1. Introduction
Far right parties often promote extreme rhetoric on social and cultural issues. In theory, these messages should alienate voters who ascribe to socially liberal and anti-authoritarian norms. However, many far right parties have recently expanded their voter base to include more socially moderate individuals (e.g., Akkerman, Reference Akkerman2015; de Lange and Mügge, Reference de Lange and Mügge2015), challenging the idea that such parties succeed exclusively from a cultural backlash among disaffected social conservatives (Inglehart and Norris, Reference Inglehart and Norris2017). When are far right parties rejected for their deviant positions and when are voters willing to overlook them?
This paper articulates and tests a theory of discounting, which sheds light on why the far right’s extreme position-taking can be costly in some circumstances but not others. We posit that the effects of extreme position-taking are moderated by existing stigma surrounding a party. When a far right party is highly stigmatized, voters are likely to take its deviant statements as authentic indicators of extremist policy preferences, inducing a penalty on the party. However, when a party is widely treated as normal, voters are likely to discount and downplay the authenticity of its extreme statements, reducing their penalties.
We empirically test the key propositions of this argument with a preregistered survey experiment focusing on Vox, the first far right party to gain prominence in Spain’s post-Francoist democracy.Footnote 1 We randomly assign survey respondents to a bundled treatment that frames Vox as either a stigmatized or a normal party and also varies exposure to Vox’s extreme statements on either LGBTQ+ issues or Franco’s dictatorship.
Our findings support the validity of the discounting theory for some, but not all, issues. When respondents are primed with Vox’s anti-LGBTQ+ statements, sympathy for the party decreases. Yet, this penalty is erased when the party is explicitly framed as a normal (non-stigmatized) party. Notably, these results are driven by nonextremist right-wing voters, offering one explanation for why more moderate conservatives holding liberal sociocultural values may be increasingly willing to support the far right. However, we find that when citizens are primed to consider Vox’s pro-Franco statements, sympathy toward the party does not move, regardless of Vox’s degree of normalization.
This evidence suggests that party normalization can, but does not always, alter penalties that a party faces for its extreme positions. This difference in the discounting effect across issues could be tied to the ability of the far right to retain uncertainty in its issue positions: Vox’s closeness to authoritarian nostalgia may be more deeply embedded in its identity than its anti-LGBTQ+ credentials, as the party has sporadically engaged in “homonationalist” appeals (Turnbull-Dugarte and López-Ortega, Reference Turnbull-Dugarte and López-Ortega2024). Alternatively, differences in discounting effects could arise from differential norm strength around each issue. We return to these points in the discussion.
This project makes three main contributions. First, it adds a new explanation for why the far right has been able to expand its voter base without necessarily moderating its positions (Halikiopoulou and Vlandas, Reference Halikiopoulou and Vlandas2020; Lancaster, Reference Lancaster2020). Second, it complements recent research that emphasizes the role of social norms in explaining far right support (Blinder et al., Reference Blinder, Ford and Ivarsflaten2013; Bursztyn et al., Reference Bursztyn, Egorov and Fiorin2020; Valentim, Reference Valentim2024), shedding light on intermediate mechanisms through which this process unfolds. Third, it contributes empirical findings to literature specifically assessing how voters react to the far right’s messaging on LGBTQ+ issues (Turnbull-Dugarte and López-Ortega, Reference Turnbull-Dugarte and López-Ortega2024) and authoritarian history (Martín et al., Reference Martín, Paradés and Zagórski2022).
2. Normalization, discounting, and moderate voters
Far right parties’ extremism is likely to alienate moderate voters. This assertion, however, is at odds with emergent trends highlighting the increasing prevalence of socially liberal voters within the far right’s electorate (Halikiopoulou and Vlandas, Reference Halikiopoulou and Vlandas2020; Lancaster, Reference Lancaster2020; Turnbull-Dugarte and López-Ortega, Reference Turnbull-Dugarte and López-Ortega2024). How do these parties avoid backlash from their deviant (and public) statements?
Our argument foregrounds the role of social norms at the party-level in explaining how voters perceive extreme statements. Social norms are constituted by both empirical expectations and normative evaluations of peer behavior (Bicchieri, Reference Bicchieri2016). A social norm against the far right is present when individuals perceive peer support for a far right party as both unlikely and widely considered “unacceptable.” Literature on social norms suggests that as far right groups become visibly entrenched in the political mainstream, individuals are increasingly comfortable expressing support for them and their issue-positions (Blinder et al., Reference Blinder, Ford and Ivarsflaten2013; Harteveld et al., Reference Harteveld, Dahlberg, Kokkonen and Brug2019; Bursztyn et al., Reference Bursztyn, Egorov and Fiorin2020; Valentim, Reference Valentim2021; Gul, Reference Gul2023). This stream of research accompanies previous work highlighting the role of contextual factors, including the behavior of entrenched actors such as center right parties (Vrakopoulos, Reference Vrakopoulos2022; Krause et al., Reference Krause, Cohen and Abou-Chadi2023) and social democrats (Bale et al., Reference Bale, Green-Pedersen, Krouwel, Luther and Sitter2010), or existing welfare institutions (Vlandas and Halikiopoulou, Reference Vlandas and Halikiopoulou2022), in conditioning opportunities for far right success.
However, we still lack theory and evidence on how exactly this context alters intermediate barriers to far right success. We posit that party-level social norms against the far right can affect their support by shaping how voters assess their extreme issue statements—and whether they penalize them. Our focus on the intermediate role of issue evaluation allows us to deepen our understanding of the causal chain between party-level social norms and the success of those parties.Footnote 2
Social stigma surrounding a party sends messages to citizens about threats posed by said parties. When individuals perceive that peers, media, and political elites reject the far right, they may internalize the idea that such parties are inherently opposed to liberal democratic politics. As a result, voters are likely to punish its extreme issue positions, which are taken at face value. This is particularly true for more moderate right-wing voters who may share some policy stances with the far right but are still wary of supporting a party they perceive as genuinely threatening the status quo on ingrained social and cultural issues.
This logic can shift as a far right party sheds its stigma and is increasingly seen as “normal.” As the far right gains an institutional voice and is seen as an actor worthy of cooperation, voters may update their expectations that the far right will break with the status quo. Following normalization, extreme statements from the far right may be interpreted as symbolic positioning against political correctness or inconsequential appeals to particular groups of voters. When normalized, voters may, therefore, discount (or downplay) the authenticity of the far right’s extreme positioning and fail to punish the party for violating issue-based norms of appropriateness.Footnote 3 Thus, we expect that the normalization of a far right party reduces the costs it faces from extreme issue statements.
The theory requires two scope conditions. First, it primarily applies to issue stances that are widely perceived as extreme or deviant. Second, discounting is most likely among moderate conservative voters, as left-leaning voters are less likely to seriously consider supporting the far right (Golder, Reference Golder2016) and extreme-right voters are unlikely to penalize extreme positions, as they may be dismissive of liberal norms surrounding those issues. This leaves moderate conservative voters, who may share some policy stances and grievances with the far right but are not primarily attracted to the party by more marginally held cultural grievances (Halikiopoulou and Vlandas, Reference Halikiopoulou and Vlandas2020), as the most likely to discount extreme positions. Accordingly, this theory can help explain how far right parties have expanded their voter base to capture more moderate voters without moderating their discourse: party normalization helps voters discount the authenticity of the party’s most extreme statements, lowering inhibitions against the far right.
3. Research design
We test whether the normalization of the far right reduces the costs that it faces from its extreme statements with an online experiment fielded in Spain between June 22 and July 13, 2023. The sample included 4,402 center- and right-wing Spanish citizens using quotas to match national demographics on age, gender, and region.Footnote 4 The survey focused on Vox, the first far right party with electoral representation in Spain after the Francoist dictatorship. Consistent with other far right parties, Vox provides a variety of examples of deviant and provocative statements on different political issues, which we use in the experiment to randomly manipulate the salience of the party’s extreme positions.Footnote 5 When fielding the survey, Vox was the third largest party in the Spanish Congress and had participated in one regional government coalition (Castile and León).
We use a survey experiment to randomly manipulate party normalization and exposure to extreme position-taking.Footnote 6 We focus on Vox’s extreme anti-LGBTQ+ and pro-Franco statements, which violate dominant norms against homophobia, transphobia, and authoritarian nostalgia.Footnote 7 We use a 2×3 factorial design, allowing us to test the effects of exposure to extreme issue statements, party normalization, and their interaction (which reflects our main theoretical claim).Footnote 8
First, we vary whether a respondent is primed to consider Vox as a normal or stigmatized party. Multiple indicators may alter perceived social norms surrounding a party. As such, we use a compound treatment that uses different framings of Vox, none of which implied deception.Footnote 9 Each treatment included differential language to describe ideological profiles (far right vs. right of the center-right), alternative perspectives on coalition formation (mostly excluded vs. have formed one pact), varied election performance (weak local results vs. stronger national results), and disparate descriptions of public sentiment (mostly reject vs. mostly consider normal):
Second, we randomize participants’ exposure to Vox’s extreme statements, which remind respondents of its deviant positions and increase the salience of the issue-position when evaluating the party. We vary whether a respondent is exposed to Vox’s extreme statements on LGBTQ+ issues,Footnote 10 Franco’s regime, or no second prime (resulting in a 2×3 design with six possible treatment groups):
To assess post-treatment sympathy for the far right (Vox), we use a feeling thermometer (our primary outcome, on a 0:100 scale).Footnote 11 We also included questions in the survey to gauge whether individuals discount Vox’s extreme statements when the party is normalized (“Vox is hostile toward LGBTQ+ people” and “if Vox had it their way, they would reinstall a dictatorship”).
Our main theoretical claim is that the normalization of the far right reduces the electoral costs of deviant position-taking. We, therefore, estimate the equation:
\begin{equation}
Y_i = \beta_0 + \beta_1 N_i + \beta_2 P_i + \beta_3 (N_i * P_i) + \epsilon_i
\end{equation}where Yi represents sympathy toward Vox, Ni indicates the normalization prime, and Pi is a factor variable indicating exposure to Vox’s anti-LGBTQ+ statements, pro-Franco statements, or no prime at all (reference category). Importantly, because both issue statements are compared to a null reference group, we can think of these as essentially separate experiments for each issue (using the same control group). We also replicate this model with a measure of the perceived authenticity of the Vox’s positions (see wording above) as an intermediate outcome to explore the discounting mechanism.
A positive coefficient on β 3 would be consistent with our theoretical expectations, suggesting that normalization reduces the penalties incurred from the party’s extreme statements. We report sample average treatment effects throughout the paper, identifying effects among center- and right-leaning survey respondents (our sample) who roughly match the population on key demographics (see Appendix A1.1).Footnote 12
4. Results
Our theory suggests that the normalization of the far right reduces the costs it faces from extreme position-taking.Footnote 13 Let us first consider this possibility for Vox’s extreme anti-LGBTQ+ statements. Figure 1 plots the point estimates and 90 and 95% confidence intervals for the effects of party normalization, exposure to Vox’s extreme anti-LGBTQ+ statements, and their interaction.Footnote 14
Treatment effects from Vox’s anti-LGBTQ+ statements.
Panel A plots these effects on sympathy for Vox, showing, in the second row, that exposure to Vox’s anti-LGBTQ+ statement induced a penalty. However, this penalty is reduced when Vox is presented as a normal party, as the positive interaction term in the third row suggests.Footnote 15 Panel B complements this evidence by exploring the treatment effects on perceptions of Vox’s hostility toward LGBTQ+ people. Respondents regard the party as more hostile toward the LGBTQ+ community when exposed to its extreme statements but discount Vox’s anti-LGBTQ+ credentials when its statements are presented alongside an explicit normalization of the party.Footnote 16 Taken together, these results support our main hypothesis, as both antipathy toward Vox and perceptions of Vox’s hostility toward the LGBTQ+ community triggered by extreme anti-LGBTQ+ statements are reduced when the party is normalized.
Do these results travel to other issues where Vox offers a deviant position? Figure 2 examines the effects of Vox’s pro-Franco statements. Panel A plots these effects on sympathy for Vox and Panel B plots effects on perceptions of Vox’s authentic commitment to democracy.
Treatment effects from Vox’s pro-Franco statements.
We do not detect the same effects: Panel A shows that neither normalization nor priming the salience of Vox’s extreme positions is enough to move sympathy for the party. Panel B shows that while raising the salience of Vox’s pro-Franco position increases perceptions that Vox would prefer to reinstall a dictatorship (second row), normalization does not reliably cause respondents to discount the party’s authenticity, as suggested by the interaction term’s small coefficient and lack of statistical significance (third row). However, we note that the point estimate on the interaction is in the expected direction, and when analyzing a moderate-only subsample (ideological self-placement between 5 and 7), this interaction effect is significant at the 0.10 level (see Figure A6(B)). Thus, we are hesitant to fully rule out the possibility of discounting on authoritarianism, though this possible discounting would still fail to translate into direct effects on sympathy for Vox.
A second observable implication of our theoretical framework is that these results are mainly driven by moderate conservatives. To explore this proposition, we split our sample into two groups: moderates who position themselves between 5 and 7 on the 0–10 left-right scale and more extreme respondents ranging from 8 to 10.Footnote 17 Figure 3 shows the treatment effects among these two groups on sympathy for Vox by ideological group.
Treatment effects on sympathy for Vox: moderates and far right.
While we fail to detect effects on sympathy for Vox when priming its pro-Franco positions, we find support for our theoretical intuition that more moderate conservatives are driving the main results for anti-LGBTQ+ statements in Figure 1. In addition, Figure A6, which replicates Figure 3 using perceptions of Vox’s authenticity as the outcome, reinforces the conclusion that moderate conservatives discount the degree of Vox’s anti-LGBTQ+ hostility after it is framed as a normal, rather than stigmatized, political party. Overall, that discounting on anti-LGBTQ+ extremism is primarily driven by moderate voters suggests a new mechanism through which far right parties can expand to capture more moderate voters (e.g., Halikiopoulou and Vlandas, Reference Halikiopoulou and Vlandas2020).
5. Discussion
The results of our survey experiment indicate that the normalization of the far right can influence the costs it faces for its extreme positions. We find evidence that moderate voters discount Vox’s hostility toward LGBTQ+ people when its statements are paired with a normalizing frame of the party. Consequently, the penalties incurred from its extreme positions are erased when the party moves from stigmatized to normal. Interestingly, and contra our expectations, we find no evidence of Vox’s pro-Franco statements affecting sympathy toward the party, regardless of normalization.
One explanation for this discrepancy is that, in Spain, the far right’s anti-LGBTQ+ positions are relatively more uncertain than their stance on other issues (Turnbull-Dugarte and López-Ortega, Reference Turnbull-Dugarte and López-Ortega2024).Footnote 18 However, Vox’s ties to Franco face little doubt, possibly limiting the degree to which normalization can affect how the party’s statements on the issue are assessed. Discounting of extreme statements may thus be most likely when multiple positions can be consistent with a party’s identity. Alternatively, differences in effects could reflect differential norm strength: if pro-LGBTQ+ norms were stronger than anti-nostalgia norms, discounting could be more consequential in the former case, explaining why we see it translating into effects on sympathy in one branch of the experiment but not the other. Future research should do more to disentangle the issues and scope conditions that are conducive to discounting.
This project sheds light on previously unidentified pathways through which party stigmatization and normalization can reshape political opportunity structures for the far right (Bursztyn et al., Reference Bursztyn, Egorov and Fiorin2020; Valentim, Reference Valentim2024). Normalization can cause voters to discount and downplay a party’s authentic policy desires, facilitating opportunities for the far right to expand its voting base and capture more socially moderate voters. At the same time, these findings highlight the power of maintaining stigmas against extreme parties through strategies including “cordon sanitaires,” limiting attention, and (non-)accommodation (Pardos-Prado, Reference Pardos-Prado2015; Spoon and Klüver, Reference Spoon and Klüver2019; Krause et al., Reference Krause, Cohen and Abou-Chadi2023). Stigmatizing the far right can reduce the likelihood that moderate voters discount a party’s policy goals, ensuring that such parties are accountable to the costs of extremism.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/psrm.2025.29. To obtain replication material for this article, https://doi.org/10.7910/DVN/CMPWS1.
Funding statement
This study is part of the Georgetown University project “Inequality and Governance in Unstable Democracies – The mediating Role of Trust,” implemented by a consortium led by Institute of Development Studies (IDS). The support of the UK Economic and Social Research Council (ESRC grant ES/S009965/1) is gratefully acknowledged.
Acknowledgements
We are grateful for comments from Oliver Lang, Adeline Lo, Evgeniia Mitrokhina, Dawn Teele, Vicente Valentim, and Matt Winters, as well as feedback from attendees at the Princeton Research Workshop in Experimental Social Science, SAIS Bologna research seminar, the panel “Right Parties, Populist Parties, and Norm Transgressions” at the 2024 Swiss Political Science Association annual conference, and the panel “Stigma and the Far Right” at the 30th International Conference of Europeanists.
Competing interests
None.
Ethical standards
The study was approved by Institutional Review Boards at Georgetown University [ID: STUDY00006690] and University of Wisconsin-Madison [ID: 2023-0778].
Author contributions
All authors contributed equally to the manuscript.
Open access
