Populist Attitudes and Direct Democracy: A Questionable Relationship

: Earlier research links citizens’ populist attitudes with the support for referendums. However, the foundations and meaning of this relationship remain unclear. This research note proposes a theoretical, conceptual and methodological discussion that identiﬁes three main problems: studies linking populist attitudes with support for referendums have a rather narrow theoretical framing limited to populist studies, there is much ambiguity surrounding the role of direct democracy in the political system, and there is a tautology in studying the relationship between populist attitudes as measured through various indices and the preferences for direct democracy. Our goal is to discuss such problems and to propose several avenues to circumvent them. In particular, we believe that connecting to adjacent literatures beyond populist studies could be an important improvement.

R esum e: Les travaux ant erieurs font ressortir un lien fort chez les citoyens entre les attitudes populistes et le soutien au r ef erendum. Cependant, les fondements th eoriques et la signification de cette relation ne sont pas totalement clairs. Cette note de recherche propose une discussion th eorique, conceptuelle et m ethodologique des travaux ant erieurs. Elle identifie trois probl emes principaux : le cadrage th eorique des travaux liant attitudes populistes et d emocratie directe m eriterait d'être elargi, la fonction exacte de la d emocratie directe dans le syst eme politique doit etre clarifi ee, et il y a une certaine tautologie dans les indicateurs mobilis es pour etudier les attitudes populistes et le soutien au r ef erendum. La note de recherche propose des solutions permettant de contourner ou r esoudre ces probl emes, notamment en elargissant la discussion audel a des etudes sur le populisme, et en s'ancrant dans les litt eratures adjacentes.

Introduction
Several works have recently tried to connect citizens' populist attitudes with support for referendums. They build upon longstanding reflections among scholars working on ideationalas opposed to strategic or discursivepopulism regarding the conception of democracy promoted by populist actors. They describe the people-centric nature of populism (Canovan 1999;Mair 2002;Mudde 2004) that promotes a model of democracy in which the people should lie at the core of democracy and "politics should be an expression of the volont e g en erale (general will) of the People" (Mudde 2004: 543). Ideational populism is considered to be closely connected with the support for direct democracy (referendums) in particular (Mudde 2007: 151-152). The empirical studies seeking to gauge this relationship use survey items. Their findings indicate that populist attitudes are positively associated with the support for referendums either in terms of use or in terms of providing them a more decisive role in the decisionmaking process. Evidence comes from the Netherlands (Jacobs et al. 2018;Zaslove et al. 2020), a comparison of France, Germany, Switzerland and the UK (Mohrenberg et al. 2019), and 17 European countries (Rose and Wessels 2020). These studies have the concept of populist attitudes in common, but they operationalize it differently. They either use the scale developed by Akkerman et al. (2014), the measurement proposed by Schulz and colleagues (2018), or propose to measure them through an item referring to how the will of the people should be unconstrained by minorities.
We welcome the contribution made by these studies towards a better understanding of what model of democracy populist citizens support. A recent work illustrates that citizens holding populist attitudes in Turkey, five Latin American, and nine European countries were supportive of democracy. However, we know little about what kind of democracy they support, or how they want it to be organized (Kaltwasser and van Hauwaert 2020).
In spite of their merits, we believe that these studies are far from closing the debate. In our opinion, they merely pave the way for more empirical studies on populism and democracy. This research note identifies three important ways in which studies about populist attitudes and direct democracy can be improved and proposes solutions to do it. The first section outlines the need to enlarge the theoretical foundations of recent studies beyond the mere literature on populism. The second section focuses on the lack of conceptual clarity regarding the role of direct democracy in a political system. The third section identifies a tautology in some of the methodologies used to relate populist attitudes and direct democracy. The final section outlines several avenues for further research along the lines identified this note.
Our work deviates from the standard approach of many articles. It does not propose the empirical test of hypotheses derived from an established body of literature because this is beyond our goal. Instead, we propose a conceptual and methodological discussion of recent publications on populism and direct democracy. The topic has raised significant interest recently, and publications are multiplying. Research projects show that this agenda is far from being closed. It is therefore important, we believe, to allow for a broader discussion regarding how future studies could be improved. More precisely, we discuss three dimensions of recent publications on populism and direct democracy that, in our opinion, invite for scholarly debate. We then examine how research from several other strands of literature could be used to develop fruitful new avenues for research on populism and democracy.

A Problem of Theoretical Framing: Citizens' Preferences About Who Should Govern
We argue that studies linking populist attitudes with support for referendums suffer from a problem of theoretical framing. Their theoretical foundations are almost exclusively rooted in studies on populism, while largely ignoring two other strands of literature that could be directly relevant for their topic of interest. First, studies on public support for referendums, such as Bowler, Donovan and Karp (2007) or Schuck and de Vreese (2015), are hardly mentioned. Second, they also tend to leave aside the broader and burgeoning literature on citizens' preferences for different models of democracy, including direct democracy. Insights from these two strands of literature could enrich studies on the link between citizens' populist attitudes and support for referendums.
First, let us look at support for direct democracy where one important variable is the role of education. The literature on public support for referendums has underlined that the link between education and support for direct democracy is complex to capture. Support for direct democracy may come from both the lower educated but also from some higher educated citizens (Bowler et al. 2007). In other instances, there is no effect of education, which indicates that direct democracy is equally supported by low and high educated citizens (Gherghina and Geissel 2020). Among the higher educated, the support for direct democracy is mediated by how citizens apprehend the limits of direct democracy to protect minorities (Anderson and Goodyear-Grant 2010). These studies highlight that the effect of education on support for referendums is not univocal or linear. In this respect, it would be interesting to examine if and how education interacts with populist attitudes in shaping citizens' attitudes towards direct democracy. For instance, several works control for the effect of education on support for a greater use of referendums (Jacobs et al. 2018;Mohrenberg et al. 2019). Yet, the above-mentioned studies on the effect of education on support for democracy among the broader population invite to go further than simply controlling for the effect of education. One might, for instance, expect education to be a mediating variable that differentiates the effect of populist attitudes on support for referendums among the more and the less educated populist citizens.
Another example of how the literature on public support for referendums may feed into studies on citizens' populist attitudes and referendums are insights from recent studies on instrumental motivations for support for direct democracy (Werner 2020). Public support for referendums is not only driven by ideological attitudes (like populist attitudes). Attitudes may also be instrumental in citizens' desire to call for the organization of a referendum. They would only support a referendum when they believe it increases their chances to see their most preferred policies adopted. Building on these findings, one could wonder whether citizens holding populist attitudes really support referendums because of their conception of democracy. Otherwise, this support might be an effect of a feeling that the referendum would make it easier for people to push for their most preferred policies than through representative institutions.
Second, the growing literature on citizens' preferences for alternative models of decisionmaking could also help to consolidate studies on the link between populism and democracy. This literature has shown that citizens in Western democracies promoted various models of democracy. In particular, three models prevail: a representative model of democracy, a technocrat model of governance (Bertsou and Caramani 2020; Hibbing and Theiss-Morse 2002), and models that would increase citizens' participation via either instruments of participatory and direct democracy (Bengtsson and Christensen 2016;Gherghina and Geissel 2017). Recently,  showed that these models should not be perceived as mutually exclusive. Rather, many citizens want to combine their different elements. We therefore argue that this literature on citizens' support for various models of democracy may enrich recent work on populist citizens and their demands for a greater use of referendums in a twofold way.
It may be relevant to examine how support for direct democracy among populist citizens relates to attitudes towards other models of democracy. Some of the studies we cite here do this to some extent. For example, Zaslove et al. (2020) examine in parallel the link between populist attitudes and support for referendums and deliberative assemblies. Mohrenberg et al. (2019) connect support for direct democracy and for technocracy. Research could go one step further and aim to connect populist attitudes with the preferences for the various models of democracy identified by the literature: direct democracy, deliberative democracy, representative democracy and technocracy.
Moreover, the two strands of literature can feed each other in linking process preferences with participatory behaviors of citizens. In this regard two recent studies on populist citizens and direct democracy provide puzzling findings. Jacobs et al. (2018), for example, found that populist citizens in the Netherlands were more supportive of referendums but not more likely to vote when referendums were held. Broadening their study to voting in elections and to protesting, (almost) the same team found no effect of populist attitudes among Dutch citizens on the propensity to vote and a negative effect on protest ). The literature on citizens' process preferences has already examined the relationship with participatory behaviours, however in other countries. For example, Gherghina and Geissel (2017) find that the propensity to vote in German elections is higher among representative democrats but also among those oriented towards technocrats. Participatory democrats, by contrast, are less likely to vote in elections but more likely to vote in referendums. The latter group is more likely to get involved in participatory processes. Looking at Finnish citizens, Bengtsson and Christensen (2016) show that representative democrats are the most likely to vote in elections. Participatory democrats are the most engaged in institutional (i.e. voting, contacting politicians, being member of a party) and non-institutional participation (i.e. petition, buycott, boycott, demonstrations), while technocrat-oriented citizens are the least politically active.
These examples show, we believe, that a way forward for studies on populism and direct democracy could be to enrich its theoretical foundations by broadening the literature on which it has been built so far. It would help avoiding one of the pitfalls in studies on populism, which is "that it remains too detached from adjacent literatures" (Rooduijn 2019: 367).

A Conceptual Problem: Complementing vs. Replacing Representative Democracy
The conceptual problem of the works connecting citizens' populist attitudes with support for direct democracy is the ambiguity surrounding the role of direct democracy in the political system. These works measure the support for direct democracy through survey questions about citizens' willingness to have more frequent referendums (Mohrenberg et al. 2019;Rose and Wessels 2020;Zaslove et al. 2020) or whether citizens would like referendum results to have a greater political weight or to be binding (Jacobs et al. 2018). Conceptually, they look at referendums as instruments of direct democracy that "are people centered, reduce the power of the elite and are a means to keep the corrupt elite in check" (Jacobs et al. 2018: 520); for a similar approach, see also Mudde (2007: 151-152).
This approach has several problems. First, it is unclear to what extent direct democracy should become central in the political system. It is nowhere explained whether citizens with populist attitudes desire referendums to become the most frequent way to adopt new policies, e.g. like in Switzerland or in several Western American States (Morel and Qvortrup 2017), or rather prefer referendums to be held occasionally, for major political decisions or in the case of popular dissent with decisions by public authorities. If the latter applies, populists would be very much like stealth democrats (Hibbing and Theiss-Morse 2002) who do not want to get too much involved in policy-making but still want popular votes to be held in critical junctures or when elected politicians are to be kept under popular control.
Second, it is unclear in these studies whether direct democracy procedures should prevail over decisions from representative institutions or not. Over recent decades, in many countries, direct democracy has become an element of the institutional architecture of representative democracy instead of replacing the institutions of representative democracy (Altman 2018; Set€ al€ a 2006). Elected politicians, governments and parliaments play a role in direct democracy: they can (and do) initiate referendums, they are the object of referendums (e.g. the abrogative referendums or recall elections) and they have to implement decisions taken by citizens in referendums. It remains unclear whether this is what citizens have in mind when they declare that they want more referendums to be held. The support for direct democracy may mean that citizens demand a direct say in the important decisions, with their vote being decisive for subsequent actions, i.e. implementation (Budge 1996).
This nuanced relationship between representative and direct democracy is to some extent already discussed by some of the studies seeking to connect populist attitudes and support for direct democracy. For example, Jacobs et al. (2018) analyze support for binding referendums, which means that direct democracy supersedes elected representatives. Mohrenberg et al. (2019) go one step further by exploring whether citizens would accept constitutional constraints to the will of the people expressed via referendums. Yet, we would argue that research needs to go further in understanding precisely how direct democracy instruments would coexist with institutions of representative democracy, and what role elected representatives would have in direct democracy processes (Allen and Birch 2015). Here as well, inspiration can be found in the literature on process preferences. For example, a recent study illustrates that citizens supporting citizens' assemblies advising elected representatives are not the same than those in favor of replacing elected bodies with assemblies composed of citizens selected by lot (Bedock and Pilet 2020). As such, research can dig deeper into the articulation between direct and representative democracy and how citizens, and especially those holding populist attitudes, view it.

A Methodological Problem: Addressing a Tautology
The third problem we identify in recent research testing the effects of populist attitudes on the demand for referendums lies in the way the link between the two variables has been operationalized in some studies. This section argues that there is a tautology in studying the relationship between populist attitudes as measured through various indices and the preferences for direct democracy. The tautology consists in a similar measurement of the left and right sides of the equation. Many statements included in the index of populist attitudes 1 refer explicitly to the central role of citizens in the process of decision-making and representation. These statements oppose citizens and politicians and they read as follows: "The people, and not politicians, should make our most important policy decisions", "I would rather be represented by a citizen than by a specialized politician" (Akkerman et al. 2014;Geurkink et al. 2020;Hawkins et al. 2012), "The opinion of ordinary people is worth more than that of experts and politicians" (Elchardus and Spruyt 2016) and "The people should be asked whenever important decisions are taken" (Schulz et al. 2018). These four items belong to different indices of populist attitudes, but they are all subject to the same general problem of tautology discussed below.
We do not deny that these items are appropriate to capture core components of populist attitudes, people-centrism and anti-elitism. It is indeed true that they distinguish quite well populist and non-populist citizens. Yet, it becomes more problematic when such items are used to explain support for referendums. The reflection becomes then tautological because these items refer to essential features of direct democracy, as illustrated in the literature. Direct democracy is a publicly recognized, institutionalized process that gives citizens a direct say in the decision-making process. In direct democracy citizens can "decide or emit their opinion on issuesother than through legislative and executive electionsdirectly at the ballot box through universal and secret suffrage" (Altman 2011: 7). Earlier research shows that direct democracy instruments are used across democracies to give citizens a say on a wide variety of major policy issues. The importance of policy proposals on which citizens have a direct influence requires special consideration. Policy proposals reach the ballotor can be submitted to the legislature as it is the case of the agenda initiativesif they are salient for the public (Lupia and Matsusaka 2004). This salience is demonstrated through a number of signatures for citizen-initiated direct democracy procedures or by a decision in the legislative or executive if they are elite-initiated. A comparison of all national-level referendums organized in Europe since 1793 indicates that the vast majority of popular votes were on major policies related to the political and electoral system, interior policies, economy, foreign affairs and constitution-making (Silagadze and Gherghina 2019).
Direct democracy aims to empower citizens and to reduce the role of elected officials in policy-making. It does not mean that elected representatives lose any role. They may still be involved in setting up the referendum, in the referendum campaign or in the implementation of the popular vote. Yet, the central moment in the policy-making process in the direct democracy model is the popular vote via referendum. In the US, the direct democracy procedures were introduced at the end of the 19 th century to restrict the power of political parties over legislation (Bowler and Donovan 2006). In contemporary times, many citizens' initiatives in the US are used by non-political groups, e.g. social movements, to submit issues to popular votes without the intervention of the state governments and parliaments, which are controlled by political parties (Kriesi and Bernhard 2011). In general, the citizen-initiated referendums allow ordinary citizens to bring on the political agenda policies that are ignored by politicians and institutions of representative democracy. For example, in 2016 a citizen-initiated referendum was organized in Bulgaria on topics that could harm most political parties: limitation of state funding for parties, the introduction of compulsory voting and the change of the electoral system.
These features indicate the essence of direct democracy: important policy decisions are taken by citizens and the role of politicians is constrained. The four statements included in various indices of populist attitudes gauge thoroughly this essence. As such, the people are in charge of decisions, the representation in direct democracy happens through the voice of other citizens, and the opinion of citizens matters (as opposed to that of politicians). If these statements measure various components of direct democracy, it makes little sense to investigate how they are related to preferences for direct democracy. Some of the papers we cite acknowledge this problem of tautology. For example, Jacobs et al. (2018) run robustness checks for their results, excluding one item too closely related to popular sovereignty in their index of populist attitudes. Mohrenberg et al. (2019) opt for an index of populist attitudes excluding items tapping too directly into popular sovereignty. Yet, we still believe that the methodological problem should be addressed more upfront in any future study in order to make clear that one should avoid the tautology we underline here. Reflection is therefore needed to propose alternative statements that might capture populist attitudes in their people-centrism and anti-elitism but that would be more neutral towards the use of referendums.

The Way Forward
The three weaknesses identified in this research note show that the studies linking citizens' populist attitudes with direct democracy can be a useful starting point for more research on populism and democracy. This section makes a few suggestions about how to circumvent the problems and challenges outlined in the previous sections, and how to advance the research agenda.
To begin with the theoretical framing of studies on populist attitudes and support for direct democracy, we believe that the literatures on populist attitudes, public support for direct democracy and citizens' preference for decision-makers can feed each other. One way to proceed would be, for example, to focus on how populist attitudes are linked to various preferences for different models of decision-making. The literature on citizens' democratic preferences (see section 2) shows that citizens in contemporary democracy hesitate between several main models of decision-making: centered around elected politicians, directly governed by citizens, and led by independent experts or technocrats. This broad picture is absent from the studies connecting populist attitudes with support for referendums. For example, it may be relevant to understand how citizens with populist attitudes conceive referendums vis-a-vis these alternative models of decision-making.
Another possibility is to combine the preferences for broad decision-making processes (who should govern) with support for specific institutional mechanisms such as referendums, deliberative mini-publics, technocratic governments, elections etc. Zaslove et al. (2020) make a first step in that direction and examine the link between populism and support for deliberative citizens' forums. Such a broad approach of the various models and instruments of democracy would directly contribute to enriching our understanding of what kind of democracy populist citizens endorse. Populists 'campaign for a modification of democratic procedures' (Mudde and Kaltwasser 2017: 95), but the alternative model that they propose continues to remain unclear.
Second, the conceptual weakness that we have underlined can be addressed by looking beyond the assumption that direct democracy replaces representative democracy. The practice indicates that more often it complements it. As such, the studies looking at populist attitudes must explore other arguments than the opposition between direct and representative democracy. More efforts are required to explaining when, where and how populist citizens desire to use directed democracy: with what frequency, who initiates it, on which policy issues. Equally important would be to ask for the role elected politicians should keep relative to the decisions made by citizens (e.g. none, limited or full). Politicians lie at the core of the system and direct democracy may be used for major policy decisions and to keep them under control.
Thirdly, the methodological weakness can be avoided by removing the tautological indicators. The straightforward approach to this is to use scales or indicators of populist attitudes that do not tap into items that are close to the essence of direct democracy such as "people and not politicians should decide"; for an example of a potential alternative, see Rose and Wessels (2020). Another possibility is to decompose the scales to separate people-centric attitudes and other components of populism (Mohrenberg et al. 2019). The various propositions that we have made are an invitation for scholarly discussion. Our goal is not to propose their immediate empirical tests. Further scientific exchanges with authors in the field of (direct) democracy and populism should be open following this research note before going to new surveys or other empirical protocols.
Finally, we also believe that digging deeper into how populist citizens may conceive direct democracy could enrich our understanding of non-populist citizens attitudes on the topic. There are many non-populist citizens who support a greater use of referendums. They may have different reasons to support this mode of decision-making, and they would assess the virtues and vices of referendums differently. A comprehensive understanding of the link between populist attitudes and direct democracy should therefore go hand in hand with increasing our knowledge of how all citizens relate to referendums and to their articulation with representative democracy in general.

Funding Information
Jean-Benoit Pilet's involvement in this research note is part of the project POLITICIZE-Cure or Curse? This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No 772695).