A new Iberian exceptionalism? Comparing the populist radical right electorate in Portugal and Spain

ABSTRACT

Since their transition to democracy in the 1970s, Spain and Portugal have been ‘immune’ to the success of populist radical right (PRR) parties. This exceptional situation, however, came to an end: Chega’s leader, André Ventura, was elected in the Portuguese parliament, while VOX has become the third most voted political party of Spain. Using new online survey data from the Spanish and Portuguese national elections in 2019, we find that the Iberian PRR electorate is mostly in line with the characteristics of the PRR electorate in Western Europe when it comes to socio-demographics, political dissatisfaction, media diet, and the rejection of immigration and feminism. Interestingly, however, the support for Chega and VOX does not come from economic losers of globalization. Finally, both parties capitalize on country-specific issues —national unity in Spain and welfare in Portugal— but PRR parties might struggle to establish themselves within the party system of the two Iberian countries.

KEYWORDS:

• VOX
• Chega
• radical right
• exceptionalism
• Iberia

Introduction

Between 2018 and 2019, Iberian exceptionalism came to an end: VOX gained 12 seats in the regional parliament of Andalusia and subsequently became the third most voted party at the national level, while Chega was the first radical right party to obtain a seat in the Portuguese parliament and its leader came third at the 2021 presidential elections. For the first time since the end of the authoritarian regimes of Francisco Franco and António Salazar, populist radical right (PRR) parties obtained representation in the political systems of the two Iberian countries. The electoral success of PRR parties is common in Europe, but since their return to democracy in the mid-1970s Spain and Portugal were two countries considered ‘immune’. With the end of Iberian exceptionalism, it is time to analyze the demand-side factors driving the vote for VOX and Chega to understand whether their electorate presents the characteristics described in the relevant literature and to what extent the electoral breakthrough of PRR parties in formerly negative cases is linked to peculiar factors.

No country is immune to nativism, authoritarianism, and populism —the key ideological elements of PRR parties (Mudde 2019) – hence, the end of Iberian exceptionalism should not come as a complete surprise. However, Portugal and Spain represent particularly interesting cases for the study of PRR parties for at least three reasons. First, they both expressed a very strong resilience against PRR parties (Alonso and Rovira Kaltwasser 2015; da Silva and Salgado 2018), a fact that has only a few parallels in Europe.¹ Second, the end of this exceptional situation followed an almost identical timing, with VOX gaining representation for the first time in 2018 in the regional elections of Andalusia and confirming its performance at the national level in the following April, and Chega electing its leader André Ventura in parliament in October 2019, later coming third in presidential elections in January 2021.² Third, Spain and Portugal are two comparable countries because they followed similar political trajectories; they have a similar political and electoral system; and they both suffered the consequences of the Great Recession and the subsequent socio-economic crisis.

Empirically, we use new online survey data from the 2019 national elections in both countries to determine how different sets of variables affect respondents’ self-reported probabilities to vote (PTV) as well as vote choice for VOX and Chega. We demonstrate that on the one hand Chega and VOX attract supporters with expected characteristics: young, religious men with low education from (in the case of Chega) rural areas. Further, political dissatisfaction and a media diet based on Facebook and tabloids predict the vote for the two PRR parties. But while VOX and Chega voters are clearly opposed to cultural openness, and show strong anti-immigration, anti-globalization, and (in the case of VOX) anti-feminism feelings, they cannot necessarily be characterized as economic globalization losers. Another element that makes the supporter of VOX and Chega stand out is their position on very salient issues in the respective countries — territorial unity in Spain, and welfare in Portugal.

We study the type of electorate of VOX and Chega to assess if it corresponds to the kind of PRR voters usually identified in the literature. Our aim is to understand whether the electorate of VOX and Chega is comparable to the electorate of PRR parties across Europe, and to what extent the end of the Iberian exceptionalism relies on specific elements.³ For this purpose, we structure the article as follows: first, we describe the two PRR parties and their background, to then review the relevant literature and develop our hypotheses on which factors could explain the vote for both Chega and VOX. Next, we describe the data we use and proceed to the empirical analysis, concluding with the implications and limitations of this study.

VOX and Chega

This section sketches a brief history of VOX and Chega, their leaders, their ideological characteristics, and their position on particularly relevant issues, arguing that these two parties are comparable. First, both Chega and VOX show the three characteristics common to every PRR party (Mudde 2007): the use of a populist rhetoric, separating society in a Manichean way between the pure people and corrupt elites; the presence of nativism, promising to protect native born inhabitants against migrants; and authoritarianism, advocating for a strictly ordered society and therefore placing strong emphasis on the importance of law and order. Both parties show these characteristics and therefore they have been classified as belonging to the family of the populist radical right (Rooduijn et al. 2020).⁴

Concerning the use of a populist discourse, both Chega and VOX articulate a classic ‘us vs. them’ rhetoric that separates Portuguese and Spaniards from alien groups such as migrants and mainstream political elites while challenging basic features of liberal democracy such as social diversity and respect for minority rights. At the same time, both parties claim to represent the ‘voice of the people’ in opposition to the political correctness of mainstream politics (Mendes and Dennison 2020). The two parties present a slightly different type of nativism, because Roma communities are Chega’s main target while VOX is against migrants in general and Muslim ones in particular.⁵ Both Chega and VOX focus on ‘crime and security’ as well as ‘law and order’, proposing an authoritarian vision of society (Ferreira 2019), which insists on traditional gender norms and opposes feminism (Bernardez-Rodal, Requeijo Rey and Franco 2020; Mendes and Dennison, 2020). In particular, VOX compares feminism to a violent ideology —‘the feminist jihad’— where feminists —‘Feminazis’— are part of a ‘progressive dictatorship’ (Ribera Payá and Díaz Martínez 2020). Similarly, Chega considers abortion as tyranny, and claims to fight against the ‘dictatorship of gender ideology’ (Chega 2019).⁶

The leaders of VOX and Chega followed similar political trajectories as well: they both started within the mainstream right but eventually decided to form a more radical party. André Ventura has enjoyed a significant visibility as football commentator for a national TV channel between 2014 and 2020, and he was a member of the mainstream —and despite the name right-wing— Social Democratic Party (Partido Social Democrata – PSD). Ventura resigned in October 2018 and created Chega in April 2019, becoming only a few months later Member of Parliament for the Lisbon constituency. Similarly, Santiago Abascal was city councillor and member of the Basque Parliament representing the People’s Party (Partido Popular – PP), until he grew disillusioned with the PP and created VOX in December 2013. The fact that both leaders have a background in mainstream right-wing parties might have helped them to receive greater visibility and a less stigmatized image from the media (Mendes and Dennison 2020). In conclusion, VOX and Chega are two comparable PRR parties that experienced an electoral breakthrough in two countries traditionally considered ‘immune’, and the next section advances several hypotheses concerning their electorate.

Hypotheses to explain the vote for VOX and Chega

In order to identify the demand-side factors linked to the support for VOX and Chega, we formulate several hypotheses based on the literature on PRR parties. We recognize that the electorate supporting PRR parties, the demand-side, interacts with supply-side factors, including how these parties act within the political opportunity structure and the political space left open by the competitors (Eatwell 2003; Golder 2016). This exploratory study, however, does not try to provide an all-encompassing picture of the political opportunity structure available to VOX and Chega or a global account of their electoral performance. We focus on the potential electorate of VOX and Chega to understand to what extent it is comparable to the electorate of similar parties across Europe, and to what extent it shows peculiar elements.

To formulate our hypotheses, we rely on a strand of literature that considers the demand for PRR parties as a reaction to phenomena such as modernization (Betz 1994) and globalization (Kriesi et al. 2008). Citizens who feel ‘left behind’, or who believe that their social status is threatened (Elchardus and Spruyt 2016; Spruyt, Keppens and Van Droogenbroeck 2016; Gidron and Hall 2020), are often found to be among the electorate of PRR parties because these parties intercept their ‘grievances’ (Rydgren 2007). As a result of deindustrialization a large number of less qualified jobs in manufacturing were lost and the manual working class increasingly constitutes the core constituency of the radical right, indicating a process ‘proletarianization’ of the radical right’s support base (Ignazi 2003; De Lange 2007; Minkenberg and Perrineau 2007). In particular, the relative decline in the social hierarchy has been found to make (male) routine workers susceptible to the nativist platforms of PRR parties (Kurer 2020). Furthermore, voting for PRR parties is considered as an expression of distrust towards the political elite (Fieschi and Heywood 2004), dissatisfaction with the way democracy works (Betz 2002), and a rise in public disenchantment with traditional parties (Rooduijn, Van der Brug and de Lange 2016).

In addition, we consider a complementary strand of literature on ‘mediatization of politics’ (Esser 2013): the media-logic and the political-logic are converging in providing emotional and conflictive stories, and political actors are increasingly aware of how to ‘use’ the media to gain visibility, a process called ‘self-mediatization’ (Strömbäck 2008). In this scenario, supporters of PRR parties have a particularly low trust in mainstream media (Fawzi 2019), and for this prefer online news (Tsfati 2010). The use of social media can influence the electoral performance of populist parties for two reasons: first, social media are especially suited for the communication of populist radical right messages because they allow bypassing the intervention of elite gatekeepers (Engesser et al. 2017), and second, populist parties employ populist content and style more frequently on Facebook and Twitter than in political talk shows (Ernst, Blassnig, et al. 2019). Political information is a key aspect of voting behaviour and mobilization (Brady, Verba and Schlozman 1995), and knowing that political misinformation relates positively to support for right-wing populist parties (van Kessel, Sajuria and van Hauwaert 2020), we consider this aspect as very relevant to understand the electorate of PRR parties in a comparative perspective.

Combining these strands of literature and relying on existing empirical studies testing these theories, we derive our hypotheses on the socio-demographic characteristics of citizens supporting PRR parties and their levels of political dissatisfaction, investigating to what extent they can be characterized as economic and cultural ‘losers of globalization’, as well as their media diet. Finally, we formulate a hypothesis about the impact of country-specific issues because the support for PRR parties partially depends on contextual factors (Muis and Immerzeel 2017a) and on the levels of politicization of particularly salient topics in the public debate (Grande, Schwarzböz and Fatke 2019).

First, we focus on the sociodemographic profile of traditional supporters for PRR parties. The vast literature on the topic suggests that, despite the impossibility to define every characteristic of PRR voters (Rooduijn 2018), several elements are consistently present while others depend on context (Muis and Immerzeel 2017a). Voters for PRR parties tend to be men more often than women (Harteveld et al. 2015; Mudde 2019)⁷, young (Tillman 2021; Zagórski, Rama and Cordero 2021)⁸, religious (Arzheimer and Carter 2009; Zúquete 2017)⁹, and living in rural rather than urban areas (De Lange and Rooduijn 2015; Rodríguez-Pose 2018).¹⁰ Lastly, PRR voters also tend to have a low education (Ivarsflaten and Stubager 2013; van Hauwaert and van Kessel 2018), which is also in line with the arguments of Kriesi et al. (2008).

H1: Young, religious men with low education and living in rural areas are more prone to cast a vote for VOX and Chega.

Moreover, political dissatisfaction and democratic malaise can be explained by a repulsion toward every established political party, thus producing a radical rejection of mainstream politics and a disenchantment with political parties (Mair 1998) that can translate in voting for anti-system and protest parties. In particular, political dissatisfaction is known to increase the probability to cast a vote for extreme right parties (Lubbers, Gijsberts and Scheepers 2002). Voters of right-wing populist parties are normally among the most dissatisfied with how democracy works (Bowler et al. 2017), and disenchantment with existing political parties and the way in which democracy works are factors commonly associated with the success of PRR parties (Betz 2002). The erosion of democratic legitimacy can lead voters to cast their preference for PRR parties in order to express their distrust towards the political elite (Fieschi and Heywood 2004), a behaviour in line with Cas Mudde’s definition of populism as an ideology separating ‘the pure people’ from ‘the corrupt elite’ (2004). Moreover, voters prefer a new party if they consider all other parties to be corrupt (Pop-Eleches 2010), and new parties are more successful when they politicize corruption (Bågenholm and Charron 2014).

H2: Those who are more politically dissatisfied are more prone to cast a vote for VOX and Chega.

Globalization introduced a new integration-demarcation cleavage that does not necessarily follow the classical cleavage divisions based on class or religion (Lipset and Rokkan 1967), pitting those who see their opportunities in life diminished as a result of globalization against those who profit from it (Kriesi et al. 2008; Bornschier 2010). PRR parties are considered more likely to attract those voters that have been described as ‘losers of globalization’ such as manufacturing and low skilled workers (Helbling and Jungkunz 2020) given their position of vulnerability in a complex and global labour market. These voters feel marginalized and impoverished: in turn, this creates a type of resentment that left-wing parties traditionally channelled, but that has been increasingly exploited by PRR parties (Kitschelt and McGann 1995). In sum, those voters that believe that the national economy is performing poorly tend to vote for PRR parties (Mols and Jetten 2017), as well as those that suffer economic hardship in a favourable context at the national level (Rooduijn and Burgoon 2018). These factors make us expect that —in addition to those with low education already mentioned in H1— people who have a negative perception of their economic situation tend to vote for Chega and VOX.

H3a: Economic ‘losers of globalization’ are more prone to cast a vote for VOX and Chega.

Beyond an economic motivation, the ‘losers of globalization’ oppose social change like multicultural cosmopolitanism because they perceive globalization and the opening of national borders as a threat (Teney, Lacewell and De Wilde 2014). This is a cultural conflict based on values (Inglehart 1977), where a group of people feels unfairly treated by society (Elchardus and Spruyt 2016) and opposes post-materialistic values such as immigration, multiculturalism and feminism. Indeed, anti-immigration stances are a common feature of PRR parties (Mudde and Rovira Kaltwasser 2018), and opposing immigration is a relevant predictor of the vote for the extreme right (Lubbers, Gijsberts and Scheepers 2002). Delegitimization and ridicule of feminism are also core elements in the PRR ideology (Sauer 2020): insisting on traditional gender values allows PRR parties to oppose ‘progressive elites’ and to identify an alleged enemy threatening traditional values. Tellingly, anti-feminism is often combined with nationalism and, in particular, anti-Muslim gender nationalism (Bernardez-Rodal, Requeijo Rey and Franco 2020).

H3b: Cultural ‘losers of globalization’ are more prone to cast a vote for VOX and Chega.

Next, we focus on the media diet of those who are more likely vote for PRR parties. The growing literature connecting media studies and populism shows that citizens with populist attitudes read more tabloids and consume more entertainment-based media (Hameleers, Bos and de Vreese 2017). Furthermore, citizens with populist attitudes rely more on alternative media and social media to procure political information (Müller and Schulz 2019). In particular, it seems that Facebook is extremely suitable for the propagation of radical right populist messages (Zulianello, Albertini and Ceccobelli 2018), significantly more than its competitor Twitter (Ernst et al. 2017). Therefore, it seems plausible to argue that those who vote for PRR parties also tend to have an ‘unhealthy’ media diet particularly focused on social media (in particular Facebook) and tabloids compared to voters of traditional or mainstream parties.

H4: Those who rely on tabloids and social media to gather political information are more prone to cast a vote for VOX and Chega.

Given that voter profiles are context dependent (Burbank 1997) we also analyze country-specific issues. When political parties converge towards the centre of the political spectrum, this provides expanding political opportunity structures for new right-wing populist parties (Ignazi 1992; Kitschelt and McGann 1995). Moreover, far right parties are expected to do well when they are seen to own the issues they promote and when these issues are salient (Golder 2016). When PRR parties like VOX and Chega politicize very salient topics, this can translate into an electoral boost because it allows them to capitalize on values and feelings that otherwise would remain without representation. For these reasons, when it comes to particularly significant topics or issues that are central in the public debate, we expect VOX and Chega to benefit from positioning themselves in a way that clearly distinguishes them from the other parties.

In Spain, the centre-periphery cleavage has always been relevant (Morlino 1998) and, after having been ‘frozen’ during the Franco era, it re-exploded in recent years, culminating in the Catalan Independence Referendum and the consequent constitutional crisis. As a result, the issue of territorial unity was highly salient in the 2019 elections (Mendes and Dennison 2020). Crucially, VOX’s core values include Spanish nationalism, territorial unity and the refusal of regional autonomies, systematic repression of pro-independence movements (Turnbull-Dugarte, Rama and Santana 2020), calling for the suspension of the autonomous independence of the Catalan region and the constitutional prohibition of any party that seeks separatist objectives (Turnbull-Dugarte 2019). Recent studies found that criticizing the PP for being too tolerant towards regional nationalisms (Vampa 2020) VOX was able to gain votes (Rodríguez-Teruel 2021).

H5a: Those who are in favor of national unity and against regional separatism are more likely to vote for VOX.

In Portugal, political competition still revolves largely around socioeconomic issues (Ferreira da Silva and Mendes 2019). The most salient issue is the system of welfare provisions because the Great Recession did not create specific concerns about migration or terrorism like in other European countries, but rather concerns about unemployment, the economic situation, pensions, and public debt (Gómez-Reino and Plaza-Colodro 2018). Indeed, the 2019 election largely centred around socio-economic issues, more specifically, around health and social security (Fernandes and Magalhães 2020). This does not mean that a PRR party like Chega cannot thrive. In fact, Chega is the only party openly proposing to privatize the health system and to restrict welfare measures for Portuguese people (Chega 2019). This position clearly separates Chega from the rest of its competitors, and it might increase its electoral success.

H5b: Those who are in favor of liberalizing and privatizing welfare while reducing public spending are more likely to cast a vote for Chega.

Data

To test our hypotheses, we make use of online survey data collected in 2019, when both Spain (April 28) and Portugal (October 6) held legislative elections. The survey is a representative two-wave panel online survey (one pre – and one post-electoral wave) with a sample of ∼3000 respondents in Spain and ∼1500 respondents in Portugal and fulfils a crossed quota of gender (2 categories), age (3 categories) and education (3 categories).¹¹ While nonprobability online surveys are less established than probabilistic face-to-face surveys and tend to differ in their marginal distributions, they have been shown to yield very reliable results especially when it comes to causal inferences and explanatory models such as vote choice (Dassonneville et al. 2020), which is what we do in this paper.

Following Turnbull-Dugarte et al. (2020), we use two dependent variables: self-reported probability to vote (PTV) and vote choice (recall) for VOX and Chega. The PTV item asks respondents to indicate the probability that they would vote for a party on a scale from 0 (definitely would not vote for this party) to 10 (definitely would vote for this party), thus offering a more long-term perspective into the potential electorate of both parties. Figure 1 shows the distribution of this variable for the two PRR parties we analyze. We can see that Chega is less known amongst respondents than VOX, while VOX has a higher percentage of respondents indicating that they would definitely not vote for this party. Given that Chega is a very new party that was only founded a few months before the 2019 elections, this is an expectable pattern.

Figure 1. Probability to vote (PTV) for VOX and Chega in 2019 (authors own online survey, wave 2).

Indeed, we decided to use the second wave of the online survey that was fielded right after the elections, as the prominence of Chega —and hence the validity and comparability of the PTV measure— is higher than before the elections. For the analysis, we recoded PTV to a 0–1 scale. Compared to actual vote choice, PTV has the advantage that it allows us to analyze a larger N, given that we have data from the whole sample.

Second, and to make sure our results are valid, we replicate the analysis using vote recall (equally from the second wave) as a dependent variable (0 – voted for other parties, 1 – voted for Chega/VOX).¹² This variable captures short-term, election-specific support for both parties (Turnbull-Dugarte, Rama and Santana 2020). Especially in the case of Chega, this variable is relying on a rather small set of voters: in the post-electoral survey, we only have 46 respondents who indicated that they had voted for this party, out of 987 who casted a valid vote. In the case of VOX, we have 257 (out of 2243) respondents having voted for them. However, as the next section will show, results are very comparable between PTV and vote recall, with some effects (expectedly) losing significance for Chega when using vote recall, most likely due to the small sample.

Our independent variables are grouped in five categories, and to ease comparison between effect sizes, we equally recode all of them to a scale ranging from 0 to 1 in order to make effect sizes comparable. To test the first hypothesis about sociodemographic factors, we use gender, age (in years, from old to young), education (8 categories, from high to low), religiosity (4 categories), and the area where respondents live (urban to rural, 4 categories). To test our second hypothesis, we use different variables that capture political dissatisfaction: respondents dissatisfaction with the performance of the current government (PS in Portugal, and PSOE in Spain), and for Spain their dissatisfaction with the performance of the previous PP government, given that the government changed only four month prior to the elections; dissatisfaction with democracy (all on a 5-point scale); and the perception of corruption amongst political parties (11-point scale). We also control for ideology (left-right-placement, 11-point scale) in this model. To test the third hypothesis about ‘losers of globalization’, we use two sets of variables: for the economic dimension, respondents’ household income (5-point scale, high to low); their ownership of assets (4 categories; owning a residence, a business/property/land, stocks, or savings; reversed to indicate no assets); and their assessment of the national economic situation as well their personal economic situation (5-point scales). For the cultural dimension we use respondents’ opinion on globalization being bad for their country; their position on immigration (support for a more restrictive immigration policy) and on feminism (the right to abortion in Portugal, acceptability of violence against women in Spain), all 5-point scales.

To test the fourth hypothesis, we use different variables that capture respondents’ media diet: how often they use different online tools to obtain political information (forums, Facebook, and Twitter, all on a 4-point scale); as well as their preferred newspaper, where we code tabloids as 1 (Correio da Manhã in Portugal, ABC in Spain) and other newspapers as 0. To test our fifth hypothesis, we use respondents’ positions on country-specific political issues. For Spain, we use respondents’ positions on Catalan independence, which we measure with two variables: the rejection of autonomy for Catalonia (‘Catalonia has reached too much autonomy’, dichotomous variable) and the support for a centralized state (5-point scale). For Portugal, we use two questions on welfare, asking respondents if they agree that public services should be cut in order to reduce taxation, and if they want the national health system to be privatized, both on an 11-point scale. For descriptive statistics of all variables, see table A4 in the online appendix.

Analysis

Using the data described in the previous section, we test our hypotheses by running simple regression models in the two countries to determine how different sets of variables affect respondents’ self-reported probabilities to vote (PTV) as well as their actual vote (vote recall) for VOX and Chega. Given the relatively small sample size (especially in the case of Portugal), and as we are using a rather large set of independent variables, we decided for a step-wise approach, by running separate models for each hypothesis and each set of predictors to avoid multicollinearity and an over-fit model. All models, however, include the initial set of sociodemographic variables (gender, age, education, rural-urban, religion) as controls. Moreover, full models including all independent variables at once are available in the online appendix (Table A5).

The following figures plot the resulting coefficients (i.e. effect sizes) and 95% confidence intervals for the two parties and for both dependent variables. Note that the models with PTV as a dependent variable are linear regressions, and report regression coefficients. The models with vote recall as a dependent variable are logistic regressions, and coefficients are reported in odds ratios to allow a substantial interpretation. The full models can be found in Table A1 (PTV) and Table A2 (vote recall) in the appendix.

Figure 2 shows how sociodemographic factors influence PTV and vote: we can see that in accordance with our first hypothesis, young men with lower education are more likely to sympathize with and vote for both VOX and Chega.¹³ Moreover, VOX and Chega sympathizers and voters tend to be above-average religious, and Chega supporters are more likely to live in rural rather than in urban areas. For Spain, these findings are mostly in line with a recent study by Turnbull-Dugarte et al. (2020), which finds VOX voters to be male, younger, and more religious than average. Contrary to their findings, however, we cannot confirm that VOX supporters are more urban and more educated. The positive effect of low education in both countries is in line findings from Kriesi et al. (2008), and points to the importance of educational differences in triggering ‘cultural voting’ for the radical right (Bornschier 2018).

Figure 2. Sociodemographic factors (H1).

Linear regression coefficients (PTV) and odds ratios from logistic regression (vote recall). Full models in Table A1 and A2 in the Appendix. Data: authors own online survey, wave 2.

Figure 3 shows the effects of political dissatisfaction. As the following figures, these models include all previously discussed sociodemographic variables as controls. The results mostly confirm our second hypothesis, which assumes that political dissatisfaction increases the likelihood to support VOX and Chega. Indeed, negatively evaluating the performance of the current government (PSOE in Spain and PS in Portugal, both centre-left) strongly increases the likelihood to vote for both VOX and Chega, although for Chega this effect loses significance in the vote recall model, and the effect is generally stronger for VOX. Interestingly though, negative evaluations of the previous PP government actually decrease the PTV for VOX and do not affect vote choice, implying that VOX supporters were rather satisfied with the performance of the PP when it was in power and suggesting that many former PP voters are now sympathizing with —or even voting for— VOX. Dissatisfaction with democracy, as expected, increases the PTV and vote for both VOX and Chega. Despite both parties regularly insisting on the corruption of the political system, the perception that corruption amongst political parties is widespread is not significant in its effect on PTV, and only slightly significant for the vote recall for VOX.¹⁴ When it comes to ideology, as expected, being right leaning is the strongest predictor of support and vote for both parties. In the case of Spain, these results fully support previous findings by Turnbull-Dugarte et al. (2020) as well as Rama et al. (2021), pointing to a strong effect of political dissatisfaction, especially dissatisfaction with democracy, on support for VOX.

Figure 3. Political dissatisfaction (H2).

Linear regression coefficients (PTV) and odds ratios from logistic regression (vote recall). All models additionally control for gender, age, education, religion and rural vs. urban. Full models in Table A1 and A2 in the Appendix. Data: authors own online survey, wave 2.

Figure 4 shows our third hypothesis which claims that ‘losers of globalization’ are more likely to cast a vote for VOX and Chega. We test two dimensions: economic as well as cultural ‘losers of globalization’. This model, as all others, controls for socio-economic factors, most importantly education, which Kriesi et al. (2008) found to be a key variable identifying the losers of globalization. In the economic dimension, the results cannot fully confirm this hypothesis: a lower household income and owning no assets actually decrease the PTV, and (to a lesser degree) the vote choice for VOX, while the PTV for Chega is not significantly determined by income or assets. The vote for Chega, however, is actually positively affected by not owning any assets. Both Chega and VOX voters are, on the other hand, strongly convinced that the national economic situation got worse in the past year, but they do not think that their personal economic situation got worse.

Figure 4. Globalization losers (H3).

Linear regression coefficients (PTV) and odds ratios from logistic regression (vote recall). All models additionally control for gender, age, education, religion and rural vs. urban. Full models in Table A1 and A2 in the Appendix. Data: authors own online survey, wave 2.

In the cultural dimension, the results are clearer: anti-globalization and anti-immigration feelings have a strong effect. Moreover, opposition to feminism also increases the likelihood to vote for VOX, but not significantly for Chega. In conclusion, VOX and Chega voters are not necessarily clear economic globalization losers themselves, but they are critical of the economic situation, and strongly feel as cultural losers of globalization. A positive relationship between a high economic status and support for VOX is in line with the results of by Turnbull-Dugarte et al. (2020). However, contrary to other studies (Ortiz Barquero 2019; Turnbull-Dugarte 2019) we find a significant relationship between opposition to immigration and the vote for VOX. Importantly, those results hold while controlling for the (positive) effect of low education, which, as previously discussed, is an important factor in explaining economic and cultural attitudes of demarcation (Kriesi et al. 2008).

Figure 5 displays the effects of ‘media diet’ on respondents’ voting behaviour, and mostly confirms our fourth hypothesis: using Facebook and (in the case of VOX) internet forums as a source of political information increases the probabilities to vote for a PRR party, while, at least for VOX, using Twitter to obtain political information decreases this probability. Moreover, reading tabloids rather than other newspapers also significantly increases the likelihood to support and vote for both Chega and VOX (however becoming non-significant for Chega’s vote recall). This again confirms the findings of Turnbull-Dugarte et al. (2020) about VOX voters being more likely to follow the election campaign on social media.

Figure 5. Media diet (H4).

Linear regression coefficients (PTV) and odds ratios from logistic regression (vote recall). All models additionally control for gender, age, education, religion and rural vs. urban. Full models in Table A1 and A2 in the Appendix. Data: authors own online survey, wave 2.

Finally, Figure 6 shows the results for country-specific issues that we expect to drive the vote. In accordance with our last hypotheses and previous findings from Turnbull-Dugarte et al. (2020) country-specific issues significantly affect the vote likelihood for the two PRR parties: in Spain, rejecting Catalan autonomy and supporting a centralized state strongly increases the likelihood to vote for VOX. And similarly, support for a privatized health system and for a reduction in public services increases the likelihood to vote for Chega in Portugal —although to a lesser extent than the Catalan issue in Spain— confirming both H5a and H5b.

Figure 6. Country-specific issues (H5).

Linear regression coefficients (PTV) and odds ratios from logistic regression (vote recall). All models additionally control for gender, age, education, religion and rural vs. urban. Full models in Table A1 and A2 in the Appendix. Data: authors own online survey, wave 2.

Overall, the results are very consistent using both vote recall and PTV as a dependent variable. The models for Spain replicate almost perfectly the results using PTV, while the results for Portugal are in some cases slightly weaker in their significance than the PTV models. This was to be expected, considering the very small sample of Chega voters. However, the fact that even with the small Portuguese sample all effects still hold their direction, and most remain significant, shows that our findings are robust. Robustness tests using all independent variables at once (Table A5 in the online appendix) show equally consistent results, with directional effects remaining the same despite sometimes lower levels of significance due to the relatively small N. Moreover, we can confirm most of the findings from previous work on the determinants of support for VOX (Turnbull-Dugarte 2019; Turnbull-Dugarte, Rama and Santana 2020), while adding new findings and being able to compare, for the first time, VOX voters to the Chega electorate.

Conclusions

Since their transition to democracy, the electorates of Spain and Portugal seemed to reject PRR parties, reminiscent of long-lasting right-wing dictatorships under Francisco Franco and António Salazar. Between 2018 and 2019 the so-called ‘Iberian exceptionalism’ ended with VOX and Chega breaking through the Iberian political system and getting their representatives elected. This marked a watershed moment in Portuguese and Spanish democratic history, generating a series of questions about what made this unprecedented electoral breakthrough possible. A well-established strand of literature has been studying the electorate of PRR parties, and our aim was to establish whether the support for VOX and Chega follows a similar pattern. Using online survey data from the 2019 legislative elections in Spain and Portugal, we run country-wise regression models to determine how different sets of variables affect respondents’ vote choice as well as probabilities to vote (PTV) for VOX and Chega.

The results of this analysis indicate that on the one hand Chega and VOX attract voters with expected characteristics, while on the other hand their support also shows some peculiar traits. Concerning socio-demographic elements, those more likely to vote for VOX and Chega follow a pattern in line with previous studies on PRR parties: they are young, highly religious men with low education, and in the case of Chega, they live in rural areas. The fact that PRR parties in Spain and Portugal attract young voters might have a long-term impact and suggests that VOX and Chega do not simply attract old-style voters, nostalgic of the authoritarian regimes of Franco and Salazar, but a more modern electorate.

Political dissatisfaction is, as expected, a strong predictor of support for the two PRR parties, although corruption perception is not a driving reason.¹⁵ Also, VOX voters interestingly are only dissatisfied with their current PSOE government, but not with the previous PP administration. This might indicate that many of those who used to vote for the PP switched to VOX or are at least considering it. In addition, their media diet is in line with the expectations generated by the relevant literature: those more likely to vote for Chega and VOX use Facebook and tabloids as sources of political information. Additionally, VOX supporters are more likely to use online forums as a source of political information, but less likely to use Twitter. Finally, in line with other PRR parties, VOX and Chega voters can also be qualified as cultural globalization losers: they are driven by anti-globalization, and anti-immigration feelings, and —although only in in the case of VOX and not Chega— reject feminism.¹⁶

Interestingly, however, we also found that support for VOX and Chega does not follow the script in every regard, thus suggesting that the end of Iberian exceptionalism might be partially linked to country-specific reasons and not only to a generic wave of support for PRR parties that is characterizing most countries in Europe and across the world (Mudde 2019). For example, it is impossible to classify those more likely to vote for Chega and VOX as obvious economic ‘losers of globalization’: while they do believe that the general economic situation is worsening, their personal economic situation is not affected. To the contrary, those supporting VOX cannot be classified as poor by any means; in fact, their income and assets are clearly above average, supporting findings by Turnbull-Dugarte et al. (2020). While Chega voters actually do show below-average assets, but still seem unaffected by personal economic circumstances. In sum, our evidence does not clearly support the hypothesis that economic hardship drives the vote for PRR parties, and rather points to the cultural dimension of globalization skepticism, as well as to the strong influence of education. This is in line with recent studies claiming that middle-income and high-status groups also vote for PRR parties because anxiety about losing subjective social status proves to be more important than actual decline (Engler and Weisstanner 2020), and pointing to the important role of education in PRR support (Bornschier 2018). Overall, we do not find empirical support for the idea that the radical right attracts economic globalization losers, while the cultural dimension of globalization is definitely linked to the support for PRR parties. Whether this is specific to the two Iberian countries, or a general European trend, should be the object of future research.

Another interesting characteristic about the support for VOX and Chega is their position on very salient issues in the respective countries, confirming that voter profiles are context dependent (Burbank 1997) and that the support for PRR parties partially depends on contextual factors (Muis and Immerzeel 2017b). Being against Catalonia’s independence and in favour of the centralization of the state excellently predicts support for VOX. Since this was the most salient issue at the time of the 2019 elections, it is possible to argue that it can explain not only the end of ‘Spanish exceptionalism’ but also the much bigger success enjoyed by VOX compared to Chega. In Portugal, the most salient issue lies on the socio-economic axis of competition: public services and the welfare system. While we show that support for Chega is indeed linked to a preference of privatization over public services, one could argue that welfare is comparatively less salient than the Catalan issue and does not polarize the public debate and the party landscape to the same degree. Together with the fact that VOX, but not Chega, capitalizes on strong anti-feminism feelings, this could explain the sharp differences in the electoral success of the two parties.

This study has several limitations that need to be highlighted. First, there are factors that would deserve to be considered in order to offer a more comprehensive explanation, but that go beyond the scope of this study. For example, future research should consider the role of stigmatization of the authoritarian past and its impact on the electoral breakthrough of Chega and VOX. Interestingly, in both Spain and Portugal there is a strong evidence of the so-called antidictator bias, which brings people as well as parties to self-identify as left-wing because of the stigma associated with the past authoritarian regime (Dinas and Northmore-Ball 2020). For this reason, right-wing parties in Portugal and Spain have been constantly treated by their electorates as the most right-wing in Europe, despite these parties holding relatively moderate positions according to their manifestos (Dinas 2017). Future studies should investigate whether the stigmatization of the authoritarian past is fading in both countries with a similar timing, opening up opportunity structures for the success of PRR parties (Manucci 2020).

Moreover, we focused on demand-side factors and did not analyze supply-side factors. Future studies might therefore explore the role of the media in legitimizing PRR parties in Spain and Portugal, which is certainly a relevant opportunity structure for PRR parties (Ernst, Esser, et al. 2019). Another direction for future research is the role of charismatic leadership, given that populists are more likely than other candidates to exhibit traits that are associated with it (Nai and Martínez i Coma 2019). We also lacked items to measure nationalism as well as populist attitudes, which recent studies confirm being present in both Spain and Portugal (Marcos-Marne, Plaza-Colodro and Freyburg 2020; Santana-Pereira and Cancela 2021).

In sum, our findings imply that VOX’s and Chega’s electorates follow patterns that are largely in line with what the literature on PRR parties predicts, but with two important caveats concerning the necessity to distinguish between economic and cultural ‘losers of globalization’, and the relevance of salient topics in the public debate that can transfer votes from the mainstream to the radical right. These results contribute to the literature on populist and radical right parties beyond the Iberian countries in two ways. First, our results indirectly illuminate the different electoral performance of the two parties suggesting that the saliency of different topics in the public debate matter: it is true that VOX exists since 2013 and had a more established voter base by 2019 than Chega, but crucially Chega, contrary to VOX, could not (yet?) exploit a very salient and polarized issue like the national unity in Spain, and thus remains electorally weaker. Second, our results show that PRR parties in Spain and Portugal do not attract economic losers of globalization, suggesting that the success of PRR parties depends on their ability to connect voters with cultural grievances over immigration to the group of voters with economic grievances over immigration (Halikiopoulou and Vlandas 2020). The Iberian exceptionalism came to an end, but PRR parties might fail to establish themselves after their initial breakthrough for very ‘Iberian’ reasons. Those who voted VOX might choose to support the mainstream PP in case it takes a more decisive position about Catalonia, and Chega will thrive only if the public debate will revolve around highly salient cultural rather than socio-economic issues as in the rest of the continent.

Acknowledgements

This article results from data collected and research undertaken in the MAPLE project (ERC-Consolidator Grant, 682125) as well as the POPULUS project (financed by the FCT, PDTC/SOC-OC/28524/2017). We are grateful to the participants of the SPARC seminar at the ICS-ULisboa for their valuable feedback on this project, especially to Marina Costa Lobo, Pedro Magelhães, Enrico Borghetto, Mariana Mendes, Stuart Turnbull-Dugarte and Steven Van Hauwaert. We are also grateful to the three anonymous reviewers for their helpful comments.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Fundação para a Ciência e a Tecnologia [PDTC/SOC-OC/28524/2017]; European Research Council [ERC-Consolidator Grant, 682125].

Notes

1 Ireland, Luxembourg, and Malta might be the only remaining countries in Europe that did not experience the electoral breakthrough of PRR parties.

2 The electoral success of VOX might have constituted an advantage for Chega’s breakthrough, whose voters received signals of electoral viability from a party that is ideologically very close. This is in line with recent studies that propose a model based on the transnational and interdependent diffusion of far right parties and voters (Kuyper and Moffit 2019; Van Hauwaert 2019). However, the direct contacts between the two parties have been unsystematic, also because they belong to different European groups, and until now (September 2021) the two leaders never had an official meeting.

3 The success of any political party is linked, at least to some extent, to specific or peculiar elements. At the same time, as it will be clarified below, theories and empirical studies allow to identify demand-side explanations for the success of PRR parties that are generalizable.

4 Other datasets based on expert surveys (Meijers and Zaslove 2020; Norris 2020) consider VOX as a PRR party, but do not have data available for Chega yet, which is however described as populist and extreme right by Fernandes and Magalhães (2020). The only book entirely devoted to VOX (Rama et al. 2021) considers it a PRR party, and a recent study concludes that populism in VOX is present although subordinated to nationalist and traditionalist elements (Marcos-Marne, Plaza-Colodro and O’Flynn 2021). While a binary classification of parties ignores possible nuances, we are confident that VOX and Chega can be defined as PRR parties.

5 A video from 2018 in which VOX warned against the potential dangers of Islam caused much controversy, conjuring a future in which Muslims had imposed sharia in southern Spain, turning the Cathedral of Córdoba back into a mosque and forcing women to cover up. During the Covid-19 pandemic, Chega’s leader Ventura took the opportunity to define Roma people as a threat to public health, proposing a plan based on confinement and surveillance for Roma communities.

6 At the 2020 Chega convention, the party leader André Ventura tried to pass a motion calling for the removal of ovaries from women who have abortions.

7 Men vote for PRR parties more than women because of gender differences regarding socio-economic position and lower perceptions regarding the threat of immigrants, as well as differences in socialization and populist attitudes (Spierings and Zaslove 2017; Harteveld and Ivarsflaten 2018).

8 Studies indicate both old (Lubbers, Gijsberts and Scheepers 2002) and young people (Arzheimer 2012) as those more likely to vote for PRR parties, but also the youngest and oldest citizens (Arzheimer and Carter 2006), while a meta-analysis shows the relationship is complex and situational (Stockemer, Lentz and Mayer 2018). We expect younger voters to support VOX and Chega for two reasons. First, younger voters are less likely to have formed party attachments and this increases the likelihood that they will vote for newer parties such as VOX and Chega (Tillman 2021). Second, young people tend to vote for PRR parties when the job market is less promising and more precarious for their cohort (Zagórski, Rama and Cordero 2021) and because young voters are less biased against PRR parties given the fading stigmatization of the authoritarian past (Dinas 2017).

9 We are aware that religiosity does not explain voting for PRR parties in every context (Lubbers, Gijsberts and Scheepers 2002; Norris 2005). However, given the relatively high levels of religiosity in Spain and Portugal, and in line with the existing studies on VOX (Turnbull-Dugarte, Rama and Santana 2020; Rama et al. 2021) we expect VOX and Chega to attract religious voters.

10 Non-urban areas usually show higher unemployment rates, lower education and income levels, and lower shares of immigrants.

11 See table A3 in the online appendix for more information on the survey, and table A6 for a representativeness analysis of the online survey data. Table A6 compares the fit of the survey’s unweighted sociodemographic distribution with population benchmarks and adds the same information for two standard face-to-face surveys, the ESS and the EB, from the same year. Our online survey is significantly more representative than both face-to-face surveys, especially when it comes to age and education.

12 Abstention, null and blank vote are coded as missing in order to compare PRR voters only to those respondents who made a valid vote choice.

13 In addition to the direct effect, we also tested for a curvilinear effect of both age and education but did not find any evidence for either.

14 This low significance might be due to a very low variance on the corruption variable: most respondents tend to believe that parties are corrupt.

15 Given the repeated corruption scandals that plagued the political systems of both countries, and particularly Spain, this finding is surprising. Our results seems to confirm studies that consider the politicization of corruption an element which is not unique to PRR parties (Engler 2020). Moreover, as already explained, the lack of statistical significance might be due to the fact that voters of most parties think corruption is widespread.

16 The difference between VOX and Chega could be linked to two factors: first, as explained in the data section, the questions used in the countries are different (acceptability of violence against women in Spain, and rejecting the right to abortion in Portugal), and second, anti-feminism might be more salient for VOX and less relevant for Chega.

Appendix

Table A1. Explaining the probability to vote for VOX and Chega (full models to figures 2–6, left panel).

DV: PTV	Model 1 VOX	Model 1 Chega	Model 2 VOX	Model 2 Chega	Model 3 VOX	Model 3 Chega	Model 4 VOX	Model 4 Chega	Model 5 VOX	Model 5 Chega
Gender (male)	0.0757***	0.0218	0.0409***	0.0291^+	0.0255*	0.0264	0.0543***	0.0285	0.0604***	0.0256
	(6.88)	(1.26)	(4.41)	(1.65)	(2.13)	(1.41)	(4.89)	(1.64)	(4.73)	(1.27)
Age (young)	0.0650*	0.187***	0.0577*	0.169***	0.0616*	0.174***	0.0994***	0.138**	0.0602^+	0.173***
	(2.28)	(4.53)	(2.37)	(3.92)	(2.02)	(3.73)	(3.39)	(3.21)	(1.76)	(3.56)
Education (low)	0.0925**	0.102*	0.131***	0.0910*	0.0554^+	0.0643	0.0862**	0.0848^+	0.0478	0.121*
	(3.17)	(2.35)	(5.38)	(2.02)	(1.68)	(1.28)	(2.92)	(1.92)	(1.42)	(2.41)
Religious	0.235***	0.126***	0.0505**	0.107***	0.112***	0.119***	0.157***	0.116***	0.249***	0.120***
	(12.82)	(4.34)	(2.98)	(3.46)	(5.55)	(3.58)	(8.20)	(3.93)	(11.82)	(3.56)
Rural area	−0.00473	0.0762***	−0.00410	0.0597*	−0.00743	0.0687**	−0.00346	0.0806***	−0.00109	0.0659*
	(−0.30)	(3.34)	(−0.31)	(2.54)	(−0.45)	(2.77)	(−0.22)	(3.49)	(−0.06)	(2.49)
Ideology (right)			0.444***	0.215***
			(20.35)	(5.93)
Gov. performance bad			0.314***	0.0987*
			(14.46)	(2.12)
Previous gov. perf. bad			−0.110***
			(−4.89)
Dissat. with democracy			0.0565**	0.116**
			(3.16)	(3.23)
Corruption parties			0.0268	0.0160
			(1.07)	(0.45)
Low household income					−0.0496*	−0.0202
					(−2.39)	(−0.54)
No assets					−0.0701**	0.00874
					(−2.72)	(0.22)
Economic sit. got worse					0.234***	0.153**
					(7.79)	(2.80)
Pers. econ. sit. got worse					−0.0262	−0.0152
					(−0.90)	(−0.29)
Anti-globalization					0.0577*	0.109*
					(2.30)	(2.53)
Anti-immigration					0.183***	0.154***
					(8.77)	(4.90)
Anti-feminism					0.238***	0.0303
					(11.41)	(1.09)
Against Catalan autonomy							0.0949***
							(7.16)
More centralized state							0.288***
							(13.89)
Privatize national health								0.116***
								(3.93)
Reduce public services								0.106**
								(3.14)
Political inform.: forums									0.0890***	0.0618
									(3.46)	(1.27)
Political inform.: facebook									0.0393*	0.100***
									(2.00)	(3.66)
Political inform.: twitter									−0.0414*	0.00301
									(−1.97)	(0.07)
Newspaper pref.: tabloid									0.104***	0.0756**
									(3.32)	(3.12)
Constant	−0.0427	−0.0261	−0.289***	−0.207***	−0.193***	−0.182**	−0.233***	−0.0714	−0.0377	−0.0834
	(−1.38)	(−0.59)	(−7.85)	(−3.73)	(−5.21)	(−3.30)	(−7.13)	(−1.57)	(−1.06)	(−1.64)
Observations	2847	1355	2543	1197	1951	1114	2433	1306	2030	994
R^2	0.070	0.041	0.443	0.116	0.305	0.094	0.222	0.062	0.095	0.083

Linear regression models, t statistics in parentheses. ⁺ p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001. Data: authors own online survey, wave 2.

Table A2. Explaining the vote choice for VOX and Chega (full models to figures 2–6, right panel).

DV: Vote recall	Model 1 VOX	Model 1 Chega	Model 2 VOX	Model 2 Chega	Model 3 VOX	Model 3 Chega	Model 4 VOX	Model 4 Chega	Model 5 VOX	Model 5 Chega
Gender (male)	1.785***	2.518**	1.743**	2.556*	1.289	3.059**	1.560**	2.597**	1.414*	2.577*
	(4.17)	(2.70)	(3.02)	(2.53)	(1.27)	(2.68)	(2.81)	(2.71)	(2.04)	(2.40)
Age (young)	3.261***	4.130^+	2.972*	3.536	3.802**	4.719	6.280***	3.703^+	2.784*	4.301^+
	(3.39)	(1.90)	(2.36)	(1.54)	(2.65)	(1.59)	(4.33)	(1.65)	(2.33)	(1.74)
Education (low)	6.284***	3.484	7.184***	3.876	4.184*	1.542	6.435***	2.627	4.195**	4.509
	(4.66)	(1.41)	(3.88)	(1.46)	(2.47)	(0.40)	(4.23)	(1.10)	(2.99)	(1.46)
Religious	5.086***	1.552	1.227	1.115	2.079*	1.357	3.770***	1.614	5.503***	1.277
	(7.19)	(0.83)	(0.64)	(0.18)	(2.29)	(0.46)	(4.89)	(0.88)	(6.24)	(0.41)
Rural area	0.873	1.449	0.717	1.159	0.685	1.220	0.760	1.319	0.826	1.315
	(−0.69)	(0.93)	(−1.29)	(0.35)	(−1.41)	(0.41)	(−1.21)	(0.68)	(−0.79)	(0.59)
Ideology (right)			127.8***	7.386**
			(10.47)	(2.90)
Gov. performance bad			48.92***	1.547
			(9.40)	(0.54)
Previous gov. perf. bad			1.320
			(0.69)
Dissat. with democracy			3.694***	4.519*
			(3.75)	(2.29)
Corruption parties			2.497^+	2.376
			(1.76)	(1.18)
Low household income					0.967	0.277
					(−0.10)	(−1.62)
No assets					0.484^+	4.526^+
					(−1.77)	(1.82)
Economic sit. got worse					21.28***	7.111*
					(6.50)	(2.00)
Pers. econ. sit. got worse					0.646	1.410
					(−0.92)	(0.34)
Anti-globalization					3.315**	1.943
					(2.99)	(0.85)
Anti-immigration					11.23***	10.08**
					(5.05)	(2.97)
Anti-feminism					9.954***	0.880
					(6.77)	(−0.24)
Against Catalan autonomy							3.102***
							(4.76)
More centralized state							44.73***
							(10.12)
Privatize national health								1.633
								(0.95)
Reduce public services								3.512*
								(2.03)
Political inform.: forums									1.774^+	1.100
									(1.93)	(0.13)
Political inform.: facebook									1.971**	6.315***
									(2.75)	(3.50)
Political inform.: twitter									0.707	0.428
									(−1.27)	(−1.25)
Newspaper pref.: tabloid									2.431**	0.800
									(3.14)	(−0.51)
Observations	2243	987	2113	900	1627	807	1971	952	1674	731
Pseudo R^2	0.056	0.039	0.429	0.121	0.327	0.153	0.231	0.057	0.082	0.089

Logistic regression models, coefficients are odds ratios; t statistics in parentheses. ⁺ p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001. Data: authors own online survey, wave 2.