How the 2017 Federal Election in Germany Affected Satisfaction with Democracy among AfD Voters

Abstract

Dissatisfaction with democracy is among the central determinants of voting for a populist radical right party, such as the German Alternative für Deutschland (AfD). At the same time, elections have been shown to increase political satisfaction, especially when voters see themselves as winners of the election. In this article, we ask, first, whether voters of the AfD have partly been reconciled with democracy during the 2017 federal election in Germany in which the party entered the Bundestag for the first time since its foundation. Second, we investigate which individual level factors conditioned the change in democratic satisfaction among AfD voters. Our empirical analysis, which is based on panel survey data from the 2017 Germany Longitudinal Election Studies (GLES), shows that AfD voters experienced a stronger increase in support for the political regime than any other voter group. Moreover, policy congruence with the AfD on socioeconomic issues as well as identifying with a party other than the AfD enhanced this boost in satisfaction with democracy. Our findings point to the importance of political representation and opportunities for protest-based voting behaviour regarding public support and the legitimacy of democratic political systems.

ACKNOWLEDGEMENTS

A previous version of this article was presented at the German Political Science Association (DVPW) conference (Frankfurt, 2018) and the EUROLAB authors’ conference (Cologne, 2019). We would like to thank the participants and our discussants Heiko Giebler, Mario Quaranta and Sergio Martini for their insightful feedback. Moreover, we thank our student assistants Hanna Doose and Sarah M. Lehmann for their valuable assistance as well as the two anonymous reviewers for their helpful comments and suggestions.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

SUPPLEMENTAL DATA AND RESEARCH MATERIALS

Supplemental data for this article can be accessed on the Taylor & Francis website, https://doi.org/10.1080/09644008.2020.1741550.

Notes

1 In addition, the party gained seats in several regional elections in Eastern Germany during the same year.

2 In July 2015, the party founder and former leader, Bernd Lucke, left the party after struggle about the future direction of the party, thereby paving the way for a more right-wing political programme under the new leaders Frauke Petry and Jörg Meuthen.

3 Moreover, while being male is one of the rather consistent factors of AfD support (Bieber, Roßteutscher, and Scherer 2018; Schmitt-Beck, van Deth, and Staudt 2017), the socio-demographic profile of the partisanship has shifted from a ‘party of the self-employed’ to an over-representation of workers and people with lower education (Bieber, Roßteutscher, and Scherer 2018, 448–449).

4 The disdain for liberal, representative democracy among AfD voters is often accompanied by the wish for more populist conceptions of democracy, i.e. more majoritarian, direct and anti-pluralist forms of democracy (Steiner and Landwehr 2018). However, as Bowler et al. (2017) conclude for right-wing populist party supporters in general, these are largely ‘dissatisfied, but not direct democrats’.

5 While these studies have solely concentrated on the German AfD, their findings are corroborated by comparative research. Oesch (2008), for example, finds that dissatisfaction with the way democracy works is an equally strong predictor of support for right-wing populist parties in Western Europe as cultural grievances.

6 In the words of Anderson (1998, 583), SWD ‘does not refer to democracy as a set of norms, but to the functioning of the democratic political system’.

7 By doing so, we deliberately leave aside a vast amount of literature that deals with other determinants of political support and democratic satisfaction here. It has been consistently shown that many individual-level factors and contextual conditions determine people's SWD. For example, citizens tend to be more satisfied with democracy, when they evaluate the personal and national economy positively (Christmann 2018; Nadeau, Daoust, and Arel-Bundock 2020). Moreover, SWD is shaped by the electoral system (Aarts and Thomassen 2008), social provisions (Lühiste 2014) as well as attitudes towards public administration and political institutions (Ariely 2013; Wagner, Schneider, and Halla 2009; Zmerli, Newton, and Montero 2007).

8 Since most populist parties mobilise by convincing their followers that the elite dominates the people in the democratic decision-making process (Rooduijn, van der Brug, and de Lange 2016, 33), many populist voters believe that they are underrepresented in these processes and, thus, become dissatisfied with the way representative institutions work.

9 In the words of Singh, Karakoç, and Blais (2012, 202), ‘winners of a match tend to enjoy the sport more and attend future games when their team does well, while those on the losing side become embittered and discontented with their team and even the entire sport’.

10 Arzheimer (2018, 147) claims that the central reason for this interdependence is that RRP voters ‘support the Radical Right because of their anti-immigrant claims and demands, and their sense of frustration and distrust may very well result from their political preferences on immigration not being heeded by the mainstream parties’.

11 In their literature review, Birch and Dennison (2019, 111–112) identify three common features of the ‘protest voting’ concept. First, protest voting is an expressive behaviour (see also Brennen and Lomasky 1993), aimed to send a signal by voting for a minor party. Second, it is targeted at mainstream parties that are or have been in government and, thus, implies an expression of discontent with the elite's consensus on certain issues and/or the incumbents as such. Third, the protest vote is usually casted for an outsider party that is not yet represented in parliament.

12 Alvarez, Kiewiet, and Núñez (2018) understand ‘insurgent protest voting’ and ‘tactical protest voting’ as two distinct forms of protest voting, whereas Birch and Dennison (2019) integrate both features into their definition of protest voting.

13 More specifically, Bieber, Roßteutscher, and Scherer (2018) show that almost half of the AfD voters identified with an establishment party prior to the federal election. Mader and Schoen (2019) conclude in their panel study during the 2017 electoral campaign that ‘the immigration-friendly policies of Chancellor Merkel during the refugee crisis led to an alienation of immigration-critical CDU supporters. At the same time, the AfD gained support among supporters of restricting immigration – some of whom were presumably prior supporters of the CDU/CSU’ (Mader and Schoen 2019, 17).

14 Since some survey items were only asked once or irregularly, not all variables are extracted from the same panel wave.

15 With a share of around 10 per cent, AfD voters are somewhat underrepresented in our sample.

16 The exact distribution of the dependent variable is shown in Figure A-1 (see appendix).

17 However, we ran additional robustness checks controlling for the effect of the first vote choice in the 2017 federal election on voters’ SWD, which does not affect our results.

18 The survey item (w7) asks (on 7-point scale) to which degree the respondents/parties are in favour of facilitating (1) or in favour of restricting (7) migration.

19 The survey item (w7) asks (on 7-point scale) to which degree the respondents/parties are in favour of lower taxes and less social benefits (1) or in favour of additional social benefits and higher taxes (7).

20 For these two variables, we rely on the pre-electoral wave W7 (September 2019).

21 To test if the effect of protest voting on SWD is affected by citizens’ level of political interest, we run a robustness check adding this variable to our model. The results show that political interest does not change the findings of our analysis. Protest voting and political interest are only correlated by 0.0357.

22 Both congruence measures and the variable capturing the protest vote correlate with each other on a low level (<r = 0.11), which can also be seen by the joint distribution of these variables (Table A-7). An overview over all correlation coefficients for AfD voters can be found in Table A-1 (see appendix).

23 The exact operationalisation of all variables is laid out in Table A-2 (see appendix).

24 It is worth noting that there still exist significant differences between East and West Germany regarding SWD. On average, citizens in East Germany are more dissatisfied with the way democracy works compared to the Western population (Niedermayer 2009). However, more recent data shows that the East and the West are converging. Even though the West still shows higher levels of satisfaction, the gap in 2014 was smallest since reunification (Campbell 2019).

25 This is rather a methodological issue than a theory-driven choice. When using the changes in SWD as dependent variable, the initial level of SWD influences the possibility to ‘move’ up or down the scale of the dependent variable (i.e. controlling avoids ceiling effects). This is particularly relevant when analysing only one voter group (such as AfD voters), whereas it seems less necessary (or even obstructive) when comparing different voter groups that start from very different (average) levels of SWD.

26 In the appendix, we display the summary statistics for the dependent, independent and control variables for all voter groups together (Table A-3) as well as per voter group (Table A-4).

27 A Bonferroni multiple-comparison test yields that the change for AfD voters is significantly different (on 95 per cent level) to the changes of voters of CDU/CSU, SPD, Greens and non-voters, but not significantly different from voters of FDP, the Left and other parties.

28 Unfortunately, the question has not been asked again between August 2016 and August 2017, i.e. in the panel waves W2, W3 and W4.

29 We acknowledge that the R² values of these regression models are relatively small. However, it is not our intention to fully explain the changes in SWD, but to show that there is a significant difference between the AfD voters compared to all other groups.

30 As the dependent variable could also argued to be measured on an ordinal scale, we additionally computed ordered logit regression models. The results are exemplified in Table A-6 in the appendix. As the results resemble our OLS regression outcomes, they indicate the robustness of our empirical findings.

31 Although our research interest lies in detecting differences among AfD voters, we also checked whether the results vary from those for other voter groups. Therefore, we computed the same regression model and included interaction terms between our main variables of interest (congruence with the party voted for and identification with another party) and the vote choice (see visualised results in Figures A-2 and A-3 in the appendix). As it turns out, the findings are not easily transferable to other voter groups, but rather unique for voters of the AfD.

32 As representative values for high and low policy congruence on the socio-economic policy dimension, we use one standard deviation (sd = 1.2) above and below the mean (mean = 4.8), i.e. ‘low congruence’ = 3.6, ‘high congruence’ = 6.0.

33 In line with our theoretical hypothesis, we consider protest voters those AfD voters who identify with another political party. In contrast, ‘no protest’ refers to AfD voters who identify with the AfD itself or do not possess any party affiliation.

Additional information

Notes on contributors

Ann-Kathrin Reinl

Ann-Kathrin Reinl is a doctoral candidate at the GESIS-Leibniz Institute for the Social Sciences in Cologne, where she works in the SoliKris project studying solidarity and de-solidarisation in times of European crises. Her work has been published in edited volumes and peer-reviewed journals, such as Comparative European Politics and Representation.

Constantin Schäfer

Constantin Schäfer is a postdoctoral researcher in the RECONNECT project and research associate at the University of Münster (WWU). His research interests include political psychology and political behaviour in comparative perspective as well as party competition over European integration, with a special focus on Euroscepticism and European Parliament elections. His work has been published in edited volumes and peer-reviewed journals, such as the Journal of European Public Policy and West European Politics.