In a global context of growing distrust in political parties and a hollowing of democracies, political parties are adapting movement frames and strategies across the world. From Africa to Latin America and across Europe, “movement parties” have combined party and movement characteristics in an attempt to reach groups that have become disconnected from institutional politics (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Kitschelt Reference Kitschelt, Katz and Crotty2006). Such parties seek to bring new issues onto the political agenda and to encourage previously disenfranchised citizens to take political action. Moreover, in countries such as the United States, where the two-party system limits the prospects of new parties, social movements have infiltrated both the Democratic and Republican parties, leading them to incorporate social movement characteristics and frames (Heaney and Rojas Reference Heaney and Rojas2015; Skocpol and Williamson Reference Skocpol and Williamson2012; Tarrow Reference Tarrow2021).
The growing cross-pollination between political parties and social movements invites social scientists to advance encompassing analyses of the processes whereby new social issues are framed, placed onto the political agenda, and, finally, sedimented into laws. While scholarship on social movements and, separately, electoral studies has each traditionally eschewed the interaction present in their objects of study (McAdam and Tarrow Reference McAdam and Tarrow2010), there have been notable productive attempts to remedy this oversight over the last decade (e.g., Borbáth and Hutter Reference Borbáth and Hutter2021; Bremer, Hutter, and Kriesi Reference Bremer, Hutter and Kriesi2020; Hutter and Vliegenthart Reference Hutter and Vliegenthart2018; McAdam and Tarrow Reference McAdam and Tarrow2010; Tarrow Reference Tarrow2021). However, most efforts have focused on understanding and conceptualizing organizational dynamics between political parties and social movements, in turn paying less attention to the demand side (i.e., of citizen support for them).
In this paper, we ask: what drives citizens to vote for movement parties in Europe? We pose this question in light of scholarship painting movement parties as an organizational innovation responding to an erosion of party linkages to society (Kitschelt Reference Kitschelt, Katz and Crotty2006). That response has encompassed a range of organizational policies predicated on a closer relation between the party leadership and its support base that would, inter alia, widen the candidate pool to include movement activists and other party outsiders (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017). Accordingly, movement parties have seized on digital technologies—for example, the Rousseau platform of Italy’s Movimento 5 Stelle (M5S)—as a means to enable candidate selection by “ordinary citizens” and, in that way, signal a step change away from more established candidate selection processes (Deseriis Reference Deseriis2020, 1770; Reference Deseriis2021). Likewise, kindred research has begun the work of characterizing movement party voters, namely as an electorate distinguished by its positive regard for protests and their influence, as well as its appetite for participating in them (Mosca and Quaranta Reference Mosca and Quaranta2017; Santos and Mercea Reference Santos and Mercea2024). Thus, this article builds on the growing understanding of the movement party organization, its candidates, and their selection, as well as of some of the proclivities of its voters, to shed light on the reasons that this specific electorate has for choosing candidates for political office.
We focus our analysis on European countries because they have been home to a variety of movement parties, including green/left-libertarian, radical right, centrist, and eclectic types (Santos and Mercea Reference Santos and Mercea2024). Using a choice-based conjoint experiment embedded in a survey of six European countries (Denmark, Germany, Hungary, Italy, Romania, and the United Kingdom), we test which attributes voters find more important when choosing to support electoral candidates (Prentoulis and Thomassen Reference Prentoulis, Thomassen, Fominaya and Feenstra2020). Building on the theoretical framework we present in the next section, we asked survey respondents to choose between two potential candidates in parliamentary national elections with dissimilar characteristics with respect to their previous involvement in nonelectoral activities, their institutional experience, their stances on what their job as representatives would be, as well as their positions regarding environmental and migration policies. We randomized the candidate characteristics, allowing us to isolate the causal effects of each attribute on respondents’ candidate choices. Importantly, we did not display the contenders’ party membership, forcing participants to make a decision based on candidates’ characteristics (Kirkland and Coppock Reference Kirkland and Coppock2018). In a final analysis, we used data on voting during the last general election in each country, obtained from elsewhere in the survey, and typologies of movement parties previously developed in the literature (Caiani and Císař Reference Caiani and Císař2019; Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Kitschelt Reference Kitschelt, Katz and Crotty2006; Santos and Mercea Reference Santos and Mercea2024) to compare the choices of conventional and movement party voters. We relied on this delineation of the electorate—which we qualify below—to understand what distinguishes movement party voters from the rest of the public.
We find that the most important feature driving movement party supporters’ voting choice is candidates’ policy positions. Moreover, our results show that movement party voters prefer anti-elitist candidates. However, surprisingly, we also find that they favor mainstream candidates who want to ensure the smooth running of the existing political system. Contrary to our expectations, candidates’ previous experiences of protest participation and whether they were institutional outsiders did not have an effect on respondents’ choices.
We conclude that support for movement parties is mainly motivated by policy and ideological factors, with movement party voters exhibiting stronger policy preferences compared to traditional party voters. Dissatisfaction with the perceived lack of decisive policy action from existing institutional elites prompts movement party supporters to seek alternative political actors promising policies aligned with their beliefs. Nonetheless, they still wish the current political system to function such that those changes can be delivered. While voter support for movement parties may not be guided directly by their roots in the protest field or their outsider position vis-à-vis institutions, the organizational strategies of these parties preserve a closer connection to social movements and their demands that in turn results in more clear-cut positions on emerging societal claims that have been shunned by established political elites. Such policy positions ultimately guide movement party support.
The paper is structured as follows. First, we explore previous scholarly work on movement parties, highlighting the little attention paid to the demand side of this phenomenon. Second, we elaborate on the possible sources of support for movement parties to formulate our hypotheses. Third, we present our novel survey data and experimental design. Finally, we explore the results of the conjoint experiment and discuss the drivers of electoral support for movement parties.
Conceptualizing Movement Parties
Movement parties are “coalitions of political activists who emanate from social movements and try to apply the organizational and strategic practices of social movements in the arena of party competition” (Kitschelt Reference Kitschelt, Katz and Crotty2006, Reference Kitschelt, Katz and Crotty280). In this way, they “straddle the conceptual space between ‘party’ and ‘movement’” (Gunther and Diamond Reference Gunther and Diamond2003, Reference Gunther and Diamond188). Parties are political groups organized to compete in elections by fielding candidates for public office (Sartori Reference Sartori1976, Reference Sartori64); social movements are networks of groups and individuals who share a collective identity, and who act with some degree of organization and continuity as they seek to effect or resist social change through nonelectoral means (cf. Della Porta and Diani Reference Della Porta and Diani2006; Snow Reference Snow, Myers and Cress2004). Importantly, however, not all authors restrict their understanding of movement parties to those political parties originating from social movements (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017). For instance, Minkenberg (Reference Minkenberg2019) contends that radical-right movement parties are less likely to emanate from social movements than their progressive counterparts. Notwithstanding this, a distinctive characteristic of movement parties is their strong link to social movements and their practices (Kitschelt Reference Kitschelt, Katz and Crotty2006).
Equally, movements have permeated established parties. As the party organization became increasingly disconnected from civil society, some authors suggested that party elites became office maximizers to the detriment of their “representative function” (Cohen Reference Cohen2019, 1092). Against this backdrop, the “movementization” of established parties represented the transfer of movement-like characteristics by factions seeking to rejuvenate the party through renewed connections to civil society (Cohen Reference Cohen2019, 1093). The UK Labour Party under Jeremy Corbin was an example of a long-standing party that embedded a transient, movement-like organization in its fold (Dennis Reference Dennis2019). While relatively short-lived in the broader political arc of the Labour Party, the movement–party nexus has remained an enduring feature of parties that emerged as counterweights to the status quo—a feature that has been entrenched by movement parties (Butzlaff Reference Butzlaff2023; Hutter, Kriesi, and Lorenzini Reference Hutter, Kriesi, Lorenzini, Snow, Kriesi, Soule and McCammon2018; Peña Reference Peña2021).
Even though scholarship on movement parties extends over more than two decades, social scientists will readily notice that existing academic studies have concentrated on the “supply side” of their relationship with the electorate, paying more attention to party characteristics and their place in the party system than to what leads citizens to support them. Following the debate started by Katz and Mair’s (Reference Katz and Mair1995; Reference Katz and Mair1996) cartel party thesis, in which these scholars noted an increasing estrangement of political parties from civil society, movement parties have been conceptualized as innovative organizational attempts to rejuvenate linkages between institutional politics and society (Kitschelt Reference Kitschelt, Katz and Crotty2006; Santos and Mercea Reference Santos and Mercea2024). As such, they represent a drive to ease citizen participation in the party organization as well as a new avenue for channeling demands by civil society into party programs (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017).
Thus, first, some authors have presented movement parties as the outcome of strategic decisions by political entrepreneurs to overcome problems of social choice (Kitschelt Reference Kitschelt, Katz and Crotty2006). Movement parties make little investment in formal internal structures. Requirements to enter the party are low (Deseriis Reference Deseriis2020; Mosca Reference Mosca2020), to the extent that it is sometimes unclear to individuals in the networks contiguous with these parties whether they are part of the party or not (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017). Progressive movement parties emphasize bottom-up decision making through assemblies and other types of grassroots initiatives (Kitschelt Reference Kitschelt1988; Reference Kitschelt, Katz and Crotty2006), while radical-right parties utilize plebiscitarian forms of engagement with their membership that solicit its allegiance to the positions of the leadership (Pirro and Gattinara Reference Pirro and Gattinara2018).
Second, scholars who have examined movement party repertoires emphasize that, in addition to party-like actions, these parties also engage in activities associated with social movements. Movement parties have pushed for direct democracy, which they have embraced in their own internal party organization through internet technologies (Deseriis Reference Deseriis2020; Mosca Reference Mosca2020). Additionally, movement parties may organize demonstrations and other types of activities outside institutions to show strength and gather support (Borbáth and Hutter Reference Borbáth and Hutter2021; Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Pirro and Gattinara Reference Pirro and Gattinara2018). Despite this, Mosca and Quaranta (Reference Mosca and Quaranta2017) have warned against equating movement parties with protest parties and understanding support for them—referred to in their paper as “the vote of the protesters”—as a protest vote. More recent research has indicated that while radical-right movement party voters share common characteristics with protest voters, the rest of movement party voters display distinct attitudes (Santos and Mercea Reference Santos and Mercea2024).
Nonetheless, their connection to the grass roots arguably allows movement parties to be better able to detect emerging changes in societal demands and to adapt their programs quickly (Santos and Mercea Reference Santos and Mercea2024). Indeed, some observers link the prominence of movement parties—notably, those parties that have gained ground since the Great Recession of 2008 and the economic austerity following it—back to a “crisis of representation” (Hutter, Kriesi, and Lorenzini Reference Hutter, Kriesi, Lorenzini, Snow, Kriesi, Soule and McCammon2018). That crisis manifested itself as a series of popular mobilizations by people disenchanted with the response of democratic institutions and the political actors leading them to the plight of ordinary citizens. Movement parties, these authors argue, have come to embody new cleavages—for example, the cleavages separating the “winners” and “losers” of globalization (Hutter, Kriesi, and Lorenzini Reference Hutter, Kriesi, Lorenzini, Snow, Kriesi, Soule and McCammon2018, Reference Hutter, Kriesi, Lorenzini, Snow, Kriesi, Soule and McCammon324–Reference Hutter, Kriesi, Lorenzini, Snow, Kriesi, Soule and McCammon26)—which stand them in contrast to established parties. However, as parties adapt to positions inside the institutional arena, movement activities may become less prominent in their repertoires (Deseriis Reference Deseriis2020).
Third, other observers have paid attention to how movement parties bridge their party and movement ethos through their communication strategies. While movement parties use complex communication strategies on multiple platforms (Della Porta Reference Della Porta2021), it has been noted that many use “populist communication”—simplifying complex policy matters, and emphasizing a monolithic understanding of a “people” opposed to “the elites” who have failed to represent them (Engesser, Fawzi, and Larsson Reference Engesser, Fawzi and Larsson2017). Moreover, movement parties are avid users of unmediated forms of communication, such as social media, with the objective of connecting directly with their electorate (Mercea and Mosca Reference Mercea and Mosca2021). Indeed, empirical research has found that active social media users are more likely than others to vote for movement parties such as M5S and La France Insoumise (Mosca and Quaranta Reference Mosca and Quaranta2021).
Fourth, given their greater ideological capacity to incorporate emerging social issues, movement parties have tended to appear at critical historical junctures when key policy issues that are thrust forward are neglected by the electoral establishment (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017). During these periods, the greater flexibility of social movements has allowed them to be at the forefront of social change, representing new positions on emerging issues (Kriesi et al. Reference Kriesi, Koopmans, Duyvendak and Giugni1995). Some observers have therefore submitted that movement parties “are more likely to be driven by ideological militancy than by pragmatic political considerations” (Tarrow Reference Tarrow, Porta and Diani2015, Reference Tarrow, Porta and Diani95) because they look to place claims advanced by social movements within the institutional arena that lacked sustained representation prior to these new parties entering the political fray. In this way, while some movement parties share common characteristics with other types of parties, they represent a distinct category.
Yet other authors have portrayed movement parties as a transient response to a demand for collective action that permeates institutional politics in the early stages of their formation (Kim Reference Kim2023), highlighting their status as political challengers to conventional parties. Nonetheless, one of the characteristics of challenger parties is their lack of experience in government (De Vries and Hobolt Reference De Vries and Hobolt2020), which is not always the case for all movement parties. Some movement parties—such as the German Greens (Bündnis 90/Die Grünen), the Romanian Uniunea Salvaţi România (USR), and M5S—have been in government, in coalition with established parties. Moreover, despite the aforementioned use of populist communication by many movement parties, not all of them can be considered populist parties per se.
Fifth, it is notable that several movement parties from across the ideological spectrum—including green parties, centrist movement parties such as Romania’s USR and Hungary’s Momentum, as well as the Danish radical-right movement party Nye Borgerlige—are not identified as populist parties in PopuList, the widely used typology of European populist parties developed by Rooduijn and colleagues (Reference Rooduijn, Andrea, Halikiopoulou, Froio, Van Kessel, De Lange, Mudde and Taggart2023). Ultimately, while being depicted as a transient phenomenon (Kitschelt Reference Kitschelt, Katz and Crotty2006), movement parties have been able to endure the test of time, representing a distinct organizational form that has permeated a diversity of party types (Tarrow Reference Tarrow, Porta and Diani2015, Reference Tarrow, Porta and Diani95).
Movement Parties in the European Context
In recent years, we have witnessed examples of parties emanating from social movements and social movements infiltrating political parties across the world. In the US, the Tea Party movement boosted the ranks of the Republican Party and remade American conservativism, moving its policies toward the Far Right (Skocpol and Williamson Reference Skocpol and Williamson2012). Moreover, the Tea Party played a key role in Donald Trump’s victory in the 2016 Republican primaries and in the presidential election later that year (Herman and Muldoon Reference Herman and Muldoon2018). On the other side of the political divide, the Democratic Party has also been heavily influenced by social movements. The antiwar movement against the conflicts that the US started in Iraq and Afghanistan played a key role in the election of Barack Obama as the 44th US president (Heaney and Rojas Reference Heaney and Rojas2015). Furthermore, Occupy Wall Street reinvigorated the left wing of the party, strengthening the party’s connections with the grass roots and advancing progressive candidates such as Bernie Sanders and Alexandria Ocasio-Cortez, who have commanded considerable party influence and social support (Day and Uetricht Reference Day and Uetricht2021; Heaney and Rojas Reference Heaney and Rojas2015). Despite the impressive influence that movement parties have had in the US as well as in Latin American politics (Laurent Reference Laurent2022), it is in Europe where one can find the greatest variety of ideologies and strategies. Due to that diversity and the specific focus of this paper, in this section we place the spotlight on that region.
In Europe, since the turn of the century, movement parties have risen to prominence in all corners of the continent. Rooted in mass mobilizations that took place during periods of deep social transformation, movement parties have foregrounded the movement as a platform for rebuilding links with civil society. In eastern, western, as well as southern and northern Europe, such parties have sought to reimagine the party organization. They have done this through the innovative use of the internet, giving alienated voters new and greater opportunities for political involvement and for the representation of policy positions largely falling outside the political mainstream and well-rehearsed ideological alignments (Dragoman Reference Dragoman2021; Heinze and Weisskircher Reference Heinze and Weisskircher2021; Reference Heinze and Weisskircher2022; Husted Reference Husted2020; Mosca Reference Mosca2020; Pirro and Róna Reference Pirro and Róna2019). Furthermore, these parties share an anti-elitism to which they counterpose not only a rejuvenated support base that includes newly mobilized disillusioned voters (Mosca and Quaranta Reference Mosca and Quaranta2021; Herman and Muldoon Reference Herman and Muldoon2018), but also protocols for the participatory development of party manifestos or the selection of party candidates and leaders (e.g., primaries) (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Gherghina and Grad Reference Gherghina and Grad2021; Höhne Reference Höhne2021; Husted Reference Husted2018; Muldoon and Rye Reference Muldoon and Rye2020).
Occupying electoral spaces opened up by economic modernization and crisis, globalization, a realigning of sociocultural cleavages, or international migration (Dragoman Reference Dragoman2021), movement parties have embodied a “systematic opposition to mainstream politics” (Pirro Reference Pirro2018, Reference Pirro446). They exemplify an electoral expression of a break with entrenched politics along a spectrum of issues, and with the political elite espousing them. As such, green/left-libertarian movement parties are rooted in societal cleavages that originated during the critical junctures of the 1968 protests that took place worldwide (Kitschelt Reference Kitschelt1988) and the mass mobilizations that unfolded after the 2008 financial crisis (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017). They have challenged the postwar institutional settlement between capital and labor, highlighting the ecological and societal costs of the economic growth underpinning it (Kitschelt Reference Kitschelt1988).
Similarly, more eclectic movement parties, such as the Italian M5S, have been a vehicle for questioning the political system, a process that expanded following the 2008 crisis of neoliberalism (Mosca Reference Mosca2014; Mosca and Tronconi Reference Mosca and Tronconi2019). Furthermore, some authors have argued that political organizations of the new Far Right, including movement parties such as the German Alternative für Deutschland (AfD) and the Hungarian Jobbik, have been shaped by the sociocultural transformations commencing in 1968 in western Europe, and the collapse of state communism in central and eastern Europe after 1989 (Karsai Reference Karsai and Ramet1999; Pirro Reference Pirro2019). Finally, centrist movement parties championing anti-corruption frames, such as Momentum in Hungary and USR in Romania, have appeared across central and eastern Europe as a critique of the unfinished reforms during the transition from state communism to democratic systems and market economies (Dragoman Reference Dragoman2021).
Another common feature of movement parties across Europe is the charge against the establishment that has been spearheaded by protest movements. In southern Europe, M5S channeled an eclectic assortment of concerns, roused through popular mobilization, to instrumentally cut across left–right dichotomies—concerning the neoliberal economic consensus, postmaterialist cultural values, and immigration, for example—for electoral gain (Pirro Reference Pirro2018, Reference Pirro445). In Romania, USR was founded by activists whose call for a fresh approach to stemming corruption in public life, voiced through mass demonstrations, was free from any ideological commitments other than an “anti-system orientation” (Dragoman Reference Dragoman2021, Reference Dragoman310). Likewise, in Denmark, Alternativet drew a stark contrast between itself and established political parties responsible for a “crisis of representation” (Prentoulis and Thomassen Reference Prentoulis, Thomassen, Fominaya and Feenstra2020, Reference Prentoulis, Thomassen, Fominaya and Feenstra354).
Antiestablishment frames were, moreover, combined with nativism and anti-immigrant discourses by far-right movement parties. In Hungary, Jobbik fused extrainstitutional movement mobilization in mass protests with party organization, producing an electoral platform for antiestablishmentarianism, anti-globalism, and anti-corruption (Pirro Reference Pirro2019). In Germany, the AfD similarly espoused a radical right, anti-immigrant, anti-refugee, and anti-Islam agenda (Arzheimer and Berning Reference Arzheimer and Berning2019). While initially a political vehicle of an alienated, Eurosceptic, socially conservative elite challenging economic liberalism, the AfD nurtured relations with social movements ranging from the anti-immigrant Pegida (Weisskircher and Berntzen Reference Weisskircher, Berntzen, Caiani and Císař2019) to the Querdenken movement opposing COVID-19 public health restrictions (Heinze and Weisskircher Reference Heinze and Weisskircher2022).
Beyond exploring the characteristics, strategies, and typologies of movement parties, as illustrated above, we need a better understanding of what drives support for these political platforms. In the context of declining conventional parties and falling participation—both with regard to party and trade union membership as well as to voting in elections (Mair Reference Mair2013, Reference Mair201)—movement parties have successfully mobilized citizens at the voting booth as well as on the streets. However, there is a dearth of empirical research on what drives support for movement parties (but see Mosca and Quaranta Reference Mosca and Quaranta2021; Passarelli and Tuorto Reference Passarelli and Tuorto2018). We review it in the following section.
The Drivers of Support for Movement Parties
To date, insights into movement party voting are drawn principally from observational and country-specific studies. Beyond studies focused on individual parties, in this section we draw on a body of work to delineate expectations about movement party voters’ preferences. The literature on party candidates shows that candidates’ previous experience influences whether voters support them. In the absence of party labels, voters use other informational shortcuts to choose a candidate (Bullock Reference Bullock1984; Kirkland and Coppock Reference Kirkland and Coppock2018). For instance, in their study, Kirkland and Coppock (Reference Kirkland and Coppock2018) show that in the US, Republicans give greater preference to candidates’ previous work experience while Democrats prioritize their political career. As movement parties are less likely to have governmental experience and often present themselves as institutional outsiders (Borbáth and Hutter Reference Borbáth and Hutter2021; Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017), we posit the following hypotheses:
H1a: Movement party voters have a stronger preference for candidates with no institutional experience than those who do not vote for movement parties.
H1b: Movement party voters have a weaker preference for candidates with institutional experience than those who do not vote for movement parties.
Furthermore, movement parties have been depicted as beneficiaries of “the vote of the protesters” (Mosca and Quaranta Reference Mosca and Quaranta2017). Movement parties have recruited numerous activists and protesters to their ranks (Caiani and Císař Reference Caiani and Císař2019; Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017). They have often placed social movement leaders at the top of their candidate lists or designated them as spokespeople (Dragoman Reference Dragoman2021; Martínez and Wissink Reference Martínez and Wissink2022; Tournier-Sol Reference Tournier-Sol2021). Moreover, leaders of these parties often emphasize their connection to the grass roots, either discursively (Della Porta Reference Della Porta2021), by joining protesters during their mobilizations (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Dragoman Reference Dragoman2021; Passarelli and Tuorto Reference Passarelli and Tuorto2018; Pirro and Róna Reference Pirro and Róna2019), or by creating strategic alliances to shape policies (Bazurli Reference Bazurli2019). Unsurprisingly, then, other studies have suggested that protest participation increases the likelihood of voting for movement parties (Mosca and Quaranta Reference Mosca and Quaranta2017). Consequently, we hypothesize the following:
H2a: Movement party voters have a stronger preference for candidates with protest experience than those who do not vote for movement parties.
H2b: Movement party voters have a weaker preference for candidates with no experience of extrainstitutional participation than those who do not vote for movement parties.
Beyond candidates’ experience, we expect that the ideology of movement parties also has an influence on the support they receive from voters. Building on the idea of populism as a “thin ideology” (Mudde Reference Mudde2004; Stanley Reference Stanley2008; Urbinati Reference Urbinati2019), observers have highlighted the anti-elitism of some movement parties. The M5S vote was depicted as a “protest vote” insofar as it represented the electoral expression of dissatisfaction with mainstream parties and, as such, was an anti-elitist vote (Passarelli and Tuorto Reference Passarelli and Tuorto2018). This insight tallies with the assessment of the rise in the vote share of USR from 2016 to 2019, following a large wave of anti-corruption protests in Romania (Dragoman Reference Dragoman2021), a focal topic for anti-elitist political rhetoric, in east-central Europe (Engler Reference Engler2020). Therefore, we expect anti-elitism to have an influence on the propensity to vote for movement parties.
H3a: Movement party voters have a stronger preference for anti-elitist candidates.
H3b: Movement party voters have a weaker preference for mainstream candidates.
At the same time, although researchers have pointed to similarities among movement parties, they have also recognized that policies and ideology set them apart (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Kitschelt Reference Kitschelt, Katz and Crotty2006; Pirro Reference Pirro2019). Concerns for the environment and immigration are key issues that have galvanized new political actors on the opposite ends of the ideological spectrum in recent decades (Hutter and Borbáth Reference Hutter and Borbáth2019). Previous work has noted that radical-right voter support is mostly driven by nativist and anti-immigration attitudes (Kirkizh, Froio, and Stier Reference Kirkizh, Froio and Stier2022). For instance, both Germany’s AfD and Hungary’s Jobbik mobilized xenophobic attitudes and restrictive policies directed at immigrants, especially those from majority-Muslim countries (Arzheimer and Berning Reference Arzheimer and Berning2019; Pap and Glied Reference Pap and Glied2018). Accordingly, we make the following hypothesis:
H4a: The issue preferences of radical-right movement party voters are skewed toward restrictive migration policies.
Green/left-libertarian party voters, by contrast, place environmental issues at the core of their agenda (Abou-Chadi Reference Abou-Chadi2016; Borbáth and Hutter Reference Borbáth and Hutter2021; Van Haute Reference van Haute2016). Equally, environmental issues are integral to a more complex set of progressive policy preferences (Kirkizh, Froio, and Stier Reference Kirkizh, Froio and Stier2022). Even political parties not commonly placed under the green banner, such as the UK Labour Party under Jeremy Corbyn and Denmark’s Alternativet, advanced a clear progressive agenda, placing environmental issues at the center of their demands (Husted Reference Husted2018; Pickard Reference Pickard, Pickard and Bessant2018). Hence, we anticipate the following:
H4b: The issue preferences of green/left-libertarian movement party voters are skewed toward pro-environmental policies.
With regards to centrist and eclectic movement parties, both have rejected positioning themselves along preestablished political divides. For instance, it has been argued that a centrist movement party such as USR formed as a medium for an “anti-system orientation” expressed through anti-corruption claims, seeking to appeal to voters across ideological lines (Dragoman Reference Dragoman2021, Reference Dragoman310). Similarly, authors have posited that the Italian M5S combines progressive positions on the economy and international humanitarianism with conservative discourses on securitization and migration, wrapped up in an antiestablishment rhetoric (Mosca and Tronconi Reference Mosca and Tronconi2019). For these reasons, we expect the following:
H4c: Environmental and migration policies do not have an effect on the issue preferences of centrist and eclectic movement party voters.
H4d: Centrist and eclectic movement party voters have a stronger preference for anti-elitist candidates than voters for other movement parties.
Data and Methods
To test the aforementioned hypotheses, we employed a preregistered conjoint experiment embedded in a survey fielded in Denmark, Germany, Hungary, Italy, Romania, and the UK.Footnote 1 Conjoint experiments consist of presenting respondents with a series of forced choices between two candidates with a variety of attributes that are randomized simultaneously. In contrast to analyses using observational data, which most commonly look at correlations between an outcome variable of interest and a number of predictors, conjoint experiments allow researchers to approximate real-life, multidimensional scenarios in which individuals’ decisions are driven by a multiplicity of factors. For this reason, they have been used to study complex social predispositions in relation to a variety of topics, including drivers of support for political candidates (e.g., Lemi Reference Lemi2021; Ono and Yamada Reference Ono and Yamada2020), police officers’ biases in the selection of cases to investigate (e.g., Boittin, Fisher, and Mo Reference Boittin, Fisher and Mo2024), and preferences for immigrant profiles (e.g., Hainmueller and Hopkins Reference Hainmueller and Hopkins2015), among other things. Moreover, given the randomized nature of the data presented, conjoint designs allow researchers to disclose hidden preferences in the population and interpret them causally (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014).
We used the diverse case selection method (Seawright and Gerring Reference Seawright and Gerring2008, Reference Seawright and Gerring297) for our choice of countries and parties. Party linkages to citizens have continued to erode in late twentieth and early twenty-first century Europe as parties have become increasingly professionalized electoral vehicles with declining memberships (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Mair Reference Mair2013). In that environment, movement parties have come to perform similar roles in various countries, bringing new issues to the fore of the political agenda (Schwartz Reference Schwartz, Thomspon, Birnir, Kuhonta, Lewis and Smyth2016). Yet, within this broad climate, these six countries represent distinct electoral and party systems in Europe with varying electoral rules and thresholds, as well as differing alliance and cooperation strategies among parties (Sitter Reference Sitter2002) and contrasting levels of electoral support for conventional (e.g., conservative or social democrat) and new (e.g., green or far-right) party families (Bolleyer and Correa Reference Bolleyer, Correa, Bukow and Jun2020, Reference Bolleyer, Correa, Bukow and Jun45).
Furthermore, country selection was informed by our aim to cover the distinct political cultures that previous research has identified in northwestern, southern, and eastern Europe (Borbáth and Gessler Reference Borbáth and Gessler2020). Within these regions, we selected countries that allow us to capture the great variety of movement parties present in Europe (Weisskircher, Hutter, and Borbáth Reference Weisskircher, Hutter and Borbáth2022). Movement parties in our case-countries span a broad ideological spectrum from the radical Right (the AfD, Jobbik) through centrism (USR) to green/left libertarianism (Alternativet, Bündnis 90/Die Grünen) and a recombinant eclecticism in the case of M5S. Moreover, some of these parties have been in government in Germany (Bündnis 90/Die Grünen), Italy (M5S), and Romania (USR), or have preserved a challenger status outside government and in opposition to it, with different degrees of success (AfD in Germany; Alternativet and Nye Borgerlige in Denmark; Jobbik, Lehet Más a Politika [LMP], and Momentum in Hungary; and the United Kingdom Independence Party [UKIP] in the UK). Most importantly, as highlighted above, demonstrations and relations with social movements have enabled movement parties in the six countries to develop linkages with civil society and an electoral base.
While movement parties may be a transient phenomenon (Kitschelt Reference Kitschelt, Katz and Crotty2006) and voter preferences may vary over time, for the purpose of this analysis we categorize as movement party voters those individuals who casted their ballot in support of a movement party during the most recent general election in their country. The parties identified as movement parties in each country, based on categorizations previously suggested in the literature (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Kitschelt Reference Kitschelt, Katz and Crotty2006; Pirro and Gattinara Reference Pirro and Gattinara2018; Santos and Mercea Reference Santos and Mercea2024), are presented in table 1, while a table with the parties labeled as conventional can be found in section 3 of the appendix. As can be observed, our analysis includes movement parties from different ideologies (left, center, and right wing) as well as different levels of institutionalization and success (ranging from Bündnis 90/Die Grünen, the USR, and M5S, which have experience in government, to Sinn Féin, which has been part of the government of Northern Ireland, and Jobbik, which has governed some city councils, to most other parties that have never occupied any positions of institutional power). Hence, our cases provide a comprehensive view of the diversity of movement parties in Europe. On this note, while our case selection covers a range of countries and parties, it should be noted that we restrict our analysis to the European context and three broad types of movement parties. Therefore, the results and conclusions presented in this paper should be interpreted as bounded by these scope conditions.
Selection of Movement Parties Included in the Study
The Cross-National Survey
Surveys were fielded online by the public opinion company YouGov between February 21, 2022, and March 11, 2022, using YouGov’s “active sampling” methodology. While some studies have pointed to issues of selection bias for internet-based surveys (Elliott and Valliant Reference Elliott and Valliant2017; Ferri-García and Rueda Reference Ferri-García and del Mar Rueda2020), Miratrix and colleagues (Reference Miratrix, Sekhon, Theodoridis and Campos2018) have shown through simulations and experiments administered online by YouGov that their survey data offers a representative picture of populational attitudes that alleviates the need to use population weights for survey experiments.Footnote 2
We surveyed a total of 10,347 respondents (1,001 in Denmark, 2,024 in Germany, 2,051 in Hungary, 2,101 in Italy, 946 in Romania, and 2,224 in the UK). Variations in sample sizes and length of fieldwork arise from YouGov’s methodology for procuring nationally representative samples for their online panels in different countries. Additionally, during survey setup, we utilized an English-language master version of the questionnaire, which YouGov then translated into the countries’ respective languages using its internal team. These translations underwent subsequent review by native speakers with academic and social science backgrounds to ensure maximum similarity of questions and response items across countries. In Romania, we offered Romanian and Hungarian versions of the questionnaire to respondents that the platform identified as connecting from Romania’s central and northwestern regions, as well as to those connecting from Bucharest, the capital city. In all three regions there is a significant proportion of Hungarian speakers. More details about the demographics against which YouGov created the survey samples and how they compare to those in our sample, as well as further information about the survey methodology, can be found in section 1.2 of the appendix.
The Conjoint Experiment
As part of this survey, we included a preregistered choice-based conjoint experiment (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014; Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020) in which respondents were asked to make a forced choice—that is, a binary selection between one of two hypothetical candidates in parliamentary elections. Each respondent was presented with five tasks of this type, resulting in a total of 51,735 observations (Mercea & Santos Reference Mercea and Santos2024).
At the beginning of the exercise, all respondents read the following introduction:
Please imagine that there are national parliamentary elections taking place next week. We would like to show you 5 pairs of profiles of potential candidates for entering into parliament. You will see 5 similar screens, one after the other, where candidates will have different attributes. It is important that you read the descriptions of each candidate carefully so, please, take your time. We would like to know, based on your preferences, which one of the two candidates you would prefer to have representing you in [country’s] parliament. You may have to choose between two candidates that you do not fully agree with but we would like to ask you to make an effort to think which candidate of the two you would prefer nonetheless. People have different preferences on these issues, and there are no right or wrong answers.
The aim of this introduction was to accommodate electoral systems divided into single-member districts, such as that used by the UK, alongside systems in which party competition revolves around candidate lists. Across the electoral systems in our country cases, candidates are always clearly identifiable, either because they stand in single-member districts under their party’s banner or because they are leading their party list in their electoral district.
For respondents, each task involved making a forced choice among two candidates on the basis of five attributes. Each attribute could take one of three levels, which were randomly allocated for both candidates and for each task. The attributes and possible levels are presented in table 2. First, we included an attribute on the candidates’ experience in institutional politics. The levels in this attribute included experience in government, experience in institutional positions but not in government and, finally, no institutional experience. Second, with regards to candidates’ experience outside institutions, in addition to protest experience and no previous participation in nonelectoral activities, we included a level for experience in a noncontentious, nonelectoral activity. Third, we presented an argument on the candidates’ reasons for running using anti-elitist, mainstream, and neutral framing. Finally, we included two policy positions that are salient for green/left-libertarian and radical-right movement parties. The primary issue for green/left-libertarian movement parties was environmental policies. Candidates could propose more ambitious climate change policies, a reversal of climate change policies, and the upholding of the status quo. As for the radical Right, its salient issue was migration policies. Candidates could propose to open or close the country’s borders to most migrants, or to maintain current policies.
Candidates’ Attributes and Levels
Attribute levels were designed to ensure that all possible combinations were plausible. Hence, there were no limitations to attribute combinations, allowing for a complete randomization. To eliminate the possibility that respondents’ selections were dependent on the order in which attributes were presented, we randomized the order of attributes in the first task and maintained that order through the whole exercise to minimize respondents’ confusion. Finally, we categorized respondents according to how they voted in the latest national parliamentary elections.
In addition to the forced choice, we invited respondents to evaluate each of the candidates on a one-to-seven scale, where one indicated they strongly disapprove of the candidate and seven that they strongly approve of the candidate. We used this measure to perform several robustness checks. First, we considered whether the comparison between the ratings of the two candidates was in line with the forced-choice selection. In only 7.31% of cases, the candidate selected in the forced choice had a lower rating than the candidate who was not selected. Second, we ran the same analysis on two subsets of the data: one in which we exclude the observations in which the candidate selected in the forced choice has a lower rating than the candidate not selected, and another where we also exclude instances in which respondents gave an equal rating to both candidates. Third, we ran our analysis on the whole sample but did so using the rating of each candidate as an outcome, instead of the binary choice. All these analyses, which we report in section 6 of the appendix, confirm the robustness of our results and give us confidence in the quality of our data and the strong empirical pillars of our conclusions. Over and above these robustness checks, we also report the country models in section 7 of the appendix. They show by-and-large similar results across the countries.
Results
Following recent debates about how to analyze conjoint survey experiments, we estimated results using marginal means (MMs) (Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020). We prioritized the calculation of marginal means over average marginal component effects (AMCEs) (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014), as the former provide a more straightforward interpretation of subgroup differences in preference (Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020). MMs reflect the marginal probability that an individual profile is selected, given that a specific attribute level is present. An MM of 1 means that a candidate with that feature was always selected by respondents. Conversely, an MM of 0 means that respondents never selected a candidate who attained such a level. Instead of using a reference category as is the case in analyses using AMCEs (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014), an MM of 0.5 indicates that the level was not significant for respondents, as when that characteristic appeared in the task there was an equal probability of selecting either of the two candidates. To compute MMs, we used the R package cregg (Leeper and Barnfield Reference Leeper and Barnfield2020). As we established no limitations to attribute-level combinations, meaning that each attribute is randomized independently of all other attributes, the estimation of treatment effects can be interpreted causally. For ease of interpretation, and in line with how previous scholarship has reported conjoint experiments (e.g., Kirkizh, Froio, and Stier Reference Kirkizh, Froio and Stier2022; Kirkland and Coppock Reference Kirkland and Coppock2018; Leeper, Hobolt, and Tilley Reference Leeper, Hobolt and Tilley2020), the results are plotted in several figures. Tables presenting the same data in numerical form are included in section 4 of the appendix.
Figure 1 shows the effect of each attribute on the probability of selecting a candidate by movement party voters and conventional party voters, respectively. First, contradicting our first hypothesis, it shows that candidates’ institutional experience does not significantly influence movement party voters’ candidate selection. Conventional party voters, however, pay attention to this attribute. As such, whereas conventional party voters prefer candidates who either have no institutional experience whatsoever or who have experience in government, movement party voters do not seem to place much importance on whether candidates are outsiders to institutions. Equally, conventional party voters do not favor those who have institutional experience but no experience in government.
Marginal Means of Candidates’ Levels for Movement Party and Conventional Party Voters
Second, results for the attribute on extrainstitutional experience are surprising and go in the opposite direction to what we hypothesized in H2. While movement party voters are slightly more likely to prefer candidates who have volunteered in their community, doing so 52% of the time, candidates who have participated in protests are marginally less likely to be selected (49% of the time). This finding indicates that, while protest participants are more likely to vote for movement parties (Mosca and Quaranta Reference Mosca and Quaranta2021), overall, if anything, movement party voters tend not to favor candidates who participate in protests, displaying a similar negative preference to conventional party voters.
Third, and just as notably, both a mainstream reason (“to support the smooth running of the current political system”) and an anti-elitist reason (“because corrupt elites too often ignore ordinary citizens”) for running increase the probability that movement party voters select a candidate. These results invite an unexpected reflection. Movement party voters are clearly anti-elitist, as the probability of them selecting an anti-elitist candidate is around 54%. However, they also seem to want the current political system to work, favoring pro-systemic candidates 52% of the time. Hence, support for movement parties should not be interpreted as a wholesale vote against the current political system. Rather, movement party voters appear to point to political elites as representing a core problem plaguing contemporary democracy. While movement party voters perceive a crisis of representation, they do not appear to see this problem as systemic. Instead, they seem to believe that political representatives, rather than institutions, are the ones failing them. Such voters therefore turn to fresh faces in the hope that new candidates challenging current elites can make the current democratic system work.
Another interesting finding from this attribute is that the preferences of conventional party voters are the same as those of movement party voters, albeit less pronounced. Conventional party voters also express a preference for candidates who display either mainstream or anti-elitist reasons for running. Therefore, there seems to be an overall desire for the current political system to run smoothly and a general perception that contemporary institutional leadership does not pay sufficient attention to societal needs. Finally, as this preliminary analysis groups together all parties irrespective of ideology, even if some results regarding policy positions seem significant, they are, a priori, not theoretically meaningful. A finer-grained analysis dividing movement party voters by party type is presented below.
The Importance of Institutional Experience for Movement Party Voters
The following figures display the marginal means of each group of voters for each attribute in the experiment. In these plots, red lines represent movement party voters, while blue lines represent conventional party voters. Abstainers are represented in black. Each symbol represents a different party type: a cross (X) is used for right-wing parties, a vertical line (|) for left-wing parties, and an asterisk (*) for centrist and eclectic parties.
Focusing first on the preferences for candidates’ institutional experience, Green/left-libertarian movement party voters, on average, prefer institutional outsiders (figure 2). On average, candidates who have no institutional experience are selected 52% of the time by green/left-libertarian movement party voters. Moreover, the probability that candidates with experience in institutions but not in government are selected by these voters is only 48%, while there is no effect of having government experience on the likelihood of receiving support from green/left-libertarian movement party voters. These preferences diverge from those of left-wing conventional party voters, who have a marked preference for candidates with institutional and governmental experience and no significant preference for institutional outsiders. The only converging attitude on this matter is that both groups of voters display a negative inclination toward candidates with institutional but no governmental experience.
Marginal Means of Candidates’ Institutional Experience by Party Vote
Voters of radical-right movement parties prefer candidates with government experience around 52% of the time. Moreover, they show no preference for institutional outsiders. Finally, these voters also show no significant inclination toward candidates who are institutional outsiders or who have been involved in institutional politics but never in government. Interestingly, conventional right-wing voters behave according to the nonconfirmed expectations we had about radical-right movement party voters, as conventional right-wing voters are significantly more likely to select candidates with no institutional experience.
On average, voters of centrist and eclectic movement parties have a negative preference for institutional outsiders. The probability that they select candidates who have no institutional experience is around 47%. Additionally, they have no significant preference for candidates with institutional experience, independent of whether they have been in government. Again, conventional centrist party voters display the attitude we expected from their movement party counterparts and have a significant and positive preference for institutional outsiders.
In conclusion, ideology matters to whether movement party supporters prefer institutional outsiders to represent them. These results indicate that H1a and H1b should be nuanced. Regarding H1a, no group of movement party voters shows, on average, a stronger preference for candidates with no institutional experience. Voters of green/left-libertarian movement parties are more likely to shun candidates who have been in institutions but never in government. Contradicting H1b, radical-right movement party voters show a significant preference for candidates with governmental experience.
Do Movement Party Voters Want to Be Represented by Protesters?
Figure 3 displays the probabilities for selecting candidates based on their extrainstitutional experience. Despite previous research arguing that movement parties are “the vote of protesters” (Mosca and Quaranta Reference Mosca and Quaranta2017, Reference Mosca and Quaranta431), based on the finding that protest participation increases the likelihood of voting for movement parties, it seems, on average, that movement party voters disregard candidates’ protest experience. Nonetheless, this attitude differentiates movement party voters from conventional party supporters who have a negative preference, across ideologies, for candidates with protest experience. Furthermore, movement party voters, on average, neither favor nor are averse to candidates with no extrainstitutional experience. In this case, the attitudes of movement and conventional party voters are more aligned, as the only group that differs from this pattern are left-wing conventional party voters, who display a negative preference for candidates with no institutional experience.
Marginal Means of Candidates’ Extrainstitutional Experience by Party Vote
The only extrainstitutional engagement by candidates preferred by movement party voters is volunteering in the local community. Radical right as well as centrist and eclectic party voters have a significant preference for this level (53% and 52%, respectively), while green/left-libertarian movement party voters have no significant opinion on this matter. This feature is also preferred by conventional party voters of all ideologies. Altogether, then, it seems that, on average, the kind of extrainstitutional experience favored by movement party voters, as well as by all conventional party voters, is noncontentious community involvement.
These results contradict our expectations derived from H2a and H2b, as they show that extrainstitutional experience has a significant effect on a small proportion of movement party voters. Moreover, the kind of extrainstitutional experience that is valued by some movement party voters (radical right and eclectic) is noncontentious and often perceived as a nonpolitical type of experience, such as volunteering (Fleischmann and Steinhilper Reference Fleischmann and Steinhilper2017; Karakayali and Kleist Reference Karakayali and Kleist2016). When it comes to protest experience, it is true that the average probability of shunning a candidate for having participated in protests is lower among movement party supporters than conventional party voters. However, these differences are not statistically significant, and the results are more nuanced and weaker than the ones we had originally expected.
Against the Elites but Not against the System
In figure 4, a candidate’s reason for running has a clear significant impact. The figure shows that movement party voters, across ideologies, are more likely to support candidates who display anti-elitist sentiments (who are running “because corrupt elites too often ignore ordinary citizens”), in line with our expectations in H3a. Among conventional party voters, anti-elitist candidates are only preferred by those on the Left. Right-wing and centrist conventional party voters display no significant preference for anti-elitist candidates. Preference for mainstream candidates (who are running “to support the smooth running of the current political system”) varies across parties. Centrist and eclectic movement party voters have a preference for mainstream candidates, selecting them around 53% of the time, contradicting our expectations derived from H4d. However, while the estimate of the marginal mean also displays a positive preference, the confidence interval for green/left-libertarian and radical-right movement party voters slightly crosses below the 0.5 line. Looking into conventional party supporters, voters of all ideologies are positive toward mainstream candidates. Hence, while the marginal means of all movement party voters show a preference for mainstream candidates, contradicting H3b, results are significant only for centrist and eclectic movement party voters. One should note, at this point, that different groups may have interpreted the meaning of “the current political system” in varying ways. While subjective readings of this notion were possible, the deliberate reference to the “current system” narrowed the scope for interpretation across subpopulations.
Marginal Means of Candidates’ Reason for Running by Party Vote
Furthermore, all voters, on average, are unfavorable toward candidates who display a neutral reason for running (who run “to deliver policy changes”). These results are particularly meaningful for radical right as well as centrist and eclectic movement party voters, who exhibit a negative preference, selecting these candidates around 41% and 43% of the time, respectively, and whose preference is also significantly smaller than those of their conventional counterparts.
Policy Positions
Moving to candidates’ policies, in figures 5 and 6 we see that movement party voters are highly ideological voters. Policy attributes are those with the largest marginal means, particularly for green/left-libertarian and radical-right movement party voters. This means that the attributes that have the greatest influence on respondents’ candidate choice are those related to candidates’ positions on migration and environmental policies.
Marginal Means of Candidates’ Position on Migration by Party Vote
Marginal Means of Candidates’ Position on Environmental Policy by Party Vote
The choice of green/left-libertarian movement party voters is driven by both policy positions displayed in the experiment. When it comes to migration, they prefer candidates who propose to open borders to most migrants or those who propose to keep the status quo. They also do not favor candidates who propose to close borders. While migration policy is important for green/left-libertarian party voters, environmental policy has the strongest influence on their candidate selection. Candidates who propose to make climate change a priority have around a 65% chance of being selected, and the probability that a candidate who wants to reverse climate change policies is selected is only 32%. Candidates in favor of maintaining climate change policies are slightly more likely to be selected. The preferences of left-wing conventional party voters are aligned with those of their movement party counterparts but less pronounced, both for migration and environmental policies and across all levels.
The policy preferences of radical-right movement party voters are only significant for candidate positions on migration. This result is in line with previous experimental studies that reveal radical-right voters to be exclusively driven by their nativist stances (Kirkizh, Froio, and Stier Reference Kirkizh, Froio and Stier2022). Migration-friendly candidates are selected only around 40% of the time while those who propose to close borders are selected around 60% of the time. At the same time, radical-right movement party voters have no preference for candidates who propose to maintain current migration policies. When it comes to environmental policies, none of the three attribute levels have a significant impact on the probability that radical-right movement party voters select a candidate. Conventional right-wing voters display more complex preferences. Similarly to their movement party counterparts, they have a negative preference for candidates who propose open borders and favor those who wish to close borders to most migrants. Nonetheless, in contrast to radical-right movement party supporters, conventional right-wing voters also favor candidates who propose to maintain current migration policies. Concerning environmental policies, conventional right-wing voters display a positive preference for candidates who propose to either make climate change a priority or to maintain current climate change policies, unlike the radical-right movement party electorate. Conversely, they have a negative preference for candidates who propose to reverse climate change policies.
As for centrist and eclectic movement party voters, policy positions are less important for them than they are for the other two movement party groups. When it comes to migration, they favor the status quo. They have a significant preference for candidates who propose to maintain migration policies and they are significantly and equally unfavorable toward those who want to open or close borders to most migrants. Their conventional counterparts also favor candidates who propose to maintain current migration policies and have no significant preference for those who propose to either open or close borders to most migrants. In relation to environmental policies, they do not support candidates who either wish to reverse or maintain climate change policies and have a positive stance toward those who want to make climate change a priority. Centrist conventional party supporters align with their movement party counterparts in their negative preference for candidates who wish to reverse climate change policies and their positive stance toward candidates who propose to make climate change a priority. Nonetheless, they have the opposite opinion of candidates who suggest maintaining current climate change policies, as in this case conventional centrist party voters display a positive preference toward them.
Overall, these results confirm H4a and H4b. Candidate selection by green/left-libertarian movement party voters is mostly driven by their stances on environmental policies on climate change, while migration policies also have an influence on their decision to favor one potential representative over the other. The choices of radical-right movement party voters are likewise driven by candidates’ migration policies to a significant degree, while environmental policies do not seem to be relevant for them. Both of these groups have stronger preferences than those who vote for conventional parties of similar ideological colors (as illustrated in figures 5 and 6).
Our results for centrist and eclectic movement party voters help us to firm up the analysis of issue preferences. Centrist and eclectic movement party voters have express preferences on environmental policies on climate change, albeit ones that are less marked than those of green/left-libertarian movement party voters. On migration, centrist and eclectic movement party voters prefer the status quo. Put differently, in line with H4c, centrist and eclectic movement party voters combine issue positions, not unlike the parties that they support (Mosca and Tronconi Reference Mosca and Tronconi2019; see also the interactions between policy and other attributes in section 5 of the appendix).
Altogether, our findings indicate that support for movement parties is mostly driven by the policies they espouse. We tested for interactions between policy and the rest of the attributes (figures 13–20 in the appendix), and our results show that policy positions masked the impact of the rest of the attributes when it comes to movement party voters’ candidate selection. Moreover, compared to their conventional counterparts, movement party voters have altogether more hard-line stances on policy issues. Hence, across ideologies, movement party voters place greater importance on policies than their counterparts who support conventional parties. In sum, these results indicate that the policy positions of movement parties are central to understanding their electoral support.
Conclusion
The line that divides political parties and social movements is becoming increasingly blurred, as political parties integrate more and more social movement strategies and frames. In some cases, such as the Republican and Democratic parties in the US (Heaney and Rojas Reference Heaney and Rojas2015; Skocpol and Williamson Reference Skocpol and Williamson2012) and the Labour Party in the UK (Dennis Reference Dennis2019), movements have infiltrated parties and activists have transformed how parties act and communicate. In other cases, such as leftist parties in Latin America (Oviedo Obarrio Reference Oviedo Obarrio2010; Somma Reference Somma2022) as well as green/left-libertarian (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017) and new radical-right parties (Pirro Reference Pirro2019) in Europe, movements have created parties to compete in elections.
Against the backdrop of the growth of these new political vehicles across the world, in this paper we asked: what are the drivers of support for movement parties? To answer this question, we employed a conjoint survey experiment embedded in nationally representative surveys in six European countries. Previous research has examined primarily the supply side of this phenomenon, conceptualizing these parties (Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017; Kitschelt Reference Kitschelt, Katz and Crotty2006; Pirro Reference Pirro2019; Schwartz Reference Schwartz, Thomspon, Birnir, Kuhonta, Lewis and Smyth2016) and their organizational strategies (Della Porta 2021; Della Porta et al. Reference Della Porta, Fernández, Kouki and Mosca2017, Reference Della Porta, Fernández, Kouki and Mosca201; Mercea and Mosca Reference Mercea and Mosca2021; Pirro and Gattinara Reference Pirro and Gattinara2018). Studies focused on movement party voters have been restricted to single case studies (Passarelli and Tuorto Reference Passarelli and Tuorto2018) or smaller comparative analyses using cross-sectional observational data (e.g., Mosca and Quaranta Reference Mosca and Quaranta2021). To the best of our knowledge, this paper represents the first experiment-based study of the drivers of support for movement parties.
Our results are notable on three distinct levels. First, they indicate that the vote of movement party supporters is not guided by candidates’ institutional or extrainstitutional experience. Movement parties have been presented as the embodiment of “the vote of the protesters” (Mosca and Quaranta Reference Mosca and Quaranta2021), but their voters may not look for the protestor among the candidates these parties field. When examining candidates’ (extra)institutional experience, we find that, on average, movement party voters favor neither candidates who are institutional outsiders nor those who actively participate in protests. Indeed, the movement party vote does not seem to be driven by candidates’ previous experience (or lack thereof) in electoral and nonelectoral politics. Most levels of the institutional and extrainstitutional experience attributes are not significant across ideologies and, for those that are, the estimates of the probability that a candidate is selected because of their previous experience are not especially striking—that is, in the 45%–55% range.
Second, our findings show movement party voters are not against the current democratic system but rather against the elites who are leading it at the present time. Whereas movement party voters have been painted as anti-systemic (Blühdorn and Szarka Reference Blühdorn and Szarka2004; Elshehawy et al. Reference Elshehawy, Gavras, Marinov, Nanni and Schoen2022; Norman Reference Norman2021; Vezzoni and Mancosu Reference Vezzoni and Mancosu2016), our analysis revealed that they tend to favor both anti-elitist and pro-systemic candidates. On average, movement party supporters have a preference both for anti-elitist candidates (who publicly characterize their motivation to run in elections as being “because corrupt elites too often ignore ordinary citizens”) and for candidates who run because they wish “to support the smooth running of the current political system.” While the estimates of the probability of selecting a candidate based on their reason for running are similarly between 50% and 55% (except for far-right movement party voters, who have a stronger anti-elitist stance), there is a clear pattern where, across ideologies, movement party voters prefer both anti-elitist and pro-systemic candidates. Therefore, and unlike earlier observational research suggesting that M5S voters are disenchanted with both the political system and the elites (Passarelli and Tuorto Reference Passarelli and Tuorto2018), it seems that movement party voters want the existing democratic system to work and link its current problems to elites, rather than to democratic structures as such.
Third, and most importantly, movement party voters’ preferences are, as highlighted above, guided by candidates’ policy positions. The policy stances of movement party voters are, on average, more marked than those of voters who prefer conventional parties; and they are especially so for the more ideologically divergent supporters of green/libertarian and radical-right movement parties. For radical-right movement party voters, the single, salient issue of migration leads them to favor candidates who defend closing borders to most migrants and, conversely, to withhold support from those who propose to open borders to most migrants. Environmental policies are immaterial to their choice of candidate. Contrariwise, the preferences of green/left-libertarian movement party voters are guided by candidates’ positions regarding environmental as well as migration policies. Green/left-libertarian movement party voters prefer candidates who propose making climate change a priority as well as those who want to open borders to most migrants. In turn, they are less inclined to select those who wish to reverse climate change policies and those who espouse a policy of closed borders. Additionally, when testing for interaction effects, the impact of movement party voters’ policy preferences masks the impact of the other attributes. These results reinforce our conclusion that policy positions drive the movement party vote.
Based on these three conclusions, we see a desire for policy-orientated action as the key driver behind our surprising finding pointing to movement party voters being both anti-elitist and pro-systemic. Having stronger policy preferences than conventional party voters, movement party voters are disgruntled with the current institutional elites, whom they consider to have failed to provide adequate policy solutions to major contemporary issues like the climate crisis or migration. This disgruntlement with a lack of decisive policy action from current institutional elites is what leads movement party voters to look for new institutional actors who promise to bring forward policies that align with their views. While these voters may diverge in their understanding of what the “current political system” is, their positive preference for candidates making this promise may indicate a desire among movement party voters to see institutions work to address the political and policy challenges with which each group is most concerned. This may be one of the reasons why movement party supporters channel their discontent through behavior that remains within the bounds of institutional democracy and pay little attention to whether candidates are institutional outsiders or have protest experience.
While this assertion may lead some to question the utility of the movement party concept, we would highlight that it is arguably this organizational form that facilitates the incorporation of starker policy choices by these parties, which their electorates seek. In other words, while movement parties have been shown to galvanize the “vote of the protesters” (Mosca and Quaranta Reference Mosca and Quaranta2017, Reference Mosca and Quaranta431), their constituencies extend beyond this social group, which can act as a vanguard for causes supported by a larger section of the population. Consequently, while the support of movement parties is not directly guided by their organizational form and its manifestation through demonstrations and other varieties of engagement in collective action associated with social movements, the more hard-line policies that earn the backing of movement party voters reflect the organizational relations movement parties have with social movements and their policy demands.
To conclude, we would like for this paper to provide renewed impetus to a flourishing research agenda on social and political initiatives that span the dividing line between electoral and nonelectoral politics. Further research could continue to explore what other policies are central to the anti-elitist but pro-systemic attitudes of movement party and other new party voters. While we have tested the effect of two major policies in this paper, we see other critical political issues across Europe that are ripe for that analysis—to wit, housing and the sustainability of the welfare system, including healthcare, education, and pensions (European Commission 2021; 2024).Footnote 3 Scholars could also consider whether studying voter preferences as they relate to party characteristics, rather than to candidate features, provides different results from the ones presented in this paper.
Another avenue for future research may be investigating whether our results apply to movement party voters beyond the ideological and geographical focus of this paper. In our case, we built on Kitschelt’s (Reference Kitschelt, Katz and Crotty2006) seminal categorization of movement parties into radical right and green/left-libertarian groupings, adding a third category we labeled centrist and eclectic movement parties. We expect this categorization to benefit similar research—including other experimental designs—comparing supporters of parties that do not fit neatly along the left–right axis (such as the Pirate Party [e.g., Otjes Reference Otjes2020]), or movement parties not covered in this study (such as Podemos in Spain [e.g., Flesher Fominaya Reference Flesher Fominaya2020] or Syriza in Greece [e.g., Stavrakakis and Katsambekis Reference Stavrakakis and Katsambekis2014]), with other voting blocs, as well as research into similar parties beyond the European continent. In the end, despite Kitschelt’s (Reference Kitschelt, Katz and Crotty2006) conceptualization of movement parties as a transient and volatile phenomenon, they have proved to be a resilient representation of a wider transformation in political systems across the world toward an increasingly blurred division between electoral and nonelectoral politics.
Supplementary material
To view supplementary material for this article, please visit http://doi.org/10.1017/S1537592724001439.
Data Replication
Data replication sets are available in Harvard Dataverse at: https://doi.org/10.7910/DVN/CBKZL1
Acknowledgments
We thank Mauro Barisione, David Doyle, Nora Kirkizh, Dániel Kovarek, Levente Littvay, Lorenzo Mosca, Sally Stares, and Alberto Stefanelli for their valuable comments on the design of our conjoint survey experiment. An earlier version of this paper was presented at the European Consortium for Political Research Joint Session “Movement Parties: Their Rise, Variety, and Consequences” in Edinburgh in 2022, where we benefited from insightful feedback. We are also grateful to the four anonymous reviewers for their constructive comments and to the coeditors of Perspectives on Politics, Ana Arjona and Wendy Pearlman, for their guidance throughout the review process. We would also like to thank Lyn Flight for carefully copyediting the final version of this paper. This work was supported by the Volkswagen Foundation (grant no. 98144); the European Union (NextgenerationEU); and the Romanian government, under the National Recovery and Resilience Plan for Romania, contract no. 760038/23.05.2023, cod. PNRR-C9-I8-CF204/29.11.2022, disbursed by the Romanian Ministry of Research, Innovation and Digitalization, within Component 9, Investment I8.
Open access
