The Preferred Governing Actors of Populist Supporters: Survey Evidence From Eight European Countries

Populist parties have been shown to attract many voters disillusioned with representative democracies. And some of these parties do indeed propose models of government that challenge contemporary democratic systems. However, we do not know exactly what the democratic preferences of populist party supporters are. We propose to fill this gap by investigating the types of actors that citizens who are more sympathetic to populist parties would like to see play a greater role in their national political system. First, we find that populists believe that citizens should be more involved, highlighting the people‐centred nature of populism. Second, they advocate a greater role for business leaders, military generals, and religious leaders, a preference found among both right‐wing and left‐wing populists. Third, left‐wing populists show a unique preference for scientific experts in government, suggesting a technocratic inclination. Conversely, right‐wing populists are particularly critical of elected politicians, underlining their deep anti‐elitist attitudes. Our findings suggest that, among citizens who are more sympathetic to populist parties, there is support for models of government that challenge representative democracy. The question is whether populist parties would be influenced by these citizens to push for institutional


Introduction
In the ever-growing literature on populism, one of the questions that has attracted attention in recent years is how populists relate to democracy and may potentially challenge it (Rovira Kaltwasser & Van Hauwaert, 2020;Zaslove et al., 2020).This debate is rooted in broader academic debates on citizens' disenchantment with representative democracy and on growing support for alternative models of government (Hibbing et al., 2023;Valgarðsson et al., in press).In the literature on populism and democracy, a first important question has been whether populist citizens and voters are democrats or rather hold more authoritarian views and would support a move away from democracy or at least some of its dimensions (especially the rule of law and respect for minority rights; Huber & Schimpf, 2017;Wuttke et al., 2023;Zaslove & Meijers, 2023).Other scholars have tried to examine the support of populist citizens for different alternatives to pure representative democracy, such as direct democracy, deliberative democracy, or technocracy (Bertsou & Caramani, 2022;Fernández-Vázquez et al., 2023;Jacobs et al., 2018;Mohrenberg et al., 2019).
However, this last strand of research has assessed populists' support for these actors in isolation.Recently, scholars have proposed a more direct comparison of support for different models of government by jointly examining citizens' process preferences (Beiser-McGrath et al., 2022;Font et al., 2015;Gherghina & Geissel, 2019;Hibbing et al., 2023;Pilet et al., 2024), that is, their preferences for how the political system should be organised and, in particular, which actors should govern.In this article, we build on this approach, and in particular on the survey battery developed by Hibbing et al. (2023), to systematically investigate how voters of populist parties want government to be organised and which actors they want to play a major role in shaping policy decisions.In particular, we examine whether populist voters-both on the radical right and the radical left-differ from voters of other parties in their process preferences.Our study follows in the footsteps of some previous studies that have taken a similar approach (see Heinisch & Wegscheider, 2020) and proposes the most comprehensive comparative study to date, based on data from a survey conducted in the winter of 2022 in eight European democracies: Belgium, Bulgaria, Czechia, Denmark, Finland, Greece, Ireland, and the Netherlands.
Our data show that both anti-politician and pro-citizen attitudes are strongly related to support for populist parties (left and right).Somewhat surprisingly, we found that right-wing populists are less enthusiastic about scientific experts than left-wing populists: In this respect, Covid-19 and right-wing scepticism towards medical experts in government may have played a role.While right-wing populist supporters are more inclined to trust non-traditional actors (such as businesspersons, religious leaders, and military generals), we would not expect this to be the case for left-wing populist supporters.However, our data show that supporters of left-wing populist parties appear to be more favourable towards these actors in government.
The article is structured as follows.In the second section, we build on previous research on citizens' process preferences and populist voters, and on citizens' attitudes towards representative democracy and its alternatives, to develop a set of hypotheses.In the third section, we present our data and methodology.
In the fourth section of the article, we empirically test our hypotheses.Finally, we discuss the implications of our findings for the literature on populism and democratic preferences.

Earlier Research and Hypotheses
In this study, we propose to examine the relationship between decision-making processes and support for populist parties.In this respect, we depart from most previous studies that have chosen to examine populist citizens, defined as citizens who score high on batteries of populist attitudes (A.Akkerman et al., 2014).
The decision to study voters of populist parties is motivated by the structural electoral growth of these parties across Europe.Populist attitudes tend to be more pronounced at the extremes of the left-right scale, but overall they are limited in public opinion (Vittori, Rojon, et al., 2023;Wuttke et al., 2023).For this reason, the overlap between populist party supporters and populist citizens is only partial: The well-documented mainstreaming of the radical right (T.Akkerman et al., 2016) and the electoral success of radical left parties is a consequence of the broader electoral appeal of these parties to non-populist citizens as well (Van Hauwaert & van Kessel, 2018).As political parties respond to their support base (Spoon & Klüver, 2014), knowing the process preferences of their supporters may be a good indicator of the parties' positions on these less documented issues.However, as the literature on the process preferences of populist voters remains scarce (see van der Brug et al., 2021 for an exception), we will mostly rely on the literature on populist citizens to build our hypotheses.A lively debate within this literature is whether populists have democratic preferences or whether they lean towards more authoritarian views.This debate stems from broader theoretical debates about the democratic or undemocratic nature of populism as an ideology (Canovan, 1999;Urbinati, 2014) and of populist parties (Vittori, 2022).Scholars have subsequently attempted to examine whether citizens with populist attitudes and voters of populist parties hold authoritarian or democratic views on how the political system should be organised.
In this article, we do not seek to contribute to this debate on the democratic character of populist citizens, nor do we insist on the democratic credentials of populist parties.Rather, we propose to build on another strand of research within the study of populism, which has examined which actors within a democratic system populists want to see play a central role in government.The questions that this study addresses are: What kind of actors do supporters of populist parties want to see play a greater role in shaping policy?And what are the differences between populist radical left and populist radical right voters in this respect?
Theoretically, we define populism as a thin-centred ideology that "considers society to be ultimately separated into two homogeneous and antagonistic groups, 'the pure people' versus 'the corrupt elite,' and which argues that politics should be an expression of the volonté générale" (Mudde, 2004, p. 543).One of the tenets of populism is its anti-elitist stance and, in particular, its anti-political elite stance: Populism has been defined as hostile to pluralism (Urbinati, 2014), as it rejects that society is made up of different groups with different interests.For populists, the homogeneity of people is translated into the homogeneity of social interests (Caramani, 2017).Therefore, it is not surprising that the literature has examined the attitudes of populist citizens towards elected politicians and political parties, which are the core actors of the contemporary representative model of democracy.The findings are very consistent in this respect.Populist citizens and voters are disaffected democrats (Rovira Kaltwasser & Van Hauwaert, 2020), meaning that they are dissatisfied with the way democracy works in their country (Zaslove & Meijers, 2023) and this is because they have very negative views of politicians and political parties (A.Akkerman et al., 2014;Rooduijn, 2018).
Building on those studies, we can formulate a first hypothesis: H1: Supporters of populist parties hold more negative attitudes toward elected politicians.
Another recurring and fairly well-studied dimension of populist citizens' and voters' views on how the political system should function and which actors should be given a key role in shaping policy decisions is that they strongly favour giving citizens a greater and more direct role.One of the three core dimensions of populism is people-centrism (A.Akkerman et al., 2014;Mudde, 2004).It is defined as support for a model of government in which core decisions are left directly to citizens, without the mediation of elected politicians, political parties, or representative institutions.In this context, several authors have shown a strong preference of populist citizens and voters for more referendums (Jacobs et al., 2018;Mohrenberg et al., 2019;Rojon & Rijken, 2020;Wuttke et al., 2023;Zaslove et al., 2020).On this basis, we can propose a second hypothesis: H2: Supporters of populist parties hold more positive attitudes toward giving citizens a greater and more direct role in policy-making.
Another area of interest that has recently emerged in studies of the types of governance that populist citizens and voters support, and the actors they would like to see empowered, is the relationship between populism and technocracy.Within this area of research, views are more mixed, both theoretically and empirically.Theoretically, as discussed by Caramani (2017), populism and technocracy share some similarities, such as scepticism towards the party model of government and the idea that there is a single and accessible "best solution" for every policy decision.At the same time, the two models of government are very different in other respects.In particular, while populism is based on a deep trust in the ability of the people to govern, technocracy is based on the assumption that only a few experts have the necessary skills to govern.This ambivalent view of the links between populism and technocracy is reflected in empirical research.First, when populist parties are in government and electorally strong, they tend to appoint more technocrats than non-populist parties (Pilet et al., 2023).At the level of public opinion, some studies find a correlation between populist attitudes and support for a greater political role for independent experts (Fernández-Vázquez et al., 2023;Heinisch & Wegscheider, 2020).In contrast, other studies have shown that populist citizens have more negative views of science and scientific experts (Eberl et al., 2023).Although the literature is divided on this point, we propose a third hypothesis that postulates a positive relationship between populism and support for scientific experts: H3: Supporters of populist parties hold more positive attitudes toward giving a greater role in policy-making to scientific experts.
So far, all our hypotheses have been based on the idea that all populist voters share common views on how the political system should be organised.However, we also know from previous research that even if populist voters share some political attitudes, they form a rather heterogeneous group, especially when comparing voters of radical right and radical left populist parties (Rooduijn et al., 2017).In this respect, Heinisch and Wegscheider (2020) have shown that there are dimensions of the democratic preferences of populist citizens that are common to all populist citizens, but also other dimensions on which there are substantial differences related to the host ideology to which populism is attached (radical right or radical left).Following their example, we therefore propose to discuss where the supporters of radical right and radical left parties may differ in their preferences for who should govern.
First, what supporters of different types of populist parties have in common, in terms of their views on how government should work, is a negative evaluation of elected politicians and support for a greater and more direct role for citizens in policy-making (Heinisch & Wegscheider, 2020;van Dijk et al., 2020).These two aspects are directly related to two of the core dimensions of populism: anti-elitism and people-centrism (Mudde, 2004).
However, there are also dimensions of process preferences on which we can expect differences between voters of radical right and radical left populist parties.A first difference could be derived from studies on the so-called stealth democracy model (Hibbing & Theiss-Morse, 2002).Stealth democrats are described as citizens who are dissatisfied with representative democracy and elected politicians, but who do not want to move to a model of greater citizen participation.Rather, they want political decisions to be taken more quickly by independent actors who have acquired skills outside party politics.Experts, but also business leaders, are seen as such actors.This strand of research is useful for this study because several authors have linked stealth democratic attitudes and populism (Mohrenberg et al., 2019;Stoker & Hay, 2017;Webb, 2013).The two are not identical, but they share some features, such as a dislike of more consensual and deliberative ways of making policy, or the idea that decisive action should be taken by political outsiders.
Support for such actors is often associated with more authoritarian views of politics and, more generally, with authoritarian regimes.In consolidated democracies, however, some citizens have been found to want to retain the core principles of democracy, but to involve these actors in political decision-making (Meyer et al., 2008).In Western countries, this link between populism and covert democratic views seems to be particularly strong when it comes to radical right-wing populism.It is less often associated with left-wing ideology (Hibbing et al., 2023;Pilet et al., 2023).This is because while both left-wing and right-wing populists are anti-elitist, their conception of anti-elitism and people-centredness is different: Right-wing populism is associated with a nationalist conception of the people, with a strong emphasis on old-fashioned traditions (Taggart, 2000) and a law-and-order approach to those who are not part of the people.Left-wing populism, on the other hand, is closer to the demands of populist social movements (Aslanidis, 2017), such as an emphasis on the inclusion of excluded minorities and lower social classes in the decision-making process.In particular, it is less likely that radical left populist voters would welcome giving business and religious leaders a greater role in politics.This is because supporters of radical right populist parties have stronger religious beliefs and hold authoritarian views (Dunn, 2015;Immerzeel et al., 2013;Tillman, 2021), while the opposite is true for supporters of radical left populist parties (Rooduijn et al., 2017;Visser et al., 2014): H4: Contrary to supporters of populist radical left parties, supporters of radical right populist parties hold more positive attitudes toward giving a greater role in policy-making to business leaders, religious leaders, and military generals.

Data and Method
We propose a comparative approach to test our hypotheses.Indeed, one of the main weaknesses of previous research on populists' preferences for how government should work is that it is mostly based on single-country case studies (and mostly in Northwestern Europe).Comparative research is scarce, while country differences may be important in shaping individuals' views on how the political system should be organised.Such cross-country differences have been observed, for example, in several papers examining support for liberal democracies among populists (van der Brug et al., 2021;Wuttke et al., 2023;Zaslove & Meijers, 2023).

Survey
For this article, we rely on an online survey fielded in January 2022 by the survey company Qualtrics, which covers eight countries: Belgium, Bulgaria, Czechia, Denmark, Finland, Greece, Ireland, and the Netherlands.
In the case of Belgium, we collected two samples, one for the French-speaking region and one for the Dutch-speaking region.In our sample, we include parliamentary democracies belonging to Eastern Europe (Bulgaria, Czechia), Western Europe (Belgium, Ireland, the Netherlands), Southern Europe (Greece), and Scandinavian countries (Denmark and Finland).We include countries with different levels of political institutionalization because it might be an important contextual variable affecting citizens' preferences for actors in government.We consider young (Bulgaria, Czechia, and Greece) and established democracies (the remaining five countries).There are countries with substantial previous involvement of technocrats in government (Bulgaria, Czech Republic, and Greece), and countries with no (e.g., Ireland) or limited previous experience of technocrats in government (Vittori, Pilet, et al., 2023).There are countries where direct democracy has often been used at the national and local level (Denmark and the Netherlands) or for major decisions (Greece), and others with limited or no experience with referenda (Hollander, 2019).There are democracies with higher trust in representative institutions (Denmark and the Netherlands) compared to those with intermediate (Greece) or low trust (Bulgaria and Czechia).As for the political system, the country selection ensures variation in party systems and the logic of government (ranging from single-party governments to broad coalitions).Finally, they present different configurations regarding populist parties, in terms of host ideology, electoral strength, and position within the party system (e.g., in power vs. ostracized challenger parties).Each country has a sample of approximately 1,500 respondents.Four stratification criteria were used to make the samples representative of the whole population in each country: age, gender, place of residence, and education.Since we could not match the exact quotas for each criterion, we weighted our sample to correct for underrepresented groups.In the Supplementary File, we provide full information about the sample in each country, the quotas, and the mismatches we have identified.
The survey duration was approximately 15 minutes and included questions related to political attitudes, voting behaviour, process preferences, and respondents' socio-demographic characteristics.Attention checks were included during the survey: Respondents who did not pass the attention checks were dropped.

Dependent and Independent Variables
The dependent variables capture the views of populist voters regarding the role of four sets of actors that could be given a key role in shaping policy decisions in their country: elected politicians, citizens, scientific experts, and a fourth cluster of actors composed of business leaders, military generals, and religious leaders.
The measures build upon the approach recently proposed by Hibbing et al. (2023) and replicated by Pilet et al. (2024).Hibbing and colleagues first developed a comprehensive battery of 21 survey items to capture citizens' process preferences (see Appendix I in the Supplementary File 1) and tested it for the case of the United States.They identified, via factor analysis, seven dimensions.The first revolves around citizens' capabilities as decision-makers; the second focuses on politicians' capabilities as decision-makers; the third is about conferring power to the people; the fourth is about transferring power to scientific experts; the fifth assumes shifting power to non-traditional actors, such military and business leaders; the sixth suggests empowering generic actors closer to citizens; and the final seventh dimension measures perceptions of the nature of governing.Based on the test of the same battery replicated in Europe, Pilet et al. (2024) identified five core dimensions that are consistent across the countries covered in this article.These dimensions are aligned with those identified by Hibbing and colleagues, except for the third dimension (conferring power to the people) and the sixth dimension (actors closer to citizens; see Table 1).We thus focus on four dimensions and the policy-making role of (a) elected politicians, (b) citizens, (c) scientific experts, and (d) business leaders, military generals, and religious leaders.We leave out the fifth dimension which is not about who should govern but how decisions should be taken (consensus vs. majoritarian democracy).
The factor scores extracted from the factor analyses are used as dependent variables in our regressions.
The main independent variable is the sympathy score for the populist parties in the eight countries under analysis.The wording of the question was the following: "How would you rate your feelings about each of the following political parties on a scale ranging from "very negative" (0) to "very positive" (100)?"Using the PopuList classification, we identified whether parties are populist or not.As we have separate expectations for left-wing and right-wing populist parties, we have considered populist left-wing parties as those that are "populist" and "far left" in the PopuList and populist right-wing parties as those that are "populist" and "far right."The full list of populist parties and their ideological inclination is available in Appendix II of the Supplementary File 1.In total, we identified 16 populist parties in our dataset: 9 radical right, 5 radical left, and 2 that are neither radical right nor radical left.Building on earlier research, we have added PTB-PVDA (The Workers' Party) in Belgium as a populist radical left party (Goovaerts et al., 2020) and excluded the New Flemish Alliance in Belgium, which is listed as a borderline case of right-wing populism.
For this study, we decided to focus on supporters of populist parties rather than citizens with populist attitudes as in other studies (Fernández-Vázquez et al., 2023;Heinisch & Wegscheider, 2020;Jacobs et al., 2018;Mohrenberg et al., 2019;Zaslove et al., 2020).Looking at populist citizens is extremely insightful, but would pose a problem for the goal of this study.Our aim is to examine whether supporters of populist parties differ from citizens who do not support these parties in terms of the actors they want to govern.
Two of the four groups of actors that we examine are elected politicians and citizens.By definition, if we were to correlate populist attitudes with support for politicians and citizens as policy-makers, we would find a strong association, as attitudes towards these two actors are two of the three constituent dimensions of populist attitudes (Gherghina & Pilet, 2021).Studying the process preferences of populist voters is more interesting as the overlap between populist attitudes and populist voting is only partial (Hawkins et al., 2020;Van Hauwaert & van Kessel, 2018).In some countries under analysis, there is more than one relevant populist party (Appendix II of the Supplementary File 1).Therefore, we stacked our dataset to have each respondent providing a score for each populist party in the country.Our main independent variable is a continuous variable ranging from 0 to 100, as indicated above.However, in order to make the analysis more convincing, we carry out two further robustness tests: In the first, instead of having a continuous variable, we have dichotomised it, distinguishing between respondents who support the populist parties (score above 50) and those who do not (score below 50).In the second robustness test, we subset our sample to include only those who scored above 50 on the scale of sympathy for populist parties.The results are robust and consistent regardless of whether we use continuous or binary variables for measuring support for populist parties (Appendices VI and VII of the Supplementary File 1).
We have also included other controls in our analyses that are relevant to the study of process preferences.Indeed, in addition to the traditional socio-demographic variables (age, gender, education, and subjective income), we included as controls several political attitudes, i.e., political interest, internal efficacy, and left-right self-positioning.These variables have been commonly used (either as explanatory or as control variables) in previous analyses of process preferences, whether in support for particular models of democracy (Bedock & Pilet, 2020;Christensen & von Schoultz, 2019;Font et al., 2015;Gherghina & Geissel, 2019;Hibbing et al., 2023;Pilet et al., 2024;Webb, 2013) or their combination (Haesevoets et al., 2023(Haesevoets et al., , 2024)).The descriptive statistics and the wording of the questions can be found in Appendix III of the Supplementary File 1.

Methods
In order to detect how citizens are split according to the actors they prefer to be in government, we first ran the same factor analysis as in Pilet et al. (2024), which is based on the 21-item battery we described above.
To do so, we relied on the factanal function in R with varimax rotation, reporting Thompson's scores.
The results of the factor analysis (pooled sample and by country) are presented in Appendix IV of the Supplementary File 1.Using the same data and items as in Pilet et al. (2024), we found the exact same five dimensions: (a) citizens' capabilities as decision-makers, (b) politicians' capabilities as decision-makers, (c) transferring power to scientific experts, (d) transferring power to non-traditional actors, and (e) the nature of governing.These dimensions are robust across countries.For our analysis, we decided to exclude the fifth dimension, as we are interested only in actors in government (Table 1).
After the factor analysis, to generate our dependent variables, we extracted each respondent's loading onto each factor.Therefore, each respondent has as many factor scores as the number of dimensions identified by the factor analysis (i.e., four, as we left aside the fifth one).For the dimensions related to the political role of citizens, scientists, and non-traditional actors, a higher factor score means that the respondent is in favour of giving a greater political role to this actor (citizens, politicians, scientific actors, non-traditional actors).
For politicians, it goes in the opposite direction because the items of the survey battery are all phrased negatively.Therefore, a higher score means being more negative about elected politicians.
We ran four ordinary least square regressions with country fixed effects, in which the respondents' factor loadings on the four dimensions are our dependent variables, while the sympathy score for the populist parties is our main independent variable.Due to the stacked nature of the dataset, we clustered the standard errors at the respondent level.For all four dimensions, we distinguish between one model with all populist parties clustered together, one model where we include countries with radical left populist parties only, and one model where we include countries with radical right populist parties only.

Empirical Analysis
The main results of our analyses are presented in Table 2 (the full model specification can be found in Appendix V of the Supplementary File 1).Our first hypothesis (H1) was based on the well-established anti-elitist sentiments of populist citizens.Reflecting their negative views of politicians and political parties and their dissatisfaction with democracy in general, we expected populist supporters to hold negative views of politicians as decision-makers (and therefore to have higher factor score on that dimension-see methods section above).Our results in the first column (Politicians, All pop) confirm this expectation: A one-point increase in support for populist parties is associated with a  = 0.003 ( < 0.001) increase in the negative evaluation of politicians.This means that H1 is fully supported: The more individuals support populist parties, the more negative their view of politicians.
However, although we hypothesized that this effect would be present regardless of the host ideology of the populist party (radical left or radical right), the outcomes of the regression models distinguishing between radical left and right populist party supporters lead to a more nuanced conclusion.Indeed, we found that the effect in the main model seems to be driven mainly by supporters of radical right populist parties (Table 2, column three: Right pop).Indeed, while the coefficient is positive for both types of populist parties, it remains statistically significant only for supporters of a radical right populist party ( = 0.004,  < 0.001).This means that the support for radical left populist parties is not associated with holding negative views of politicians (Table 2, column two: Left pop), whereas the higher the support for radical right populist parties the more negative is the evaluation of the current political elite.This finding is consistent with some findings highlighting that if radical right and left populist voters share protest attitudes (distrust of politicians and parties, dissatisfaction with democracy) to a greater extent than mainstream voters, these attitudes are more strongly correlated with support for radical right than left populist parties (Goovaerts et al., 2020).
Our second hypothesis taps into a second core dimension of populism, namely people-centrism.Indeed, the literature has emphasised that populist citizens want to have a greater say in decision-making and are positive about models of democracy that give more direct power to ordinary citizens.The results of our analysis fully confirm this view and thus support H2. Indeed, Table 2 column four (Citizens, All pop) shows positive and statistically significant coefficients for the citizen dimension ( = 0.003,  < 0.001), meaning that populist party supporters tend to score higher on the ability of citizens to act as decision-makers.Moreover, this correlation is similar for radical left ( = 0.004,  < 0.001) and radical right party sympathy ( = 0.004,  < 0.001; Table 2, columns five and six: Citizens, Left pop and Right pop).This is unsurprising, as the literature has shown that radical supporters are much more supportive of direct democracy and referendums than mainstream voters (Rojon & Rijken, 2020;van Dijk et al., 2020).
Our third hypothesis rests on the debated proximity between populism and technocracy.Using theoretical grounds of techno-populism as a benchmark, we expected that populist supporters would value the suggestion of giving a greater role to independent experts.Our results show that supporting populist parties correlates positively ( = 0.003,  < 0.001) with our dimension grouping items covering shifting greater power to experts (Table 2, column eight: Experts, All pop).This means that H3 is supported.This finding may seem at odds with populist party supporters' support for citizens as decision-makers, as in a technocratic model only experts have the right to make decisions.However, these findings are in line with the techno-populist idea which suggests that strong rejection of the political elite may lead populists to also endorse technocrats (who circumvent traditional representative decision-making processes) as a solution.Nonetheless, distinguishing between radical left and right populist party supporters leads again to a more nuanced conclusion for H3.The positive coefficient ( = 0.003) on the experts' dimension is only statistically significant ( < 0.001) for the supporters of radical left populist parties (Table 2, column nine: Experts, Left pop).For radical right supporters, the coefficient remains slightly positive but significant ( < 0.1; Table 2, column 10: Experts, Right pop).To understand this, it is important to recall that two out of the three items used to measure support for experts in government were related to the power of scientific and medical experts.
In this regard, some research investigated to what extent "science populism" overlaps with "political populism": In this regard, it has been shown that anti-science opinions are strongly predicted by political conservatism and right-wing partisanship, while liberal views generally correlate with trust in science and scientists (Blank & Shaw, 2015;Eberl et al., 2023;Remsö & Renström, 2023).We believe that what we observe might reflect this specific political division in the European context (while most studies on the topic focus on the United States), which became more salient in the context of the Covid-19 pandemic (Eberl et al., 2021).
Finally, turning to H4 and preferences for giving a greater role to non-traditional actors (business leaders, military generals, and religious leaders), we expected that such actors would be more supported by respondents feeling closer to right-wing than to left-wing populist parties.Findings from Table 2, however, do not confirm this expectation.We indeed find (Table 2, column 12: Non-traditional actors, Right pop) that supporting radical right populist parties correlates positively and significantly ( = 0.006,  < 0.001) with this dimension.But we also observe the same effect among supporters of radical left populist parties ( = 0.004,  < 0.001; Table 2, column 11: Non-traditional actors, Left pop) and among supporters of populist parties in general ( = 0.005,  < 0.001; Table 2, column 10: Non-traditional actors, All pop).In other words, supporters of populist parties, both on the left and on the right, hold more positive views than the rest of the citizenry regarding giving a greater political role to business leaders, military generals, and religious leaders.
This finding regarding supporters of left-wing populist parties is, however, puzzling and hard to reconcile with earlier studies.One interpretation might relate to authoritarianism among supporters of radical left populist parties.Indeed, some studies show that left-and right-wing authoritarians do not differ extensively in terms of psychological predispositions, as they both support "conservation" values favouring security, conformity, and tradition (Federico et al., 2017).Moreover, left-wing authoritarianism predicts a taste for political violence and disruptive order (Costello et al., 2022).Finally, another line of interpretation might be the relatively soft attitudes towards secularization of radical left populist parties in countries like Greece (Syriza) or Ireland (Sinn Fein), especially once they reached power and emerged as dominant electoral forces.Hence, their supporters might be more open toward non-traditional actors than what might be expected from traditional radical left voters (Ramiro, 2016).Looking at bivariate correlations between support for left-wing populist parties and the three questions about non-traditional actors, the highest association is (surprisingly again) with religious leaders (0.15; 0.11 for business leaders and generals).All in all, this counterintuitive finding would deserve more research and refinement, as it appears difficult to fully make sense of it.

Conclusion
In the current context of a "democratic malaise" and potential illiberal backsliding, the literature on populism and how it relates to the design of democratic systems has expanded over the last years (Rovira Kaltwasser & Van Hauwaert, 2020;Wegscheider et al., 2023;Wuttke et al., 2023;Zaslove et al., 2020).Our study aimed to contribute to this debate by disentangling what supporters of populist parties want in terms of democracy, and more specifically what kind of actors they would like to play a greater role in shaping policy decisions.
Our work specifically connects recent studies on citizens' process preferences with the literature on populism and democracy (Hibbing et al., 2023;Pilet et al., 2024).It also feeds into broader contemporary debates on citizens' disenchantment and on the risk that eroding democratic satisfaction would lead to support for models of government challenging representative democracy.
This article proposes a comparative study of the preferences of supporters of populist parties (both right-wing and left-wing) across eight European democracies (Belgium, Bulgaria, Czechia, Denmark, Finland, Greece, Ireland, and the Netherlands).Their views on the political role that should be given to elected politicians, citizens, scientific experts, and non-traditional actors (business, army, and religious leaders) are compared to those of the supporters of mainstream parties.The goal was twofold: (a) to see whether supporters of populist parties have unique process preferences compared to the rest of the electorate, and (b) to compare the views of both left-wing and right-wing populist party supporters.
The first and main finding is that we indeed observed, across the nine countries covered, some specificities in how populist party supporters want government to be organized in their country.In particular, they differ from supporters of mainstream parties in two respects.First, they think that citizens are highly capable of being closely associated to policy-making.This finding confirms the importance of people-centrism as a major trait of populism (Canovan, 1999;Heinisch & Wegscheider, 2020;Jacobs et al., 2018).The second specificity of populist party supporters is more puzzling.It is that they support giving a greater role to business leaders, military generals, and religious leaders.We expected to observe it among supporters of right-wing populist parties who have been shown to hold more authoritarian and conservative political views, but we find it as well for supporters of left-wing populist parties.This last finding remains harder to explain, even if some studies have also suggested authoritarian inclinations among supporters of radical left populist parties (Federico et al., 2017).
On the other hand, we have also observed some differences in the preferences of radical right-and left-wing party supporters.When it comes to supporters of left-wing populist parties, they appear to be more in favour of giving a greater role to scientific experts in policy-making.This finding gives some credit to earlier works that have connected populist and technocratic views (Bertsou & Caramani, 2022), but they only hold for left-wing populism, not for their right-wing counterpart.
Turning to supporters of radical right populist parties, the literature suggests that they differ from mainstream voters and from supporters of left-wing populist parties in being (even) more critical of elected politicians (Rojon & Rijken, 2020).This is in line with the core definition of populist attitudes (A.Akkerman et al., 2014) as a combination of people-centric and anti-elitist views.We can confirm this finding: They are (even) more people-centric and anti-elitist, as our study also reveals that these two views on how government should function are widely shared, beyond populist party supporters.
These findings also have limitations.In particular, it is not entirely clear what the new mix of actors that would be empowered at the expense of elected politicians would look like.We may observe support for models that appear contradictory.Actually, we observe support for both more democratic developments (empowering citizens) and more authoritarian views (giving a greater role to military generals or religious leaders).Such puzzling findings highlight the need to dig deeper into populists' process preferences.In particular, the way forward seems to be to look more closely at preferences for mixed actors rather than for different actors taken separately.This approach would allow us to see more precisely what the ideal model of government might be that is preferred by supporters of populist parties across Europe.Another important next step would then be to take more global approaches to comparing the views of populist party supporters in Europe and on other continents.
Nevertheless, our results have important implications.They confirm that citizens who feel closer to populist parties, and especially to far-right populist parties, have a complex relationship with representative democracy.They are disillusioned with elected politicians and representative institutions.And they are open to giving a greater role to other actors in policy-making: citizens, but also experts.The question is whether these preferences might pose a challenge to contemporary democracies.Two conditions must be met.First, we know that supporters of populist parties have different process preferences, but it is not clear that these preferences are very high on their list of political priorities, so that they become a game changer, for example in terms of vote choice (Rooduijn, 2018).Second, populist parties should care about these preferences and push for democratic reforms within the institutions.However, this does not seem to have been very much the case in recent years (Bedock et al., 2023;Gherghina & Pilet, 2021).

Table 2 .
Ordinary least square regression with country fixed effects (standard errors clustered at respondent level).