Introduction

Negative attitudes toward immigration are among the most direct triggers of the Populist Radical Right Parties’ (PRRP) vote in Europe, with two main groups of theories dominating the literature: economic and cultural explanations. Among the former, scholars have contended that labor competition between native and migrant populations might be at the root of this link, albeit mediated by social class. It has been argued that the poor would fear the consequences of the presence of a high percentage of low-skilled and unemployed immigrants, who would compete for the same jobs (Jetten, 2019). However, in this article we do not focus on economic explanations based on job competition, but on those related to social spending and social benefits; accounts that have received much less attention in the literature despite the promising findings of previous contributions (see Facchinni and Mayda, 2006; Dustmann and Preston 2007; Hainmueller & Hiscox, 2010; Edo et al., 2019). Could social spending related to immigration be a reason behind the vote for PRRPs? Some previous studies have found correlations between the mere presence of immigration and the vote for PRRPs, but is it in those contexts where immigration represents a greater expense (especially for the higher and middle classes) and competition for limited social benefits (especially for the lower classes) that the natives tend to develop a greater propensity to vote for the far-right (Alesina et al., 2019; Di Tella et al., 2017; Fong & Luttmer, 2009; Kuziemko et al., 2015)?

Regarding the second set of explanations, the literature has shown how cultural heterogenization linked to the arrival of migration is perceived by some citizens as a threat to the local culture and “traditional way of life”, becoming a trigger to vote for PRRPs (Mudde, 1999; Card et al., 2009). Thus, the defense of “the nation”, or “the us” against a vague cluster of “the others”, has been employed as a core element in PRRP discourses, especially in those contexts where the main flows of migration have a different religious or cultural background (Halikiopoulou et al., 2012).

These two ‘families of explanations’ have been traditionally studied independently, despite academics highlighting the need to analyze the interaction of economic and cultural variables (Mols and Jetten, 2021; Chen, 2020; Halikiopoulou & Vlandas, 2020; Edo et al., 2019; Golder, 2016). Previous works argued how cultural heterogenization derived from the arrival of immigrant populations might affect social trust and cohesion, eroding intergroup solidarity and support for the welfare state as a result (Alesina & La Ferrara, 2002; Breznau & Eger, 2016; Habyarimana et al., 2007; Luttmer, 2001). In this line, we expect that in contexts where immigration implies greater heterogeneity and higher social expense, support for PRRPs will be higher (Schmidt-Catran & Spies, 2016). Focusing on the effects of contextual-level unemployment, a factor for which research findings to date are mixed (Sipma & Lubbers, 2020), we expand existing accounts by simultaneously assessing unemployment among different groups (in the general population, EU migrants and non-EU migrants) and their impact on the likelihood of supporting PRRPs across social classes.

Our results confirm that only in contexts with large concentrations of non-EU unemployed migrants—Third Country Nationals (TCNs)—the propensity to vote for PRRPs is higher. Nonetheless, the effects of contextual-level unemployment seem to depend on the social profile of the voter. We find that while upper classes (according to Oesch’s classification, those belonging to “higher-grade service”) tend to vote for PRRPs to a greater extent in contexts of high unemployment (among immigrants irrespective of their country of origin, but also among nationals), for the lower and middle classes (lower-grade service, small business owners, but especially skilled and unskilled workers), only TCN unemployment boosts the likelihood of supporting these parties. Our findings point to both economic and cultural explanations, dependent on natives’ social class, regional migration and economic contexts. Thus, while for the richest (and therefore those who contribute most with their taxes) the additional economic burden would be the strongest explanation for their vote, for the middle and lower classes it is rather a combination of economic and cultural explanations that fosters their support for PRRPs. These findings call for including tax burden and social benefits’ competition explanations in future studies of the effects of unemployment on PRRP voting.

Economic Self-interest or Cultural Threat?

If postmaterialist values; specific policies related to environmental defense and preservation; and other left-wing policies were behind the emergence of Green Parties during the 80s (Mair, 1997), most scholars have linked the success of Radical Right Parties (RRPs) to immigration. According to Husbands, (1988), immigration is their raison d'être, and several authors (e.g., Van der Brug et al., 2000) even label RRPs ‘anti-immigrant’ in order to emphasize the importance of immigration in the radical right’s electoral strategy (see Fennema, 1997 for a comprehensive conceptualization). For other scholars (e.g., Mudde, 1999), anti-immigrant feelings are, above all, a manifestation of nationalism (combined with a kind of xenophobia against the “other”, i.e., the non-native) that constitutes the core ideology of the radical right (nativism). Mudde, (1999) contests this “single-issue” theory, arguing that, while RRPs gained ground in the late 80s and early 90s by emphasizing the relevance of immigration, it was in the late 90s that these parties started to develop a message focused on patriotism and security (both related with immigration); but also against the established elite. Thus, Mudde argues that, rather than rejecting the political party per se, contemporary RRPs “produce a constant stream of populist anti-party sentiments” (p. 191), which adds populism to nativism and authoritarianism as core characteristics of PRRPs (Mudde, 2007). In sum, it seems clear that one of the main drivers of support for PRRPs is anti-immigrant sentiment (Eatwell & Goodwin, 2018).

At individual level, there is a great amount of empirical evidence pointing towards a clear relationship: negative attitudes toward immigrants are positively associated to PRRP voting (Ivarsflaten, 2008; Rooduijn, 2018; Zhirkov, 2014). However, when the literature has analyzed the contextual determinants of PRRP support in relation to immigration, the results are frequently inconsistent (Amengay & Stockemer, 2019). For example, Werts, Scheepers and Lubbers (2012) in their comparative analysis of 18 European countries find a positive relationship between immigration rates and vote for PRRPs. However, Mudde (1999: 185) argues that “there seems to exist consensus regarding the absence of a clear-cut relation between the number of immigrants and the electoral success of Extreme Radical Parties (ERPs) in a certain territorial unit”. Daniel Stockemer (2016, et al., 2018, 2021) seems to agree that the mere presence of immigrants does not lead to greater support for far-right parties, pointing to individual explanations based on the perceptions of immigration, economic deprivation and dissatisfaction with the political regime. A recent study of the French case by Vasilopoulos et al., (2022), confirms the mixed results of extant literature, showing that while immigration share at neighborhood level was negatively associated with voting for Marine Le Pen in the 2017 Presidential elections, at department level, there is a positive and significant relationship between migrant presence and support for the National Front (now Rassemblement National (RN); or ‘National Rally’ in English).

Economic Self-interest

But, why does the presence of immigrants lead to greater support for PRRP in certain contexts? Economic and cultural explanations are related to migration given that “peaks in immigration/asylum-seeking not only increase competition for scarce resources (so-called realistic conflict threat), but also fears that the host community’s culture and identity might become overshadowed (so-called symbolic threat)” (Mols and Jetten, 2021). Regarding the economic approach, the most common line of research has focused on labor market competition, arguing that under negative economic conditions both natives and immigrants might compete for the same scarce jobs, and given these negative circumstances, natives might be more prone to blame immigrants for this situation, lured by the nativist messages of PRRPs. It has been argued that it is economically deprived citizens, the so-called “losers of globalization” (Betz, 1993; Im et al., 2019; Kriesi et al., 2006; Kurer, 2020), who are more likely to resent having to compete with immigrants over resources and treat them as scapegoats, blaming them for their economic marginalization (Lubbers et al., 2002). In this context, PRRPs successfully mobilize both cultural and economic grievances over immigration (Hobolt & Tilley, 2016; Ivarsflaten, 2008). One example of such populist electoral propaganda would be Jean Marie Le Pen’s classic National Front statement: “One million unemployed is one million immigrants” (Billard, 2017). However, neither empirical evidence from the field of economics has shown a direct link between immigration and unemployment, nor have analyses of electoral behavior shown that economic deprivation automatically drives an increase in the vote for PRRPs (Lubbers et al., 2002; Mols & Jetten, 2020). In particular, the predicted positive effect of contextual unemployment on PRRP voting is far from established in the literature and there seems to be no evidence that the effect of unemployment on PRRP voting is greater when migrant populations are larger (Sipma & Lubbers, 2020).

However, the set of economic factors behind PRRPs’ success (on which we focus on in this article) is not related to labor competition, but to a much less studied phenomenon: the effects of immigration on the tax burden and competition for limited social benefits. In fact, in their analysis of the determinants of anti-immigrant attitudes, Dustman and Preston (2005) show that negative perceptions of social spending generated by immigration are even more important determinants than negative perceptions on the effects of this immigration on the labor market, especially in regions with a high presence of immigrants.

From the discipline of economics, the traditional strategy when studying the impact of immigration on macroeconomics has been to determine the balance between its contribution and its cost (Facchini and Mayda, 2009). This model has conventionally assumed that low-skilled immigration tends to generate a negative balance in public accounts. Thus, low-skilled immigration would increase social spending (in the form of education, healthcare, and unemployment benefits and subsidies, among others) since these immigrants have, on average, less income than the native population. However, the arrival of highly-qualified immigrants contributes to enriching receiving countries’ human capital, productivity and innovation, with a positive effect on both salaries and the native tax burden (Peri, 2016). Despite the vibrant debate on the balance of immigration on local economies in the specialized literature (Borjas, 1999a, 1999b; MaCurdy, Nerchyba and Bhattacharya, 1998; Peri, 2016), PRRP voters tend to agree that immigration has a negative impact on spending and the use of social benefits (Simon, 1989; Hanson, 2007; Borjas, 1999a, 1999b; Halla et al., 2017). In order to test this idea, we analyze the extent to which—in contexts of higher tax burdens and competition for social benefits related to immigration (specifically, where high rates of unemployed immigrants are registered, see Facchini and Mayda 2006)—natives are more likely to support PRRPs.

However, economic theory has shown how the relationship between immigration and economic preferences is neither linear nor automatic (see Hainmueller & Hiscox, 2010: 65). In fact, the expectation of “the fiscal burden model” is that the higher and middle classes would only be opposed to the arrival of low-skilled immigrants that entail higher costs, as this would lead to an increased tax burden (Dustmann and Preston 2007: 2; Facchinni and Mayda, 2006). As said, it has also been argued that the poor would also fear the consequences of a high proportion of low-skilled and unemployed immigrants, as they ‘compete for the social benefits of the former’ (Jetten, 2019). Furthermore, while some authors have shown that the unemployed are more likely to opt for PRRPs (Rama & Cordero, 2018), others have found mixed results (Rooduijn, 2018; Siedler, 2011; van Elsas, 2017) or even that the unemployed are the least willing to support this family of parties (Siedler, 2011). With a survey experiment, Hainmueller and Hiscox, (2010) show that both the concerns associated with the loss of the poor’s social rights and that of the rich about rising taxes are relevant in predicting negative attitudes towards immigration.

Thus, economic self-interest in relation to immigration might affect both the higher and lower social classes, as its potential costs might lead to increased taxes or the cutting of social expenditure. Citizens with lower socioeconomic status are the greatest beneficiaries of social benefits and, in general, the most in favor of economic redistribution (Dustmann & Preston, 2007). However, this inclination might be different in contexts where native workers must share these limited benefits with immigrants. More so given that, due to ‘welfare chauvinism’ (Andersen & Bjørklund, 1990), support for redistribution already goes hand-in-hand with disagreement that immigrants should receive benefits among PRRP supporters.

In an experiment in Norway, Bay and Pedersen, (2006) asked participants whether they favored a minimum income and next, whether they favored this minimum income if it also applied to immigrants. Those with lower socioeconomic levels were the most likely to change from a positive to a negative attitude in such a trade-off. In turn, the upper social classes are the biggest contributors to economic redistribution and, generally, those who show negative attitudes towards redistribution; especially if it benefits non-nationals (Hanson, 2007). Fachinni and Mayda, (2006) demonstrate that attitudes towards immigration among the richer harden in fiscal scenarios where immigration involves greater welfare costs. Similar results were reached by Dustmann and Preston (2004), who studied the impact of different scenarios of immigrant flows and its effects on the higher classes.

This double association in the literature on attitudes towards redistribution has also been transferred to that on electoral behavior. PRRPs, it has been shown, receive votes from both the vulnerable sectors of society and those in more favorable economic circumstances (Jetten et al., 2015). In fact, the “wealth paradox” (Mols & Jetten, 2017) highlights that PRRP manage to attract a high proportion of voters from both lower and upper-middle classes (Grofman and Muller, 1973; Guimond and Dambrun, 2002). More specifically, Ivarsflaten (2008) investigates the “strange bedfellows” (blue collar workers and the petty bourgeoisie), which represent the leading sources of votes for PRRPs in France and Denmark. Despite their differences (blue collar workers tend to agree on greater social welfare coverage, whilst the petty bourgeoisie argue for less state intervention in the economy), both share a pejorative view of immigration, which makes the former opposed to immigrants obtaining better social benefits, and the latter negative about tax burdens related to these benefits.

In this sense, following the economic self-interest explanations, we expect a greater presence of unemployed immigrants to increase the vote for PRRPs among high class natives, as the tax burden generated by unemployed immigrants falls on them more heavily. At the same time, in the same contexts, we might expect higher levels of support towards PRRPs among lower class natives, as they compete with immigrants for scarce social benefits.

H1a

The higher the presence of unemployed immigrants in a given region, the greater the likelihood to vote for a PRRP among higher social classes.

H1b

The higher the presence of unemployed immigrants in a given region, the greater the likelihood to vote for a PRRP among lower social classes.

Cultural Threat

However, economic explanations generally fail to account for the success of PRRPs alone, as cultural factors also need to be included in the equation (Sniderman et al., 2004). In their analysis of 21 democracies, Card, Dustmann and Preston (2009) conclude that explanations based on the compositional effect (the greater diversity generated by immigration from different cultures) have more relevant effects on attitudes towards immigration than the perceived economic effects of the arrival of immigrants (on wages, taxes and benefits). However, a number of enlightening contributions in the literature on PRRP voting have insisted on considering not the addition, but the interaction between cultural and economic explanations. Chen, (2020) reaches interesting conclusions for the United States: those who suffered most during the Great Recession tended to vote for the democratic socialist Bernie Sanders in contexts where migration from Central America is less common, while those who lost their employment in areas with high immigrant flows from Central America tended to opt for Trump. Edo et al., (2019) have also contributed to the study of this interesting interaction in the French presidential elections. Using panel data from 1988 to 2017 they determined that the presence of low-skilled immigrants from non-Western countries had an impact on the vote for extreme right parties—mainly the populist RN (Edo et al., 2019).

In the same vein, Halla et al., (2017) studied the case of Austria and found a generally positive correlation between aggregate levels of immigration and support for far-right parties, but a stronger effect for low-skilled immigrants, immigrants from Turkey and the ex-Yugoslavian countries; and in areas with many high-skilled natives. This means that the determining factor might not be immigration per se but the type of immigration. These case studies are not isolated contributions, as illustrated by Brunner and Kuhn, (2018) in the case of Switzerland, and Dinas et al., (2019) in the Greek case. While in the former study, the main finding was that the presence of culturally different immigrants increased support for the Swiss People’s Party (SVP); in the latter, it was the so-called ‘refugee crisis’ that boosted support for the extreme-right Golden Dawn in the September 2015 Greek elections.

In this regard, it is worth underlining the relevance of the 2015 refugee crisis, as the largest migration and humanitarian crisis that Europe had faced since World War II. This occurred when a large flow of people fleeing conflict (especially the civil war in Syria), persecution and extreme poverty in the Middle East, Africa and Asia, attempted to enter Europe in search of refuge. Countries such as Germany and Sweden received a large number of refugees, prompting an immediate response from PRRPs, who perceived them as a threat to both the local economy and way of life.

Following these analyses, we do not assume culturally similar immigrants to have the same effects on electoral behavior as immigrants from third countries (Bjørklund 2007: 248). However, are these perceptions and behaviors also mediated by natives’ social classes? Scheepers et al., (2002) show that the poor, manual workers and less educated, feel more threatened on the cultural dimension than the rich. In fact, these authors find a positive relationship between the percentage of non-European immigrants in close proximity and the higher percentage of manual workers against immigration (2002: 29). The scholars concluded that these feelings are due to a double threat: economic and cultural. Thus, the interaction between the economic and symbolic threat would trigger the PRRP vote, especially among the less skilled and less educated (Gidron & Hall, 2017; Golder, 2016; Halikiopoulou & Vlandas, 2020).

Zhirkov (2014: 292) also found that social class played a fundamental role in explaining the vote for PRRPs. His analysis compared abstainers and moderate right voters and found that blue-collar citizens were most likely to support radical right parties. In the context of why this particular social class has a predisposition to vote for the radical right, Oesch, (2008) analyzed working class support for far-right parties in Austria, Belgium, France, Norway and Switzerland. The study clearly demonstrates that cultural, rather than economic, explanations drive support for these kinds of political formations. In this sense, members of the working class do not seem to buy the xenophobic messages of radical right leaders because they believe that immigrants put their economic subsistence at risk, but because they pose a threat to their national way of life. Assuming that cultural explanations are the most powerful in explaining working class support for the radical right and, as we previously discussed, the presence of immigrants in a country is positively associated with the tendency to support PRRPs, we expect that if these immigrants are more culturally diverse (i.e., TCNs); blue-collar voters might be more prone to support PRRPs.

On the other hand, Edo et al., (2019: 100) show how TCN flows contributed to the success of PRRPs “due to public finance concerns”. Contexts with high ethnic diversity tend to correlate with lower levels of support for redistributive policies (Luttmer, 2001), a finding corroborated by evidence based on experiments in which subjects were found to be more charitable if beneficiaries corresponded to the same ethnic group (Fong & Luttmer, 2009; Hanson, 2007) demonstrate that the rich in the United States tend to show negative attitudes towards immigration in those states in which immigration from Central American countries has caused an increase in taxes. Magni (2020) further argues how the rich are less concerned about the poor in contexts where inequality is high. In other words, the greater the distance between the richest and the poorest, the lesser the feelings of solidarity among the richest. Finseraas (2012) also finds that the rich are less supportive of redistribution in contexts with high ethnic diversity. However, when analyzing the causal mechanisms, the author finds that this lack of solidarity is not due to a lack of solidarity specific to diverse populations, but because in racially heterogeneous contexts social mobility is harder, and the rich are less afraid of them socially thriving and becoming beneficiaries of redistribution.

In sum, according to our expectations, high levels of TCN unemployment would increase the vote for populist parties among lower social classes, as they feel more threatened on the cultural dimension than the rich, which might indicate a more relevant trigger to vote for a PRRP. At the same time, we might expect that in societies with high levels of economic inequality (as in contexts of high unemployment) the higher classes will be more likely to vote for PRRPs, as these tend to show lower levels of solidarity when ethnic diversity is higher. Considering this, our last hypotheses are:

H2a

The higher the presence of unemployed immigrants from non-EU countries in a given region, the greater the likelihood to vote for a PRRP among higher social classes.

H2b

The higher the presence of unemployed immigrants from non-EU countries in a given region, the greater the likelihood to vote for a PRRP among lower social classes.

Data

In this article, we go beyond the existing literature—based mostly on case studies—by including over 60 regions from 10 European countries (Austria, Belgium, Czech Republic, Finland, Germany, Italy, Norway, Sweden, Switzerland, and the United Kingdom) in our analyses. This wide sample of regions is limited by the availability of data at individual (ESS, Round 8) and regional levels (OECD, 2018).Footnote 1 To study the fiscal burden of the immigrant population at regional level we use data from the Organization for Economic Cooperation and Development (OECD, 2018).Footnote 2 More specifically, our key independent variables are the share of the unemployed (among the general population), the share of unemployed among migrants from EU member states, and the share of unemployed among TCNs, all at regional level. In this regard, it is important to note that not every migrant from outside the EU (TCN) might be perceived as “culturally different”; nor every European immigrant as “culturally similar” to the recipient country nationals. However, given the subjective nature of this notion, the comparative nature of our analyses, and the fact that TCNs themselves feel more discriminated against (on grounds of ethnicity, nationality, or race) compared to EU nationals (OECD/European Union, 2018: 256), considering “TCNs” as culturally different is the best proxy available in the regional database, and has been used as such in previous literature (Scheepers et al., 2002).

Carrying out detailed analyses of the ‘fiscal balance’ of immigration is not an easy task either. Previous economic literature has employed calculations that consider the expenses generated by migration (mainly in the form of social services) and its economic benefits (mainly through taxes). This group of theories tends to agree that low-skilled immigration—being below the economic average of the population—drives higher costs than when immigration is skilled (Dustmann & Preston 2007: 2; Facchinni & Mayda, 2006; Dadush, 2014). However, these calculations are extremely complex, depending on the tax burden in each context; and an agreement in the literature on the most appropriate way to calculate it has not been reached. In the same vein, different countries, or even regions, vary on social benefits for the unemployed migrant population, and PRRPs tend to highlight these benefits (independently of their coverage) to justify their stances against immigration. Therefore, in this paper we have chosen to isolate the expenditure side of the balance sheet by analyzing a more direct indicator, namely the unemployment rate among the general population, the EU immigrant population, and the TCNs. We also control for the proportion of foreign-born, to account for immigrants at regional level.Footnote 3

At individual level, we use data from the 2016 European Social Survey (ESS, Round 8).Footnote 4 The dependent variableFootnote 5 considers voting in the last parliamentary elections and is binary, coded 1, for having voted for a PRRP, and 0 for casting a ballot for another party (abstainers, those who do not know whether they voted or not, and those who do not answer the question, are omitted from the analysis).Footnote 6 Objective social class based on Oesch’s (2008) 5-category classification (1 = higher-grade service; 2 = lower-grade service; 3 = small business owners; 4 = skilled workers; 5 = unskilled workers) is the key independent variable at individual level. For the sake of robustness, models were also rerun using household income (Household total net income from all sources, in deciles) instead of social class.

In all models, we control for gender (1 = female), age (7 categories; 18–24; 25–34; 35–44; 45–54; 55–64; 65–74; 75 or more), habitat (1 = a big city; 2 = suburbs or outskirts of a big city; 3 = town or small city; 4 = country village or farm), education (1 = low; 2 = medium; 3 = high); ideological self-placement on the left–right scale (0 = left; 10 = right); anti-immigration attitudes (0 = ‘immigrants make the country a better place to live’; 10 = ‘immigrants make the country a worse place to live’); Euroscepticism measured by attitudes towards EU unification (0 = should go further; 10 = gone too far—note that we inverted the original scale) distrust in political parties (0 = complete trust; 10 = complete distrust); and negative economic sociotropic evaluation (0 = extremely satisfied with the present state of the economy in the country; 10 = extremely dissatisfied—we inverted the original scale); being unemployed and looking for a job in the last 7 days (1 = unemployed); perception of the tax burden of social benefits (“Social benefits/services cost businesses too much in taxes/charges”; 1 = disagree strongly; 2 = disagree; 3 = neither agree nor disagree; 4 = agree; 5 = strongly agree—we inverted the original scale). Table A1 in the Online Appendix gathers the descriptive statistics for all variables.

Table 1 Impact of regional share of unemployment (in the general population, among EU migrants, and among TCNs) on PRRP voting by natives’ social class
Full size table

Results

Given the dichotomous nature of our dependent variable, the hierarchical setting of the data and the necessity of including cross-level interactions, we use multilevel binary logistic regression models with random intercepts by regions nested into countries.Footnote 7 Table 1 gathers binary logistic regression coefficients of six models.Footnote 8 The odd ones are the additive models testing for the effects of regional share of unemployed among the general population (1), share of unemployed among migrants from other EU members states (3), and share of unemployed among TCNs (5). The even ones are interactive models that add the interaction term between each of the regional unemployment characteristics and natives’ social class (2, 4 and 6). Data is weighted with the poststratification and population ESS weights. We omitted the individual-level control variables in Table 1—for the full table see Table A2 in the Appendix.

We find a significant negative effect of regional unemployment among migrants from EU member states on PRRP voting (model 3), while in the case of the proportion of unemployed TCNs (model 5), the association is the opposite. Also, there seems to be no significant impact of unemployment among the general population (model 1). Thus, voting for PRRPs is a more likely outcome in a context of high unemployment only among TCNs. The diverse effects of unemployment, depending on the origin of the unemployed in a given region, seem to point more toward a combination of both cultural and economic factors.

Regarding the conditional effects of regional unemployment rates, as the significant negative coefficients of the interaction terms in models 2, 4, 6 show: the natives’ social class matters. The lower and middle classes’ likelihood to support PRRPs seems less affected by regional unemployment and unemployment among EU migrants. It is the higher classes’ propensity to vote for PRRPs that is boosted by regional unemployment rates, irrespective of the origin of those without employment – all (including natives), EU migrants, or TCNs. However, regarding lower social classes, only unemployment among TCNs seems to boost their support for PRRPs.

It is easier to interpret these interaction terms’ effects by looking at their marginal effects in Fig. 1.Footnote 9 The continuous line with circles represents the higher-grade service class. With an increase in all three measures of regional unemployment—general (left hand panel), among EU migrants (center panel), among (TCNs)—the likelihood of higher-grade professionals to support PRRPs clearly rises. This confirms hypotheses H1a and H2a and offers support for the fiscal burden approach. In the case of general unemployment and EU migrant unemployment, the relationship with PRRP voting is inverse for the remaining social classes—the more unemployment, the lower the PRRP vote. However, as with the upper classes, TCN unemployment seems to foster PRRP voting also among the other classes (particularly among unskilled workers). Hence, given the importance of the origin of unemployed migrants for the lower classes, while H2b finds empirical support, H1b has to be rejected.

Fig. 1
figure 1

Source: Own elaboration based on ESS8 and OECD, 2018. Notes: Marginal effects of the interaction terms from models 2, 4, and 6 from Table 1

Marginal effects of unemployment rates—general (left) among EU migrants (center) and TCNs (right)—on PRRPs’ support by natives’ social class.

Full size image

Lastly, to account for the fact that social classes may include multiple groups in which their self-interests regarding immigrant unemployment may differ (e.g. petit bourgeois running a small business seeing immigration as a cause of higher taxes, and big CEOs seeing immigration as an opportunity that puts pressure on the job market and favors lower wages beneficial for their business), as a robustness check, we rerun the analysis with household income instead of social class (see Table A4 and Figure A1 in the Appendix). The results are robust to this specification. First, unemployment among TCNs drives support for PRRPs across the income spectrum (there is only a slightly significant interactive effect). Second, income exerts negative effects in contexts of low general unemployment and unemployment among EU migrants, and boosts support for PRRPs where these indicators are high.

Conclusions and Discussion

This paper contributes to the literature on the relevance of economics behind PRRP voting. Previous literature has shown that explanations based on labor competition between natives and migrants are especially applicable to the lower classes and less qualified workers where jobs are scarce, pointing to a mediating relationship between economy and social class. In general terms, these contributions have shown how, in contexts of economic scarcity and when demand remains constant, migrants compete with the native population for jobs and social services. This “economic threat” combines with the “cultural threat” to explain the growth of PRRPs in Europe. However, in this paper we do not focus on labor market competition, but on an economic aspect that has received much less attention, despite its potential in recent analyses: the tax burdens generated by immigration. Specifically, we analyzed the extent to which in those territories where immigration generates higher levels of expenditure (where unemployment rates among the immigration population are high and subsequently the fiscal balance is more negative) the propensity to vote for PRRP is greater.

When examining unemployed migrants in more than 60 regions in 10 European countries, we find that higher concentrations generate greater tendencies to vote for PRRPs, although this relationship is mediated by social class, as was the case with labor competition. More specifically, among the upper classes—those that contribute most with their taxes –, the presence of unemployed immigrants is positively related to voting for PRRPs, regardless of the country of origin of the migrant population. However, among the lower classes this behavior only occurs when the unemployed migrant population is from third countries (namely non-European).

This interesting interaction adds to studies that have highlighted the need to interrelate economic and cultural explanations. For those voters who contribute most with taxes, the relationship seems to be purely economic. Thus, higher-grade professionals tend to support PRRPs to a greater extent in those contexts where unemployment is higher; no matter if these unemployed are from the general population, migrants from the EU or from third countries. On the contrary, among the middle and lower social classes support for PRRPs decreases in regions with higher unemployment rates among the general population and among immigrants from other EU member states. However, this support increases where the unemployed are from third countries. This finding might point to an interaction between economic and cultural explanations, since the fiscal burden generated by unemployment as well as the ‘competition’ for social benefits only drives support for PRRPs when the unemployed are non-European.

This study supports an increasingly established line in the literature analyzing the contextual factors behind PRRP support. Neither the “fiscal burden” nor the “cultural threat” in isolation can explain the growth of PRRP parties in Europe, as explanations considering the interaction between both are necessary. Additionally, in this article we have gone one step further by showing that these explanations might depend on both the voters’ profile (i.e. social class) and context, which indicates a complex explanation behind the vote for this family of parties. Lastly, it also provides insight into the link between contextual-level unemployment and PRRP voting, calling for a concurrent analysis of how unemployment rates across different social groups affect the likelihood of supporting PRRPs by voters’ social class.