Abstract
Introduction
The 2018 regional elections in the autonomous community of Andalucía in Spain signalled a substantial shake-up of the country’s party system, bringing about the end of Spain’s exceptional status 1 as a country without a radical right-wing political party (Alonso and Kaltwasser, 2015). The new right-wing challenger, Vox, gained electoral success for the first time, winning 11% of the votes cast and gaining a total of 12 seats in the regional parliament. The party’s support was also essential in ousting the socialist-led regional government, which had been in power in Andalucía for more than 36 years, supporting the right-wing coalition between the Partido Popular (PP) and Ciudadanos.
In this paper, I provide the first empirical assessment of the individual-level factors that explain the motivations behind electoral support for Vox. Relying on post-electoral survey data 2 from 1413 voters collected shortly after the election took place, it is argued that the political conflict regarding the separatist movement in Catalonia and the inability of national government to resolve the same had a large impact on explaining individual support for Vox.
Empirically, I demonstrate that those voters who express support for reducing the political independence of the country’s autonomous communities are more likely to vote for Vox. In significant contrast to the understanding of the radical right-wing vote across other Western European countries that view right-wing support to be driven by concerns over immigration, distrust with the political establishment, and voters from the lower end of the socio-economic distribution (Ford and Goodwin, 2014; Mudde, 2007; Norris, 2005; Oesch, 2013; Rydgren, 2008; Zhirkov, 2014), I find that concerns over radical right-wing issues such as immigration are
Election context
The elections of 2 December 2018 in Andalucía took place three months ahead of schedule after the regional president, Susanna Díaz, called early elections in the October. The announcement came only a few months after the successful no-confidence motion brought forward by the Spanish Socialist Worker’s Party (PSOE) replaced the PP’s Mariano Rajoy with their own leader, Pedro Sánchez, as president of Spain. Díaz’s call for early elections was widely viewed as a political gamble that the regional president thought would be advantageous given that she would be able to monopolise on the PSOE’s political momentum after taking over the national executive (Martín-Arroyo, 2018).
The elections took place, however, during a time of heightened territorial tensions. Alongside the traditional partisan conflict across liberal-left and conservative-right regarding economic and social issues, Spanish political competition is also marked by a second dimension involving centre-periphery conflict over devolution (Alonso et al., 2015; Pardos-Prado and Sagarzazu, 2019; Sánchez-Cuenca and Dinas, 2012), specifically regarding the political competences of Spain’s autonomous regions. On 1 October 2017, the Catalan government, led by a coalition of separatist parties, held an unauthorised 3 referendum and unilaterally declared independence from Spain shortly afterwards on 27 October. The national government, led at the time by Rajoy, acted to suppress the separatist movement and invoked Article 155 of the Spanish Constitution, which led to the cessation of the Catalan government and other regional authorities, and the temporary direct rule over the region from Madrid until new elections were called. Little improvement has been made to the political conflict since then (Orriols, 2018), with the former leader of the Catalan government, Carles Puigdemont, currently in self-imposed exile in Brussels whilst a cohort of his former government colleagues undergo trial for charges of rebellion against the state in the country’s supreme court. Currently, a separatist coalition remains in power in Catalonia whilst political conflict over Catalan separatism has noted an expansion beyond the confines of the region’s own electorate into the national arena (Kingsley and Minder, 2017; Sánchez-Cuenca, 2018).
In addition to the national domestic conflict caused by the Catalan question, the elections in Andalucía took place at the end of a year when Spain had become the new focal point of the European migrant crisis, with the number of migrants arriving to Spain from Africa in 2018 surpassing those that landed in both Italy and Greece (Keeley, 2018). The decision of the Sánchez-led socialist government to adopt a stance solidarity with incoming migrants and to accept the disembarking of migrants rescued on the Aquarius ship after the Italian government refused to receive them was also not without criticism from the press and opposition parties (Sanmartín, 2018).
The political success of Vox has largely been claimed by political commentators to be the result of the party’s ability to mobilise voters on both the Catalan and immigration themes (Jabois, 2018; Llaneras et al., 2018). I argue that whilst there is empirical support for the former, the claim for the latter is not supported by the individual-level data.
Vox
The popularity of Vox and the electoral potential of the party was first made apparent after it gathered a crowd in surplus of 9000 people to an arena in Madrid (Lambertucci, 2018) in October 2018. Like the other recent partisan entrepreneurs in Spain (Anduiza et al., 2014; Turnbull-Dugarte, 2019), Vox had been able to garner much of its initial support through a grassroots-based communication strategy carried out over social media (Viejo and Alonso, 2018) where the party has been able to expand its support base without relying on the traditional mainstream media outlets.
Vox is a party with a political ideology firmly grounded in authoritarian conservativism and nationalism. Far from being a single-issue anti-immigrant party (Art, 2011), it advocates policy positions including opposition to same-sex marriage, abortion, gender quotas, gender violence protection ordinances, and general social welfare provision, marking the party’s ideological parity with other Western European radical right-wing parties (Art, 2011; Mudde, 2004, 2007; Norris, 2005). Indeed, it’s populist approach of promising to “make Spain great again” alongside its nativist authoritarianism signal its fulfilment of Mudde’s (2004) classification as a populist radical right party. Of note in the specific case of Vox, however, is the addition of a strong anti-feminist message (Simón, 2019) brought about in response to recent waves of feminist advancement in Spain as evidenced by the historical 8M demonstrations in 2018.
Echoing parties of a similar ideological and populist flare such as France’s
On the Catalan question, Vox promotes policies that reflect both its preference for authoritarianism and focus on Spanish nationalism. In response to the separatist threat, the party calls for both the suspension of the autonomous independence of the Catalan region – policy proposal number one in its programme of 100 measures (Vox, 2018) – as well as the constitutional prohibition of any party that seeks separatist objectives. 5 These objectives, particularly the latter, mark Vox’s extremism and spatially political distinctiveness vis-à-vis the other right-wing parties.
Election result
The election resulted in the worst electoral outcome observed by the PSOE in the region, with the party losing 14 seats and any path to the region’s executive. The regional variant of Podemos, Adelante Andalucía (AA), also saw a reduction in its electoral fortunes, losing almost 6% of its vote share. Across the leftist bloc, PSOE and AA lost a total of 17 seats to the right. The mainstream right-wing parties of PP and Ciudadanos claimed victory, arguing that the people of Andalucía had voted for change and political renewal despite the PP actually worsening its position vis-a-vis the previous regional elections in terms of both vote share and seats (Table 1).
Regional election results in Andalucía from 2015 and 2018.
PSOE: Spanish Socialist Worker’s Party; PP: Partido Popular; Cs: Ciudadanos; AA: Adelante Andalucía.
In 2015 = shared vote share and seat count of Podemos and Izquierda Unida Los Verdes-Convocatoria por Andalucía before uniting under the AA regional brand.
Despite the mainstream right’s claims of victory, the national press focused on the main news story of the night, which was that of the political eruption of Vox into the electoral arena, with political pundits making clear that some form of political agreement with the radical right party was very likely given the legislative arithmetic of the results. In the end, a new PP–Ciudadanos coalition government was installed with support of Vox, bringing to a close 36 years of PSOE-led governments in the region.
Not altogether surprisingly, the vast majority of Vox supporters came from parties already ideologically placed within the right (Figure 1). Almost half of the voters came from the PP whilst one fifth came from Ciudadanos. Votes from parties on the left represented a far smaller proportion of their electoral revenue, providing, together, less than 10% of Vox’s votes in 2018.

Partisan vote revenue of Vox support.
In the following section, I examine the individual factors that are associated with electoral support for Vox. Given the strong link between support for radical right parties and immigration observed across much of Western Europe (Ford and Goodwin, 2014; Mudde, 2007; Norris, 2005; Rydgren, 2008; Zhirkov, 2014) alongside the party’s own hostilely critical position towards the issue, the link between immigration and support for Vox represents a rational assumption. The results of the following, however, demonstrate that Spain remains different: I find a robust non-effect between concerns over immigration and individual support for Vox. Given the saliency of the Catalan question, I hypothesise that support for Vox will be a function of voters’ views on the political powers of the autonomous communities given the strong message of Vox against regionalism and aggressive pro-nationalist messaging. Support for this hypothesis is borne out in the data.
Testing vote choice
Dependent variable and estimation technique
In order to assess the individual-level determinants of vote choice for Vox, I rely on the post-electoral survey data completed by the CIS between 10 December 2018 and 3 January 2019. The dependent variable relies on self-reported vote choice for each of the five main parties (PSOE, PP, Ciudadanos, AA, and Vox) in the 2018 election. Given the nominal nature of the dependent variable, I estimate a multinomial logistic regression model (baseline: abstention 6 ).
Independent variables
To test the individual explements of electoral support for Vox, I focus on assessing opposition to territorial devolution – measured in support for the independence of the autonomous communities – and concerns over immigration. The dichotomous independent variable
Controls
The model conditions for a vector of politically relevant variables as well as socio-economic indicators. At the political level, the model includes five controls: support for the incumbent (0–10) at the regional and national level; ideological placement across the left (0) versus right (10) axis; distrust towards politics (1 = respondent’s impression of politics is of distrust; 0 = otherwise); and former political activity (1 = respondent had previously attended a political protest; 0 = otherwise).
The model also caters for two demographic indicators: gender (1 = male) and age. At the socio-economic level, the model includes four controls: religiosity (min: never attends church; max: frequently (at least once per week) attends church); education (min: primary education; max: tertiary education); income; and employment. Income is operationalised on an 11-point ordinal scale 7 and employment status is a categorical variable (baseline: working) indicating those who are unemployed, retired, studying, or not in the labour force 8 (NILF).
Analysis
For ease of interpretation, the average marginal effects (AMEs) of the main explanatory variables are visualised in Figure 2 (for detailed regression and AME output, see the online appendix). The analysis provides a number of significant findings. Firstly, concerns over immigration do not exhibit any independent effect on the probability of voting for Vox, or any other party, at levels of traditional statistical significance. Against popular assumptions and empirical evidence to support this expectation, immigration concerns are not operating in Spain as an influential factor for individuals’ voting behaviour or specifically in support of Vox.

AMEs of main explanatory variables.
As a robustness check, 9 I test that the non-significance of the immigration variable is not conditional on socio-economic status. Since concerns over immigration have been found in some cases to be driven by economic anxieties amongst the working class and other “economic losers” (Ford and Goodwin, 2014; Kriesi et al., 2008; Mudde, 2007; Oesch, 2013), if this is at work in the Spanish case one might expect there to be a significant relationship between immigration and support for Vox amongst those at the lower end of the income distribution. This does not appear to be the case: re-estimating the main model to include an interaction term between immigration and income reveals that immigration exhibits no effect on the probability of voting for Vox, regardless of individuals’ income. Figure 3 illustrates the effect of the interaction, which highlights the consistency of the non-effect of immigration. Notably, the effect is absent at income values where the majority of the population belong (see mean and standard deviation values in Figure 3).

Robustness check on (un)conditionality of immigration.
Moreover, whilst the effect does not provide results of significance (
Greater clarity is provided by views on Spanish devolution. Those who believe that the independence of the autonomous regions should be reduced observe a five-percentage-point increase in the probability of voting for Vox. The effect is both substantive and significant. For both the PP and Ciudadanos, individuals’ desire to castrate autonomous independence does not appear to have provided them with any electoral support: the null-effect lies well within the confidence intervals of the variable for both parties.
Blame game
Significant variation is observed in the effect of support for the political incumbents at different levels of the political system (Figure 4). Whilst positive evaluations for both the regional (Díaz) and national (Sánchez) president increased votes for the PSOE, individual assessments of Díaz did not impact electoral support for Vox. In other words, Vox votes were not driven by discontent with the leader operating at the subnational level, but of negative assessments of the president at the national level. Since the Catalan question is the contemporary political issue of greatest national saliency at the time, it is easy to draw a link between Sánchez and the incumbent national government’s management of the separatist crisis.

AMEs of incumbency.
Interestingly, evaluations of the incumbents produce heterogenous effects on the probability of voting for AA, the other main leftist party alongside PSOE. Those who positively evaluated Sánchez observed an increase in the probability of voting for either of the left-wing parties. The same was not observed by those who positively assessed Díaz, which actually significantly decreased voting for AA. This suggests that Sánchez was able to mobilise voters across the left both for his own party as well as for AA whilst Díaz could not.
Finally, noteworthy amongst the control variables is that political distrust does not display an effect of significance in relation to support for Vox (see Table A2 and Table A4 in the online appendix) whilst it does for a number of other parties, particularly AA. This is interesting because it marks Vox as distinct from other European radical right parties, particularly right-wing populists, who have garnered support from voters who express a lack of trust in the political system (Mudde, 2004). Vox voters are not sourced from those who are politically disenchanted but those driven by the territorial conflict.
Disloyalty on the right
Since over two thirds of Vox’ votes came from voters that had previously voted for the other main right-wing parties, PP and Ciudadanos (Figure 1), I also run a logistic regression model estimating votes cast for Vox compared to the other right-wing parties amongst those who voted for right-wing parties in the 2016 general election. Table 2 reports the AMEs as percentage points (for detailed logistic regression output, see the online appendix). Of note, again, is that concerns over immigration do not explain why right-wing voters moved their support to Vox. The results show that former PP and Ciudadanos voters are driven to Vox largely by their left–right ideological preferences as well as their support for reduced autonomous independence. Right-wing voters who were in favour of reducing the competences of the autonomous regions observed an increase in the probability of voting for Vox of eight percentage points. The explanatory role of nationalistic concerns related to devolution, therefore, exhibit both a significant and politically substantive effect on the individual-level vote choice for Vox and explains why right-leaning voters jumped ship from the mainstream right to the more extreme and nationalist option. Given that the PP had been in power nationally until the summer of 2018, it may well be that the exodus of former PP voters with anti-devolution views is the result of the party’s own mismanagement of the territorial conflict during their term in office at the height of the Catalan crisis.
Average marginal effects of support for Vox in 2018 amongst right-wing voters in 2016.
NILF: not in the labour force.
Conclusions
The 2018 regional elections in Andalucía represent both a turning point in the regional government but also the regional party system (and likely nationally after the general election in April 2019) from the four-party system that emerged in 2015 (Orriols and Cordero, 2016) into a five-party system, with Vox joining on the extreme end of the rightist bloc previously dominated by the PP and more recently in tandem with Ciudadanos.
Apart from the obvious ideological link between those who place themselves on the right of the ideological spectrum, the results of the analysis here report that voters are driven to vote for Vox when they believe in reducing the independence of the country’s autonomous regions. It is, therefore, clear that the political conflict engendered by the Catalan question that drives Vox’s pro-Spanish and nationalist message is what attracts its supporters. Importantly, support for castrating devolved powers also explains the support for Vox amongst former voters of Ciudadanos and the PP who make up the largest source of Vox’s 2018 vote revenue.
Concerns over immigration do not have an independent effect on support for Vox. The significance of this finding is paramount as it marks Spain (still) as being different from the rest of Western Europe. Whilst Vox’s success represents the end of the country’s exceptional status as a country free from an electorally successful radical right-wing party, the motivations for this party’s success do not appear to be dependent on the individual-level determinants – mainly that of immigration – that explain support for right-wing populists elsewhere (Ford and Goodwin, 2014; Lubbers et al., 2002; Mudde, 2007; Norris, 2005; Zhirkov, 2014). The radical right may have dawned in Spain, but the influential political role of immigration on voting has not.
Supplemental Material
3_vox_appendix_RR – Supplemental material for Explaining the end of Spanish exceptionalism and electoral support for Vox
Supplemental material, 3_vox_appendix_RR for Explaining the end of Spanish exceptionalism and electoral support for Vox by Stuart J. Turnbull-Dugarte in Research & Politics
Footnotes
Declaration of conflicting interests
Funding
Supplemental materials
Carnegie Corporation of New York Grant
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
