of the associated developments, while the demarcation pole prefers to keep the own community shielded from these transformations and to maintain the character and identity of their community (Hooghe & Marks, 2018). This integration-demarcation opposition is visible on various dimensions: cultural (pro- vs. anti-migration and diversity), economic (free-trade vs. protectionism), and political (pro- or anti-EU) (Emanuele et al., 2020; Helbling & Jungkunz, 2020; Hooghe & Marks, 2018; Kriesi et al., 2008). The victims of globalisation are more likely to choose the demarcation side and to hold RWP attitudes and support RWP parties because of the problems, frustration, fear, unease, uncertainty and status loss blamed on globalisation (Betz, 1994; Bornschier & Kriesi, 2013; Kriesi et al., 2006; Rydgren, 2007; Swank & Betz, 2003; Teney et al., 2014; Wheatley, 2016).
However, the evidence for a globalisation-RWP link is not conclusive. Although many studies link RWP to globalisation in specific countries, comparative studies are less supportive. For instance, Teney et al. (2014) observed that, among EU-countries, globalisation strengthens the demarcation pole – the communatarian dimension – of the new cleavage, but the link with de structural component is quite weak, i.e. these attitudes do not coincide with the victims and beneficiaries of globalisation, and subjective factors play an important role in explaining these attitudes. Likewise, Rooduijn (2018) found that the electorates of populist parties not always consist of the victims of globlization and varies widely over the parties. Some other studies (Bergh & Kärnä, 2020; Polyakova, 2015) failed to find a relationship between the level of economic globalisation and support for populist parties. Bergh and Kärnä (2020), however, did find that EU membership leads to increased support for populist parties and suggests that political globalisation might strengthen populism.
The victims and beneficiaries of globalisation are defined by how globalisation affects the life-chances of various groups in society. According to the globalisation cleavage literature the main victims are lower and lower-middle class persons with low education and precarious employment or in formerly protected but now increasingly threatened sectors (Betz, 1994; Emanuele et al., 2020; Fervers & Schwander, 2015; Geishecker & Görg, 2008; Grande & Kriesi, 2012; Harteveld, 2016; Hessami, 2011; Karadağ & Uygur, 2019; Kimmel, 2017; Kriesi et al., 2006; Langsæther & Stubager, 2019; Rooduijn, 2018; Teney et al., 2014; Van der Brug & Van Spanje, 2009). They not only face most economic insecurity, but also feel socially and culturally most threatened. The old social contract is broken and their prospects for a better life and upward mobility are thwarted (Helbling & Jungkunz, 2020; Hooghe & Marks, 2018; Rydgren, 2007). For instance, Kimmel’s (2017) ‘angry white men’ not only see their socioeconomic position threatened but also feel culturally victimised and alienated. Their traditional roles in society, identities, value patterns and understanding of the world are less and less in tune with the new cosmopolitan, universalistic and individualistic cultures that gain prominence in postmodern society.
Education is a major factor distinguishing victims and beneficiaries (Enyedi, 2008; Grande & Kriesi, 2012; Harteveld, 2016; Helbling & Jungkunz, 2020; Kapstein, 2000; van Elsas, 2017). Higher educated persons are better equipped to benefit from the employment opportunities in a globalised and post-industrial economy and to deal with the cultural changes and increased diversity and therefore are more likely to embrace cosmopolitan and postmaterialist attitudes (Emanuele et al., 2020; Kriesi et al., 2008; Langsæther & Stubager, 2019). Persons with low education, however, face declining employment opportunities with the disappearance of stable industrial jobs and, in a dualised labour market, are increasingly outsiders on the secondary labour market, fearing competition from new groups entering the labour market (Bornschier & Kriesi, 2013; Geishecker & Görg, 2008; Grande & Kriesi, 2012; Hooghe & Marks, 2018; Kapstein, 2000; Kriesi et al., 2006). Migrants are seen as a threat to their social position, status, and identity. Ethnic competition arguments (Rydgren, 2007) state that migrants compete with them for employment, housing, neighbourhoods, social services, etc., and thus are perceived as a threat to their livelihood. Increased diversity is experienced as a danger to their way of life, traditions, attitudes, and values, i.e. as a challenge to the society they were familiar with (Kriesi et al., 2008; Lucassen & Lubbers, 2012; Oesch, 2008; Rydgren, 2007). In addition, the neo-liberal turn in politics, the decline of the welfare state and the perceived loss of national sovereignty make the victims feel abandoned by politicians, politics, and the state (Harteveld, 2016; Kitschelt, 2002; Morselli & Passini, 2018; Oesch, 2008; Rooduijn, 2018). To sum up, the uprooting of the victims provides a fertile soil for the development and diffusion of RWP attitudes. In WNE counties it is mainly those with low education and precarious employment who are victims of globalisation and who should be most attracted to RWP attitudes.
The debate about the effect of globalisation on welfare and labour market policies is dominated by two competing hypotheses: the compensation hypothesis which states that open, i.e. globalised economies develop more extensive welfare provisions and labour market regulations to compensate for the fluctuations of the international markets; and the efficiency hypothesis that claims that the globalisation-driven neo-liberal austerity policies led to welfare retrenchment (Anderson & Obeng, 2020; Bergh, 2021; Dorn et al., 2018; Hopkin & Blyth, 2019; Katzenstein, 1985; Meinhard & Potrafke, 2012; O’connor, 2017; Potrafke, 2015; Starke et al., 2016; van der Veen, 2009). Both views share the underlying idea that welfare and labour market arrangements can buffer the effects of globalisation for the vulnerable groups in society. To the extent they lessen the impact and prevent the worst effects of globalisation, they limit the uprooting and associated uncertainty and frustration, and thus also the spread of RWP attitudes.
Globalisation may be a global process but it does not uniformly affect all countries and regions. This is very obvious if one compares how it affects low income countries versus high income countries, but even within the EU countries are differently exposed to and affected by globalisation (Beck, 2007). The WNE experience is often taken as the norm, but the path followed by the EU members varies widely. In particular, the more recent EE EU-members for much of the post-WWII period experienced authoritarian communist regimes that severely restricted the opportunities for the old cleavages to institutionalise (Deegan-Krause, 2007; Enyedi, 2008). They only started their transition towards a capitalist market economy and a liberal democracy after the fall of communism, experiencing multiple transformations simultaneously and quite rapidly (Dorn et al., 2018). Their communist legacy affects their experience of globalisation and (post-)modernisation (Bustikova & Kitschelt, 2009; Enyedi, 2008).
Table 1 confirms the regional differences among EU-members. In 2008 the average GDP per capita in EE was still only about half of that in WNE, although the average growth in GDP per capita over the past decade of the former was more than double that of the latter. And although all EU-members are highly globalised, the KOF-indices show that WNE is still substantially more globalised than SE and EE, although globalisation proceeds at a much higher rate in the latter two regions. Similar trends are observed for the economic, social, and political dimensions, although the political globalisation differences are non-significant. That globalisation means something different for EE, is also shown that while other EU countries are deindustrialising, the EE-members are still industrialising. They also attract fewer migrants than WNE-members, but, to the contrary, experience considerable emigration.
Differences between Eastern European, Western & Northern European and Southern European EU members on globalisation and development indicators.
. | Region . | . | . | ||
---|---|---|---|---|---|
. | Eastern Europe . | Western & Northern Europe . | Southern Europe . | Total . | E2 . |
N | 10 | 11 | 6 | 27 | |
Development | |||||
Growth GDP per capita, 1998–2008 | 5.54 | 2.63 | 2.43 | 3.66 | 0.575*** |
GDP per capita, 2008 | 21404.57 | 44431.92 | 31228.48 | 32969.18 | 0.543*** |
Globalisation | |||||
KOF Globalisation Index, 2008 | 79.45 | 87.70 | 80.77 | 83.11 | 0.703*** |
Change in KOF Globalisation Index, 1998–2008 | 12.44 | 3.96 | 8.72 | 8.16 | 0.580*** |
KOF Economic Globalisation Index, 2008 | 76.32 | 83.78 | 75.48 | 79.17 | 0.311* |
KOF Social Globalisation Index, 2008 | 77.78 | 86.38 | 80.57 | 81.90 | 0.623*** |
KOF Political Globalisation Index, 2008 | 84.24 | 92.96 | 86.26 | 88.24 | 0.146 |
(De-)industrialisation | |||||
Industrial employment, 2008 | 34.31 | 23.08 | 26.24 | 27.94 | 0.676*** |
Change in industrial employment, 1998–2008 | 0.02 | −0.15 | −0.11 | −0.08 | 0.604*** |
Migration | |||||
Migrant stock, 2005 | 6.24 | 12.38 | 8.55 | 9.26 | 0.178 |
Population growth, 2003–2006 | −0.41 | 0.65 | 0.79 | 0.29 | 0.458*** |
Inequality | |||||
Gini coefficient, 2008 | 30.37 | 28.25 | 31.65 | 29.79 | 0.114 |
Politics | |||||
Liberal democracy index, 2008 | 0.74 | 0.84 | 0.80 | 0.79 | 0.317* |
Corruption perception index, 2008 | 5.02 | 8.22 | 5.72 | 6.48 | 0.754*** |
Government expenditures | |||||
General government expenditures on social protection, 2008 | 12.73 | 18.22 | 14.58 | 15.38 | 0.497*** |
Change in general government expenditures on social protection, 1998–2008 | −0.51 | −0.53 | 1.87 | 0.01 | 0.238* |
. | Region . | . | . | ||
---|---|---|---|---|---|
. | Eastern Europe . | Western & Northern Europe . | Southern Europe . | Total . | E2 . |
N | 10 | 11 | 6 | 27 | |
Development | |||||
Growth GDP per capita, 1998–2008 | 5.54 | 2.63 | 2.43 | 3.66 | 0.575*** |
GDP per capita, 2008 | 21404.57 | 44431.92 | 31228.48 | 32969.18 | 0.543*** |
Globalisation | |||||
KOF Globalisation Index, 2008 | 79.45 | 87.70 | 80.77 | 83.11 | 0.703*** |
Change in KOF Globalisation Index, 1998–2008 | 12.44 | 3.96 | 8.72 | 8.16 | 0.580*** |
KOF Economic Globalisation Index, 2008 | 76.32 | 83.78 | 75.48 | 79.17 | 0.311* |
KOF Social Globalisation Index, 2008 | 77.78 | 86.38 | 80.57 | 81.90 | 0.623*** |
KOF Political Globalisation Index, 2008 | 84.24 | 92.96 | 86.26 | 88.24 | 0.146 |
(De-)industrialisation | |||||
Industrial employment, 2008 | 34.31 | 23.08 | 26.24 | 27.94 | 0.676*** |
Change in industrial employment, 1998–2008 | 0.02 | −0.15 | −0.11 | −0.08 | 0.604*** |
Migration | |||||
Migrant stock, 2005 | 6.24 | 12.38 | 8.55 | 9.26 | 0.178 |
Population growth, 2003–2006 | −0.41 | 0.65 | 0.79 | 0.29 | 0.458*** |
Inequality | |||||
Gini coefficient, 2008 | 30.37 | 28.25 | 31.65 | 29.79 | 0.114 |
Politics | |||||
Liberal democracy index, 2008 | 0.74 | 0.84 | 0.80 | 0.79 | 0.317* |
Corruption perception index, 2008 | 5.02 | 8.22 | 5.72 | 6.48 | 0.754*** |
Government expenditures | |||||
General government expenditures on social protection, 2008 | 12.73 | 18.22 | 14.58 | 15.38 | 0.497*** |
Change in general government expenditures on social protection, 1998–2008 | −0.51 | −0.53 | 1.87 | 0.01 | 0.238* |
Significance: *: p < .050, **: p < .010, ***: p < .001
Furthermore, EE-members have weaker institutions to deal with the consequences of globalisation: their regimes are less democratic and more corrupt, while they spend less on social protection and have higher inequality than their WNE partners. Dorn et al., (2018), for instance, found that globalisation in EE had more profound impact on inequality than in WNE while the welfare institutions in the former were less able to buffer this effect. Welfare and labour market policies vary widely over the EU countries as they use different policy mixes and only few countries opt for either a pure neo-liberal and austerity policy or for compensatory policies in response to globalisation and its crises (Häusermann & Schwander, 2010; Rathgeb, 2018; Rueda, 2014). Policies also vary by region and welfare state type (Leibrecht et al., 2011; Potrafke, 2015). Countries with more extensive welfare policies and less dualised labour markets will be better able to buffer globalisation and limit the uprooting it causes.
The classic globalisation cleavage model predicts that victims – low education, outsider employment – are more prone to adopt RWP attitudes. This reflects a WNE bias, as the effects of globalisation on uprooting, and thus RWP attitudes, may vary from country to country. For instance, economic openness not necessary leads to increased volatility (Bergh, 2021), and may thus lead to varying degrees and types of victims and therefore also to differences in the support for RWP ideas. The economic performance and the way in which labour market and welfare institutions create different types of outsiders (Fervers & Schwander, 2015; Häusermann & Schwander, 2010) affect who support RWP parties and to what extent they adopt such attitudes. The globalisation cleavage model predicts that prevalence of RWP attitudes will increase with a country’s degree of globalisation and the uprooting it causes (Grande & Kriesi, 2012; Teney et al., 2014) (Hypothesis 1). As the opposition between victims and beneficiaries will strengthen with globalisation, so will the RWP attitudinal gap between them (Hypothesis 2). However, the model implicitly assumes that globalisation happens uniformly in all countries. If this is not the case, and globalisation creates different victims and beneficiaries in the various countries, the traditional victim/beneficiaries opposition may explain little of the differences in RWP attitudes.
The prevalence of RWP attitudes in countries not only depends on the type and number of victims and beneficiaries but also on the uprooting it causes (Hypothesis 3). One would expect RWP attitudes to be more prevalent in countries experiencing more uprooting. The state plays a crucial moderating role by how it reacts to globalisation and protects the vulnerable segments of society. Welfare state provisions may protect against the worst outcomes, but neo-liberal policies that promote globalisation and limit the welfare state led the victims to feel abandoned by the state and politics. However, this argument is very much inspired by the WNE countries: we anticipate that the difference between victims and beneficiaries will be largest in countries most exposed to globalisation and where its consequences are least mitigated by welfare state provisions (Hypothesis 4). In countries that are less globalised the structural base for the new cleavage may be non-existing or take a different configuration. In EE, globalisation has less advanced and proceeds at a much faster pace than in WNE. This suggests that the uprooting in EE is much more widespread and less contained to specific segments of society, leading to overall higher levels of RWP attitudes and a smaller gap between beneficiaries and victims.
Methods and data
Sample
We use data from the 2008 wave of the European Values Study (EVS2008) (EVS, 2016), which contains data for the 27 European Union (EU) member countries. The samples for the 28 surveys – with separate surveys for Great Britain and Northern Ireland – amount to 40,465 observations in total. The smallest sample is Northern Ireland (N = 500), the largest is Germany (N = 2075).
Variables
At the core of RWP attitudes lie populist attitudes in combination with an anti-migrant and anti-EU sentiment. The dependent variable, the RWP attitude index (RWPAI), is constructed as the sum of five standardised principal component score variables (see Supplemental Materials (SM), Appendix A), capturing (1) the extent migrants are considered a threat, (2) anti-EU sentiments, (3) distrust in political and social institutions, (4) dissatisfaction with democracy, and 5) a preference for strong or non-elected leaders. The final RWPAI is standardised with a high score meaning strong RWP attitudes.
The social position variable captures the beneficiaries and victims of globalisation, and is based on the respondents’ migrant status, education level, and labour market status (see SM, Appendix B) Migrant status distinguishes between native, Western migrant and non-Western migrant respondents, and is determined by the country of birth of the respondents’ mother. Education level distinguishes between low (less than complete secondary education), medium (completed secondary education) and high (completed tertiary) education. Labour market position (LMP) distinguishes between a precarious LMP, a professional LMP, non-active and other. A precarious LMP is defined as employed in an outsider occupation in an increased dualised labour market: either a lower sales or service position, a lower technical or routine labour job, being unemployed, or disabled. A professional LMP is defined as either large employers, higher or lower managers or professionals.
Victims are defined as natives with a low/medium education and a precarious LMP, while the beneficiaries are natives with higher education and a professional LMP. Of the other categories it is less clear whether they are victims or beneficiaries. Migrants do benefit from globalisation, but non-western migrants, in particular, tend to have low social status and a precarious LMP. Western migrants are often better off, although there is substantial variation here as well. Otherwise active and non-active natives are also neither clearly victim nor beneficiary.
At the country level four sets of variables are included, (1) exposure to globalisation: level of globalisation, both overall and economic, political and social globalisation (2008) (Gygli et al., 2019), and globalisation trend (1998–2008), (2) development: GDP per capita (2008) (World Bank, 2018), and growth GDP per capita (1998–2008); (3) the effects of globalisation: industrial employment (2008) (World Bank, 2018), change in industrial employment (2008), migrant stock (2005) (United Nations, 2017a), population growth (2003–2008) (United Nations, 2017b); and (4) the country’s capacity to cope with the effects of globalisation: Gini coefficient (2008) (Eurostat, 2018), liberal democracy (2008) (Coppedge et al., 2018), corruption (2008) (Transparancy International, 2018), social protection expenditures (2008) (Eurostat, 2019a, 2019b), and change in social protection expenditures (1998–2008), For more detailed information, see SM, Appendix C.
Results
Right-wing populist attitudes
Social position: Victims and beneficiaries
Overall, 22.0% of the respondents are classified as victims, while 10.9% can be considered beneficiaries. The largest category is the non-active natives (36.0%), but this is a very heterogeneous group containing not only homemakers, but also retirees and students, followed by the other employed natives (21.7%). Most migrants in the sample are Western migrants (7.3%). It is common for non-Western migrants (2.1%) to be underrepresented in such surveys.
However, the distribution of the various social positions varies substantially across countries (see SM, Appendix D). WNE members tend to have relatively many beneficiaries and few victims of globalisation, while EE countries have relatively many victims and few beneficiaries. This reflects the different levels of globalisation and modernisation.
Social position and RWP attitude
Table 2 and Figure 3 show the multilevel regression results for the effects of social position and country characteristics on the RWPAI-scores. In Table 2 the overall level of RWP attitudes is examined, while Figure 3 focusses on how the difference in RWPAI-scores between victims and beneficiaries varies over the countries. All regressions are performed using R (R Core Team, 2018) and the multilevel and nlme packages (Bliese, 2016; Pinheiro et al., 2018).
Multilevel regression models for RWPAI score, social position, and country characteristics.
. | Model . | ||
---|---|---|---|
b (se) . | 2 . | 4 . | 6 . |
Intercept | −0.399*** (0.079) | −8.063*** (1.146) | −7.269*** (0.949) |
Social position (ref: Beneficiaries) | |||
Victims | 0.639*** (0.059) | 0.639*** (0.059) | 0.641*** (0.059) |
Western migrants | 0.280*** (0.052) | 0.280*** (0.051) | 0.287*** (0.050) |
Non-Western migrants | 0.005 (0.070) | 0.040 (0.068) | 0.021 (0.070) |
Non-active natives | 0.427*** (0.043) | 0.426*** (0.043) | 0.428*** (0.043) |
Otherwise employed natives | 0.391*** (0.042) | 0.390*** (0.042) | 0.392*** (0.042) |
Globalisation | |||
Globalisation | 0.078*** (0.013) | ||
Economic globalisation | 0.037*** (0.007) | ||
Political globalisation | 0.027*** (0.004) | ||
Development | |||
GDP per capita | −0.015*** (0.004) | −0.015*** (0.004) | |
Growth of GDP per capita | 0.148*** (0.026) | 0.141*** (0.028) | |
Effects of globalisation | |||
Industrialisation | 0.027** (0.008) | 0.030*** (0.008) | |
Change in industrialisation | −0.023** (0.006) | −0.026*** (0.006) | |
Capacity to cope | |||
Corruption | −0.172*** (0.035) | −0.142*** (0.031) | |
Inequality | 0.044*** (0.009) | 0.047*** (0.010) | |
Marginal R2 | 0.035 | 0.181 | 0.192 |
AIC | 107798.3 | 107768.5 | 107765.9 |
. | Model . | ||
---|---|---|---|
b (se) . | 2 . | 4 . | 6 . |
Intercept | −0.399*** (0.079) | −8.063*** (1.146) | −7.269*** (0.949) |
Social position (ref: Beneficiaries) | |||
Victims | 0.639*** (0.059) | 0.639*** (0.059) | 0.641*** (0.059) |
Western migrants | 0.280*** (0.052) | 0.280*** (0.051) | 0.287*** (0.050) |
Non-Western migrants | 0.005 (0.070) | 0.040 (0.068) | 0.021 (0.070) |
Non-active natives | 0.427*** (0.043) | 0.426*** (0.043) | 0.428*** (0.043) |
Otherwise employed natives | 0.391*** (0.042) | 0.390*** (0.042) | 0.392*** (0.042) |
Globalisation | |||
Globalisation | 0.078*** (0.013) | ||
Economic globalisation | 0.037*** (0.007) | ||
Political globalisation | 0.027*** (0.004) | ||
Development | |||
GDP per capita | −0.015*** (0.004) | −0.015*** (0.004) | |
Growth of GDP per capita | 0.148*** (0.026) | 0.141*** (0.028) | |
Effects of globalisation | |||
Industrialisation | 0.027** (0.008) | 0.030*** (0.008) | |
Change in industrialisation | −0.023** (0.006) | −0.026*** (0.006) | |
Capacity to cope | |||
Corruption | −0.172*** (0.035) | −0.142*** (0.031) | |
Inequality | 0.044*** (0.009) | 0.047*** (0.010) | |
Marginal R2 | 0.035 | 0.181 | 0.192 |
AIC | 107798.3 | 107768.5 | 107765.9 |
Models 3-6 also included as independent variables: Social globalization, Change in globalization, Migrant stock, Population growth, Liberal democracy, Social protection expenditures, and Change in social protection expenditures. As their effects proved non-significant, they were excluded from the parsimonious models 4 and 6 and are not shown in the table.
Significance: *: p < .050, **: p < .010, ***: p < .001
Seven models were estimated, an intercept-only baseline model (0), 2 models with only the social position variable (1, 2), and 4 models including the country variables as well (3–6). Table 2 shows a selection of the estimated models (for full results, see SM Table E-1). Model 1 is the random intercept model in which the effects of social position on the RWPAI are assumed to be constant over all countries and only the mean levels on the RWPAI are allowed to vary over the countries, while model 2 is a random slope model in which the effects of social position are allowed to vary over the countries. Both models not only perform significantly better than the base model (0), but the random slope model (2) also fits significantly better than the random intercept one (1). Models 3 and 5 are the full random slope models including the country variables, model 3 including the overall globalisation variable and model 5 the three subdimensions of globalisation. Models 4 and 6 are the parsimonious models for models 3 and 5, respectively, estimated by stepwise deletion of the least significant coefficients.
Victims on the average score significantly higher on the RWPAI than beneficiaries and have on the average also the highest RWPAI scores of all social positions. The beneficiaries score significantly lower on RWPAI than all other social positions, except for the non-Western migrants. Non-active and other active natives on the average score similar on the RWPAI: significantly higher than the beneficiaries but lower than the victims, illustrating their ambivalent position regarding globalisation and modernisation. Controlling for the country level variables does not fundamentally alter the effects of social position on RWPAI.
Models 4 and 6 confirm that the national context influences the overall RWPAI level. The model confirms that, ceteris paribus, the overall level of RWP attitudes in a country increases with the overall level of globalisation. According to this effect, the average EE EU member nation would score 0.64 SD lower (0.0779 × (79.45–87.70)) on the RWPAI than the average WNE EU-member. Similar results were found for economic and political globalisation, but not for social globalisation (see model 6). The speed of globalisation as measured by the change in globalisation between 1998 and 2008 did not affect the overall level of RWPAI attitudes. As the results for the other variables are similar in models 4 and 6, the further discussion is based on model 4.
However, the effect of globalisation is countered by the observation that the average RWPAI-score is lower in richer than in poorer countries. The effect of GDP per capita predicts that the RWPAI-score for a respondent from an EE country is 0.33 SD higher than for an otherwise similar one from the WNE countries (−0.0146 × (21.405–44.432)). Globalisation and GDP per capita correlate r = 0.61 (p < .001, N = 27). Likewise, countries experiencing faster economic growth also tend to have higher average RWPAI scores. An EE respondent therefore on the average will score 0.43 SD higher on RWPAI than a WNE one [0.1479 × (5.54–2.63)]. Both globalisation and faster modernisation in the poorer EU countries seem to generate uprooting and more support for RWP attitudes. The effect of globalisation is further reduced by the other intermediate variables retained in model 4. The larger the employment in industry the higher the overall RWPAI-score tends to be, but how faster such employment grows, the lower the overall score. As both indicators are strongly positively correlated (r = 0.69, p < .001, N = 27) their effects to some extent cancel each other out. Because of lower industrial employment WNE EU-members will have lower average RWPAI-scores than EE ones, but de-industrialisation raises RWPAI-scores in WNE countries. The migration variables, migrant stock and population growth, were surprisingly unrelated to a country’s average RWPAI score. Higher levels of corruption increase the average RWPAI-score. The model predicts that, ceteris paribus, respondents in the country with the highest perceived corruption (Bulgaria, CPI = 3.6) on the average will score 0.98 SD higher on the RWPAI than similar respondents from the country with the lowest (Sweden, CPI = 9.3). The inequality in a country also increases the overall level of RWP attitudes in a country. Neither the level of liberal democracy, nor social protection expenditures were found to influence the RWPAI score.
Cross-country variations
Estimated differences in RWPAI-score, ‘Victims’ vs. ‘Beneficiaries’.
Estimated attitudinal gap, differences in RWPAI-scores between ‘victims’ and ‘beneficiaries’, by country characteristics.
Estimated attitudinal gap, differences in RWPAI-scores between ‘victims’ and ‘beneficiaries’, by country characteristics.
The level of globalisation positively not only affects the overall level of RWP attitudes in a country, but also increases the gap between victims and beneficiaries (model 7). And because the level of globalisation is strongly negatively correlated with the pace of globalisation (r = −0.67, N = 27, p < .001) one also observes a negative effect of the change in globalisation on the RWPAI-gap (17). Further supporting this finding is that the attitudinal gap also increases with the levels of economic and social globalisation, although not with that of political globalisation. The findings for the development indicators (models 11 & 12) support this interpretation. Countries with higher levels of development have larger attitudinal gaps than lesser developed EU members. The growth of GDP per capita, which correlates negatively with the level (r = −0.42, N = 27, p = .031) has a negative effect on the RWPAI-gap. The results also support the de-industrialisation hypothesis (14 & 15): lower levels of industrial employment and a decline in industrial employment over the previous decade both widen the attitudinal gap between victims and beneficiaries.
The analysis in model 22 shows that immigration fuels the gap in RWP attitudes: population growth has a positive effect on the gap in RWP attitudes. For instance, the model predicts that the RWPAI-gap in the country with the highest population growth (Ireland: 2.08) is 0.68 SD higher than in the country with the lowest population growth (Latvia: −1.19). The interaction effect between migrant stock and the victim social position is also significant and positive but the model (21) is not significantly better than the main effects model (4).
The RWP attitudinal gap is also smaller in more corrupt countries (13) and in countries with higher inequality (16). The attitudinal gap is most pronounced in the more egalitarian countries. Higher government social protection expenditures (19) also increase rather than diminishes the attitudinal gap.
Conclusion and discussion
The globalisation cleavage argument provides an explanation for the rise of RWP attitudes and votes. However, most studies take the underlying mechanism for granted. This study examines whether the level and kind of globalisation does affect the prevalence of RWP attitudes, especially among the ‘victims’ of globalisation, in the 27 EU member states. The literature on the globalisation cleavage draws largely on the WNE experience. Globalisation, however, is not a uniform process, it means something different in WNE and EE.
The results overall support the globalisation cleavage argument: the prevalence of RWP attitudes and the attitudinal gap between victims and beneficiaries increased with the level of globalisation, and support the idea that the development of RWP attitudes a reaction is to the uprooting caused by transformations such as globalisation. For instance, the high overall prevalence of RWP attitudes in EE may not only reflect the differences in globalisation experiences but also their ongoing transitions towards a capitalist market-economy and a liberal democracy.
Cleavages require that specific attitudes are linked to specific groups in society, that the normative component is congruent with the structural one. Indeed, in all EU countries the victims (lower educated and lower-class natives) have significantly higher average RWP attitudes scores than the beneficiaries, which validates the core victims/beneficiaries categories identified in the literature, although it does not exclude that the composition of these categories varies somewhat with the degree and manner of globalisation. The adoption of RWP attitudes is a reaction to the uprooting caused by globalisation. The attitudinal gap varied substantially over countries and tended to increase with the level of globalisation. Although some of the effects of the subdimensions of globalisation, economic, social and political globalisation, proved non-significant, overall their effects were quite similar, they all to some degree seem to increase both the overall level of RWP attitudes and the attitudinal gap between victims and beneficiaries. These results therefore do not provide conclusive evidence that one dimension is substantially more important for the development of a globalisation cleavage than the others. The results also suggest that the globalisation cleavage is less developed in the poorer EE EU-members than in WNE, which may reflect both the WNE bias of the globalisation cleavage literature and the different globalisation experience of EE countries.
Contrary to expectations is our finding that strong buffering institutions, such as welfare states and democracy, do not reduce the attitudinal gap between victims and beneficiaries, but rather increase it. Lower inequality and higher social expenditures were found to lead to higher RWP scores among the victims. It is possible these effects are spurious, reflecting higher levels of globalisation, but it also is possible that a more effective welfare state increases the fear of losing out among the victims as neo-liberal austerity policies threaten welfare programmes. In countries with weaker welfare states the victim groups may rely less on welfare programmes and therefore be less affected by their retrenchment.
RWP attitudes reflect an underlying dissatisfaction with and fear for the future, one’s own social position and the way the country is evolving. These may spread, driven by political parties, movements and media, beyond the groups most heavily affected by the transformations, creating a RWP zeitgeist (Mudde, 2004). The results suggest that such zeitgeist in the EU members can be attributed to the inequality in the country, corruption, increasing diversity, but also by high emigration in response to the poorer living conditions and limited opportunities in a country.
A first limitation of this study is that the data dates to 2008. Between 2008 and now, much has changed. Right-wing populism has become more influential: in some countries – Poland, Hungary, Italy – populist parties provided the head of government, while in others it supported government and in most EU countries right-wing populist parties are now represented in parliament. Whether that means that the new cleavage has become more intense, or, to the contrary, whether RWP attitudes have become mainstream and have weakened the new conflict axis, cannot be determined based on this data. Nevertheless, this study provides a base against which post-2008 developments can be evaluated. Globalisation is an ongoing process and both its impact on society or how a society reacts to it are bound to evolve over time. Its effects on political attitudes therefore can also change over time.
A second limitation is the coarse operationalisation of social position. The typology used in this study built upon the definitions of victims (or losers) and beneficiaries (or winners) commonly used in the globalisation cleavage literature. Although the literature predicts that lowly educated natives in precarious employment and work conditions run the highest risk of being victimised by the transformations, they certainly are not all victimised, nor are other groups immune to victimisation. As these transformation processes are multidimensional, they create a wide range of victims and beneficiaries. More fine-tuned indicators that capture the degree to which one is victim of beneficiary on multiple dimensions are needed to provide a more detailed and nuanced view of the dynamics underlying the formation and adoption of RWP attitudes. Victimisation on different dimensions may have different consequences for RWP attitudes.
This study also suffers from a shortcoming inherent to this kind of studies. The results largely reflect the differences between the WNE EU members on the one hand, and the EE ones on the other, and with the Southern EU members in between. Therefore, any country-level variable that reflects these geographic differences will to some extent explain the differences in RWP attitudes and the attitudinal gap between victims and beneficiaries. Although the evidence supports a hypothesis, the actual causal mechanism may involve related but not-included factors.
Proportion ‘Very high’ on RWP Attitude Index (RWPAI), 2008, vs. Change in Liberal Democracy Index (LDI), 2008–2016.
Proportion ‘Very high’ on RWP Attitude Index (RWPAI), 2008, vs. Change in Liberal Democracy Index (LDI), 2008–2016.
Their dissatisfaction with democracy and with the functioning of the political system means that they are more attracted to non-democratic ideologies, parties, and politicians, and more likely to support non-democratic policies. The turn towards illiberal democracies and the attacks on liberal freedoms and rights, mainly but not exclusively in several East-European countries, relies to a considerable extent on the support of this alienated and dissatisfied crowd for populist politicians, parties, and policies. The neo-liberal policies promoted by, among others, the EU are ineffective combatting RWP attitudes. To the contrary one needs policies that remedy the threats and uprooting posed by globalisation. Future research, therefore, should focus on policies that alleviate the victims’ fears and perceived threats and that provide positive perspectives for the future.
Disclosure statement
No potential conflict of interest was reported by the author(s).
References
Author notes
Supplemental data for this article can be accessed online at https://doi.org/10.1080/23254823.2022.2121299.