1 Introduction

The results of the fall 2017 German federal elections have seen the far right populist “Alternative für Deutschland” (“Alternative for Germany” – AfD) surging to become the third largest party behind the center-right Christian Democrats (CDU/CSU) and the center-left Social Democrats (SPD). Surprising many observers, the success of the AfD was prove to the fact that the populist surge in Europe and elsewhere is far from over. The representation of the AfD in the federal parliament with 92 “Mitglieder des Bundestages” (“Members of the Federal Parliament”, MdB) also marks a new phase in the history of post-war democratic Germany. The parliament of the Federal republic had been without a far-right party throughout its history of almost 70 years, reflecting the country’s Nazi past and its continuing remembrance in official and societal memory.

So how did this happen? In this paper, I argue that the AfD’s success can be explained by the party’s ability to mobilize an electorate that in past elections had increasingly abstained from voting. Employing a spatial logic and using data form the Chapel Hill expert survey, I show that the proportion of the policy space that is covered by the positions of parties who gained representation in the federal parliament has increased two-and-a-half-fold between 2013 and 2017. In the next step, employing district level data, I show that the AfD was especially strong in those districts where the increase in voter turnout was particularly high. At the same time, this increase in electoral participation that increased support for the AfD, primarily hurt the Social Democrats, but also the CDU/CSU. I also show that the AfD is the only party whose electoral fate in 2017 was positively and significantly influenced by the far-right neo-Nazi party NPD in the previous elections in 2013. This means that the AfD was able to tap into the potential of a party that the Federal Constitutional Court in a landmark 2017 decision has described as promoting a political concept which“seeks to abolish the liberal democratic order of the Federal Republic”, “disregards human dignity and is incompatible with the principle of democracy” (Bundesverfassungsgericht 2017).

To further corroborate my findings, I draw on micro-level data from the German Longitudinal Election Study (GLES) (Roßteutscher et al. 2018; Rattinger et al. 2017a, b) and show (a) that of those voters who had not participated in 2013 but went to the polls in the 2017 elections, a plurality voted for the AfD, suggesting that the AfD was able to mobilize previous non-voters; (b) that the AfD was viewed disproportionately favorably among voter who abstained in 2013 but turned out in 2017; and (c) that the centrist shifts of the CDU were perceived much more strongly by supporters of the AfD than by either the general electorate or by experts.

The AfD was successful by positioning itself in an unpopulated area of the policy space. Recent data from the Chapel Hill Expert survey show that the AfD is more conservative on a societal dimension than any other party in the German political system. On a state-market dimension, the AfD holds the most unclear positions of all parties in the Federal Parliament. Some party representatives have tried to considerably soften the party’s pro-market profile and are now advocating policies that are in line with nativist pro-welfare state stances of parties like the far-right Sweden Democrats. The AfD also holds distinct positions on European integration; it is the party that holds the most pronounced anti-EU integration positions. At the same time, the EU issue has very high salience for the AfD. I will show that the AfD was able to enlarge the policy space to include these issues, especially immigration, that were perceived as important by many voters in the 2017 elections. Using GLES data, I show that for over 40 percent of voters, refugee and immigration issues featured as the most important or the second most important political problem. For AfD voters, this number is 57 percent. These findings are in line with recent results showing that AfD voters are strongly motivated by anti-immigrant sentiments (Goerres et al. 2018).

But why was the AfD able to mobilize so many voters just now? I argue that the perceived centrist turn of the CDU/CSU together with a depoliticized climate during the 2013–2017 Grand Coalition has contributed to the AfD’s success. By moving to the political center on issues like immigration, minimum wages, and, more recently, marriage equality, Angela Merkel’s CDU has won approval from centrist voters. At the same time, this has rendered the center-left SPD and the center-right CDU more alike. On top of that, the Grand Coalition that has governed the country from 2013 to 2017 commanded 80 percent of the votes in the Bundestag. This has contributed to the perception that politics had been depoliticized because all or almost all of the parties supported the policies of the government, especially in issue areas like EU integration and during the so-called refugee crisis of 2015. The AfD was able to re-politicize the discussion, albeit in an ugly way. This is reflected in the fact that 81 percent of AfD supporters said that the “AfD does not solve any problems, but they are telling it like it is”Footnote 1.

The results presented in this paper have implications that reach beyond the 2017 German elections. From a normative perspective, the increase in electoral participation is a good thing for democracy. However, since many of the new or returning voters voted for a political party outside the mainstream consensus, the ability of the political system to provide functionally adequate policies to the demands of the electorate has decreased. Therefore, the support for a far-right protest party could undermine the output legitimizing dimension of politics or what Sven Steinmo has referred to as the “efficacy” of a political system – a government’s ability to “get things done.” (Steinmo 2013, 85). This, in turn, would further undermine citizens’ trust in the system, thereby giving rise to more votes that go to forces outside the established political consensus. Against this backdrop, it is questionable whether an increase in turnout is an unequivocally desirable development.

In the remainder of this paper, I will first situate my argument in the context of the larger literature on electoral participation, de-politicization and far-right party support. In the empirical part, I first present evidence on the AfD’s unique position in the policy space. I will then show results from regression models that connect turnout to party support in the 2017 German elections, before presenting micro-level empirical evidence. A final section concludes.

2 Theoretical considerations: Indifference, alienation and the politics of de-politicization

Far-right parties may be a new phenomenon as a presence in the Federal Republic’s parliament. Yet, they are anything but new in the political arena of most Western democracies. However, their significance for electoral participation has been scantly explored. This is somewhat surprising given that we have fairly solid results on the turnout-enhancing effect of a higher number of parties competing for votes (Martin and Plümper 2005; Cancela and Geys 2016). Furthermore, as Crepaz (1990) has noted more than a quarter century ago, the presence of parties competing on a dimension that is different than the one other parties are competing on, can significantly enhance electoral turnout.

While Crepaz was referring to post-materialist parties that had at the time of his writing been only recently established as political competitors in the electoral arenas of the West, these parties are analytically not different from contemporary right-wing populist parties. Just like the post-materialist parties of old, the right-wing populist and authoritarian parties do not mainly compete on a classic left-right dimension of government intervention and redistribution. Rather, they position themselves most distinctly on the opposite end of the societal dimension that their post-materialist competitors had been opening up so successfully as a new cleavage in electoral politics (Müller-Rommel 1999). Paul Taggart (1995) describes both the emergence of “New Politics” parties on the left and of “New Populism” parties on the right as a reaction to the “demise of the postwar settlement” (Taggart 1995, S. 47).

While the New Politics party have since been absorbed into the mainstream and, especially in the case of the German Greens, see themselves as the true heirs to the values of liberal democracy, the right-wing populists are competing as anti-system parties and as political forces that are in opposition to the established parties and “corrupt elites” (see, for many, Mudde (2016). Two strands run through the rhetoric of the authoritarian populist. The first one is a rejection of the system as a whole. It is perceived as corrupt and broken, not in line with “the will of the people”. In Germany, the AfD claims that the rule of law has been compromised and that

“the secret sovereign in Germany is a small, powerful political oligarchy that has spread in the existing political parties. […] A political class has formed whose primary interest is their power, their status and their material well-being. This oligarchy controls the levers of power in the government (“staatliche Macht”), political education and the influence of the media on the population.” (AfD 2017, point 1.3, author’s translation).

This is fairly typical right wing populist parlance. The people are manipulated by an elite conspiracy. If only the people could take back decision making powers, all problems could be swiftly overcome.

The second strand of far-right rhetoric is the sameness of all parties in the system. Not only the political system itself is rejected, but the political parties in that system are cast as indistinguishable. The AfD manifesto of 2017, cited above, is full of references to “the parties”, “all parties”, “different parties who all promise the same”. The only party that is unlike all the other parties is, of course, the AfD.

These two strands, anti-system rhetoric, and indistinguishable-parties rhetoric, align with two theoretical explanations for vote abstention in the literature, alienation and indifference. In a spatial voting model, alienation occurs when all the parties or candidates are too far away from a voter’s ideal point, thus rendering the cost-benefit calculation of voting negative. Indifference stems from the impression on the part of the voter that parties/candidates are too similar to warrant the effort of voting. The authoritarian populists position themselves both as alternatives to other parties, thereby addressing the indifference motive, and as political forces outside the system, thus addressing the alienation motive. This is in line with theoretical expectations from simulation studies that show that an increased propensity to abstain will have a radicalizing and polarizing effect on the party system (Plümper and Martin 2008). This effect should be stronger for abstention from alienation motives than for abstention that is due to indifference (Erikson and Romero 1990; Adams et al. 2006, 1120–1121).

According to Dassonneville and Hooghe (2016), both indifference towards parties as well as alienation from all parties and therefore the party system as a whole has increased over time in the established democracies of Western Europe. Using a time series of survey data from 1989 to 2014, Dassonneville and Hooghe (2016) show that the de-alignment of voters from parties, i.e. the weakening of the bond to a specific party that stems from socialization and cultural practices, has led to an electorate that is both more open to other parties and alienated from all parties. In such a context, a party that is unlike all the other parties in the system, like the German AfD, can hope to garner substantial support, both from voters who are interested in the policy positions the party stands for and from voters who have been alienated from all parties and chose not to participate in elections in which the new party was not present.

Kitschelt and McGann (1997, 48) seemed to have had the German case in mind when they wrote, 20 years before the fact, that the electoral prospects of the “New Radical Right” will be markedly improved “where moderate left and right parties have converged toward centrist positions and may even have cooperated in government coalitions”. Abedi (2002) tests this prediction against the expectation that right wing parties’ electoral success might be improved not by an increasing similarity of established parties but rather by an increasing polarization of the party system, as Ignazi (1992) has famously claimed. Abedi finds empirical support only for the similarity thesis, not for the polarization thesis.

Both indifference and alienation as reasons for vote abstention invoke a spatial logic in which parties either hold positions that are too similar or all parties are perceived as too far away from a voter’s ideal point to make voting worthwhile. Both developments have their historical and theoretical underpinnings. Against the backdrop of the “government overload” discussion of the mid 1970s to mid 1980s (Streeck and Schäfer 2013, 169–195), policy platforms in advanced economies generally shifted towards more market friendly positions. As policies throughout the OECD countries became more economically liberal and more open towards increasingly globalized markets, the problem-solving capacity of national governments became increasingly more doubtful. Hellwig (2014) presents evidence that electorates today ascribe limited responsibility for managing the economy to national governments. This is a result reflected in Steiner and Martin (2012) who find that (a) partisan positions have become more similar as a result of global integration and (b) the increase in similarity has decreased turnout – which brings us back to our initial discussion about the relationship between indifference, turnout, and the electoral success of right wing authoritarian parties. The convergence of platforms, i.e. the perceived increased similarity of political parties is one source of declining turnout.

Globalization has been invoked by, among others, Adams et al. (2009); Ward et al. (2011); Burgoon (2009) as an explanation for an increased similarity of party platforms. Mudde (2004, 555) draws a direct connection between the “actual or perceived” limitation that globalization puts on the ability of policy makers to influence outcomes and the “Populist Zeitgeist”. Crouch (2004), on the other hand, has argued that Western countries have moved to the age of “post-democracy”, in which government functions are delegated to market actors and where government agencies themselves function not according to a political logic but rather in a mimicry of the market model. Such a post-democracy de-politicizes the political by re-commodifying previously de-commodified aspects of the lives of its citizens. In such an environment, a rhetoric of “There is No Alternative” (TINA) – famously associated with former British Prime Minister Margaret Thatcher – can be used politically to convey the inevitability of adhering to market forces, if not surrendering to them.

A powerful echo of the TINA ceterum censeo could be heard during the world financial and economic crisis and even more so during the sovereign debt crisis in Europe. In 2010, German Chancellor Merkel called a bailout package for Greece “alternativlos” – “without alternative” (FAZ 2010). Without the German parliament consenting to a law committing Germany to a loan guarantee of 22.4 billion Euro, a new financial crisis would threaten prosperity and the viability of the European project. This was not the last vote in the German parliament during the looming crisis and it was not the last time that the TINA rhetoric was invoked.

But globalization not only renders the positions of established parties more alike, it may have also changed the structure of political competition. Kriesi et al. (2006) argue that the new cleavage between winners and losers of globalization processes becomes “embedded” in the old cultural dimension of a two-dimensional policy space. In this process, new issues will be integrated into the cultural dimension, prime among them immigration and EU integration (Kriesi et al. 2006). Along similar lines, Bornschier (2010) argues that globalization indirectly increases conflict on a cultural dimension by weakening the economic cleavage. This is important for my argument: Applying a temporal logic, a weakening of the state–market dimension that precedes the increase in conflict on the cultural dimension can be expected to first lower electoral participation and then, after the conflict on the cultural dimension has intensified, to increase electoral participation to the benefit of those actors that hold the most outspoken positions on this dimension.

For the German case, it took until April 2013 for such an outspoken new player to enter the political arena. The “Alternative for Germany” explicitly opposed “bailouts” for Greece and other countries in the Eurozone. From the beginning, the party positioned itself as an integration skeptic party that challenged the post-war consensus in Germany. West Germany, since the early days of the first Adenauer government which was established in 1949, sought to firmly embed itself in Western cooperation and integration arrangements. The transfer of sovereignty to supranational entities was viewed as a means to regain the ability to act on the international stage in a situation when Germany was a pariah among nations. This strategy has proved extraordinarily successful and has since defined foreign policy in Germany. By challenging this consensus, the AfD was also challenging the notion that there was no alternative – to European integration, to a bailout for Greece, to a further transfer of sovereignty when eventually fiscal policies would have to be delegated to the supranational level, to more redistribution within Europe. The de-politicization that for the European context had first been driven by a functional logic of integration in the shadow of a permissive consensus (Hooghe and Marks 2009) and then by the purported inevitability to advert the worst fallout of the triple crises had backfired. It had brought a new actor into the political arena that was thoroughly re-politicizing the discussion; a strategy that was eventually rewarded at the polls.

In the next section, I will put these propositions to the empirical test. I will show that the AfD was successful because it was able to mobilize the electorate and that the electorate had previously abstained because of indifference and alienation motives that were fueled by a feeling that voting made no difference.

3 Empirical Evidence

In this section, I will present empirical evidence for the claim that the AfD was able to mobilize an electorate that had not participated in previous elections. The evidence will be presented in three steps. First, I show descriptive findings at the party system level and at the level of electoral districts. Macro data at the district level is then used to draw inferences about the impact of turnout changes on the electoral results of the AfD and other parties in the 2017 elections. I then turn to micro level data from different waves of the GLES. I show that the individual level data is broadly in line with the theoretical claims outlined above.

3.1 Macro level data and descriptive findings

3.1.1 The policy space

Fig. 1a shows the political positions of the larger German parties on two dimensions, based on data from the Chapel Hill Expert Survey (Polk et al. 2017). The figure depicts parties’ positions on two aggregate dimensions. The horizontal axis is a state-market dimension and can be described as the “classic” left-right dimension ranging from a policy stance that advocates welfare spending, redistribution and government intervention to pro-market, low tax, low spending policy preferences. The second dimension is the “new politics” dimension, the materialist/post-materialist dimension (Inglehart 1989), or the GAL/TAN dimension – “Green, Alternative, Libertarian”/“Traditional, Authoritarian, Nationalist” (Hooghe et al. 2002).

Fig. 1
figure 1

Political parties in Germany, a 2017 and b 2014 positions

Full size image

The AfD is by far the most conservative party on this dimension. Disregarding the CSU – the CDU’s sister party that only competes in the state of Bavaria – the party commands a unique position in the top right corner of the policy space. The CDU is very much a party of the middle, both on the economic and the societal dimension. Prior to the 2017 elections, the CDU/CSU voted for a minimum wage and a re-regulation of the labor market; a shift that did away with some of the pro-market reforms of the Schröder government. On the GAL/TAN dimension, the CDU/CSU opened up the vote to allow for the introduction of marriage equality, including the possibility for gay and lesbian couples to adopt children. Maybe most importantly, Chancellor Merkel’s 2015 decision to open Germany’s borders to about 1.1 million refugees and migrants was an unprecedented shift. The move has garnered the support of the opposition Green party and earned the Chancellor approval far into the left voting base.

Fig. 1b shows party positions in 2014, again based on CHES measures. The CDU changed its position on the GAL/TAN dimension only moderately, from 6.0 in 2014 to 5.8 in 2013. However, as will be shown below, the extent of this shift is perceived as much larger by voters of the AfD than by voters who didn’t support the AfD. There is another difference between the situation in 2014 and 2017: The proportion of the policy space that was covered by parties represented in the federal parliament increased markedly between the two points in time. Fig. 2a and b shows the policy space with the ideological space covered by parties in parliament shaded in grey. After the 2013 elections, with the AfD and the FDP narrowly missing the representation threshold, a mere 12 percent of the ideological space was covered by the positions of the parties represented in parliament. After the 2017 election, this number increased to almost 30 percent. It seems plausible that some voters who did not vote in the 2013 election chose to abstain because of alienation reasons. With the AfD as a credible competitor, these voters may well have found an alternative more to their liking in 2017. In line with this, and as will be shown below, a plurality of those that didn’t vote in the 2013 elections but participated in 2017 voted for the AfD.

Fig. 2
figure 2

a 2014 and b 2017 party positions with ideological space covered by parties in parliament

Full size image

Comparing Fig. 1a and Fig. 1b also shows that the AfD has moved somewhat to the left on an economic dimension. It is worth pointing out that the economic position of the AfD is less straightforward than the data might suggest. Of all the parties covered by the 2017 CHES data, the AfD is by far the least clear in their economic position. On a 0 to 10 scale for clarity of the economic left-right position, the party scores a 3.5. By contrast, the Green party is at 6.3 and the CDU at 6.1 for the clarity of their respective economic stances. The AfD is experiencing an intra-party struggle between its market-liberal origins and welfare-nativist forces. As Lefkofridi et al. (2014) point out, voters who combine preferences for redistribution with authoritarian views, represent a “considerable size” of the electorate across Western Europe. Some in the AfD, especially in east Germany, seem to make a push to become more attractive to those voters by softening their market-liberal profile.

The position of the AfD is unique in another respect as well. In Fig. 3, we see the positions parties take regarding EU integration and the salience of this issue in the respective party manifestos. Again, the AfD stands out from all the other parties with its emphasis on the EU and the negative attitude towards EU integration. The tight cluster of CDU, SPD, and the Greens in the pro-integration, moderato-to-high salience region of the figure reflects the post-war consensus among established political parties in GermanyFootnote 2. The AfD departs from this consensus, offering itself as an alternative to voters who do not share the pro-integration stances of other parties in the political arena.

Fig. 3
figure 3

Parties’ 2017 positions on European integration

Full size image

3.1.2 Voter shifts and regional turnout patterns

But was the AfD actually able to mobilize voters with these policy stances? To answer this question, we can first look at voter shifts. Fig. 4 is based on data from the German Longitudinal Election Survey post-election survey that was conducted between Sept. 25 and Nov. 30, 2017 (Roßteutscher et al. 2018). It shows that previous non-voters constituted the second-largest voting block of the AfD among voters who had previously not voted for the AfD. The AfD won 5,878,115 list votes. Of these, more than 13 percent came from former non-voters.Footnote 3

Fig. 4
figure 4

Voter shifts in the 2017 federal elections based on GLES post-election data. Source: Roßteutscher et al. 2018

Full size image

The AfD won the largest chunk of votes from former voters of the CDU/CSU. More than two million CDU/CSU voters turned to the AfD. This is in line with the expectation that the AfD was able to mobilize voters because (a) the party occupies an unpopulated area of the policy space and that (b) the party that has been closest to this region, the CDU/CDU, would suffer from a move to the center to the benefit of the AfD.

We can, therefore, cautiously infer from these observations that the AfD was, indeed, able to mobilize voters and that this mobilization effort increased voter turnout. Of course, the AfD did not mobilize all of those who abstained in 2013. In fact, with an increase in participation rates of 4.6 percent between 2013 and 2017, the 14.7 million non-voters still constitute the largest group when compared to the results of individual parties. As will be shown below, however, the AfD was uniquely successful in their mobilization effort and was the plurality winner among those who said that they had abstained in 2013 but claimed to have voted in 2017.

The pattern of an increase in participation benefiting the AfD can already be seen in one of the earliest elections in which the AfD competed. Coinciding with Federal elections, the state of Hesse conducted elections to the state legislature on Sept. 22, 2013. The AfD, having been founded less than eight month before that date, competed in these elections and won 4.1 percent of the votes. Analyzing where those votes came from shows that only 80,000 of the AfD’s roughly 127,000 votes in the state came from voters who had voted for other parties in the previous, 2009, election. Almost 47,000 votes (37 percent) came from former non-voters.Footnote 4

Interestingly, the short-lived Pirate party did not have the same effect on electoral turnout. In the 2012 elections in Germany’s most populous state, North Rhine-Westphalia, the Pirate party surged to 7.8 percent of votes, guaranteeing them 20 seats in the state’s 237 seat parliamentary assembly. However, electoral participation was virtually unchanged from the previous elections in 2010. The Pirate’s success was not one of mobilization but one of voters switching from other parties. This in in line with theoretical expectations from spatial voting theory: The Pirates positioned themselves in a crowded area of the policy space. In many respects, their position was largely identical with that of the Green party and bore strong resemblance to the Left party’s positions. In contrast, the AfD occupies a position in the policy space that is virtually unpopulated, especially unpopulated by contenders with a reasonable expectation to make it into the federal parliament. This is also the reason why the electoral fate of the Pirates will probably not be shared by the AfD.

A final piece of macro-level descriptive empirical evidence comes from the regional distribution of turnout rates. Turnout increased overall to 76 percent from 71.5 percent in 2013. However, turnout did not increase everywhere to the same extent. Fig. 5 shows a map of the change in turnout rate between 2013 and 2017 on the level of electoral districts. The Eastern part of the country stands out, with turnout rates in some district in Eastern Bavaria jumping by 10 percent and more.

Fig. 5
figure 5

Change in voter turnout between 2013 and 2017 elections

Full size image

We can look at a map that shows the change of AfD vote share from the 2013 elections (Fig. 6). While there is no perfect congruence with turnout changes, it seems clear that the AfD did particularly well in those regions that saw pronounced changes to electoral turnout.

Fig. 6
figure 6

Change in AfD vote share from 2013 elections

Full size image

3.2 Macro level inferences

I will now analyze the electoral success of the parties represented in the German parliament using multivariate models. The question asked here is what determines the electoral success of parties and, most importantly, how does this success relate to electoral participation? In taking advantage of the variation in both variables at the electoral district level, we can draw conclusions about the importance of electoral participation. By relating participation to the electoral success of different parties, we can draw inferences about which party benefited the most. I will show that the increase in participation has disproportionately benefitted the AfD.

The data come from the Federal Returning Office and are at the district level. There are 299 electoral districts in Germany. In each of these districts, voters cast two votes. The first vote determines the plurality winner of a competition among candidates in the district. The second vote goes towards the party list and is the more important one since deviations from the second-vote based result through overrepresentation of list candidates will be equalized in the composition of the Federal parliament.

As my core independent variable, I include the change in electoral participation from the previous election. Modeling turnout as change rather than as level allows for an implicit comparison between structurally different districts with different baseline voter turnout rates. Additional independent variables are population density to account for the influence of a possible urban/rural cleavage, the percentage of people with “migration background”Footnote 5 in a district to gauge the idea that exposure to migrant decreases far right success, disposable income per capita and the unemployment rate to test theories of economic deprivation as an explanation for far-right success and a dummy variable (“East”) that is set to 1 if the district is located in the former GDR. Table 7 in the Appendix (available online) gives an overview and descriptive statistics for these variables.

I estimate two sets of multivariate OLS models for the vote shares of each of the six parties CDU/CSU, SPD, AfD, FDP (the pro-business liberals), the Greens and the Left Party.Footnote 6 In each of these models, the vote share of the respective party in the 2017 elections is the dependent variable. The two sets of models are distinguished by their treatment of past election results. The models shown in Tables 3 and 4 include the respective party’s vote share in 2013, while the models shown in Tables 1 and 2 do not. The interpretation of the coefficients changes accordingly. While those models that do not include 2013 vote shares model the level of 2017 vote shares as a function of the covariates included in the model, the models that do include 2013 vote shares effectively model the change in vote shares from 2013 to 2017. The “Vote share 2013” variable leaves only that part of the variance to be picked up by the other covariates that is not modeled by that variable. This is just the difference between a party’s result in 2013 and its result in 2017.

Table 1 Determinants of party success, not including past election results
Full size table
Table 2 Determinants of party success, not including past election results
Full size table
Table 3 Determinants of party success, including past election results
Full size table
Table 4 Determinants of party success, including past election results
Full size table

Considering the level models in Tables 1 and 2 first, we find a number of interesting results. The change in turnout from 2013 (a 4.6 percent increase on average) benefitted the AfD and hurt both the SPD and the FDP. It was barely significant for the CDU/CSU and the Left Party. Turnout change did not matter for the Greens.

Substantively, increasing the change in turnout (i.e. mobilization) from its mean by one standard deviation leads to a one percentage point increase in AfD votes. This is a relatively large effect, given that the AfD fell short of clearing the threshold for parliamentary representation by just 0.3 percent in the 2013 federal elections. On the other hand, the SPD massively lost from an increase in turnout, showing that the party was not able to speak to its voters. Of the other variables, we observe a number of significant effects. Most importantly the AfD is stronger in the East. On average, going from a Western district to one in the East increases AfD vote share by more than 11 percent.

Secondly, the AfD is stronger in less populated areas, i.e. outside the big cities. The AfD is also stronger where voters are less educated and where unemployment is higher. Finally, the AfD is stronger in districts where the share of the population with a “migration background” is higher. Given the definition of “migration background” (see above), immigrants from Russia who came to Germany in the early 1990s and who disproportionately support the AfD could explain this effect. Fig. 7 shows a z-score standardized coefficient plot (beta coefficients) that allow for a more accessible comparison.

Fig. 7
figure 7

Coefficient plot of z‑score standardized independent variables for AfD support

Full size image

If we analyze the AfD as a “New Populism” party (Taggart 1995) and compare it to its competitor at the other end of the GAL/TAN continuum, we find that, there, the Green party is almost a mirror image of the AfD when it comes to the sources of its electoral support. The Greens are stronger in more densely populated areas (i.e. in the cities), they benefit from a highly educated population and they are stronger where unemployment is lower. Interestingly, they are – just like the AfD – stronger in districts where the population share with “migration background” is higher. I would venture the guess that this effect has different sources for the Greens and for the AfD. For the Greens, it seems plausible to expect that this result reflects their success in the multi-cultural neighborhoods of the big cities. Finally, the Greens are weaker in the East, but this result is only marginally significantly.

In fact, the AfD is unlike any other party – there is no competitor that matches the AfD’s profile in terms of the combination of variables that predicts party success. And the AfD is unique in yet another, more troubling respect – its strong predictive power of a variable that captures the NPD’s vote share in 2013. As mentioned above, the NPD is a party with a clear neo-Nazi profile. Table 5 shows regression results on AfD (model 1) and CDU/CSU (model 2) vote shares that include the district results of the NPD in the 2013 elections. The AfD is the only party whose success is positively and significantly related to NPD vote shares in 2013. By contrast and for example, the CDU/CSU’s share is negatively and significantly related to NPD vote share in 2013 (the same is true for FDP, the Greens, and the Left party, while the result is insignificant for the SPD)Footnote 7. The magnitude of this effect is large, too: Comparing the district with the lowest NPD support in 2013 to the district with the largest support, the result for the AfD would increase by more than 2 percent through that effect alone (with everything else being equal). What is more, AfD support in 2017 is positively and significantly related to NPD support in 2017 as well. The difference between the district with the lowest NPD support and the district with the highest NPD support translates into an increase of AfD votes of almost 2 percent. Again, the positive and significant relation between NPD support in 2017 and AfD support in 2017 is unique to the AfD.

Table 5 Neo-Nazi vote share and electoral success
Full size table

I will now turn to the discussion of the results shown in Tables 5 and 6. Recall that these regressions include vote shares from the 2013 elections as independent variables. We can thus interpret the effect of the change in turnout variable as evidence for the success in mobilizing voters. Put differently, the change on turnout here models the difference between 2013 and 2017 vote shares.

For the AfD, this variable is positive and significant, while it is negative and significant for the CDU/CSU, the SPD and the FDP. This means that the AfD was able to turn a change in turnout into a change in support over 2013, i.e. that it was able to disproportionately mobilize voters. By contrast, the CDU, whose vote share level benefited from an increase in turnout (all else being equal), was hurt if the difference to the last election is considered. The coefficient on the turnout change variable can be interpreted as an indicator for the degree to which a party was able to asymmetrically mobilize supporters. Increasing the difference in turnout between 2013 and 2017 by one standard deviation from the variable’s mean accounts for more than 15 percent in the difference between the AfD’s vote share between 2013 and 2017. Going from the minimum turnout change to the maximum accounts for almost 80 percent in the AfD’s increase in vote share over 2013. For the CDU/CSU who lost more than 8 percent in vote share between 2013 and 2017, the change in turnout accounts for more than a fifth of its losses. For the SPD, this number is almost 30 percent. Comparing z‑score standardized beta coefficients, we find that the change in turnout is the second most important predictor for the change in AfD support, behind the East dummy variable.

In the next section, I will employ empirical evidence at the level of the individual voter in order to corroborate these macro-level findings and relate them to the theoretical discussion.

3.3 Micro level empirical evidence

Macro-level evidence can only tell us so much. While the results at the macro level outlined above seem to be broadly in line with the theoretical expectation of a mobilization success of the AfD based on motivating voters who did previously not participate in elections, individual-level data can provide a more complete picture. Here, I augment the macro-level information with inferences based on individual level data. Using different waves of the German Longitudinal Election Study (GLES), we can first address the question whether a change in indifference motives is responsible for the AfD’s success. The GLES contains a question that asks whether respondents think that “Voting can make a big difference” or if they think that “Voting won’t make any difference to what happens in politics” (Rattinger et al. 2017b) In 2009, the percentage of respondents who opined that voting can make a big difference, was 31.3. This number was even lower in 2013, where only 28.7 thought that voting makes a big difference. In 2017, on the other hand, 41.7 percent of respondents ascribed a big impact to voting.Footnote 8

To see if there is a connection between the perception of the importance of voting and electoral participation, respondents’ opinions are related to their (self-reported) actions. Fig. 8 uses the GLES post-election surveys from the years 2009, 2013 and 2017 to gauge whether there is a consistent pattern that connects the perception of the impact of voting with self-reported electoral participation. Despite the over-reporting of voting, there is a clear pattern across all three federal elections. Electoral participation increases in the importance respondents ascribe to voting. What is noteworthy about this is the increase in participation rates among those who believe that voting “won’t make any difference” from 2009/2013 to 2017. Almost 70 percent of those who say that voting doesn’t make a difference claim that they participated in the 2017 elections. Part of this is without doubt due to over-reporting of electoral participation. But it is also in line with the the expectation that the AfD was able to mobilize in particular those who feel distanced and alienated from the established party systemFootnote 9 (Schwarzbözl and Fatke 2016).

Fig. 8
figure 8

Attitudes on importance of voting and self-reported electoral participation

Full size image

How does this relate to the increase in support for the AfD? In the GLES survey, voters are asked to recall whether they have participated in the previous federal elections. Comparing these data between 2013 and 2017, we can construct a variable “BECAMEVOTER” that takes on the value 1 if someone reported not to have voted in 2013 but said that they participated in 2017. All other combinations of the two electoral participation variables are coded as 0. We can then ask how the distribution of this variable is different between self-reported voters of the AfD and those who either voted for another party or refused to answer. Among AfD voters, the rate of those who had become voters in the 2017 elections, i.e. those who did not participate in 2013 but did participate in 2017, was 14 percent. By comparison, among all other respondents, that rate was just over 3 percent. Looking at this the other way round, the AfD is the most voted-for party among those who became voters in 2017. Of the 72 voters who did not participate in 2013 but claimed to have done so in 2017, 21 reported having voted for the AfD (29.2 percent), while 16 (22.2 percent) said that they had voted for the second strongest political force, CDU/CSU. Despite the relatively small number of respondents who switched from non-voting to voting, this is a striking result and shows at the individual level that the AfD benefitted the most from increased voter turnout. Mobilization success was skewed towards voters who ended up voting for the AfD.Footnote 10

A different observation corroborates this conclusion. We can also look which party a voter “feels close to”. Such closeness is associated with voting decisions but does not fully determine them. In fact, a plurality of voters (31.49 percent) report not feeling any closer to any party than to others. Among those who switched from not voting in 2013 to voting in the most recent election, this number is 41.67 percent. Among those that do feel close to a party, however, the AfD is the second-most mentioned party (the SPD is the most mentioned party). We can interpret this finding as evidence that many of those who decided to participate in the 2017 elections but had not participated in 2013 have become dissociated from party politics and feel no attachment to any party. Most of these votes, as seen above, went to the AfD. This observation is in line with both an indifference and alienation motive for abstention.

What were the topics these voters cared about? The GLES 2017 data contain an open question regarding the most important and the second most important political issue. I coded all the responses that contain the terms “Flüchtlinge”, “Ausländer”, “Einwanderung”, “Asyl” or “Islam” and derivations thereof (including misspelled terms like “Flüchlingskriese”) as 1, zero otherwise. There is no big difference in the importance respondents ascribe to the refugee topic between those who had not participated in 2013 but did in 2017 and other groups of voters. Among both groups, refugees, immigration and foreigners are mentioned by about 40 percent of respondents as the most or second most important political problem. For the subgroup of those who did not vote in 2013 but did in 2017 AND reported having voted for the AfD in 2017, this number stood at 57 percent, suggesting that an important motivational factor can be found in the migration and refugee topic. This observation is compatible with the notion of the AfD as a party that is unlike other parties in their outright opposition to allowing refugees into the country. These findings corroborate recent studies on the positions of AfD supporters with their emphasis on immigration (Goerres et al. 2018; Lengfeld and Dilger 2018), but also – in light of the findings presented above – their disenchantment with the party system as a whole (Schwarzbözl and Fatke 2016).

An additional piece of evidence for the theoretical claim outlined above comes from looking at the perception of voters. The GLES data contain information on how voters view parties. They are asked to voice their opinion on parties by using an 11 point scale with a range from \(-5\) to \(+5\) where \(-5\) stands for “I don’t think anything of this party at all” and +5 denotes “I think very highly of this party”. The AfD is the least liked party with an average value of 2.77. A full 60 percent of voters chose the worst possible judgment for the AfD. By comparison, the SPD, an established party that took a bad beating in the 2017 elections and reached its worst electoral results in the party’s history, has a mean favorability value of 7.4. Only 3.27 percent of respondents say that they think absolutely nothing of the party. Comparing these numbers with those among voters who did not vote in 2013 but came to the polls in 2017, we find that the share of those who think absolutely nothing of the AfD drops to 44.4 percent and the average favorability rating increases to 4.25. In fact, the AfD is the only party (with the notable exception of the LEFT party) who fares better among those who came out to vote in 2017 but didn’t vote in 2013. Again, the mobilization effect is strongly skewed towards the AfD.

Table 6 Perception of political positions
Full size table

As outlined above, the shift of the CDU on a GAL/TAN dimension as measured by the CHES data was fairly moderate. However, how is this shift perceived by the voters? GLES data from 2017 and 2013 show that the shift of the CDU to the left is most pronounced (a) in the perception of AfD voters and (b) on a GAL/TAN dimension. While non-AfD voters placed the CDU at 6.79 and voters of the AfD saw the CDU at 6.48 on the libertarian-authoritarian dimension dimension after the 2013 elections, this perception changed massively in 2017. AfD voters placed the CDU at 4.98, while non-AfD voters placed the party at 6.1. This perception of a  pronounced leftward shift of CDU on a new politics dimension by AfD voters also renders the CDU and the SPD more alike. In 2013, the perceived difference between the parties on the GAL/TAN dimension was 1.74 among non-AfD voters and 1.99 among AfD voters. In 2017, these differences changed to 0.87 and 0.58, respectively. The difference between AfD voters and non-AfD voters in the perception of the CDU in comparison to the SPD thus increased between 2013 and 2017. This is evidence of a successful mobilization that portrayed the CDU as being “too far” away from AfD voters’ ideal points and the AfD taking up the slack.

4 Conclusion

The increase in electoral participation in the 2017 federal German elections is in part responsible for the strong result of the far right authoritarian and populist AfD. The party was able to asymmetrically mobilize the electorate, at the expense of almost all other parties. All else being equal, the difference between the district with the smallest change in turnout and the district with the largest turnout change accounts for 80 percent of the AfD vote increase between the 2013 and the 2017 elections. Furthermore, the party was stronger in the East, in less densely populated areas, in districts where the population is less educated and where unemployment is higher. Additionally, the effect of a variable that captures the percentage of the population with a “migration background” was positively and significantly related to the AfD’s vote share. This result is most likely due to the strong support the party enjoys among the group of Russian-German immigrants. Maybe most disturbingly, AfD vote share is positively and significantly related to support for the neo-Nazi NPD. This is true for NPD vote shares in both the 2013 and the 2017 elections and controlling for structural and socio-economic factors.

The micro-level evidence presented in this paper is in line with these macro-level observations. The mobilization of voters who had not participated in the 2013 elections but did so in 2017 was heavily skewed in favor of the AfD. The AfD was the strongest party among this group of voters. Furthermore, AfD voters perceived a much more pronounced shift of the CDU to the left than either the general electorate or as is portrayed in expert data on the party’s political positions.

Theoretically, I have argued that the mobilization success of the AfD can be explained by two factors. First, I have invoked the logic of an increasing similarity of established parties’ positions. With the CDU’s perceived moves towards the electoral center, especially on a societal dimension, the AfD was left without competitors in the pronouncedly conservative or even reactionary area of the policy space. Secondly, I have argued that electoral competition in Germany (and elsewhere) has been depoliticized with purported functional necessities supplanting political debate. The AfD has re-politicized the debate with its opposition to further EU integration and especially its stance against refugees. The AfD’s success can be interpreted as an effect of having reduced abstention from indifference and alienation motives by providing voters with an alternative that is both distinct from other parties on policy grounds and promised to be outside the established political consensus, thereby offering a re-politicization of politics.

These findings have implications that reach beyond the 2017 German elections. Normatively, the increase in electoral participation is good news for democracy. Apparently, the AfD was able to fill a representation gap with political positions (and a distinct style of politics) that had otherwise not been available. It remains to be seen whether the AfD will be able to deliver on the hopes of its voters. The AfD is a political party outside the mainstream consensus. This will make it harder for the political system to develop functionally adequate policies in order to meet the demands of the electorate. This effect could undermine output legitimacy of the system, thus feeding back into the source of support for the a party outside the political mainstream. The increased turnout could then – somewhat paradoxically – work to ultimately reduce the overall legitimacy of the political system.