The democratic interface: technology, political organization, and diverging patterns of electoral representation

ABSTRACT

Democracies are experiencing historic disruptions affecting how people engage with core institutions such as the press, civil society organizations, parties, and elections. These processes of citizen interaction with institutions operate as a democratic interface shaping self-government and the quality of public life. The electoral dimension of the interface is important, as its operation can affect all others. This analysis explores a growing left-right imbalance in the electoral connection between citizens, parties, elections, and government. This imbalance is due, in part, to divergent left-right preferences for political engagement, organization, and communication. Support on the right for clearer social rules and simpler moral, racial and nationalist agendas are compatible with hierarchical, leader-centered party organizations that compete more effectively in elections. Parties on the left currently face greater challenges engaging citizens due to the popular meta-ideology of diversity and inclusiveness and demands for direct or deliberative democracy. What we term connective parties are developing technologies to perform core organizational functions, and some have achieved electoral success. However, when connective parties on the left try to develop shared authority processes, online and offline, they face significant challenges competing with more conventionally organized parties on the right.

KEYWORDS:

• Connective parties
• hybrid organization
• democracy and technology
• communication and organization
• movements and parties

Democratic systems are suffering disruptions in many nations. Perhaps the most important disconnect between citizens and institutions is in the area of parties and elections. The weakening of modern era social institutions such as unions and other group membership organizations has eroded traditional bases of party support. In addition, global business pressure to privatize public goods and expand market solutions have reduced policy and ideological differences among center parties, both left and right (Crouch, 2004). As the political center becomes ‘hollowed out’ – to use Mair’s (2013) term – voter identification weakens, electoral marketing communication becomes less effective, and competition for votes increases from new parties (and resurgent older ones) on the political extremes. Many of the new parties have taken unconventional or ‘hybrid’ (Chadwick, 2013) organizational forms because of factors such as scarce resources, adoption of social media communication strategies, or thin professional staffing beyond dominant founding personalities.

A puzzling element in the midst of this disruption is that radical right parties have done better than the left in electoral terms. Our data show just how large this imbalance at the electoral interface has become, despite the comparable numbers of voters, both left and right, who have been adversely affected by social and economic change, and who have doubts about the legitimacy of center parties and about democracy itself. This article explores the underpinnings of this electoral puzzle through the lenses of organization, communication, and technology.

The argument revolves around the idea that divergent preferences are emerging between citizens on the left and the right for how to organize the electoral linkage between society and state. A vertical bureaucratic model has long dominated conventional party politics, and has been adapted by most parties on the radical right. By contrast, substantial numbers of voters, primarily on the left, are developing preferences for horizontal models that favor direct, participatory, deliberative democracy – even in the electoral arena. This demand for more distributed party and campaign organization puts parties that supply such horizontal organization at a disadvantage in electoral contests where conventional organizations are often better able to coordinate communication, use conventional media, develop campaign strategy, and mobilize voters.

With so many center parties feeling pressure from citizen defection, many have experimented with uses of communication technologies to boost engagement. However, many observers remain skeptical about how meaningful such outreach affordances really are, either for party members and voters or for parties (e.g., Chadwick & Stromer-Galley, 2016; Gauja, 2016; Gibson, Greffet, & Cantijoch, 2017). In this context of organizational experimentation and change, we note a special set of cases in which technologies of engagement are not just grafted onto existing bureaucratic organization, but actually replace core bureaucratic functions. What we call ‘connective parties’ replace such core operations as agenda setting and candidate selection with a mix of in-person assemblies coordinated through technology platforms and affordances. In some cases, these party ‘operating systems’ have developed from necessity, as happened when Beppe Grillo’s Five Star Movement (M5S) in Italy grew quickly into a national political force beyond the management capacity of Grillo and a few allies. In other cases, connective parties have attempted to strike a balance between movement demands for direct democracy and the need for central leadership, as with Píratar in Iceland, Alternati v et in Denmark, and Podemos in Spain. We explore this tension in all of these cases, noting that in particular Podemos represents an interesting example of this persistent tension between central leadership and connective organization, resulting in splits in party leadership and the defection of some of its movement base.

These tensions between traditional bureaucratic organization and technology-enabled distributed democracy play out differently across cases, but often become more limiting on the radical left, where organizational preferences for direct democracy are most pronounced. These developments have implications for the nature of democracy itself and for the role of parties and elections in its future. We navigate this argument with caution, as the current electoral landscape is filled with anomalies and uncharted developments, such as the Brexit vote in the UK, the election of Donald Trump in the US, the victory of Emmanuel Macron’s unconventional En Marche! Party in France, and the rise of anti-democratic parties in various democracies.

The current era of economic turbulence and political turmoil has generated new patterns of political organization and communication, few of which may turn into stable new institutional arrangements at the democratic interface. However, against this backdrop of change, our data from European democracies reveals a left-right divergence of citizen engagement preferences that favors political parties on the radical right. It is true that center parties are still in government in most nations, but it is also clear that many voters are searching for alternatives. Even more disturbing is that support for democracy itself is in decline across Europe and the US, particularly among younger age cohorts (Foa & Mounk, 2016). In the context of this crisis of democracy, disproportionate electoral gains on the right become more worrisome, as many parties advocate restrictions on immigration, and generally promote intolerance for diversity and basic rights among established minority populations. Even more extensive threats to established democratic principles have emerged in nations where right wing parties have entered government (e.g., Poland, Hungary, and Turkey) and imposed serious restrictions on the press, academic freedom, and other basic liberties. Thus, our focus on politics beyond the center, along with our assessment of why center parties have had trouble re-establishing voter loyalty, offers a full spectrum look at the future of democracy itself.

The democratic interface in the electoral arena

The democratic interface can be understood as the communication and organization processes that engage citizens with institutions of collective self-governance. It is possible to describe citizen interfaces with the press, the courts, civil society organizations, social movements, and other institutions thought to constitute democratic societies. In this analysis we focus on the electoral interface, which may affect how all of the others operate. In particular, we examine the demand for different kinds of political organization from party followers, and the role of communication technology in meeting (or not meeting) those demands as parties experiment with various forms of organization.

The electoral interface in modern representative democracies hinges on the notion of institutional linkages between society and government. The organization and communication of linkage is commonly perceived to be one of the most important functions that political parties perform: they are seen as key mechanisms for articulating and aggregating the ‘raw’ flow of interests and ideas into comprehensive programs for collective governance (Almond, 1960, pp. 35–36; Sartori, 1976). At present, two distinct models are emerging for what electoral linkage entails and how it is achieved.

Although the electoral interface is theoretically malleable, there are strong pressures to conform to similar organizational forms during any given historical period. In the currently dominant model, parties organize linkage in a top-down or centre-out manner that increasingly detaches citizen engagement from core organizational processes, except through communication-based appeals during elections. This marks an historical departure from the mass parties of the mid-twentieth century in which channeling the interests of cleavage-based membership groups was the benchmark of party democracy. The breakdown of traditional party cleavages, and the movements that defined them, led to what have been termed ‘catch-all’ parties that de-emphasized ideology and offered marketable programs to different groups (Kirchheimer, 1966). More recently, with the rising influence of business, and the defection of voter support bases, the center has shifted toward so-called ‘cartel parties’ that collude to secure state benefits and coalition advantages (Katz & Mair, 2009). Recent comparative mapping reveals surprising uniformity in party structures within countries (Poguntke et al., 2016). There are, of course, variations in party culture and ideology, as when parties prioritize inclusiveness or strong leadership, or connections with (or distance from) social movements (Chiru, Gauja, Ghergina, & Rodriguez-Tenuel, 2015; Gauja, 2016; Hutter, 2014; Scarrow, 2015, chapter 2).

As these transitions in party organization pushed citizens farther from the ideational and organizational work of parties, election campaigns came to rely increasingly on marketing strategies to assemble voters through professionalized party communications and personalized focus on party leaders (Blumler & Kavanagh, 1999). The reliance on professional marketing, branding and spin all reinforce vertical models of linkage and engagement: leader-focused, centralized, bureaucratic organizations that market finished programs to the public. Both the vertical organizational forms, and the communication processes that mirror them, reinforce what has been criticized as the ‘hollowing out’ of the established parties in terms of public ideas and citizen engagement (Crouch, 2004; Mair, 2013).

As campaign communication from the center appears hollow and ineffective to growing numbers of citizens, grassroots preferences for new democratic linkages are introducing instability into elections. Indeed, the array of organizational preferences spans a spectrum from more expert-led efficiency to growing demands for more direct participation in democratic processes (Bengtsson & Mattila, 2009; Close, Kelbel, & Van Haute, 2017; Font, Wojcieszak, & Navarro, 2015; Webb, 2013). Many conventional political organizations have trouble satisfying preferences for flexible and autonomous engagement that blur distinctions between member-participant and voter-supporter within and beyond electoral politics (Bimber, Flanagin, & Stohl, 2012). In particular, traditional bureaucratic parties struggle to satisfy demand for very different kinds of democratic linkages that can loosely be termed horizontal, based on the shared generation of ideas, flexible yet meaningful engagement, and organizational processes that draw input from the periphery as well as center.

What seems clear is that, despite being well aware of this changing mix of voter demands, few center parties have abandoned the hierarchical structures that are beneficial to maintaining internal leadership and power hierarchies, access to state resources, and attractiveness to coalition partners that define the benefits associated with their political cartel status. As a result, those parties have experimented with a variety of relatively superficial communication technologies aimed at developing flexible ties with followers (Chadwick, 2013; Chadwick & Stromer-Galley, 2016; Gauja, 2016; Scarrow, 2015; Vaccari & Valeriani, 2016). Following on Margetts (2001) prediction that ‘cyber parties’ would use technology to strengthen the relationship directly with voters, parties further blur the distinction between member and supporter to offer diversified affiliation structures (Gauja, 2016) or ‘multispeed’ memberships (Scarrow, 2015). Some pursue hybrid repertoires aimed at switching between conventional party and social movement modes of communication and engagement (Chadwick, 2013; see also Chadwick & Stromer-Galley, 2016; Gibson, 2015; Vaccari & Valeriani, 2016). Some innovation has even addressed areas of intra-party democracy, such as leader and candidate selection and policy development processes. Such moves can be observed across the center left and right, although established social democratic parties seem more active, particularly in reforms of policy development processes (Chiru et al., 2015; Gauja, 2013, 2015, 2016).

Intriguingly, however, much of this technology-enabled innovation serves to reinforce the conventional model of vertical linkage. Like other organizations facing similar challenges in what has been termed a post-bureaucratic society (Bimber et al., 2012; Heckscher & Donnellon, 1994), many parties use technology mainly to reduce the costs of organization and communication. In many instances, they embrace enhanced engagement, but keep it peripheral to principal party operations. Studies in the context of US election campaigns observe strategies of ‘managed interactivity’ (Kreiss, 2012) or ‘controlled interactivity’ (Stromer-Galley, 2014; see also Jensen, 2017). Invitations for broader input are often limited to non-binding consultation or decisions on already established alternatives (Invernizzi-Accetti & Wolkenstein, 2017). In all, there is little evidence that such peripheral technology-enabled innovation is reversing long-term declines in membership or voter confidence (van Biezen, Mair, & Poguntke, 2012; Poguntke et al., 2016). Indeed, some suggest that half-hearted party organization and participation reforms further disillusion citizens (Della Porta, Fernández, Kouki, & Mosca, 2017, p. 179).

Direct social media communication with voters may produce some marginal advantages for conventional parties, as when UK Labour used social media strategically to partially circumvent the traditional media in the 2017 elections. However, socially mediated contact between candidates and supporters may be perceived as more genuine (and replace bureaucratic functions) when coming directly from politicians with thin formal party organizations, as with Donald Trump’s regular tweeting to growing legions of followers (around 30 million) through the 2016 election and afterwards.

The limited effectiveness of center parties chasing discontented voters has either pushed growing numbers of citizens out of the electoral process, or sent them searching for alternatives. While some of these emerging parties such as Macron’s En Marche! or Grillo’s Five Star Movement have spanned the center of the spectrum, many more operate at the political extremes. To the extent that preferences for direct democratic participation are more concentrated on a reconfigured left, corresponding forms of party organization attempting to meet those demands face challenges of interest aggregation, communication coherence, and scale of voter mobilization – all of which may contribute to a left-right imbalance of electoral outcomes. This problem of unequal representation (which we demonstrate empirically below) more generally reflects ways in which preferences for different forms of organization play out in terms of the character, scale, stability, and coherence of various political mobilization processes (Dahlberg-Grundberg, 2015; Kavada, 2013). Indeed, we may be entering an era in which different organization and participation preferences operate as forms of ideology that better distinguish voters on the far left and right than earlier distinctions about class or economic positions (Bengtsson & Mattila, 2009; Gauja, 2016; Lavezzolo & Ramiro, 2017).

The relatively smaller number of parties (concentrated more on the left) that attempt to meet the demand for horizontal democratic organization face more challenges than parties on the radical right that more easily satisfy voter preferences for hierarchy, central leadership, and simpler ethno-nationalist messages. One result of this is that voting rates are higher on the radical right than the radical left over the last decade, despite higher levels of most other forms of political participation on the left. We suggest that a contributing factor in these trends is a short supply of organizationally acceptable parties on the left. The following two sections consider these differences in the demand and supply of different kinds of political organization in the European context.

Democratic discontent and diverging organizational preferences

As suggested earlier, the factors underlying recent democratic discontent are complex, interconnected and affect citizens across the ideological left – right spectrum. They include the continuing crisis of neoliberal capitalism, the growth of economic and political inequality, economic austerity and market solutions that render life in society precarious for many citizens, and the overlaid problems of war, climate change, mass migrations and refugee crises (Crouch, 2004; Harvey, 2007; Klein, 2015; Kriesi, 2012). Such factors underpin continuing shifts in the European left-right ideological space around the economy and social values, reinterpreting the traditional economic dimension of class and the role of government through divergent left and right responses to globalization, nationalist and racial identity politics, and welfare nationalism (Kriesi et al., 2006, 2008).

This section explores the demand side of the European electoral interface as it plays out in organizational and communicational terms. As noted above, citizens that self-identify on the left or right can be broadly expected to differ in terms of preference for horizontal or vertical forms of political organization (with exceptions, for example of left authoritarians or possibly older generation ideological left voters). The following discussions examine the underpinnings of different preferences for organizational engagement on the right and the left.

Organizational preferences on the right

Preferences for vertical or hierarchical organization on the right have long been associated with higher respect for authority, and related emotional appeals of authoritarianism (Parker & Barreto, 2013). There is less pressure for internal democracy and participation on the party organizations, which are assumed to have an easier time maintaining centralized organization (Beyme, 1985; Chiru et al., 2015; Gauja, 2016, chapter 6). The focus on a delimited set of issues, e.g., concentrated on in-group, nationalist, or racist issue positions, also favors centralized organization (Mudde, 2007; Parker & Barreto, 2013).

Citizens on the radical right also disproportionately prefer national and cultural uniformity anchored in rules, traditions and customs, which makes for a readily available and consistent set of party positions on the right. We have explored these patterns using European Social Survey data (ESS 7, v 2.1, 2014, Israel excluded).¹ As explained in more detail in the next section, we defined radical left and radical right according to the left-right scale break points beyond which people identified with parties on the center left and center right (e.g., beyond Labour/Social Democrats and Christian Democrats/Liberals). Party positions along a left-right spectrum were scored independently using the Chapel Hill Expert Survey, also described more fully in the next section. We combined measures for ‘importance to do what is told and follow rules’ and ‘importance to follow traditions and customs.’² Agreement on these combined measures was roughly 40% on the radical right compared to just 22% the radical left. Even higher levels of agreement approaching 60% occur on the radical right for the statement that ‘it is better for a country if almost everyone shares customs and traditions.’ In contrast, agreement with this statement on the radical left was just above 30%. And 73% of respondents on the radical right agreed that it is important for immigrants to be committed to the way of life in a country, contrasted with around 50% on the left. All of these differences are highly significant (p ≤ .0001). These positions play into the simpler nationalist and exclusionary appeals of parties on the radical right.

Many parties on the radical right began as movements, and some have included calls for direct democracy (Pauwels, 2014). However, recent studies suggest that citizens on the right prefer ‘stealth democracy’ (Hibbing & Theiss-Morse, 2002) that favors efficiency and expert input over debate and direct participation (Bengtsson & Mattila, 2009; Bowler, Denemark, Donovan, & McDonnell, 2016; Lavezzolo & Ramiro, 2017; Webb, 2013). This would imply that engagement preferences continue to be compatible with the leader-centered and centralized organization of the vertical model. There are of course tensions between movements and parties, and within parties when factions contest leadership. For example, in Germany, the anti-Islamic grassroots movement PEGIDA shares a considerable Facebook user base and set of issue frames with the Alternative für Deutschland party. (Stier, Posch, Bleier, & Strohmaier, 2017). Early stages of the movement-party relationship suggested that the party grew by mirroring selective messages of the movement (Rucht, in press). However, leading into the 2017 national elections, party membership rejected calls for a more moderate political course and doubled down on nationalist and racist positions. This led to losses of some center right support in polls, but consolidated the national leadership around an even more extreme agenda. Internal splits have characterized other radical European parties on both ends of the spectrum, including Podemos in 2016 and the Finns Party in 2017.

Organizational preferences on the left

Movements associated with left parties have long raised demands for intra-party democracy, as happened with the ‘fundi’ faction of the German Greens in the 1980s (Gauja, 2016; Poguntke, 1987). Despite such democratic tensions, the party leadership gradually adopted cartel strategies aimed at becoming attractive coalition partners, even at the expense of movement democracy. More recent generation of left parties have faced even more persistent demands for direct engagement of supporters in core bureaucratic processes (Bengtsson & Mattila, 2009).

One source of demands for horizontal political organization stems from the globalization protest cultures dating from the 1990s. What might be termed a meta-ideology of diversity and inclusiveness in issue agendas and engagement styles emerged as prominent themes in the global justice movement and in the World Social Forum and various regional forums (Smith, 2016). These direct democratic values also spread in later protest waves such as Occupy Wall Street and the Spanish indignados (15M). Many activists in these movements resisted high-visibility leadership and bureaucratic organization in favor of shared authority and direct, deliberative engagement (see Castells, 2012; Della Porta, 2013; Anduiza, Cristancho, & Sabucedo, 2014; Smith, 2016; Tormey, 2015; Szolucha, 2017).

More broadly, studies of younger, more racially and culturally diverse generations on the left also indicate disdain for older models of bureaucratic membership and delegated representation. Young citizens on the left typically seek faster and more immediate political gratification than typically found in conventional party participation (Bang, 2016; Kreiss, 2014; Loader, Vromen, & Xenos, 2014; Polletta, 2014). Accordingly, recent movements on the left have not combined as easily with conventional parties and the electoral arena as have those on the right. Some anti-austerity and anti-corruption protests have shifted away from conventional national-level party politics, and some have fuelled ‘movement parties,’ often with uneasy relationships to the movements, such as Podemos in Spain (Della Porta et al., 2017; Casero-Ripollés, Feenstra, & Tormey, 2016).

The question at this point is whether these divergent preferences for political participation and organization have traveled into the electoral arena, where democracy has its most important institutional interface with society. The following section explores electoral outcomes in Europe, confirming that there is substantial underrepresentation on the left in countries for which both ESS data and election results are available. We also show that a number of counter-hypotheses that might challenge our explanation of divergent democratic engagement preferences cannot explain this imbalance.

Exploring the electoral imbalance in Europe

This section investigates the skewing of the electoral interface in Europe, beginning with constructing a representation index: a rough indicator of the extent to which different segments of the electorate connect with and secure proportional representation from parties during elections. To begin with, we tallied the most recent electoral outcomes for parties beyond the center (we limited the analysis to European nations with populations of 5 million or more). Parties beyond the traditional center are classified as radical left and radical right by the Chapel Hill Expert Survey.³ According to this independent expert rating system, at the time of this writing, radical right parties were in parliament in 16 European countries with populations of 5 million or more, and formed or supported governments in 4 of those (Poland, Switzerland, Denmark, Finland).⁴ Parties on the radical left were in parliament in just eight comparable nations, and formed governments in only one of them (Greece).⁵

An obvious counter-hypothesis that might explain this imbalance of electoral outcomes is that various national populations have shifted dramatically to the right. Addressing this possibility requires determining how to reliably measure national distributions of left-right political identification and to see how those distributions match up with electoral outcomes. This entails establishing a baseline measure of political identification that maps reliably onto party position. We, therefore, constructed a political identification measure to identify citizens on the radical left and right. Radical right and radical left here refers to citizens who typically do not identify with parties of the center left (e.g., Labour and Social Democrats) or center right (e.g., Christian Democrats and Liberals), and who locate themselves on the ends of the political spectrum beyond these parties. This is tricky because, as noted, the political space in Europe is shifting away from classic left-right distinctions, meaning that left-right self-identification may be perceived differently on individual or country basis, and some parties reject such labels. Nevertheless, people continue to self-identify along a left-right continuum. The salient question here is whether those political identifications map reliably onto the spectrum of radical left and right parties, as established independently by the Chapel Hill Expert ratings.⁶ In other words, we seek to establish whether self-identification as radical left or radical right on the ESS survey also corresponds to ESS measures of identification with parties that are independently judged radical left or right on the Chapel Hill survey.

Our political identification measure, and other measures presented in this analysis, are based on the same ESS data introduced above (round 7, 2014, ed. 2.1),⁷ which is a cross-national survey of European electorates, conducted bi-annually via extensive face-to-face interviews. We included 20 countries and a total N of 37,623 in our analysis.⁸ As is recommended in the ESS Documentation Report, we applied the ‘post-stratification weights’ as well as ‘population Weights’ to the data in order to take differences in sample sizes as well as sampling errors into account.⁹ The political identification measure builds on respondents’ answers to a question asking them to place themselves on a 0–10 scale of left to right.¹⁰ We then identified break points for political identification score ranges that fell outside center left (e.g., Social Democrats or Labour) and center right (e.g., Christian Democrats or Liberals) party identifications. This produced a scale in which 0–2 designates as radical left, 3–7 as center, and 8–10 as radical right.¹¹ In this scheme, individuals who merely tilt slightly left or right count as ‘Center.’ We then conducted a validity test to ascertain that this self-ascribed measure maps reliably onto political parties along the same continuum by identifying the party with which each respondents most identified, and scoring that party along an independent left-right party continuum developed by the Chapel Hill Expert Survey of parties, which is based on party issue position. The test confirmed that the self-ascribed party identification measure correlated strongly (r = .874, p < .0001) with the independent Chapel Hill party scale.¹²

The next step was to determine whether the citizens self-identifying on the radical left and radical right achieve equal representation in terms of party vote shares. We focused on the above-mentioned nations with populations over 5 million in which parties independently classified as radical left (8 nations) or radical right (16 nations) gained more than 5% of the vote in the most recent election. We constructed a representation index using the percentages of citizens falling into the radical left and right segments of the political identification scale and the vote shares of the radical left and right parties for these nations. We divided the percentages of votes for radical left or right parties with the percentage of the public who self-identify either as radical left or right based on our scaling of the ESS data. In this representation index, a national score of 1 theoretically indicates perfect representation, meaning that the percentage of votes in a particular country equals the portion of the electorate that self-identifies at the corresponding place on the political spectrum. National scores below 1 indicate underrepresentation, while results greater than 1 can be understood as overrepresentation. This is not a perfect measure of electoral representation, but it offers a rough indicator of the extent to which different segments of the electorate connect with parties during elections.

We calculated the representation index for the countries included in the 2014 ESS survey.¹³ On the radical right, the mean representation ratio of self-identification to party-vote-share is 1.31, compared to a mean representation ratio of .91 on the radical left. This represents a 40-point difference in a scale that had an actual range between .41 and 3.37, depending on the country.¹⁴ This suggests a substantial electoral imbalance, in the sense that the radical right is overrepresented and the radical left underrepresented in relation to citizen self-identification.

Why would there be under and overrepresentation in the electorates across Europe? We identified and tested four possible counter-hypotheses that might explain these differences. All four were eliminated. A fifth hypothesis involving lower voting rates on the left than the right was also tested. The results support our explanation that the electoral imbalance reflects organizational and engagement preferences on the left that are not matched with a supply of parties offering the requisite organizational and engagement opportunities.

1. There are generally fewer citizens on the radical left than on the radical right.

The first counter-hypothesis addresses the straightforward notion that there might simply be fewer citizens on the one side than the other. Looking across our entire 20 nation ESS sample of 37,623, we find that roughly 12.06% identify on the radical left and 12.34% identify on the radical right, which produces a non-significant difference (p < .278) in a binomial test. (The country-by-country analysis of proportional representation in nations with the radical left and right parties in parliaments was conducted in the representation index reported above.)

2. Trust/confidence in politicians and parties is higher on the left.

The second counter-hypothesis pertains to a general feeling of trust towards political actors and institutions. It rests on the idea that the radical left may be more satisfied with the status quo, and therefore more prone to complacency. In order to test this claim, we cross-tabulated the left-right placement scale with an aggregate of two of the questions found in the ESS survey, which focus respectively on levels of trust in political parties, and levels of trust in politicians.¹⁵ Respondents who self-identified as radical left were significantly more likely to hold low levels of trust (63%) than those on the radical right (54%). Conversely, the radical right was significantly more likely to exhibit high levels of trust (10%) than the left (6.5%). Both comparisons are significant at p < .0001. Based on these criteria, it does not seem to be the case that the left is more trusting of established political actors and institutions; the finding suggests just the opposite. Since the right is more present in government, this could be expected to boost levels of trust. However, this still does not explain why the radical left has not produced similar levels of party electoral support as the right.

3. Satisfaction with government/economy/democracy is higher on the radical left.

Relatedly, the assumption underpinning the third counter-hypothesis is that the radical right feels more politically and economically marginalized, and therefore less satisfied, than the left. If the radical left is content with the state of the government and the economy, we might expect them to be less engaged in elections. We created an aggregate of three questions tapping government, economic and democratic satisfaction after determining that all three measures showed similar valences.¹⁶ We find that citizens on the radical left are significantly more likely to hold low levels of satisfaction (56%) than the right (41%). Conversely, the radical left is far less likely to exhibit high levels of satisfaction, at 10%, compared to 20% on the radical right. All of these differences are statistically significant at p≤.0001, with similar differences observed for each of the individual items in the scale.

4. Political participation is lower on the left.

The fourth counter-hypothesis proposes that generalized disillusionment on the radical left may produce lower participation rates in various other political activities beyond voting. Like the others, this hypothesis receives no empirical support. When asked about participation and civic engagement activities outside of the realm of electoral politics, the left is far more active than the center or the radical right. Overall, 22.3% of those identifying as radical left had participated in demonstrations in the past 12 months, compared to just 6.6% on the radical right. We ran these numbers for various age groups (14–29; 30–49; 50+) and the protest trends were amplified among younger members of the public who self-identify as radical left, with 28% participating in demonstrations, compared to just 11% for young citizens on the radical right. Looking at another measure of participation, we found that working in a social organization was significantly higher on the radical left than on the right (25% compared with 17%), and higher still for citizens under 30 (27%, contrasted with 15% on the radical right). Again, all of these differences are significant at p < .0001.

This preliminary analysis suggests that the electoral imbalance implied by the representation index is not explained by obvious counter-hypotheses about citizen ideology, trust and satisfaction, or political participation. We have found one interesting data trend that supports our explanation that citizens are developing preferences for types of democratic linkage that run counter to what the relevant parties offer. Formulated as a hypothesis, we predict that citizens on the radical left vote at lower rates (or abstain at higher rates) than citizens on the radical right.

5. Voting rates are lower (and nonvoting is higher) on the radical left

Confirming this hypothesis, we find that nonvoting is higher on the left than the right, or to put it in positive terms, voting rates are persistently lower on the radical left over time. We compared respondents’ reports of whether they had voted in the last national election in all ESS rounds from one (2002) to seven (2014).¹⁷ As shown in Figure 1, voting rates have been lower on the left, and the gap has increased slightly. For reference, the most recent difference of 83% vs. 78% is significant at (p < .001). We also looked at regional variations in these trends. Due to understandable variation in which countries may be holding more or less contested elections at different points in time, we find the following regional patterns: the L-R voting gap has closed recently in Central Europe (Czech Republic, Poland, Hungary, and Slovenia); the gap has grown in Central Western Europe (Austria, Switzerland, and Germany) after 2008; the gap has been relatively small (3–5%) but persistent in Spain and Portugal; the gap is large (8%) and most persistent in North Western Europe (Denmark, Finland, Norway, and Sweden); and it is large and growing larger (10%) in Western Europe (Belgium, France, UK, Ireland, Netherlands).

Figure 1. Voting rates reported in ESS data, 2002–2014.

These data, in the context of the above hypotheses and the earlier analysis suggest that the supply of parties on the radical right better meets voter demand for types of organizational engagement than parties on the left. Some of this gap may be due to greater range and complexity of issues on the left, but this factor is difficult to disentangle from the participatory preferences on the left for more diversity and inclusiveness and deliberative engagement aimed at democratically addressing the diversity of issues. Indeed, the pressures emanating from globalization and austerity protest movements for parties that reflect those engagement values has led to the creation of interesting forms of what we term connective parties that aim to bring followers into core bureaucratic processes of agenda setting and candidate selection via mixes of offline and online routines. As discussed in the next section, although some of these technology-driven parties have found success, others have struggled to achieve both scale and credibility among followers.

Connective parties and their challenges

This section turns to the supply side of the electoral interface to explore attempts to organize voter-party-campaign linkage in more horizontal forms, with emphasis on organizational and ideational inclusiveness. Pioneering parties are implementing such models on the basis of technology-enabled organization, which offers the potential for organizational innovation at large scale. However, the early evidence indicates that such parties face old dilemmas with new twists. It is difficult for them to meet internal demand and at the same time compete successfully in the current electoral arena which still favors organization and communicational forms from earlier eras of party democracy. To the extent that the supply side cannot deliver on the horizontal model in the contemporary context, this can be expected to further skew the electoral interface.

This section focuses on connective parties, which go further than other organizationally hybrid parties to develop technology infrastructures that are integral to core party operations. We define connective parties as organizations in which technology platforms and affordances are indistinguishable from, and replace, key components of brick and mortar organization and intra-party functions. These components include affiliation management, policy generation, leader and candidate selection, and public communication.

Other scholars have discussed the growing role of technology in new parties that face various organizational challenges. We note in particular, Haberer’s (2017) pioneering work in identifying the origins and workings of what she terms ‘net-parties’. We prefer the designation of connective parties to distinguish a particular set of organizations that display distinctive theoretical properties. In particular, connective parties: (1) cannot function organizationally without technology platforms (in contrast with more conventional parties discussed earlier that use internet technologies in more peripheral ways); (2) engage supporter networks in connective action that affects the core operations of the party organization (i.e., not just mobilizing votes); and (3) can be located as an extensions of organizationally enabled forms of connective action in the earlier theoretical framework (Bennett & Segerberg, 2013).

While there are growing numbers of these kinds of organizations, space prohibits an exhaustive review. We have selected illustrative cases that include national-level parties that display diverse comparative properties, ranging from northern to southern Europe, and occupying different positions on a left-right continuum (although, for reasons outlined earlier, these parties tend to skew left). Most importantly, all of our cases have adopted technology to serve core organizational functions, although they differ in terms of particular organizational technology adoptions and the precise manner in which they combine horizontal and vertical elements. These cases are: Spanish Podemos (established 2014, entered national parliament 2015); Italian Five Star Movement (M5S) (established 2009, entered national parliament 2013); Danish Alternativet (established 2013, entered national parliament 2015); the Icelandic Píratar (established 2012 following the Pirate Party model, entered national parliament 2013). We also discuss a local connective party, Barcelona en Comú, to assess whether some of the organizational issues confronted on the national scale can be better addressed in smaller scale deployments. It is important to note that these parties are all moving targets with respect to their ideological positioning, their organizational structure and infrastructure, and, crucially, the extent to which they implement or dismantle the mechanisms for inclusive engagement and sharing authority that defines their location in a horizontal organization model over time.

All of these parties purport to pursue a new way of doing (party) politics. As part of this, they explicitly reject ideological positioning on the conventional left-right scale (Alternativet, 2017a; Five Star Movement, 2016, 2017a; Podemos, 2016). This said, the two larger parties have come to attract support concentrated to particular parts of the ideological spectrum. Podemos increasingly attracts left-identifying supporters, while the M5S has stabilized with support across a broader spectrum of left-center-right, with those differences registering on various issues such as immigration, labor market policies and the EU (Ceccarini & Bordignon, 2016; Della Porta et al., 2017).

The public documents of these parties emphasize the importance of generating new political ideas from the bottom up (or from the citizen periphery to the party center). Inclusiveness, participation and transparency are central (Alternativet, 2017b; Five Star Movement, 2016, 2017a; Píratar, 2016, 2017, principles 6.1–6.3; Podemos, 2016, pp. 7–10). While Alternativet aims to develop a new political culture and ‘living everyday democracy’ (Alternativet, 2017a), the others stress the importance of keeping political representatives within democratic control. The M5S in particular promotes the notion of an unfiltered process in which citizens enact politics without intermediaries. Party representatives, drawn from the citizenry, act as the direct extension of popular will, with extensive obligations to follow the party line and report to their constituencies (Five Star Movement, 2017b; see also Podemos, 2016, pp. 9–10; Píratar, 2016, 2017). The priority is to facilitate an association of citizens rather than build a team of professional politicians.

The parties, therefore, pursue radically revised ways of organizing linkage and engagement. All have different ideas and systems for how to share authority and draw in peripheral input. M5S and Píratar place more emphasis on direct democracy, while Podemos and Alternativet stress the formative and deliberative development of ideas. They combine vertical and horizontal elements in complex ways. Table 1 outlines key mechanisms as the parties describe them in Spring 2017, showing that all purport to implement mechanisms that are designed to enable inclusive input in party processes, and shared authority between members and party leadership in party functions such as policy development, and candidate selection. All require membership or the equivalent for decision-making, but ‘membership’ requires very little, and non-members can contribute to discussions and ideas along the way. As noted below, there has been significant debate among supporters about the degree to which these connective mechanisms are genuine, or whether they mask a more controlling central leadership (such concerns have created tensions in Podemos and M5S).

Table 1. Technology-enabled organization in four connective parties.

	Affiliation	Candidate selection Policy process	Key dedicated online platform
Alternativet	Online registration and membership (small minimum fee)	Semi-online primaries permitted PoLab proposals: Face-to-face workshops > Online hearing > Online vote > (Continuous online validation) Online consultation	www.alternativet.dk Link to AlleOS operating system. AlleOS operating system; including Dialog (discussion and voting platform); Appgree (consultation)
M5S	Online registration (no fees)	Online primaries Online idea and discussion forum Online consultation and validation (particular focus legislation proposals)	www.Movimento5.stella.it; including Forum; Login to Rousseau operating system. Rousseau operating system; including Lex (legislation platform), Meetups
Píratar	Multilevel online registration (voluntary fee)	Online primaries Policy proposals: In-house meetings > Online discussion > Online vote	https://piratar.is; Links to Discourse.piratar.is (discussion platform); X.piratar.is (voting platform)
Podemos	Multilevel online registration (no fees)	Online primaries Citizen proposals: Compilation phase > Online endorsement phase > Development phase > Online vote	https://podemos.info; Including the Participa portal with link to Plaza Podemos 2.0. Plaza Podemos 2.0; including ICP (Citizen proposals); Debates y Opinión (debate forum)

The point (and challenge) for all of these parties is to build the capacity for flexibility, shared authority, and breadth and depth of engagement in a robust form at scale. To this end, the parties aim to build technology solutions to enable collaboration and decision-making at large scale and in combination with executive action. The range of custom-built or commercial digital affordances (including some overlap with less connective parties), includes: (a) platforms for general party communication such as party websites, wikis, and blogs; (b) social media, which make it possible for party actors to maintain controlled accounts while enabling supporters to leverage their own social networks; (c) specialized applications for proposing, discussing, and decision-making such as Discourse (discussion platform), Appgree (large-scale communication and consensus software); and (d) online platforms or ‘operating systems’ to organize the key technology outlays.

For most of these parties, their technology-enabled organization is a moving target both in terms of the technology development and organizational process. The operating systems, the individual mechanisms and processes they underpin are still under construction and experimentation. Alternativet provides good examples of the rapid changes. The design and technology of the idea and policy generating process has been tweaked several times since the public launch of the party, including an early decision to shift from party-seeded to grassroots-initiated idea incubation groups (Ringblom, n.d.). The national policy process as set out in the party Handbook in 2017 (Alternativet, 2017b) moves across three phases. The initiation phase starts at grassroots level, including open offline Political Laboratories (PoLabs) that workshop an issue. If the initiating group develops a proposal, the ensuing hearing phase involves a period of open digital commentary and validation through the Dialog platform (built on Discourse software) in the AlleOS operating system. Policy proposals that receive sufficient member support pass directly to the Political Catalogue for political representatives to consult, and remain continuously open to member assessment online in an ongoing verification phase. Proposed contributions to the party program also require the consideration and support (at 80%) of the closed Political Forum (PoFo) committee during the hearing phase, and remain continuously open to PoFo re-assessment. In this and other cases, face-to-face components play an important part, and the process is intended to be open, accessible and transparent.

Connective parties also typically incorporate overlapping and relatively autonomous layers in party organization and communication, which contribute to the impression of a fluid structure. Our local case, Barcelona en Comú, which formed a municipal government in 2015, offers a good example of the complexity and dynamism of such arrangements. The party combines executive leadership with autonomous local groups and informal supporter networks (Haberer & Peña López, 2016). Analysis of social media communication in the 2015 election highlights how the different components can be mutually reinforcing. The Twitter accounts of the party and its candidates produced the most centralized retweet networks of any party in the local election. At the same time, highly developed crowd-enabled supporter networks engaged movement activists in party-related communication outside the formal party networks. Rather than a formal campaign strategy, this dual social media network seemed to reflect activist preferences to remain independent of formal party operations and party acceptance of this arrangement (Aragon, Volkovich, Laniado, & Kaltenbrunner, 2016). Although there was strong agreement on issue agendas between party and movement during the election, these parallel media spheres illustrate the role autonomous networks can play in connective parties. Perhaps even more, this case illustrates the desires of many movement activists to operate somewhat independently of formal party communication processes. In other cases, such as the Finns Party, the existence of parallel communication networks present greater challenges for control of party issue profiles and identity (see Hatakka, 2016).

Of all the tensions present in connective parties, perhaps the greatest source of instability is the degree to which they offer or withdraw mechanisms for inclusive engagement or sharing authority. Although still in many respects a connective party, Podemos had moved decisively away from the horizontal model by 2016. M5S to a certain extent moved in the opposite direction, although practices such as excommunicating party representatives have kept the organization within bounds set by the leadership. In part, such trajectories reflect the internal and external pressures that connective party organizations grapple with at an electoral interface that continues to favor more conventionally organized parties.

Challenges for connective parties

Connective parties that attempt to develop horizontal models potentially offer a response to voter demand for greater engagement at the electoral interface. However, their capacity to balance inclusive and meaningful participation and shared authority with coherent and robust organization is open to question. Organizational innovation easily becomes limited by the need to control the quality of ideas, sensibly aggregate preferences, monitor the legitimacy of decision processes, and coordinate the input of different groups of participants, leaders and supporters. Such problems on the supply side may contribute to skewing the electoral interface in many contemporary democracies, amplifying the underrepresentation of the parts of the electorate that develop preferences for a horizontal, direct democratic party model. We note three key areas in which connective parties face internal and external challenges.

One internal challenge has to do with the sheer complexity of building technological underpinnings for spaces and processes that inclusively and coherently operationalize shared authority, particularly at scale. It is difficult to design platforms, processes and infrastructure that incorporate specific values such as ideal deliberative standards (Borge & Santamarina, 2016; Haberer & Peña-López, 2016); appropriate expertise (Paulin, 2014); or ‘open’ spaces (Husted & Plesner, in press). Observers note that even in connective parties with horizontal aspirations, the invitation to broaden input is often limited in practice to pre-defined alternatives or particular areas (e.g., local issues but not national), and even then, there is little connection to executive outcomes (Della Porta et al., 2017; Rendueles & Sola, 2015). Moreover, complex systems of meaningful engagement in local groups are difficult to build at scale. Both participants and platforms may become overwhelmed when numbers increase. The German Pirate Party suffered collapse after its ‘liquid feedback’ and ‘liquid democracy’ platforms became overwhelmed with input that magnified differences and offered little mechanism for their resolution (Haberer, 2017; Paulin, 2014). Such issues make it difficult for parties to deliver on promises of unmediated or unfiltered processes. In more stable parties, the actual mediation involved in the technology-enabled processes may not be apparent for participants.

A second internal challenge is how to combine diverse participants and levels of participation in fluid structures. This challenge is most demanding for connective parties that emphasize open policy development and intra-party democracy. As noted above, inclusiveness in core organizational processes remains a persistent source of tension. Complicating the issue of who can participate under specific conditions, is frequent lack of agreement on what the ideal of broad inclusion actually entails. Several studies note ongoing negotiations among participants around the issues of who ought to have access to information and decision-making in internal processes, which may create conflicts between temporary supporters and dedicated members, more or less active participants, and on-site group activists and online registered supporters (Borge & Santamarina, 2016; Haberer & Peña-López, 2016). In practice, broad inclusion through technology platforms can be perceived as something that undermines the privilege of committed participants, a theme observed previously in both protests and party affiliation reforms (Bennett & Segerberg, 2013, chapter 6; Gauja, 2016; Scarrow, 2013, 2015).

These studies indicate the difficulty of designing processes and organizations, virtual and material, that temper the emergence of informal clusters and hierarchies in which the participation of some impedes the inclusion of others (Hazan & Rahat, 2010). This problem is amplified in media-rich environments with mixed modes of participation that link face-to-face meetings with digital input (see Mattoni, forthcoming). Stretching policy processes across multiple arenas with distinct norms and characteristics (size, openness, meeting structures, temporalities, online, and offline interactions) exposes proposals to various inputs and treatments, and may discourage participants who are not equally comfortable with different phases of the processes. For example, some Alternativet participants reported that the phases of technology-based deliberation disrupted what they felt had been initiated in the face-to-face settings of the PoLabs (Husted & Plesner, in press; see also Haberer & Peña-López, 2016). In general, as found in other party and movement contexts, a small number of participants are particularly active in the party and on the platform, and in some cases, broader participation has decreased over time (Casero-Ripollés et al., 2016; Della Porta et al., 2017). This said, some ‘inactive’ members report that they do not want to be actively involved, but like being part of a party that is organized in this way (Andersen, 2016).

In many ways, these organizational challenges contribute to a third challenge, that is often the most visible one: how to share authority between the membership and the executive party leadership in a stable manner. Part of this problem can be traced to the external pressures of electoral competition, which typically requires visible leaders and communication with broad voter populations in an arena dominated by conventional news media. These pressures are magnified by competition with more hierarchical parties. In this context, goals such as campaign effectiveness and shared authority easily conflict, leading to tensions regarding the autonomy of distributed groups. These pressures help account for why many of these parties are moving targets in terms of forms of horizontal organization. For example, Podemos in its early phase revolved around autonomous Círculos (local or thematic groups) as its basic organizational unit, and the public party documents still nominally assert this to be the case (Podemos, 2016, Article 6; 2017, Organizational Principle 5). Yet in practice, the 2014 Constitutional Assembly substantially revised internal roles to tone down the role of Círculos in favor of a more centralized organization aimed at waging more coherent election campaigns. These shifts away from constituent organizational models favored by many 15M movement forerunners have produced subsequent splits among leadership and followers (Borge & Santamarina, 2016; Casero-Ripollés et al., 2016; Della Porta et al., 2017, pp. 78–80, pp. 122–125; Rendueles & Sola, 2015).

Ironically, the multi-layered, flexible organization of the connective party can in practice render the role of citizen input unclear. Although implementing a fluid system with multiple layers of action is typically part of the point, the complexity of multiple spaces running in parallel can challenge coherence (see Haberer & Peña-López, 2016). The resulting fuzziness may enhance the role of the party leader instead of the grassroots. Podemos leader Pablo Iglesias quickly dominated the public agenda with a combination of publicity, television talk show appearances, a YouTube talk show, and other controlled media formats. Similarly, in M5S, unclear rules about the connections between formal and informal arenas reinforced Beppe Grillo’s role as chief megaphone at the national level (Bordignon & Ceccarini, 2015; Casero-Ripollés et al., 2016; Della Porta et al., 2017). Overall, several studies observe that party rhetoric about broad participation and technology can obscure a dearth of participatory opportunity, or at least a tension between the ideals and the practice (Bordignon & Ceccarini, 2015; Della Porta et al., 2017; Haberer & Peña-López, 2016; Treré & Barassi, 2015).

These pressures of competition and institutionalization (including expectations about party functions such as candidate development), have previously eroded meaningful engagement in movement parties (Haberer, 2017; Levitsky & Roberts, 2011). It is precisely for such reasons that the comparative party literature expects that alternative party organizations will adapt to conventional bureaucratic models or dissolve (Poguntke et al., 2016). All this suggests that connective parties trying to implement innovative horizontal organizational formats at scale offer interesting responses to citizen demands, but face significant challenges. Indeed, maintaining shared authority and inclusive, meaningful engagement become challenging in direct relation to efforts of these parties to deliver on core aims such as: inclusive participation, managing technological input coherently, and attaining enough scale to compete effectively in elections.

Conclusion

This analysis started with a puzzle about the electoral imbalance in many mature democracies. Examining the electoral interface through the lens of organization, communication and technology highlights important citizen demand-side and party supply-side dimensions of the democratic interface that are undergoing change. Preliminary evidence presented here suggests a two-fold picture. Emerging ideological preferences on the left for direct or deliberative democracy create organizational preferences for horizontal models of democratic linkage that run counter to the models typically offered by parties. At the same time, the conventional configuration of a vertical party model (often with interesting movement-driven hybrid communication processes) seems to favor the emerging ethno-nationalist right.

However, all of these developments are circling a centrist political order that is losing legitimacy. It appears that the democratic clock is ticking. As noted, growing numbers of younger citizens in Europe and the US say that living in a democracy is less important compared with previous generations (Foa & Mounk, 2016). Even among those who continue to think that democracy is the best form of government, the electoral dimension is especially criticized (Ferrín & Kriesi, 2016). Citizens are disillusioned with ‘politics as usual’ and are disconnecting from the traditional governing parties of both center left and center right. They increasingly resist the professionalized marketing style of twentieth century campaign communication that defined what Blumler and Kavanagh (1999) termed the third age of political communication. These and other developments have led observers to point to a legitimacy crisis of contemporary democracy (Castells, 2013; Della Porta, 2013; Mair, 2013).

At the very least, it appears that we are now entering a fourth age of media politics, in which familiar patterns of communication, organization and citizen engagement no longer apply – or at least do not apply very convincingly (Blumler, 2016a, 2016b). This emerging communication age is as yet ill-defined, but it seems marked by the growing chaos of multiple public spheres in which the spin of conventional parties and politicians interacts often chaotically with new forms of organization enabled by technology platforms and socially mediated information production and distribution. These hybrid formations often lack the coherent boundaries and organizational control of modern institutions such as parties and press organizations. While challenges exist on both the left and right ends of the political spectrum, party organizations on the right more closely resemble conventional bureaucratic models. Some of these are thin organizations centered around strong leaders who dominate party decision processes and communication. Such vertical models compete relatively more effectively in the legacy interfaces that remain from modern democratic electoral processes.

All of this raises questions for and about connective parties. Above all, can they scale up and still retain their defining characteristics? How can they compete more effectively against vertical model parties? Can they even compete with other hybrid forms that may emerge in particular situations. For example, in the US, Donald Trump did not have a formal party (although he entered a marriage of convenience with the Republicans), but he had a large social media following that interacted with a right wing media sphere, producing a kind of hybrid political organization in the 2016 election and afterward. Similarly, Emmanuel Macron’s En Marche! Party was something of a short-order, custom-built organization that captured broad centrist discontent with established parties in France in 2017. This suggests that in the unstable conditions faced by many democracies, the search for a new model of parties and elections may be missing the point that digital and social media, mixed with citizen discontent, offer the potential for once unlikely factions and leaders to quickly attain national prominence. Such uses of technology in emerging center-right movements may be more matters of convenience, serving as transitional phases toward other forms of organization. In this context, the quest on the left for organizational forms that create direct or deliberative democracy may prove overly idealistic. Indeed, it seems that the quest to create idealized participatory experiences may result in losing sight of other critical issues, including emerging threats to democracy itself.

Our analysis also raises larger questions for the electoral interface. The connective parties are moving targets, but so to a certain extent is the interface. The key question addressed here is, at heart, not one of political ideology, but of linkage. There is a need to further examine how a fourth age of media politics may condition the electoral interface with respect to its core organizational and communicational processes. As citizens increasingly seek alternatives to center parties (and the press systems that carry their spin), how might electoral and informational linkages be reinvented? Clearly, the fascinating experiments with horizontal connective parties suggest serious citizen reform efforts from the bottom up. However, there remains a gap between achieving alternative linkage practices based on intra-party democracy, and engaging enough voters to win elections within the constraints of current legacy election and communication processes. Perhaps new directions will emerge for more conventional parties that help them improve relations with supporters. For example, one variation that has been illuminated in the context of advocacy organizations is the practice of collecting ‘passive democratic feedback’ (Karpf, 2012; 2016). This entails organizations using advanced analytics to listen to registered supporters in order to tweak issues and infrastructure for supporters to act upon.

As democracies experience increasing turbulence due to the continuing hollowing of the center and the rise of radical alternatives, it may be that citizens will sacrifice some preferences about linkage and engagement in order to stem threats to personal liberties, social diversity, and equity-based social policies. It remains to be seen whether new technology-based organizational forms will yield stable, scalable and resilient parties that can fulfil the function of linkage while competing electorally and governing responsibly. At present, however, while citizens are engaging in a variety of creative experiments, often involving impressive numbers, there appears to be an impasse when it comes to addressing the imbalance at the electoral interface.

Acknowledgments

The authors are grateful for the helpful comments from the anonymous reviewers and many generous colleagues: Henrik Bang, Rosa Borge, Donatella della Porta, Andreas Duit, Maria Haberer, Andreas Hicketier, Emil Husted, Michael Jensen, Brian Loader, Chris Parker, Ishmael Peña-Lopez, Dieter Rucht, and Sidney Tarrow.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

W. Lance Bennett is Professor of Political Science and Ruddick C. Lawrence Professor of Communication and Director, Center for Communication & Civic Engagement at University of Washington, Seattle, USA. [email: [email protected]]

Alexandra Segerberg is Associate Professor in the Department of Political Science at Stockholm University. [email: [email protected]]

Curd B. Knüpfer is a Postdoctoral Researcher at the Political Science Department of the John F. Kennedy Institute at the Freie Universität Berlin, and Postdoctoral Researcher at the George Washington University’s School of Media and Public Affairs. [email: [email protected]]

Additional information

Funding

This research has been supported by Grants from the Swedish Research Council (Vetenskapsrådet) [2015-01835, 421-2010-2303] to Alexandra Segerberg, and from the Humboldt Foundation (Alexander von Humboldt-Stiftung) to Lance Bennett.

Notes

1 ESS Round 7: European Social Survey Round 7 Data (2014). Data file edition 2.1. NSD – Norwegian Centre for Research Data, Norway – Data Archive and distributor of ESS data for ESS ERIC. The Documentation Report is available online at http://www.europeansocialsurvey.org/.

2 The measure used here is based on a variable aggregated out of the results for ESS 7 questionnaire variables ‘ipfrule: Important to do what is told and follow rules,’ and ‘imptrad: Important to follow traditions and customs.’ Both were weighted equally at 1/2. The results were originally measured on a six-point scale and were compiled into the categories of ‘Agree,’ (1–2) ‘Ambiguous,’ (3–4) ‘Disagree,’ (5–6), and ‘NA.’

3 As based on classifications by the Chapel Hill Expert Survey (CHES; Bakker et al., 2015).

4 Poland, Law & Justice (37.6 %, 2015); Switzerland, People’s Party (29.4 %, 2015; Denmark, People’s Party (21.1 %, 2015); Austria, Freedom Party (20.5 %, 2013); Hungary, Jobbik (20.4 %, 2014); Finland, Finns Party (17.6, 2015); Slovakia, Slovak National (8.6 %) & People’s Party (8.0 %, 2016); Norway, Progress (16.3 %, 2013); France, National Front (13.6 %, 2012); UK, UKIP (12.6 %, 2015); Sweden, Sweden Democrats (12.8 %, 2014); Netherlands, Freedom Party (10.1 %, 2012; 13.1%, 2017); Serbia, Serbian Radical Party (8.1 %, 2016); Bulgaria, Patriot Front (7.3 %, 2014); Greece, Golden Dawn 7 %, 2015); and Czech Republic, Dawn of Direct Democracy (6.8 %, 2013). Source: ElectionGuide (www.electionguide.org)

5 Greece, SYRIZA (35.5 %) & Communist (5.5 %, 2015); Spain, Podemos-IU (21.1 %, 2016); Portugal, Left Bloc (BE) (10.6 %) & DemUnity (8.6 %, 2015); Netherlands, Socialist (9.7 %, 2012; 9.1, 2017); Germany, The Left (8.6 %, 2013); Finland, Left Alliance (7.1, 2015); France, Left Front, (6.9 %, 2012); and Sweden, Left Party (5.7 %, 2014). Source: ElectionGuide.

6 In this context, we trace the radical left and radical right (as categorized by CHES) and not the occurrence of populism. To the extent that populism is conceived as a thin ideology that applies across the political spectrum (Mudde, 2004), it focuses a related but distinct dimension of the puzzle at the heart of the present analysis.

7 ESS Round 7, edition 2.1.

8 Countries included in the sample were: Austria, Belgium, Switzerland, Czech Republic, Germany, Denmark, Estonia, Spain, Finland, France, United Kingdom, Hungary, Ireland, Lithuania, Netherlands, Norway, Poland, Portugal, Sweden, and Slovenia. The entire sample in the 2014 round, excluding Israel, totaled 37,623 responses.

9 We created a new weight variable, which is the product of multiplying the post-stratification and populations weight figures (variable name: ‘PSPWGHT × PWEIGHT’).

10 ‘B19. In politics people sometimes talk of “left” and “right”. Using this card, where would you place yourself on this scale, where 0 means the left and 10 means the right?’ (European Social Survey, 2014).

11 We aggregate the options of ‘Refusal,’ ‘Don’t know,’ and ‘no answer,’ as ‘NA’ in subsequent analyses.

12 We constructed the validity test as follows: First, we assigned party identification scores to the political party to which each respondent said s/he felt close, creating a score between 0 and 10 for each political party included in the ESS national surveys. (Question B18a: ‘Is there a particular political party you feel closer to than all the other parties?’ (ESS Source Questionnaire, 2014, p. 8) Next, we identified an independent reference score for those same parties by consulting CHES, which places political parties on a scale of 0 (left) to 10 (right) based on experts’ ratings of parties on given points of position, issues and ideology. When we compared these two measurements, the results showed a very strong correlation (r = 0.874, p ≤ .0001) between these two independent estimates of party position. We interpret this to mean that political identification, that is, left-center-right self-identification contained in the ESS data, maps reliably onto corresponding party positioning along the same continuum.

13 This ESS survey included 12 of the 16 nations in which parties on the radical right gained more than 5% vote share and 7 out of the 8 countries with winning parties on the radical left For the right, these countries were Poland, Switzerland, Denmark, Austria, Hungary, Finland, Norway, France, the UK, Sweden, Netherlands, and the Czech Republic. For the left, they were Spain, Portugal, Netherlands, Germany, Finland, France, and Sweden. Note that the party vote shares were obtained for national elections ranging from 2012 to 2015, but political self-identification survey data were not available for each election year.

14 For countries with left parties in parliament, the representation range ran from a low of .41 (Sweden) to a high of 1.29 (Finland). For countries with right parties in parliament, the representation range was from .47 (Czech Republic) to 3.37 (Austria).

15 We cross-tabulated the left-right placement scale with an aggregate of ESS 7 questionnaire points B5 (trust in political parties) and B6 (trust in politicians), which were weighted equally at 1/2. The original results, which were measured on a ten-point scale, were compiled into the categories of ‘Low,’ (0–3) ‘Mid,’ (4–6) ‘High,’ (7–10), and ‘NA.’

16 This variable presents an aggregate of the results for ESS 7 questionnaire points B21 (satisfaction with country’s economy), B22 (satisfaction with country’s government), and B23 (satisfaction with the way democracy works in country), which were all weighted equally at 1/3. The results were originally measured on a ten-point scale and were compiled into the categories of ‘Low,’ (0–3) ‘Mid,’ (4–6) ‘High,’ (7–10), and ‘NA.’

17 Question B9: ‘Some people don’t vote nowadays for one reason or another. Did you vote in the last [country] national election in [month/year]?’ (ESS Source Questionnaire 2014, p. 7). Here we used the subsample of countries included in ESS 7 survey (excl. Isreal), most of which were also featured in all of the previous ESS rounds: Austria (excl. ESS 4–6), Belgium, Switzerland, Czech Republic (excl. ESS 3), Germany, Denmark, Estonia (excl. ESS 1), Spain, Finland, France, UK, Hungary, Ireland, Lithuania (excl. ESS 1–4), Netherlands, Norway, Poland, Portugal, Sweden, and Slovenia