How Do Mainstream Parties Justify Their (Un)willingness to Rule with Populist Parties? Evidence from Twitter Data

Abstract

Parties usually argue in favour or against a government coalition based on party considerations in terms of projected policy implementation, power in office and vote maximization – that is, the ‘policy, office, votes’ triad. So far, however, it remains unclear which claims mainstream parties invoke to motivate their choice to rule or not rule with populist parties. Adopting the ‘policy, voter, office’ triad, this article examines mainstream parties' Twitter claims on ruling with populist parties in Austria, Belgium, Germany and the Netherlands (2006–2021, N = 1,919). Mainstream parties mainly reject ruling with (mostly radical right) populist parties. To justify unwillingness, policy-based motives referring to the populist parties' extremist nature trump motives on office-seeking and vote maximization. To justify willingness, predominantly office-seeking motivations are invoked. Party characteristics (ideology, incumbency status, size) and context, however, shape these claims. This study sheds light on mainstream parties' patterns of political communication on coalition formation with populist parties.

Bart De Wever, leader of the Flemish nationalist party New Flemish Alliance (N-VA) referred in a 2018 interview to a ‘Chinese wall’ between his party and the radical right populist party Flemish Interest (Vlaams Belang, VB). Despite this claim, he negotiated with the VB after the May 2019 elections to form a regional Flemish government. The VB had become the second largest party and together with the N-VA fell just short of a majority in the Flemish parliament. Ultimately a coalition without the VB was formed (including the Liberals and Christian Democrats instead) as – quoting informateur De Wever – ‘always a third party was needed to reach a majority. None of the other parties was prepared to deliver this majority, regardless of the formula. This is the most congruent and coherent Flemish government that could be formed with a majority in the Flemish Parliament.’¹ Political commentators are already speculating about the 2024 election where the N-VA and VB could obtain a parliamentarian majority. Despite having negotiated with the VB before, De Wever has recently declared that he would rather quit politics than rule with the VB.

This anecdote illustrates the controversy that often accompanies discussions about ruling with populist parties in Western Europe. It highlights the different arguments that can be invoked by mainstream parties to motivate their (un)willingness to rule with populist parties, and it shows that mainstream parties' positions can quickly shift. Potential government participation by populist parties is characterized by a paradox. Generally, anti-establishment politics is on the rise, transforming the political landscape with radical left-wing populist parties (LWPPs) and radical right-wing populist parties (RWPPs) systematically outperforming mainstream parties at the ballot box (Krause and Wagner 2021). Despite this electoral success, however, populist parties do not routinely become part of the government (Akkerman and De Lange 2012). Two main reasons could explain this observation. First, populist parties may be hesitant to carry government responsibility and join a coalition as this risks jeopardizing their anti-establishment profile (Fagerholm 2021). Second, mainstream parties may be reluctant to rule with populist parties for various reasons and prefer a coalition without them (Akkerman and De Lange 2012; De Lange 2012). This study focuses on the latter explanation by systematically assessing mainstream parties' claims about ruling with populist parties.

Parties' motivations for joining a government coalition are the outcome of a process in which they try to maximize various party strategies (Debus 2008). Coalition-formation theories point to tactical considerations regarding projected power in office, vote maximization and policy implementation as major reasons that shape parties' coalition preferences (Bäck 2003; Martin and Stevenson 2001). So far, however, party justifications underlying coalition formation with populist parties and the motivations underlying mainstream parties' (un)willingness to rule with populist parties remain largely underexplored. Which reasons to rule with populist parties or not are being invoked, and how do party characteristics and context shape these claims? To clarify these questions, this study reports the findings of a longitudinal and quantitative content analysis (2006–2021) of Twitter claims by 25 mainstream party actors (N = 1,919) on forming a coalition with populist parties in five Western European political contexts (Austria, Flanders, Germany, the Netherlands, Wallonia). The aim is to identify patterns in the political communication of mainstream parties and unravel their justifications for why they are willing or not to rule with populist parties based on the ‘policy, office, votes’ triad. Ultimately, the goal is to apply this typology to the case of how mainstream parties motivate their (un)willingness to rule with populist parties, comparing how often these different categories within the typology are invoked and how party and country characteristics play a role. As such, I aim to initiate a research agenda on how political actors justify their (un)willingness to rule with populist parties.

Theoretical framework

Mainstream parties' motives not to rule with populist parties

Wolfgang Müller and Kaare Strøm (2000) have advanced the policy, office, votes triad, underscoring that parties' motives for joining a government are the result of party strategy aimed at policy implementation, power in office and vote maximization. The first two (policy, office) especially are argued to be essential (Debus 2008; Pedersen 2012). Based on the policy, office, votes triad and general coalition-formation theories, various reasons that justify mainstream parties' refusal to rule with populist parties can be identified. A first set of reasons pertains to policy-seeking motives. Populist parties' ideology – often radical left or right – could be perceived by mainstream parties as extremist, which could be a reason to refrain from ruling with the party due to moral considerations. In some countries, there is deep-seated hesitation to collaborate with populist parties because of their idiosyncratic position in the political landscape, which has been labelled as ‘ostracizing’ a party (i.e. systematically ruling out political collaboration with the party at any level) (Van Spanje 2018). In some European democracies, this has been institutionalized in a ‘cordon sanitaire’ (Heinze 2018). For a long time, mainstream parties in Western Europe have excluded populist parties with communist and Marxist traditions and those with a radical right ideology. Joining a coalition with an extremist party is considered a step too far in a liberal democracy (Down and Han 2020; Valentim 2021). In this view, populist parties are ‘beyond the pale’ and challenge key democratic values and rights (e.g. rule of law, minority rights), posing a threat to the fundamentals of democracy (Van Spanje 2010). This resonates with an anti-prejudice norm that means that political actors violating this norm are not considered legitimate partners (Blinder et al. 2013). Often, parallels with questionable ideologies in the past are drawn and these parties are stigmatized (Van Heerden and Van der Brug 2017; Van Spanje and Azrout 2019).

Second, mainstream parties may believe that ruling with populist parties will not produce the desired policy direction and that ideological differences are unbridgeable. Mainstream parties may evaluate populist parties as unattractive partners from a substantive perspective due to incompatible policy priorities, policy programmes and/or ideological profiles (Bassi 2017). The minimal-range theory identifies the most successful coalition formula as the one with the lowest ideological diversity (Axelrod 1970; Martin and Stevenson 2001). When multiple coalition options are viable, formulas with populist parties may be ignored as mainstream parties may prefer partners that allow an ideologically congruent government (Deschouwer 2008).

A second group of reasons stems from considerations regarding projected power in office. First, organizational difficulties in forming a numerical majority could arise: the populist party could simply not fit in any formula to form a majority without a surplus, or other combinations may result in a stronger majority (Bassi 2017; Martin and Stevenson 2001). Similarly, a majority with a populist party may be too narrow and, hence, unstable. A second set of office-seeking motives concerns the projected behaviour of the populist party as a coalition partner. Populist parties are often newly founded parties or parties with limited governing experience (Van Kessel 2015). This could give the populist parties an image as being inexperienced, incompetent or ‘unfit to rule’ – in brief, lacking effectiveness. Perceived effectiveness or a party's perception as a viable and stable, efficient organization that can reach its (policy) goals is considered key for a party's electoral success. Perceived lack of it could impair a party's appeal as a coalition partner (Bos and Van der Brug 2010). Similarly, trust and reliability are vital qualities for coalition partners as governments are held accountable as a whole and share a joint responsibility. Earlier collaborations could signal to mainstream parties that populist parties are not reliable allies. Finally, mainstream parties may point out populist parties' reluctance to rule and tendency of self-isolation through refusing to take responsibility or by making themselves unacceptable through a lack of willingness to compromise (Han 2020).

Finally, considerations linked to vote maximization could be invoked. First, mainstream parties may fear that a populist party in the coalition risks legitimizing the populist party, allowing it to build an image as a normal, respectable party, which could consolidate its position in the electoral landscape (Down and Han 2020; Heinze 2018; Valentim 2021). Incumbent parties are likely to increase in relevance and accumulate more media attention as they carry responsibilities and are in charge of government decisions (Hopmann et al. 2018), risking normalizing them. A recent study by Vicente Valentim and Tobias Widmann (forthcoming) has found that mainstream political actors tend to distance themselves from radical right parties and ideas by explicitly reinforcing democratic norms. In this sense, mainstream parties could claim they are preserving democracy by not ruling with a populist party to prevent legitimizing it. Similarly, mainstream parties may make appeals to the will of the people and argue that – even if populist parties performed well in the elections – still a majority did not vote for the populist party, suggesting that it is democratic to exclude it and form an alternative majority. Mainstream parties could adopt promises to not rule with populist parties as an electoral strategy to entice voters.

Mainstream parties' motives to rule with populist parties

Governments with populist parties seem more prevalent in the early 21st century (Dunphy and Bale 2011; Krause and Wagner 2021). Reasons to rule with populist parties are also expected to be grounded in coalition-formation theory and to stem from the policy, office, votes triad (Müller and Strøm 2000).

First, in terms of policy-seeking motives, forging a coalition between parties that are ideologically close (e.g. centre-left and radical-left parties), even if one of them has an anti-establishment profile or is perceived as extremist, may under certain conditions be judged as desirable. Ideological congruence is generally considered as the key to a successful coalition (Martin and Stevenson 2001). Ruling with a populist party could be perceived as strengthening a political bloc and benefiting mainstream parties from a policy perspective as it may put salient issues on the political agenda and help them to realize policy goals. Second, in response to the perceived extremism of a party, mainstream parties could hope that carrying responsibility for delivering policy output may moderate populist parties' positions (Berman 2008; Bernhard 2020). This ‘inclusion–moderation’ hypothesis asserts that once populist parties are confronted with the duties of office, they will (partly) abandon or moderate their extreme position or broaden their ideological profile to maximize and consolidate their popular appeal (Akkerman and Rooduijn 2015; Schwörer 2022).

Second, office-seeking motives may play a role. First, the voters may have shuffled the cards so that no numerical majority is possible without a populist party, or there may be a lack of alternative partners that are willing to join the coalition (Andeweg et al. 2011; Döring 1995). Mainstream parties often face a choice between a ‘grand coalition’ with traditional parties or ruling with populist parties (Askim et al. 2021). Second, trade-offs between the experience, reliability and aptitude of populist parties once in office may guide mainstream parties' motives. Mainstream parties may treat populist parties – which usually have limited prior government experience – as ‘junior’ partners with low bargaining power, hoping they can dominate them and maximize their own influence (Akkerman 2012; Akkerman and De Lange 2012). This could mean that mainstream parties prioritize collaboration with a populist party over an experienced mainstream party that would be a more equal partner. Mainstream parties may judge a populist party or its leader to be reliable or competent, using this as an argument that a coalition with it will permit a stable, decisive and effective government. Possibly, the mainstream party may have collaborated well with the populist party before (e.g. on a bill), which could benefit its profile as an attractive ally, which could be used as a motive for collaboration (Krause and Wagner 2021). Finally, if populist parties – backed-up by their electoral mandate – actively claim they want to rule and assume responsibility, it could be hard for mainstream parties to deny them this opportunity. Instead they could invoke this as a justification for ruling with them.

Finally, vote-seeking motives can be invoked. First, the electoral strength of a populist party could make it difficult for mainstream parties to ignore it as a coalition partner. Unwillingness to rule with popular populist parties risks being perceived as ‘undemocratic’, and as lacking legitimacy and respect for the voters' signal, which could be anticipated by mainstream parties as harming their electoral position. Mainstream parties may hope that carrying government responsibility will erode populist parties' electoral support by making them part of the elite, undermining their niche position in the electoral landscape (Abou-Chadi 2016). Second, mainstream parties could ‘bandwagon’ or ‘hitch a lift’ on the success of populist parties, hoping that it will strengthen their own electoral position as well and secure their long-term survival. They could use the prospect or promise of change through a future coalition with a populist party as an argument in their electoral campaign (Heinze 2018).

Table 1 summarizes all theoretical reasons based on coalition-formation theories that could be used to motivate ruling with populist parties. A research question to shed light on the type of motivations that drive (un)willingness of mainstream parties is posed: Which claims do mainstream parties use to motivate their (un)willingness to rule with populist parties?

Table 1. Motives to Justify (Un)willingness to Rule with Populist Parties

The role of party characteristics: Ideology, incumbency status and size

Party characteristics and political context are arguably related to the likelihood that parties adopt certain types of justifications.

Mainstream parties' incumbency status could affect the likelihood that certain justifications will be used. Incumbent mainstream parties are expected to use office-seeking motives more easily to justify their (un)willingness to rule with populist parties than parties in the opposition. Mainstream parties in office are usually in a coalition with other (often) mainstream parties with whom they share responsibility. Hence, parties that are part of the outgoing coalition have an incentive to defend their incumbency record, which could motivate them to continue the existing coalition. The outgoing coalition also tends to be an easy alternative for otherwise difficult negotiations (Olislagers and Steyvers 2015; Strøm et al. 1994). Incumbent mainstream parties have an incentive to defend their positions and maximize their power. Collaboration with familiar parties – for example, current partners – may be more attractive than forming a government with a new, possibly unreliable and inexperienced, populist party, which could shape their justifications (Bäck 2003).

Hypothesis 1: Incumbent mainstream parties are more likely than non-incumbent mainstream parties to invoke office-based motives to justify their willingness (H1A) or unwillingness (H1B) to rule with populist parties.

Party size may shape the likelihood that mainstream parties adopt certain justifications to motivate their (un)willingness to rule with populists (Debus and Müller 2011). Larger mainstream parties could be more likely to use office-seeking than policy-seeking or vote-seeking motives. Larger mainstream parties are generally more relevant and in a key position during government-formation negotiations. Typically, the initiative for forming a government rests with larger parties. They are also more likely to have more ruling experience (Lupia and Strøm 2008). Therefore, office-seeking motives may be a more prominent objective for larger parties, guiding their justifications, while smaller mainstream parties are more likely to pursue policy- or vote-seeking goals (Deschouwer 2009):

Hypothesis 2: Larger mainstream parties are more likely than smaller mainstream parties to use office-seeking motives to justify why they are willing (H2A) or unwilling (H2B) to rule with populist parties.

The party size of populist parties may matter as well. Mainstream parties can be expected to revert more easily to vote-seeking claims when populist parties are larger. This is because the electoral weight of larger populist parties cannot be easily ignored by potential coalition partners as parties that have performed well at the elections are generally more likely to become part of the government (Bergman et al. 2021). Hence, vote-seeking motives that emphasize either that excluding (strong) populist parties may be undemocratic or that it is still democratic to build a coalition without a populist party, as the majority of citizens did not vote for that party, may be judged more pertinent to larger populist parties:

Hypothesis 3: Mainstream parties are more likely to use vote-seeking motives to justify their willingness (H3A) or unwillingness (H3B) to rule with larger populist parties than in relationship to smaller populist parties.

Next, the ideological orientation of populist parties could shape the arguments that are being invoked to justify mainstream parties' (un)willingness (Heinze 2018). While populist parties often have roots in radical right ideologies such as (neo)Nazism or (neo)fascism or originate from Marxist or communist predecessors, the former may be judged more harshly. RWPPs usually make use of blatantly racist rhetoric in which outgroups (e.g. immigrants or Muslims) are blamed for what is going wrong in society. Anti-prejudice norms exist for blatant expressions of prejudice being penalized in democratic societies (Blinder et al. 2013), which could explain why ruling with RWPPs may be considered more controversial than ruling with radical LWPPs. While LWPPs usually still cherish their communist and Marxist legacies, most of them have distanced themselves from this tradition and have transformed into socialist left parties striving for social justice for the working class (March 2012). Therefore, mainstream parties may be more likely to invoke reasons linked to perceived extremism when justifying claims for (not) ruling with RWPPs compared to LWPPs:

Hypothesis 4: Mainstream parties are more likely to invoke policy-based motives on perceived extremism to justify their unwillingness to rule with RWPPs than with LWPPs.

Mainstream parties' ideological orientation is likely to shape their motives to rule or not with populist parties. The ‘minimal range’ theory asserts that political parties strive to maximize the level of ideological congruence between coalition partners to build a majority with a clear and uniform policy direction (Andeweg et al. 2011; Axelrod 1970; Martin and Stevenson 2001). Hence, parties that fundamentally differ from a certain populist party are less likely to rule with them and will invoke this as a justification. In addition, similarity in terms of ideology and left–right position with populist parties may be used to motivate willingness to rule with populist parties:

Hypothesis 5: Mainstream parties are more likely to invoke policy–ideological congruence claims to justify their willingness to rule with ideologically closer populist parties (H5A), but are more likely to invoke policy–ideological incongruence claims to justify their unwillingness to rule with ideologically more distant populist parties (H5B).

A final set of expectations pertains to relevance, namely whether a mainstream party is negotiating with a populist party at the time of the claim and whether the populist party has ruled before. First, during actual negotiations with a populist party, mainstream parties may be more likely to invoke policy and ideological congruence to motivate willingness to rule with populist parties in order to convince their voters in line with the minimal range theory (Debus and Müller 2011). Second, when populist parties have ruled or supported a minority government before, mainstream parties may be less likely to invoke perceived extremism or policy (in)congruence to justify their (un)willingness to rule with populist parties, but may rather invoke arguments pertaining to how populist parties behave in office (Bäck 2003). Referring to prior ruling experiences offers mainstream parties an additional argument and a less controversial excuse.

Hypothesis 6: Mainstream parties involved in actual negotiations with populist parties are more likely to use motives on policy and ideological congruence to justify their willingness to rule with populist parties.

Hypothesis 7: Mainstream parties are less likely to use motives related to perceived extremism (H7A) and more likely to use office-seeking motives (H7B) to justify their unwillingness to rule with populist parties that have ruled before.

The role of the country context

Contextual factors can affect motives as well. In some countries, ruling with populist parties is seen as highly controversial as it would be a violation of a formal cordon sanitaire against certain parties. This cordon sanitaire is an agreement among all mainstream parties to not collaborate with the ostracized party at any level due to its alleged undemocratic nature (Van Spanje 2010) and can be considered as a form of ‘militant democracy’. This is the case in Belgium and Germany. Still, in other countries (e.g. Austria, the Netherlands), practices for ruling with (allegedly) populist parties are more lenient. In the former, the Freedom Party (FPÖ) has ruled before, while in the latter, populist parties have formally supported a minority government. These distinct contextual factors (a cordon sanitaire, a militant democracy approach vs prior government experience) could affect the claims of mainstream parties. Hence expectations are that in Belgium and Germany mainstream parties will be more reluctant to rule with populist parties and that their claims will be mostly informed by worries about the populists' extremist nature, while in Austria and the Netherlands reluctance will be lower and claims will mostly be informed by office-seeking motives:

Hypothesis 8: In Belgium and Germany, more claims to justify unwillingness based on policy-seeking motives linked to extremism will be made (H8A), while in Austria and the Netherlands more claims based on office-seeking motives to justify unwillingness will be made (H8B).

Data and method

Case selection

This study analyses Twitter claims by political actors in Austria, Belgium, Germany and the Netherlands. All countries are durable parliamentary democracies in which at least one strong populist party is present (see the PopuList, Rooduijn et al. 2019). Only mainstream parties belonging to the traditional party families (liberal, socialist, Christian Democrats/conservatives), the Green party families and classic regionalist parties with steady representation in parliament were included to grasp over-time trends, excluding niche parties. Following the ideational approach, we perceive populism as a thin-centred ideology, which is characterized by anti-elitism, people-centrism, homogeneity of the people and a belief that politics should be an expression of the volonté générale (Mudde 2004; Rooduijn 2014). As such, populist parties can be either radical left or right in nature, depending on the ideology they follow. Essentially, populist parties have an anti-establishment profile, combined with a specific (e.g. nativist or socialist) ideology. Three countries have a strong radical LWPP (the Workers' Party of Belgium (PTB),² the Socialist Party (SP) in the Netherlands and Die Linke in Germany) with roots in different forms of communism (e.g. Maoism or Marxism). While LWPPs have evolved and now adhere to democratic socialism, references to their communist heritage and ideals are still present in their party statutes today. Most parties define themselves in opposition to capitalism and (neo)liberal policies (March 2012). In all countries an electorally viable RWPP (Flemish Interest (VB) in Flanders, Freedom Party (PVV) and Forum for Democracy (FvD) in the Netherlands, Alternative for Germany (AfD) in Germany and Freedom Party of Austria (FPÖ) in Austria) is present. While these parties have a distinct profile and history, they share an anti-immigration stance, aversion to a multicultural society and frequently engage in outgroup rhetoric (De Jonge 2019; Van Kessel 2015; Zulianello 2020).

The countries represent distinct cases in terms of government participation of populist parties. In Belgium – Flanders and Wallonia – populist parties have never been part of a government coalition due to the cordon sanitaire (Akkerman and Rooduijn 2015; De Jonge 2019). While this cordon sanitaire is less formally enforced for the PTB, the party has never been part of a federal or regional government.³ In Germany, Die Linke and the AfD have roots in ideologies or political systems that have become discredited in Germany. While Die Linke's ancestry goes back to the former Communist Party in the German Democratic Republic (DDR), the AfD, founded in 2013, has been described as populist and adhering to views that are close to (neo)Nazism (Zulianello 2020). Given Germany's past experience with extremism, the authorities have put laws in place in the tradition of militant democracy aimed at preserving the state and its institutions from anti-democratic forces (Bourne and Bértoa 2017). Alleged radical organizations are monitored by the Bundesverfassungsgericht, which has been given the power to intervene when deemed necessary. Attempts have been made to ban parties (e.g. the National Democratic Party, NPD) in the past, making collaboration with populist parties controversial. The Netherlands and Austria have distinct traditions: populist parties have either been part of the government or have formally supported a minority government. In Austria, the cordon sanitaire that was in place has been broken twice when the FPÖ joined the federal government (Paxton 2021). After winning the elections in 1999, it became part of the government in a coalition with the conservative ÖVP for the period 2000–2005, after which the party became internally divided and lost the next general election. In 2017, however, it regained strength and joined the government for a second time under the chancellorship of Sebastian Kurz. This coalition imploded due to corruption scandals and was dissolved in May 2019 (Paxton 2021). Finally, in the Netherlands no formal cordon sanitaire is enforced. Moreover, the Netherlands has prior experience with a coalition government either where a populist party (the Ljist Pim Fortuyn, LPF, in 2002) was invited into the coalition or a situation where a populist party formally supported a minority government (the Freedom Party, PVV, in 2010) (Akkerman 2012; Akkerman and De Lange 2012). Both cabinets fell apart due to internal division.

Data selection

A quantitative content analysis has been conducted to assess Twitter claims by mainstream parties on their (un)willingness to rule with populist parties. Claims needed to fulfil a set of selection criteria: the claims must be (1) political in nature and (2) pertain to the question of ruling with a populist party at any level (local, regional or federal).⁴ This follows the approach by Marco Giugni et al. (2005) in their analysis of extreme right-wing actors' claims. Twitter has developed into a dominant tool for direct communication between political parties and the electorate. This is evidenced by the fact that all political parties in this study have a Twitter account that is used on a regular basis as a key instrument of political communication in election campaigns (Stier et al. 2018). The Twitter accounts of both the party and its leader were analysed as both are assumed to reflect official party viewpoints.

Unit of analysis

The unit of analysis was a tweet by a party or party leader. The time line of each party actor was scraped from Twitter using a search string to select all tweets on coalition preferences with populist parties via the advanced search function of Twitter.⁵ This data collection was done ex post, from 1 April to 1 May 2021, for all included accounts. In total, 1,038,115 tweets (= Total number of tweets sent from all party actor accounts since they have joined Twitter, which we obtained via applying the Tweepy program in Python, namely statuses_count) have been analysed to result in 1,919 eligible Tweets, which have been manually coded. Hence, the full population of tweets has been examined. This search goes back as far as the opening of actors' Twitter accounts.⁶ The search hits were manually checked to exclude false positives and, if they fell within the scope of the selection criteria, coded. Retweets with comments (or ‘cited tweets’) were included. Retweets of other Twitter users without any comments were not included as they do not contain a claim of the party actor (overview of the search strings and Twitter accounts in the Online Appendix).⁷

Coding and intercoder reliability

The coding has been done by the author, who has ample experience with quantitative content analysis. A specific coding scheme for this project has been developed. About 10% of the sample has been double-coded by an extensively trained PhD student. Both coders have a highly advanced proficiency in all three languages. Intercoder reliability scores were satisfactory with a Krippendorff's alpha varying from 0.72 to 0.97 and an average of 0.81.

Variables

Dependent variables

(Un)willingness to rule with populist parties. It was first coded whether the actor referred to government collaboration with populist parties and whether mainstream parties expressed either an unwillingness (i.e. non-collaboration, rejection of a coalition) or willingness (i.e. collaboration, approving a coalition) to rule with populist parties, both implicitly and explicitly. Next, the reasons given for (un)willingness were coded. Combinations were possible as political actors could refer in a single tweet to policy-, vote- and office-seeking motivations simultaneously, hence a single tweet could include multiple statements that had elements of more than one motivation.⁸ Tweets can occur in sequences (threads) – in this sample in 281 of 1,919 tweets, or 14.6% – although here they have been analysed separately.

Policy-, office- or vote-seeking motives. These binary variables were scored with ‘1’ for presence and ‘0’ for absence, based on the typology advanced in the theoretical framework. These claims needed to be present explicitly in the tweet (examples and coding scheme available in the Online Appendix). All these motives could be invoked before or after the elections. The policy-, office- and vote-seeking motives are the three broad categories that are included in the analysis, but they are based on subcategories that are described in this section. For the policy-seeking motives, we include two subcategories on perceived extremism or moderation–inclusion, and ideological (in)congruence separately in the model.

Policy-seeking motives. Policy-seeking motives are a broad category, encompassing two subcategories: (1) claims including an evaluation of how close or distant parties are in terms of policy and ideology, and (2) claims pertaining to the perceived extremism or anticipated moderation of populist parties once in office. The first subcategory (policy or ideological (in)congruence) was coded when mainstream parties referred to a populist party holding similar/opposite views on policy objectives or priorities, party programme or ideological orientation. The second subcategory (perceived extremism/inclusion–moderation) referred to claims about the alleged extremist nature of the populist party as posing a threat to democracy and the rule of law and as undermining fundamental rights or claims that ruling with a populist party could moderate its views. Both categories are taken up as two distinct dependent variables, as reliability (ɑ = 0.57) and factor analysis shows they do not load on one underlying dimension. Theoretically, it is relevant to distinguish between a mere ideological difference or claims that a party is extremist, which is a far more fundamental and severe claim as this pertains to claims about ‘parties beyond the pale’: it goes further than mere policy differences, but includes a claim about the party's undemocratic nature.

Office-seeking motives. For office-seeking motives, all claims are included that refer to constraints or opportunities regarding the projected power in office. More specifically, these are claims pertaining to numerical reasons for (not) including populist parties in government, the need for a stable, performant government and preferences for alternative coalition partners, to projected behaviour in office based on the (un)reliability, (in)competence, (in)experience of the party and prior experiences with the party – which could motivate (un)willingness to rule with a populist party and claims about whether the populist party itself is willing to rule or if it engages in self-isolation.

Vote-seeking motives. Vote-seeking motives refer to claims connecting the potential government participation of populist parties to the outcome of the elections, or the electoral base of parties to motivate (un)willingness to rule with a populist party. Here the argument is not that populist parties are undemocratic. Instead, the focus is on referring to the will of the people or respect for the voters’ signal as a way of motivating ruling with a populist party, by arguing either that a substantial share of voters did not vote for the populist party or that they did vote for the party. More generally, claims in which (un)willingness to rule with a populist party is used as an argument in an election campaign are also included.

Independent variables

Parties' political ideology. The ideologies of the mainstream and populist parties were quantified as a dynamic score (for 2010, 2014 and 2019) capturing their left–right dimension by the Chapel Hill Expert Survey (CHES) at the time of the claim (Bakker et al. 2015). This score is used in the multivariate statistical analyses, but for the descriptive analysis party families (of the CHES survey) were used. For mainstream parties, the green (Groen, Ecolo, Die Grünen, GroenLinks, Bündnis-Die Grünen), socialist (Vooruit, Parti Socialiste (PS), Sozialdemokratische Partei Österreichs (SPÖ), Partij van de Arbeid (PvdA), Sozialdemkratische Partei Deutschlands (SPD)), liberal (Open VLD, Mouvement Réformateur (MR), Das Neue Österreich und Liberales Forum (NEOS), Democraten 66 (D66), Volkspartij voor Vrijheid en Democratie (VVD), Freie Demokratische Partei (FDP)), Christian-Democrat and conservative (Christen-Democratisch & Vlaams (CD&V), Les Engagés, Österreichische Volkspartei (ÖVP), Christen-Democratisch Appèl (CDA), ChristenUnie (CU), Staatskundig Gereformeerde Partij (SGP)) and regionalist (Nieuw-Vlaamse Alliantie (N-VA), Démocrate Fédéraliste Independiste (DéFi)) party families were distinguished. The populist parties were classified based on the CHES score as being either radical left (PTB, SP, Die Linke) or radical right (VB, PVV, FvD, AfD, FPÖ).

Ideological distance. We included a difference score between the ideological left–right score of the mainstream party and of the populist party to obtain a measure of how distant the two parties are. We took the absolute difference between the score of the populist party and the score of the mainstream party, with a higher score reflecting larger distance.

Incumbency. A time-variant binary variable pertaining to whether the mainstream party is part of the opposition (coded as ‘0’) or the majority (coded as ‘1’) at the time of the claim has been included, based on the ParlGov database (Döring and Manow 2020).

Party size. The size of the mainstream and populist party was operationalized as the vote share that the party has at the time of the claim in the national or federal parliament. The data vary across time between and within party actors as our data span different elections. The data stem from the ParlGov database (Döring and Manow 2020).

Political context. Five country dummies are included: Austria, Belgium-Flanders, Belgium-Wallonia, Germany and the Netherlands.⁹

Negotiation. This binary variable (0 = ‘No’, 1 = ‘Yes’) grasps whether mainstream parties at the time of their claim were holding government negotiations with a populist party. This pertains to the N-VA (with VB) and PS (with PTB) after the 2019 elections to form, respectively, a Flemish and Walloon government in Belgium; the Dutch CDA and VVD during the 2010 formation, leading to Rutte I, formally supported by the PVV; negotiations in the aftermath of the 2017 Austrian legislative elections, resulting in a government between the ÖVP and FPÖ.

Populist party ruled before. This dummy summarizes whether populist parties have been part of a government or have formally supported a minority cabinet before. It is scored ‘1’ for the FPÖ in Austria (which was in the Schüssel I government from 2000 to 2005) and for the PVV after its support for the minority cabinet Rutte I, formed in 2010.

Level. This binary control variable grasps whether the claims pertain to the local level (0) or to the national/federal or regional level in federal states (1).

Election: pre-election, post-election. A dummy that captures claims in the pre-election period during national elections (i.e. the three months before the elections, including election day) is included and one for claims in the post-election period (i.e. up until three months after the elections).

Effective number of electoral parties (N_v). This variable reports the effective number of parties based on the election results. It is a measure of how many parties, weighted for size, are in a party system in a given election.

Results

Descriptive results

How often do mainstream parties claim on Twitter that they are (un)willing to rule with populist parties, and which types of motives do they invoke to justify their preferences? Overall, 1,919 Twitter claims were made regarding ruling with populist parties. Most claims were made in Austria (37.8%), followed by Germany (27.5%), the Netherlands (23.4%) and Belgium (11.3%). Most claims pertain to RWPPs (86.1%) and only a minority (13.9%) to LWPPs. In most claims by mainstream parties, ruling with populist parties is rejected (91.5%), while only in a small number of claims (8.5%) party actors emphasize that they are willing to rule with populist parties. In terms of electoral cycle, 27.6% of claims were articulated during the campaign running up to the election, while 6.4% of claims were articulated after the election in the negotiation period, and 65.8% were in non-election periods. Generally, in the post-election period parties tend to claim that they are more willing to rule with populist parties (i.e. 24.4%) than in the pre-election (9.9%) and non-election periods (6.4%) (see Table A14 in the Online Appendix). Of course, parties largely differ in this regard (see Table A7 in the Online Appendix).

Figure 1 lists the motives that are invoked for ruling (in percentages). The motives to justify willingness to rule with populist parties differ from those to justify unwillingness. Different motives could be combined (which is why the percentages do not add up to 100%), and Figure 1 displays (in percentages) how often a motive has been invoked to motivate either willingness or unwillingness to rule with populist parties. For unwillingness, policy-seeking motives are most prevalent (51.6%), followed by office-seeking (41.3%) and vote-seeking motives (9.9%). For the policy-seeking motives, the argument that the party is extremist is most often used (41.4%). For office-seeking motives, claims that a party is incompetent are most prevalent. The reliability of the party, alleged self-isolation (9%) and references to other alternatives/lack of a numerical majority are widespread too (10%). For vote-seeking motives, claims that ruling with a party risks legitimizing the party are most prevalent (5.3%). For willingness, the hierarchy differs: office-seeking motives are most prevalent (67.1%), followed by policy-seeking (50.6%) and vote-seeking motives (23.2%), although the last are more common than in claims to justify unwillingness. The most used office-seeking motive is that they wish to form a government with any party and keep options open for different coalitions, saying that all options for a majority should be explored. For policy-seeking motives, claims of policy or ideological congruence are most prevalent (49.4%). Moderation–inclusion is rarely invoked. Vote-seeking motives, finally, mainly emphasize that the populist party has performed well at the elections and that ruling with it would respect the voters' signal in a democratic society.

Figure 1. Mainstream Parties’ Justifications to (Not) Rule with Populist Parties (in %)

Notes: N = 1,919.

Multivariate analysis

I run seven logistic regression models with a dependent variable claim type (policy or ideological (in)congruence, perceived extremism/moderation, office-seeking motives, vote-seeking motives), separately for willingness and unwillingness to rule with populists.¹⁰

I first consider results for willingness to rule (Table 2). I expected incumbent mainstream parties to be more likely to invoke office-based motives to justify ruling with populist parties than non-incumbent parties (H1A, Model 2). The coefficient for being in the majority is positive and significant (B = 9.267, SE = 2.457, p < 0.000), hence H1A is supported. For H2A, I expected that larger mainstream parties would be more likely to use office-seeking motives to justify willingness to rule with populist parties than smaller mainstream parties. However, this is not supported, as shown by the non-significant coefficient for the mainstream parties' vote share variable in Model 2. H3A expected that to justify their willingness to rule with larger populist parties (as opposed to smaller ones), mainstream parties would be more likely to invoke vote-seeking motives (Model 3). Still, populist party size is not significant. I expected mainstream parties to be more likely to invoke policy-based motives linked to policy–ideological congruence to justify ruling with populist parties that are ideologically closer than for populist parties that are ideologically distant (H5A). Results in Model 1 show, however, that the difference score in left–right ideology between mainstream and populist parties is not significant, rejecting H5A. Finally, mainstream parties involved in actual negotiations with populist parties were expected to be more likely to invoke motives on policy and ideological congruence to justify willingness to rule with populist parties (H6). The negotiation variable (Model 1) is significant and positive (B = 2.454, SE = 1.202, p = 0.041), confirming H6.

Table 2. Logistic Multilevel Regression: Predictors of Policy-, Office- and Vote-Seeking Claims to Justify Ruling with Populist Parties

Notes: B = unstandardized coefficients; SE = standard errors.

^a p < 0.10 (one-tailed), * p < 0.05; ** p < 0.01; *** p < 0.001.

Next, I consider claims about unwillingness to rule with populist parties (Table 3). I expected incumbent (H1B) and larger mainstream parties (H2B) to be more likely to invoke office-based motives to justify their unwillingness to rule with populist parties than respectively non-incumbent and smaller mainstream parties (H1B). Expectations for incumbency (B = 7.327, SE = 1.069, p < 0.001) are confirmed in Model 6. For party size, no significant relationship is found, rejecting H2B. I also hypothesized that to justify unwillingness to rule with larger populist parties, mainstream parties would be more likely to invoke vote-seeking motives (H3B, Model 7). H3B is rejected. Next, H4 stated that mainstream parties would be more likely to invoke policy-based motives linked to perceived extremism to justify not ruling with RWPPs than LWPPs (Model 4). The coefficient of the left–right score of the populist party variable (B = 0.014, SE = 0.003, p < 0.001) confirms H4. Mainstream parties were expected to be more likely to invoke policy-based motives emphasizing policy–ideological incongruence to justify their unwillingness to rule with populist parties that are ideologically more distant (H5B). However, the difference score between left–right scores of mainstream and populist parties (Model 5) is negative and significant, suggesting the opposite direction, rejecting H5B (B = −0.028, SE = 0.010, p = 0.004). Next, I test justifications for populist parties that have ruled before (H7, respectively, in Models 4 and 6). Once populist parties have ruled or supported a cabinet, mainstream parties are not more likely to justify unwillingness to rule by invoking office-seeking motives or less likely to invoke extremism, rejecting H7A and H7B.

Table 3. Logistic Regression: Predictors of Policy-, Office- and Vote-Seeking Claims to Justify Not Ruling with Populist Parties

Notes: B = unstandardized coefficients; SE = standard errors.

^ap < 0.10 (one-tailed); * p < 0.05; ** p < 0.01; *** p < 0.001.

Finally, I consider the role of context. I expected that in Belgium and Germany mainstream parties would be more likely to invoke policy-seeking motives on perceived extremism (H8A), while mainstream parties in Austria and the Netherlands would be more likely to invoke office-seeking motives (H8B), which generally holds. In Germany, mainstream parties are most likely to invoke perceived extremism claims, and the difference with Austria and the Netherlands is significant. Still, Flanders and Wallonia differ: in Wallonia, claims about perceived extremism to justify unwillingness to rule with populist parties are more prevalent. Office-seeking motives are least prevalent in Germany, while the difference between Germany and Flanders and Wallonia is also significant.

The electoral cycle plays a role: before elections, mainstream parties are significantly less likely to justify inclusion by invoking perceived extremism and more likely to use office-seeking motives. There are fewer office-seeking claims being invoked for inclusion and exclusion after election time. Before the elections, mainstream parties are also more likely to invoke vote-seeking claims to justify inclusion. Finally, in the pre-election period, mainstream actors are less likely to use claims of perceived extremism to justify exclusion.

Discussion

Populist parties have performed well at elections throughout Europe, which feeds the debate on a potential government coalition with these parties. So far, which types of claims mainstream parties use to justify whether to (not) rule with populist parties and which factors shape these claims have been underexplored. Hence, the goal was to initiate a research agenda on how political actors justify their (un)willingness to rule with populist parties in relationship to party and context characteristics. This study systematically analysed mainstream parties' political communication on ruling with populist parties in five Western European contexts. Based on the ‘policy, office, votes’ triad, all tweets of 25 mainstream parties (N = 1,919) in Austria, Flanders, Germany, the Netherlands and Wallonia were examined via longitudinal, quantitative content analysis. Four main findings become evident.

First, despite populist parties' electoral success, mainstream parties are overwhelmingly dismissive of the idea of ruling with a populist party. In almost all claims, mainstream parties reject a coalition with a populist party. Still, RWPPs are more likely to be excluded than LWPPs. While this could be explained by anti-prejudice norms and rejection of prior legacies of these parties, it could be linked to the case selection: most countries have a right-wing authoritarian legacy, resulting in stronger rejection of the radical right (Dinas and Northmore-Ball 2020). Hence, mainstream parties' political communication on Twitter is characterized by a tendency to ostracize populist parties. This rejection suggests that mainstream parties tend to act against the mainstreaming and normalization of RWPPs, which corroborates prior evidence (Valentim and Widmann forthcoming).

Second, mainstream parties invoke distinct motives to justify willingness and unwillingness to rule with populist parties. To justify willingness, office-seeking motives are evoked most often, followed by policy-seeking on policy–ideological congruence and vote-seeking motives. To justify unwillingness, policy-seeking motives on perceived extremism are most prevalent, followed by office-seeking and vote-seeking motives. Regarding willingness, the most common office-seeking motives refer to numerical arguments, with mainstream parties emphasizing that they are open to join a coalition with different parties, including the populist party. Policy-seeking motives mostly emphasize the common ideology or similarities in party programmes as the main argument for ruling with populist parties. Vote-seeking reasons mostly involve references to the good electoral performance of the populist party and that ruling with it would respect the electorate's signal. Arguments to justify unwillingness to rule with populist parties differ: here policy-seeking motivations are most frequently invoked, followed by office-seeking and vote-seeking motives. More particularly, policy-based reasons boil down to the argument that the populist party is extremist, justifying its exclusion – which aligns with anti-prejudice norms (Blinder et al. 2013). Office-seeking motives, then, mainly emphasize the incompetence, self-isolation and unwillingness of the populist party to take responsibility, suggesting perceived effectiveness is a key consideration (Bos and Van der Brug 2010). Findings imply that mainstream parties adapt their narratives, dependent upon the situation. Overall – in line with coalition-formation theory – office- and policy-seeking motives trump vote-seeking motives.

Third, the characteristics of mainstream and populist parties shape claims of (un)willingness. The ideological orientation of both mainstream and populist parties matters. Justifications invoking perceived extremism are more likely to be used in reference to RWPPs than LWPPs, for which arguments on ideological and/or policy (in)congruence are more prevalent. More generally, the arguments on perceived extremism are most likely to be used in reference to ideologically distant parties. Hence, mainstream parties seem to prefer the ‘extremism’ argument over the argument on general ideological or policy incongruence. Importantly, this also implies that there is limited variance in this regard as mainstream parties in the ‘inclusion’ model are already rather similar ideologically to the populist parties that are included. Second, parties' incumbency status conditions the justifications of mainstream parties: incumbent parties are less likely to invoke policy-seeking motives. Instead, they are more likely to invoke office-seeking motives. Incumbent parties are responsible and occupy key positions in office, giving rise to appeals about the stability or effectiveness of a new government and the extent to which populist parties are either conducive to this goal or not (Olislagers and Steyvers 2015). Incumbent parties may prioritize being in office and holding power, which could explain why their claims to rule with populist parties are predominantly framed from an office-seeking perspective. These parties may highly value governing experience, competence or reliability of fellow partners and may put aside populist parties as unexperienced, while simultaneously portraying themselves as responsible leaders.

Mainstream parties' size also shapes their claims about ruling with populist parties, but in a different way from what we expected. Larger mainstream parties are more likely to make use of claims regarding perceived extremism, but are less likely to invoke office-seeking and vote-seeking motives. Originally, I hypothesized that larger parties would mostly invoke office-seeking motives, as these parties can be expected to be more relevant and take up a dominant role in the country's leadership and government-formation talks. Still, these parties are less likely to use office-seeking motives but favour more policy-based motives focusing on populist parties' perceived extremism. Possibly, these parties may be more concerned with their image and could fear that invoking office-seeking claims as key reasons to refuse to rule with populist parties would undermine their perceived effectiveness and emphasize their lack of leadership and commitment to actively trying to build a stable government.

Populist party size also shapes mainstream parties’ claims in an opposite way from expected, since policy-seeking motives pertaining to perceived extremism are more likely to be invoked in reference to larger populist parties than smaller ones. This, however, could be explained by the fact that in most countries under study, RWPPs have been larger than LWPPs, and claims of extremism are especially widespread with reference to the former. Mainstream parties invoke distinct claims during actual negotiations with populist parties. In these situations, mainstream parties involved in negotiations with populist parties are more likely to invoke motives on policy–ideological congruence to justify ruling with them.

Finally, context matters: in Germany and to a lesser extent Belgium (Wallonia), ruling with populist parties is more often rejected than in Austria and the Netherlands. Mostly policy-based reasons based on alleged extremism are invoked to justify this exclusion. This aligns with the militant democracy profile of these countries and their lack of experience with populists in government (Bourne and Bértoa 2017), where a formal cordon sanitaire is present and/or extremist organizations are under the observation of the constitutional court. Mainstream parties in Austria and the Netherlands are more likely to invoke office-seeking reasons to justify willingness to rule with populist parties. This suggests that prior experiences of countries conditions how open mainstream parties are to populist parties. Experience with authoritarianism seems key as Germany – where both RWPPs and LWPPs are equally rejected – has a legacy of both radical left-wing and right-wing authoritarianism (Dinas and Northmore-Ball 2020). Hence, path dependency, prior ruling experiences and institutional constraints condition mainstream parties' political communication on ruling with populist parties, implying that narratives are context-specific and idiosyncratic. In countries with distinct experiences (e.g. with communist legacies), rejection could be more strongly aimed at LWPPs too. Future studies should expand the geographical scope to assess this.

The results reported offer key insights into parties' political communication and in the type of claims and strategies that are used to motivate mainstream parties' (un)willingness to rule with populist parties. The comparative approach where claims of parties in five Western European regions are assessed presents one of the key strengths of this study. As such, this study initiates a broader research agenda on the mainstreaming and normalization of populism by disentangling mainstream politicians' justifications, claims and behaviours as well as its ramifications for voters. Many questions require further attention: When and why do politicians prioritize one justification over another? Which justifications are more likely to be successful? How do they affect voters? Do politicians also act upon these claims? Notwithstanding these limitations, these findings point the way towards future substantive work regarding justifications of preferences on ruling with populist parties and shed light on how different arguments are being articulated by distinct parties and that this is largely conditioned by party and political system characteristics.