Mainstream Voters,Non-Voters and Populist Voters: What Sets Them Apart?

Introduction

Western European democracies have changed in complex ways in recent decades. Among the most significant long-term changes is declining turnout. Despite some differences across countries, the overall picture is similar: fewer people go to the polls now than 30 years ago (Mair, 2013). Not by chance, research has shown that non-voters and voters should be thought of as two different constituencies, each with its own sociodemographic and sociopolitical characteristics. However, in recent years, this trend has been accompanied by the well-documented growth of voting for populist parties, often seen as better positioned to mobilize former non-voters than their mainstream counterparts, despite mixed evidence (Immerzeel and Pickup, 2015; Leininger and Meijers, 2021). Empirical studies that compare these three constituencies are rare, not least because non-voters are usually underrepresented in surveys and support for populist parties has only begun to receive increasing academic attention in the last few years.

While their respective characteristics and roles within Western European politics received little comparative scrutiny, the decline of mainstream voters and the rise of both non-voters and populist voters certainly represent important challenges to democracy in the region. On one hand, the growth of non-voting implies that fewer citizens participate in the political system, potentially ‘hollowing out’ the legitimacy of representative politics (Mair, 2013). Furthermore, as less educated and socioeconomically deprived citizens tend to vote less, those who win elections end up overrepresenting the ideas and interests of the well-off (Gallego, 2010). ¹ On the other hand, the expansion of populist voting involves a problematic form of political engagement, since populist forces are often at odds with liberal democracy, can foster a moralization of the political debate and may encourage polarization (Mudde and Rovira Kaltwasser, 2017, 2018). Both non-voting and populist voting can thus be seen as forms of rejecting mainstream democratic politics. Hence, it is imperative to better understand what differentiates those who remain loyal to mainstream parties from both those who abstain and those who opt for populist alternatives, which often – though not always – represent a threat to liberal democracy.

To address this research gap, in this article, we offer a detailed empirical analysis of these three different groups to examine the extent to which they should be considered separate constituencies. Taking the advantage of a national representative survey recently undertaken in Germany, in which non-voters are oversampled and a sizeable share of the electorate supports both populist radical left (PRL) and populist radical right (PRR) parties (Die Linke and the Alternative für Deutschland (AfD), respectively), we are able to compare in detail the sociodemographic and sociopolitical characteristics of mainstream voters, non-voters and populist voters.

The empirical analysis demonstrates that these are indeed three different constituencies which relate in distinct ways to representative politics. Two main findings stand out. First, we show that populist voters resemble mainstream voters in their expectations of representative democracy, but more closely match non-voters in their disenchantment with democratic practice. Second, populist voters appear mainly disappointed with a democratic and party-political system they are engaged with in principle and have unmatched expectations about. In contrast, non-voters seem both more fundamentally disconnected from and less expectant of democratic principles and representative practice. These findings have important consequences for the study of Western European democracy, because they underline that citizens may undertake two distinct forms of withdrawal from democratic politics: either disengaging from the political system or actively confronting established political parties. By demonstrating that citizens eligible to vote should be thought of as three separate groups with their own sociodemographic features and sociopolitical views, we clarify how the electorate is structured today and identify factors that influence their respective (non-)mobilization. Importantly, our findings suggest that despite their parallel emergence, structural disengagement from and indifference towards modern politics among abstaining segments of the citizenry should not be confused with the more specific subjective historical disappointment in democracy and mainstream party politics fuelling populist support. Contrary to popular belief, in a case where both leftist and rightist populist parties are electorally represented, average non-voters appear too removed from representative democratic politics and too opposed to political parties to form an untapped reserve electorate easily available for (further) populist mobilization.

The rest of the contribution is divided into four parts. In the next section, we offer a brief discussion about the relevance of distinguishing different constituencies in Western European democracies and summarize the main empirical findings of the existing literature on non-voters, mainstream voters and populist voters. After this, we explain the research design of our article, putting special emphasis on the case selection, data, operationalization and methods. Subsequently, we present the empirical analysis and the interpretation of the statistical models that help us to clarify what sets apart non-voters, mainstream voters and populist voters in contemporary Germany. Finally, we conclude by summarizing the main findings and advancing some ideas about the future research agenda on this topic.

Three Constituencies in Western Europe: Mainstream Voters, Non-Voters and Populist Voters

In his seminal contribution Ruling the Void, the late Peter Mair (2013) offers a chilling assessment of the transformations that Western Europe has been experiencing in the last decades. Specifically, he identifies declining turnout as one of the major challenges that is affecting democracy in the region. In line with the cartel party theory (Katz and Mair, 1995, 2009), Mair draws attention to the fact that growing collusion and decreasing policy differences between mainstream political parties have triggered the desertion from the political arena of an important segment of the electorate. Although this is a subtle and gradual process that is more evident in some countries than in others, almost all Western European democracies are affected by citizen disengagement from conventional forms of political participation in general and from elections in particular. At the same time, Mair argues that decreasing electoral participation has facilitated the ongoing detachment of mainstream political parties from civil society, with increasing cartelization and delegation of decision-making to non-majoritarian institutions at the national and supranational levels. Seen in this light, Western European democracies are characterized by growing:

indifference on the part of both the citizenry and the political class: they are withdrawing and disengaging from one another, and it is in this sense that there is an emptying of the space in which citizens and their representatives interact (Mair, 2013: 18, italics in original).

Interestingly, Mair’s argument does not stop here. By taking a long-term perspective, he maintains that one of the corollaries of the hollowing out of democracy is the opening of the electoral opportunity structure for the rise of a new type of political phenomenon in Western Europe: populism. In effect, populist forces do not come out of the blue. Their electoral emergence is directly related to the citizen’s perception that mainstream political parties are out of touch and work as responsible agents of international markets and supranational institutions, rather than as responsive agents of the national population (Mair, 2009, 2013; Mudde and Rovira Kaltwasser, 2017, 2018). In other words, populist forces thrive when ‘voters may come to see established parties as protectors of an elite political caste that serves its own narrow self-interests rather than looking out for – or “representing” – the broader interests of society’ (Roberts, 2017: 292). Western European populism can be seen, then, as an unexpected consequence of the growing cartelization of mainstream political parties, which by detaching themselves from their social bases and removing (international) policy decisions from electoral accountability have growing difficulties holding on to their old voters and attracting new ones.

This portrayal of Western European democracy thus leads to the increasingly common identification of three different constituencies across the region: mainstream voters, non-voters and populist voters. From a democratic point of view, these three groups relate distinctly to the political system. Mainstream voters are certainly the most loyal adherents to the liberal democratic regime, since they not only participate in elections but also opt for political parties that support the post-war consensus on what representative democracy means and how it should work in Western Europe. Quite different is the relationship that non-voters and populist voters maintain with the political system. While the former are disengaged from the political debate and do not participate in elections, the latter remain engaged in politics but back political parties that often advance agendas undermining liberal democratic institutions and norms.

In studying this phenomenon, the existing literature has identified characteristic features rendering citizens more likely to support mainstream parties, abstain from voting or support populist alternatives. We briefly summarize the main findings of this work in the following, before turning to our comparative focus.

Research Design

Case Selection: Why Contemporary Germany?

To empirically compare the three constituencies and show what sets them apart, we focus on the case of contemporary Germany. Taking advantage of a representative survey undertaken in the context of the 2017 general election, which oversamples non-voters and includes a sizeable share of respondents that declare to support populist forces, we can provide a detailed comparative empirical analysis of the three constituencies. But why does contemporary Germany represent a good case study to undertake this type of empirical analysis? Two main reasons justify this case selection.

First, there is little doubt that the structural changes affecting Western European democracy discussed above are present in contemporary Germany. In fact, one could see this case study as an ideal place to test the general argument about the subtle and gradual formation of three different constituencies, which should have their own sociodemographic and sociopolitical characteristics. As can be seen in the following figure, when looking at those citizens who are eligible to vote (rather than only those who voted), Germany has witnessed important transformations since 1990. On one hand, there is a long-term tendency towards declining mainstream party voting in the country, decreasing from around 70% of eligible voters to around 55% in 2017. On the other hand, populist parties have gradually become stronger, peaking with the support of over 15% of eligible voters. Finally, while non-voters show no clear long-term trend, their share of the electorate grew from a historical high point of 20% in 1990 to around 30% in 2013 (Schäfer, 2013: 40), and only declined significantly again in the 2017 election (Figure 1).

OPEN IN VIEWER

Figure 1.

Evolution of Three Constituencies in German Federal Elections.

These trends illustrate the puzzling relationship between the three constituencies we discussed above: as indicated by the vertical lines in the figure, we can identify distinct periods in which the relative size of the groups moves together in different ways. Until 2002, mainstream voting remained roughly stable, while non-voting decreased and increased to the apparent benefit, or expense, of populist parties. A similar dynamic is apparent from 2009 onwards: despite some losses for mainstream parties, populist gains occurred alongside significant decreases in the share of non-voters. In contrast, from 2002 to 2009, mainstream voting saw a dramatic decline, while both populist voting and non-voting increased. Given the apparent absence of a consistent relationship between these three groups’ eligible vote shares, it thus remains crucial to better understand the distinct profiles of mainstream voters, non-voters and populist voters within the German electorate.

Second, the German case study is particularly interesting because nowadays two populist forces are represented in the Bundestag. On the left, the PRL party Die Linke obtained 9.2% in the 2017 general election and has been able to establish itself at the national level from its East German origins. ³ On the right, the PRR party Alternative für Deutschland (AfD) is a recent newcomer in the German party system (Arzheimer, 2015), entering the national parliament for the first time in 2017 with 12.6% of the vote after only narrowly failing to clear the 5% threshold in 2013 with 4.7%. The existence of both a PRL party and a PRR party in the current national parliament permits examining if those who support populist forces – independent of their leftist or rightist profile – do have different characteristics to non-voters and mainstream voters. Furthermore, if (most) non-voters who might be attracted to populist parties can be considered likely to have been mobilized by either of these parties, this renders non-voters more representative than in cases where only one type of populist party is electorally present (such as Germany before 2017), thus sharpening the comparison we aim for.

Analysis

To compare the three constituencies, Figure 2 first presents our baseline model of electoral participation. As expected, sociodemographic and attitudinal factors indeed reveal stark differences in the political profiles of non-voters, populist voters and mainstream voters. Beginning with the characteristics of non-voters, political interest stands out as the sharpest dividing line between these and both types of voter groups: one standard deviation higher than average political interest renders a typical German citizen around 15% less likely to abstain from casting a ballot, while both voter groups are associated with above-average interest in politics. Strikingly, non-voters are least likely to report being highly informed about and understanding politics but simultaneously are significantly more likely to trust themselves to engage in political discussions. In terms of sociodemographics, non-voters are the least likely to report household income or education levels above average as well as the most likely to be male.

OPEN IN VIEWER

Figure 2.

Baseline Model for Non-Voters, Populist Voters and Mainstream Voters.

Mainstream voters, in turn, are unique in that they tend to more cautiously assess their ability to engage in political debates while claiming to understand and be informed about politics, in contrast to the other two groups. Sociodemographically, this constituency stands out as being much more likely western German residents and tends to consist of higher earning, more educated female citizens. Finally, populist voters appear to combine features of both other groups: they share with non-voters a self-perceived ability to engage in political debates, while sharing with mainstream voters a high interest in politics. Sociodemographically, populist voters are even more likely to be eastern German residents than non-voters and slightly less likely to report above-average incomes, but otherwise lack significant similarities.

Thus, while the three types of electorates are distinct in terms of participatory attitudes and sociodemographics, the baseline model appears insufficient to account for the specific features separating populist voters from the two other groups. Accordingly, Figure 3 presents results from a set of models which additionally includes either democratic (dis-)satisfaction, (nostalgic) deprivation, ideological views or (negative) partisanship. ⁷

OPEN IN VIEWER

Figure 3.

Extended Models for Non-Voters, Populist Voters and Mainstream Voters.

In terms of attitudes towards democratic politics, the extended models suggest that non-voters stand out from the rest of the electorate by being significantly less likely to consider democracy as the best political system. Furthermore, they share with populist voters a dissatisfaction with democracy in practice and a perception that elected leaders are not responsive to people like them but do not tend to demand greater use of referenda. While non-voters report feeling somewhat deprived today, there is no significant pattern in terms of their self-assessed deprivation 30 years ago, in contrast to both other groups. ⁸

Considering ideological and partisan leanings further crystallizes the specific characteristics of non-voters in Germany in relation to representative political practice. In contrast to both other groups, non-voters lack a clear ideological profile in any dimension of political conflict aside from immigration: non-voters tend towards rejecting immigration on average, but not as consistently as populist voters. ⁹ They are also most likely to be apartisan in the sense of not reporting positive or negative partisanship for any political party. But abstainers do not appear equally disposed towards all parties: non-voters are more likely to report anti-establishment than anti-populist partisan identities. However, this pattern flips when we control simultaneously for respondents’ ideological views, as non-voting is then significantly associated with anti-populist partisanship in addition to apartisanship (see online appendix). Together with the fact that anti-immigration views also lose significance in these models, this suggests that abstainers’ average expressed rejection of mainstream parties may stem from their anti-immigration preferences. However, it also suggests that typical non-voters in Germany who do not hold strong anti-immigration views are more likely to exhibit negative partisanship towards populist rather than mainstream parties.

These extended models also more clearly single out populist voters as a distinctive constituency. In terms of democratic attitudes, these voters retain greater belief in democracy as a political system than non-voters, but they report the greatest dissatisfaction with democratic practice and are unique in demanding greater use of direct democratic tools to supplement representative politics. Their subjective assessment of social integration also distinguishes populist voters: only this group is significantly less likely to perceive greater deprivation 30 years ago, but simultaneously feels more deprived today. ¹⁰

In distinction to non-voters, populist voters are significantly more likely to share a clear ideological profile despite loyalties to two parties that diverge on the economy, immigration and EU integration. Even further, these voters’ preferences stand in direct opposition to typical mainstream voters’ views. Political identities additionally differentiate populists: in contrast to abstainers, these voters are much less likely to report no positive partisanship and do not consistently lack a negative partisan identity. Indeed, populist voting significantly relates to strong anti-establishment partisanship, an association which seems to rest largely on their ideological disagreements (see online appendix).

Finally, these extended models also reveal further distinctive characteristics of mainstream voters. Not only are they significantly more likely to support democracy in principle compared to non-voters, but only these voters report high satisfaction with democratic practice and consider politics as responsive to people like them. Perhaps unsurprisingly, seeking greater use of referenda is negatively associated with mainstream voting.

In terms of subjective deprivation, mainstream voters almost exactly mirror their populist counterparts. They uniquely report significantly lower social marginalization today and tend to perceive themselves as more deprived 30 years ago. ¹¹ Ideological and partisan profiles further identify mainstream voters as the flipside of populist voters: despite their internal party-political competition, voters of mainstream parties exhibit shared ideological preferences in opposition to populist voters’ views and report strong anti-populist partisanship which even eclipses populist voters’ rejection of mainstream parties.

To provide a more fine-grained picture of the German electorate and especially its populist constituency, we zoom in on how left-wing and right-wing populist voters differ in their relationship to democratic and mainstream party politics in our sample. To that end, we run the same baseline and extended models described above but this time separating the group of populist voters into supporters of the AfD and Die Linke.

In a first step, our baseline factors reveal similar political interest, internal efficacy and regional residence in voters of both parties, supporting our classification (see online appendix). However, we also confirm key contrasts in income, education and gender which are in line with the literature on right-wing or left-wing populism (Rooduijn et al., 2017). ¹²

Our extended models reveal further similarities between both groups which support our categorization alongside crucial differences which set apart right-wing from left-wing populism (see Figure 4): while both voter groups differ drastically from the much more democratically satisfied mainstream voters, Die Linke voters are much less democratically dissatisfied than AfD voters. The former are not significantly more likely to report high democratic dissatisfaction and only slightly more likely to perceive lower input responsiveness to ‘people like them’ and demand greater use of referenda. This contrasts with AfD voters, who express much more radical attitudes towards the democratic and representative system. A similar picture of degrees of dissatisfaction emerges with regard to (nostalgic) deprivation. Both voter groups differ drastically from mainstream party voters and lean towards perceiving a relative social marginalization over the last 30 years, but AfD voters more strongly express such perceptions.

OPEN IN VIEWER

Figure 4.

Extended Models for Non-Voters, AfD Voters, Die Linke Voters and Mainstream Voters.

Most clearly, voters of both parties differ crucially in their ideological preferences and in their attitudes towards mainstream parties. As Figure 4 indicates, Die Linke voters are on opposite sides with AfD voters in general left–right placement and on cultural issues. In line with the key issues of PRL and PRR parties (March, 2017; Rama and Santana, 2019), Die Linke voters share a significant left-wing economic stance but no clear preference on immigration, whereas the reverse is true of AfD voters. Only on EU integration do both voter groups share sceptical positions, but again in a less drastic form in the case of Die Linke. This resonates with the notion of soft versus hard Euroscepticism, whereby PRL parties oppose EU integration practice as a ‘neoliberal’ project, while PRR parties oppose EU integration in principle on nationalist grounds (Beaudonnet and Gomez, 2017), including in Germany (Ketelhut et al., 2016).

These different preferences also mirror different types of (negative) partisanship among AfD and Die Linke supporters: AfD voters oppose mainstream voters on all ideological dimensions and report very high negative partisanship towards all established parties. This is not the case for voters of Die Linke, who tend to agree with mainstream voters in Germany on cultural issues and do not report a significant antipathy towards all established parties. As we show in the online appendix, Die Linke voters instead systematically dislike parties that are further on the right, including both the market-liberal FDP and especially the AfD. Overall, these results corroborate the view that populist voters differ systematically from mainstream supporters and non-voters, but they also underline that the PRR in Germany presents a much more drastic opposition to the existing democratic and party system than its less populist counterpart on the radical left.

Conclusion

Western European democracy has experienced structural transformations in the last decades and finds itself in the eyes of many observers in a highly uncertain period. Part of the challenge lies in the fact that more and more countries are facing the formation of three different constituencies – mainstream voters, non-voters and populist voters – that each relates differently to the political order. Even though this diagnosis is widely shared among practitioners and scholars alike, there is little empirical research about what sets apart these three constituencies. This lacuna can be explained, to a great extent, by the lack of survey data on non-voters as well as the fact that populist voting is a relatively recent phenomenon in Western Europe that has been receiving increasing attention only in recent years.

With the aim of starting to fill this research gap, in this article we analysed survey data from contemporary Germany that oversamples non-voters and includes an important number of voters for both the PRL and the PRR, providing an ideal test case to contrast citizens who do not vote with those mobilized by populist parties and those remaining loyal to mainstream parties. Our findings show that the three groups differ substantially in their relation to democratic politics. Specifically, political interest, belief in democracy as well as ideological and partisan preferences most clearly single out who tends to participate electorally in the first place. In turn, western German residence, perceptions of historically improving social integration as well as satisfaction with democratic practice and responsiveness most strongly differentiate between those who vote for established parties or rather for populist alternatives. In contrast to mainstream voters, who remain supportive of and satisfied with democracy in both principle and practice and actively reject populist alternatives which challenge it, populist voters are marked by disappointment with and ideological opposition to a democratic and party-political system they are engaged with in principle and have unmet expectations of . Populist and mainstream voters in Germany represent clearly opposed camps on all ideological dimensions except for cultural liberalism, and these frontlines are reflected in starkly polarized patterns of negative partisanship from both sides. However, in the German case, left populist opposition takes a markedly less extreme form and is more specifically directed towards right-wing ideologies and parties. In contrast, non-voters seem both more fundamentally disconnected from and less expectant of democratic principles and party politics. Yet, while non-voters in Germany seem to reject mainstream parties they ideologically disagree with on immigration, those abstainers who do not hold strong anti-immigrant views tend to feel more repelled by populist alternatives than by mainstream parties. This throws doubt on the (further) mobilization potential of such citizens for populist challengers. Overall, while mainstream voters in Germany select political parties that support the liberal democratic regime, non-voters are disengaged from the party-political sphere and democracy as a whole. In turn, populist voters remain expectant of democracy as a system and engage with the electoral arena but opt for political parties that oppose the post-war consensus on what democracy means and how it should work in practice, especially in the case of the PRR.

These results have important consequences for the study of Western European democracy, because they throw the spotlight on two alternative interpretations of the electorate. On one hand, these findings suggest that citizens may undertake two distinct forms of withdrawal from democratic politics: either disengaging from the political system and rejecting all political offers (non-voters) or actively confronting mainstream political parties and supporting new political forces that challenge the democratic system from within (populist voters). By demonstrating that citizens eligible to vote should be thought of as three separate groups with their own sociodemographic features and sociopolitical views, we thus contribute to clarifying how the electorate is structured today and identify factors that influence their respective (non-)mobilization. Importantly, our findings suggest that, despite their parallel emergence, structural disengagement from and indifference towards modern politics among abstaining segments of the citizenry should not be confused with the more specific subjective historical disappointment in democracy and mainstream party politics fuelling populist support. On the other hand, our results may also suggest that populist voters present a (temporary) midpoint between mainstream and non-voting. The former develop positive and negative partisanships, and are engaged and satisfied with democratic political practice, whereas non-voters lack positive and negative partisanships, have fully disengaged and lost all expectation of the democratic system. Seen this way, a lot may depend on how these citizens assess the medium-term ‘success’ of their populist challenge to the mainstream: if attitudes towards democratic politics and ideological representation improve as populists establish themselves in the party system, they may again resemble the mainstream electorate more closely. In contrast, if they remain frustrated from established political institutions, populist voters may eventually fully dissociate from the democratic system, further ‘hollowing out’ democracy. Given the tense relationship of populist parties with liberal democracy, balancing reconciliation of their voters with democratic politics while maintaining support for underlying principles guarding inclusive participation may thus present a major challenge facing Western European democracies.

Because of very limited existing research with a comparative focus on mainstream voters, non-voters and populist voters in Western Europe, this article makes an important contribution to a research topic that deserves much more attention. We can suggest at least three areas that are particularly important to address going forward. First, the empirical analysis undertaken here considers only one measure in time. Therefore, future research could try to work with panel data to examine citizens that from one election to another change between mainstream voting, non-voting and populist voting patterns. Particularly interesting would be to examine the extent to which the votes for PRR and PRL parties come from previous non-voters and previous mainstream party supporters. According to the empirical findings of this article, one could expect that once a PRR/PRL has been able to establish itself in the electoral arena – as in the case of Germany in 2017 discussed in this contribution – it should have very limited capacity to attract further non-voters. The reverse trajectory would be similarly interesting: under which conditions do (former) populist voters turn to abstaining from voting (again) or instead choose to (again) support mainstream parties? Our findings suggest that focusing on the drivers of greater satisfaction with democratic practice and reduced antipathy towards established parties might be promising to explore in this regard.

Second, however interesting the case, this work has focused only on one country, and in consequence, it is important to undertake comparative analyses to see whether the characteristics that we identify for mainstream voters, non-voters and populist voters in Germany are similar across Western Europe. In our opinion, the existence of both PRR and PRL parties, as well as the growing fragmentation of the electoral space into different political forces, makes the German case study not only particularly interesting but also representative of political trends one can observe in many Western European countries. Nevertheless, future studies can try to test the generalizability of the empirical findings presented in this contribution. Particularly interesting in this regard could be the comparison with other cases who contain specific territories within the country that – similar to Eastern Germany in our case study – are strongly marked by the politics of subjective status loss that should fuel the rejection of mainstream political parties and the disposition to vote for populist forces of different kind, rather than abstaining from voting altogether (Bornschier, 2010; Gidron and Hall, 2017; Mudde and Rovira Kaltwasser, 2018; Rydgren, 2013).

Third, given that the evidence we present here demonstrates that mainstream voters, non-voters and populist voters maintain a distinctive relationship with the democratic regime, practical policy responses such as mandatory voting should perhaps be more carefully weighted in regard to potential consequences for the democratic regime (e.g. Malkopoulu, 2020). Comparing our findings with cases with mandatory voting like Belgium would appear promising for studying what happens when people are forced to vote even when they do not appear to have strong liberal democratic credentials.

Supplemental Material