When the East Goes to the Poll: Explaining the Regional Differences in AfD-Voting in the 2021 Federal Election

In the 2021 German federal election, the share of East Germans who voted for the Alternative for Germany (AfD), a radical right-wing party, was more than twice as high as the share of AfD voters in West Germany. While most scholars focus on the general determinants of AfD-voting, this paper sheds light on specific variables that could explain these east-west differences. To achieve this, I develop a two-sided argument which considers both the distribution of common AfD-voting determinants and the socialization experiences of East Germans. Multivariate regression with the 2021 GLES data reveals a remaining East bonus in AfD-voting, indicating that the distribution of voting determinants alone cannot explain the regional differences in AfD vote share. Nevertheless, this paper presents a novel structured approach to the issue.


InTroDUction
On September 26, 2021, the ninth all-German federal election was held (Deutscher Bundestag 2021). While more than three decades have passed since the reunification of the country, East and West Germans still differ in their voting behavior. This phenomenon is especially pronounced for the newest political party in the German federal parliament, the Alternative for Germany (AfD) (Goerres et al. 2018: 249). Characteristics for the party are its nativist ideas and populist rhetoric (Arzheimer/Berning 2019: 6;Goerres et al. 2018: 249).
Interestingly, this East-West divide in the electoral success of the AfD was not present initially but developed over the years and coincided with a party transformation (Arzheimer 2021: 64;Bieber et al. 2018: 449). In 2013, the party achieved similar election results in both parts of Germany (Arzheimer 2021: 64). However, they did not surpass the five percent threshold to enter federal parliament (Goerres et al. 2018: 246). This changed in 2017 when the party achieved a vote share of 21.9% in the Eastern part of Germany whereas only 10.7% of West Germans voted for the emerging party (Weisskircher 2020: 644).
During the period between the 2013 and 2017 elections and specially following the so-called refugee crisis in 2015, the AfD underwent a programmatic reorientation towards a radical right, nativist party (Bieber et al. 2018: 441-42;Goerres et al. 2018: 247;Schmitt-Beck et al. 2017: 275).
In 2021, the political party lost votes and could only record 10.3% (Der Bundeswahlleiter 2021; Deutscher Bundestag 2021). Nevertheless, the East-West divide increased even further with vote shares of 20.9% in the territory of the former German Democratic Republic (GDR) and of only 8% in West Germany (Der Bundeswahlleiter 2021, own calculations). Thus, there appears to be a correlation between the voters' residence in East or West Germany and their voting behavior for the AfD that has recently or frequently been analyzed in a structured way, especially not for the 2021 federal election.
However, this divide is particularly interesting as the AfD is not as a typical regional party (Arzheimer 2021: 62) which otherwise might have been an explanation for the great differences.
Additionally, the party has attracted a lot of public attention in recent years due to its radical right slogans and statements (Goerres et al. 2018: 246). From a normative perspective, this issue demands more research and therefore raises the following question that is answered in this paper: How can one explain the regional divide between East and West Germany in AfD-voting in the 2021 federal election?
In order to address this question, the paper examines both common theoretical approaches to AfD-voting -economic and socio-demographic explanations -as well as two arguments related to the specific experience of East Germans in the former GDR by Pesthy et al. (2021). By replicating different approaches from relevant research, this paper aims to incorporate several arguments into a broader framework. When analyzing the 2021 GLES pre-and post-election data, this study identifies an 'East bonus', even after controlling for standard explanatory variables. However, the findings also indicate that different socialization in the GDR and West Germany may also play a role in explaining regional differences.
The paper is structured as follows: First, the literature regarding the East-West divide is reviewed to identify the research gap that this paper aims to fill. Second, I present a theoretical argument explaining the nature of the regional differences. Following the description of the methods and data, the theorized hypotheses are empirically tested to analyze the phenomenon. Finally, the paper concludes with a critical evaluation of its limitations.

State of Research
East-West differences in voting behavior have been an interesting phenomenon in political science since the first all-German federal election in 1990 (Arzheimer/Falter 2005: 244). However, the debate gained renewed attention in recent years with the rise of the AfD and its strong support in the Eastern region of the country (Weisskircher 2020: 614). Scholars have applied various theoretical frameworks trying to explain the success of the right-wing populist party in general and espe- Several papers offer arguments and analyses for regional differences on the meso-level: Haffert (2022) highlights the importance of historical events in shaping contemporary regional voting patterns, while Ziblatt, Hilbig, and Bischof (2022) provide insights into the influence of regional dialects on support for the AfD. Furthermore, the intergroup contact theory is an important theoretical framework in meso-level analyses. Otteni (2019) finds that low immigrant density is associated with AfD electoral success. Wagner et al. (2003: 30) also employ the intergroup contact theory to examine differences in ethnic prejudice in East and West Germany, tracing these disparities back to variations in contact with immigrants.
Multiple publications have examined the East-West divide in AfD voting but few have investigated whether these differences go beyond the general associations described above (Goerres et al. 2018: 250-51): Stroppe and Jungmann (2022) differentiate between compositional and contextual effects which is similar to the argument presented in this paper. They find an East bonus when controlling for several context factors, but their analysis is more focused on meso-level contextual variables such as regional median income. Arzheimer (2016) develops an argument for both the AfD and the Left Party on the individual level, which is analyzed more thoroughly in his 2021 paper (Arzheimer 2021). Using the GLES 2017 data, he finds no remaining East bonus for the AfD when controlling for satisfaction with democracy, the respondents' position on immigration, taxation and welfare state, and control variables such as age, gender, and education. Goerres et al. (2018) examine the supporter base of the AfD in 2016, after the reorientation of the party but before the 2017 federal election. Using the likelihood of voting for the AfD as the dependent variable, they also find no remaining effect of adolescence in the former GDR or East Germany (compared to adolescence in West Germany) (Goerres et al. 2018: 257). However, their paper only briefly discusses the causes for regional differences and suggests that the distribution of voting determinants may play an important role in this matter (Goerres et al. 2018: 250-51).
Finally, it is worth mentioning Pesthy et al.'s (2021) work. They present a comprehensive argument for the socialization of East and West Germans with regards to populist and nativist attitudes.
By comparing different generations, they discover the importance of these attitudes in influencing AfD voting in East Germany (Pesthy et al. 2021: 85). Their theoretical argument and operationalization serve as an essential foundation for this paper.
However, the state of research is still lacking a structured theoretical argument. While some scholars have briefly discussed the possible reasons for a divide between East and West German voting patterns (Arzheimer 2016: 71-73; Richter/Bösch 2017: 15), only a limited number of studies have provided a comprehensive theoretical argument for this phenomenon (see e.g. Pesthy et al. 2021;Weisskircher 2020). This constitutes the first research gap. Furthermore, there is only a limited number of analyses of the 2021 data, which creates the second research gap. Therefore, this paper fills these two research gaps by developing an argument that can analyze a possible 'East bonus' for the 2021 federal election and provide explanations for this phenomenon based on Pesthy et al.'s (2021) argument. Additionally, this paper replicates parts of their analysis and incorporates their work into a broader theoretical frame and analysis.
Given that Pesthy et al.'s (2021) argument focuses on different cultural attitudes, cultural backlash as a theoretical frame for radical right voting is not analyzed in this paper due to the overlap.
Thus, the paper primarily focuses on economic reasons and certain socio-demographic variables to minimize potential regional influences in East Germany on voting behavior.

Theoretical Framework
The question of how to explain voting differences in East and West Germany can be approached from various theoretical perspectives that are explained in the following sections.

Different Conditions but Same Assumptions in East Germany
First, different factors, such as higher unemployment rates (Lux 2018: 263), contribute to voting for right-wing populist parties. The core of this first theoretical argument is that these conditions tend to occur more frequently in East Germany than in its Western counterpart, leading to differ- To explain the rise of right-wing populist parties on the individual level, scholars often refer to modernization developments over the last three decades that created so-called winners and "los- An important aspect of this approach is the subjective perception of one's own economic status.
Individuals compare themselves to others and potentially perceive them as being economically better off which leads to negative emotions among modernization losers which in turn influences their voting behavior (Lengfeld 2017: 213; Richter/Bösch 2017: 9). Right-wing populist parties often promise "to take back control" (Schwander/Manow 2017: 5) by restricting immigration policies, which appeals to these voters (Richter/Bösch 2017: 9).
When trying to explain the association of East Germany and the AfD's success, it can be argued that the economic conditions in East Germany, resulting from decades of planned economy in the GDR and the region's integration into West Germany after reunification, contribute to the AfD's election success (Franz et al. 2018: 137;Träger 2017: 332;Weisskircher 2020: 617-18). To test this argument empirically, the first hypothesis can be proposed: H1: When controlling for the individual objective and subjective economic situation, the regional differences in AfD-voting decrease.
Another factor that may explain regional differences in voting behavior is the distinct composi- H2: When controlling for the distribution of religious denomination, the regional differences in AfD-voting is reduced.
In addition to religion, East and West Germany also differ in their age and gender composition.
Nevertheless, these differences are not distinct enough for a proper theoretical argument including hypotheses (see methods and data). Therefore, they are added as control variables to ensure that the central argument about the unique conditions in East Germany is properly tested.

Different, Unique Experiences of East Germans
A contrary approach to the one above is the assumption that East Germans have unique attitudes that predispose them to vote for the AfD. These attitudes could arise from the distinct historical experiences of East Germans whereas the theoretical approaches above produce specific attitudes in both parts of Germany (Stroppe/Jungmann 2022: 50-51). This section explores experiences before 1990. Two hypotheses are proposed based on the assumption that the differences in attitudes are not captured by the typical voting determinants discussed in the previous hypotheses.
As explained above, the AfD combines both populist and nativist aspects ( Simultaneously, the socialist culture of the GDR did not promote the concept of representation or pluralistic views of society, which stands in contrast to populism (Mudde 2004: 543-44;Pesthy et al. 2021: 75). As a result, the socialization of East Germans in the former GDR may have led to the development of populist attitudes before reunification. Moreover, during the 1989 revolution against the SED regime, East Germans' contrasting belief system of 'us-versus-them' intensified during the protests against the political elite. The slogan of the peaceful revolution, "Wir sind das Volk" ("We are the people"), even strengthens the argument for an increase in ingroup-outgroup contrasting and anti-elitism among East Germans at that time (Pesthy et al. 2021: 75). These populist thought patterns may still shape the attitudes of older generations of East Germans today (Pesthy et al. 2021: 75). Hypothesis 3 tests if populist attitudes play a significant role in the regional differences: H3: When controlling for populist attitudes, the regional differences decrease.
Nonetheless, populism as a thin ideology can also be combined with different thin or thick ideologies, such as nativism, as can be seen in right-wing populist parties like the AfD and its voters Even though the GDR government attempted to promote anti-fascist attitudes, it did not establish a constitutional identity or incorporate immigrants under one constitution as was done in West Germany (Pesthy et al. 2021: 75). In fact, immigration communities in the GDR were rare and isolated (Lemke 1991: 35; Weisskircher 2020: 619). Thus, the perception of the constitution as a common basis was comparatively unnecessary and nativist perceptions of identity, such as ingroupoutgroup thinking based on one's descent, had space to develop (Pesthy et al. 2021: 75). These may still persist among older East Germans today. Therefore, this paper tests the following fourth hypothesis empirically: H4: When controlling for nativist attitudes, the differences between East and West Germany in AfD-voting decrease.

Methods and Data
This paper examines the East-West divide in voting behavior for the AfD in the 2021 German federal election at the individual level, as aggregate-level results may lead to erroneous conclusions (Richter/Bösch 2017: 8). To achieve this, the pre-and post-election German Longitudinal Election Study (GLES) cross-section version, which compromises both online and paper-pencil interviews, is analyzed, as it is frequently used in this research field, thereby simplifying comparisons between analyses. Arzheimer (2021: 70) used the same data for 2017, while Pesthy et al (2021: 77) employed the 2017 GLES online campaign panel survey, which covers the year before the election and includes the week after the election. The GLES data contains an oversampling for East Germans that was corrected for this analysis to create some comparability with Pesthy et al. (2021: 77).
The first aim of the paper is to explain the regional influence on AfD-voting. To achieve this, various variables, operationalizing the proposed theoretical approaches, are nested in a logistic regression and average marginal effects (AME) are used. AMEs are preferred over other measures such as odds ratios since they provide a more straightforward interpretation, representing the change in the probability of the dependent variable when the independent variable increases by one unit (Best/Wolf 2014: 163).
The research question is addressed by examining the remaining (non-)significant statistical effect of the variable 'East' on the dependent variable, which is the respondent's voting decision for the AfD, coded as a dummy variable. As the dataset compromises pre-and post-election data, this variable combines voting intention and actual voting behavior in person or by mail.
The first independent variable 1 , 'east', is a dummy variable that assigns respondents to either East or West Germany, including an exact allocation of West-and East-Berlin. However, this coding may not accurately capture the experiences of all respondents as not all individuals currently living in East Germany necessarily grew up there and experienced the GDR. Although a variable that asks for the year of the respondent's move to East Germany exists in the GLES data, a more precise coding of the 'east' variable is not possible due to a substantial number of missing responses. Since only 4% of respondents moved to East Germany, it can be assumed that all respondents currently living in East Germany spent their entire lives there.
To test the first hypothesis regarding the losers of modernization theory, this paper divides the variables for this hypothesis into two groups: The first group includes objective socio-economic variables, such as household net income, education (indicated by the respondent's highest school degree) -both adopted from Lengfeld (2017: 219-220) -current unemployment, and the respondent's number of years of unemployment within the last ten years. Respondents who are currently unemployed can be considered as losers of modernization (Spier 2010: 90). Additionally, unemployment in the last ten years provides broader information on the topic of unemployment and its relationship with AfD-voting (Manow 2019: 91-94).
However, the objective economic status does not encompass the whole concept of loser of modernization theory (Lengfeld 2017: 213). Therefore, the second group of variables, which includes the evaluation of one's own economic situation, will be added to the regression. This comprises the question of class affiliation, the judgement of the current personal economic situation, the projection of the personal economic situation in one year, and the evaluation of the development of the personal economic situation in the last two years (GLES 2022a: 82-84).
To test the second hypothesis, religious denomination is added to the regression as a dummy variable. Gender and age also show different distribution patterns for both regions (see e.g. Geis/Orth 2017b) but the surplus of men in East Germany mostly applies to the group of middleaged respondents (Geis/Orth 2017a: 5). On the other hand, different age groups show divergent voting behavior regarding the AfD (Bieber et al. 2018: 449;Goerres et al. 2018: 259;Wurthmann et al. 2021: 876). Accordingly, both variables cannot be sufficiently associated with the east impact to formulate clear hypotheses. Adding them as control variables allows me to still show their relevance for a possible east effect on voting for the AfD.
The second theoretical argument with H3 and H4 is tested with descriptive analyses similar to Pesthy et al. (2021) to provide initial information on the populist and nativist attitudes of East and West Germans. However, the graphs alone cannot guarantee significant differences between both regions. To determine if both groups differ significantly in their means, weighted t-tests for independent subgroups and variance equality tests are conducted (refer to Table 4 in the appendix).
The results are interpreted as statistically significant when the p-value is below 0.05. Hence, respondents born before or in 1973 were assigned to the older age group whereas everybody born after 1973 was coded as young.
In a second step, the nativist and populist variables are added to the regression to test if they explain the regional differences. The operationalization of Pesthy et al. (2021: 77) is adopted for both variables, but with slight modifications due to limited data. For populism, the GLES 2021 survey has fewer questions than implemented in Pesthy et al. (GLES 2022a: 113;Pesthy et al. 2021: appendix). However, there are similarities between the items, and an additive index 2 is created with questions that cover the anti-elitism and sovereignty dimensions of populism 3 (see Table 1).
The index ranges from 1 to 30 and is coded into five same-sized categories for figures to provide

First Theoretical Argument: The Regional Distribution of Common Voting Determinants
To investigate whether the explanation for the East impact of living in East Germany on voting for the AfD can be attributed to the distribution of various conditions that facilitate right-wing populist voting, a regression model shows the size of the East-West divide. Model 1 in Table 2 reveals a substantial regional impact on voting for the AfD, as living in East Germany significantly increases the average probability of voting for the party by 6.9%. Thus, a considerable significant East-West divide exists which could potentially be reduced by controlling for objective socio-economic variables in the subsequent model to test the first hypothesis.
Interestingly, the regional AME on the likelihood of voting for the AfD only diminishes slightly to 6.6% and remains statistically significant (see Model 2). The disparities in economic circumstances between East and West Germany cannot solely explain the varying election outcomes for the AfD.
However, an objective perspective of the losers of modernization theory does not capture the full theory. Hence, the third model incorporates the second set of variables that pertain to the respondents' subjective assessments.
In Model 3, the average regional effect on AfD-voting decreases further, and the size of the coefficient is reduced in a more noticeable way. When controlling for both objective and subjective variables, living in East Germany raises the average probability of voting for the AfD by 5.8% compared to living in West Germany. This finding supports the first hypothesis, which suggests that worse objective and subjective economic conditions contribute to the East-West divide in AfD-voting. Nevertheless, the distribution of economic conditions cannot be seen as the driving factor behind the regional differences since the East impact remains substantial.  (GLES 2021b(GLES , 2022b; own calculations In order to test H2, religious denomination was added to Model 4 (see Table 2), resulting in a small decrease of the East AME on AfD-voting. The results indicate that living in East Germany, compared to West Germany, increases the probability of voting for the AfD by 5.1%. This finding supports H2 which suggests a decrease of the regional differences when controlled for religious denomination.
The other two socio-demographic variables, age and gender, were included as control variables in Model 5. However, this caused a slight increase in the East AME as the real regional effect on AfD-voting seems to be masked by the distribution of age and gender. When controlling for socioeconomic and socio-demographic variables, living in East Germany still increases the average probability of voting for the AfD by 5.6% compared to living in the West.
To summarize the first part of the empirical analysis, the distribution of voting determinants across regions is relevant to the success of the AfD. Even after controlling for the classical voting determinants, the average probability of East Germans to vote for the AfD still remains higher than for West Germans. While this analysis did not include all possible variables conducive to right-wing populist voting, it showed an impact beyond classic voting determinants. Therefore, the second theoretical approach is based on the plausible assumption of unique socialization experiences of East Germans leading to support for the AfD.

Second Theoretical Argument: Unique Experiences of East Germans
To test the second theoretical argument and the associated hypotheses, initial descriptive analyses shed light on the attitudes of East and West Germans before a more concrete examination through multivariate analyses.

Figure 1 illustrates the distribution of populist attitudes of respondents from East and West
Germany across different generations. For this depiction, the populist index was coded into five same-sized categories to create some similarity to Pesthy et al. (2021: 79). A first examination of the data reveals that East Germans appear more populist than West Germans, as evidenced by the different means of 14.7 (West) and 16.08 (East) (see Table 3). As the t-test shows (Table 4, appendix), East Germans are significantly more populist than West Germans. This is similar to Pesthy et al.
(2021: 79) who also report significant differences with their 2016 and 2017 data.These regional disparities may be attributed to distinct socialization experiences in the former GDR and West Germany. If this is the case, we would expect greater differences among older respondents who were born before or in 1973 and experienced the GDR. Similar to Pesthy et al. (2021: 81), older Germans show greater differences in their average populist attitudes across both regions than their younger counterparts (see Figure 1). This supports the argument that older East Germans were socialized by the populist rhetoric of the GDR government and the populist thinking of the peaceful revolution in 1989. Nonetheless, even younger generations who grew up in Germany after reunification demonstrate significantly different levels of populist attitudes, depending on their place of residence.
Upon initial inspection, this result may seem counterintuitive: Why should East Germans who were socialized in a united Germany be different to their Western peers? The answer lies in the role of parents as socialization agents. They have a significant impact on their children since they pass many of their views to their descendants (Pesthy et al. 2021: 75). It is highly probable that opinions shaped by experiences before reunification are among these inherited views which could explain the persisting differences among young Germans. Moreover, the reunification took place in 1990 but its consequences lasted years longer. East Germans who were socialized in this time still grew up under different circumstances than their Western peers.  Table 4, appendix). Source: 2021 GLES pre-and post-election survey (GLES 2021b(GLES , 2022b; own depiction  For example, as mentioned in the theory section, the absence of immigrants in the East may have influenced their opinions on migration. The column diagrams for the first nativist variable regarding the respondents' position on immigration are shown in Figure 2. The highest value stands out due to the great difference between East and West Germans. Proportionally, twice as many East Germans as West Germans chose the highest category to express their wish for a restriction of immigration. Furthermore, East Germans generally showed more reluctance by choosing the categories 7-11 more often than West Germans. These regional differences are statistically significant (see Table 4, appendix). This finding is similar to Pesthy et al.'s (2021: 81) results for their nativism index even though this analysis reports greater differences. facilitate immigration (1) -restrict immigration (11); for specific codings of all variables see Table 3 (Appendix); Means across all generations: 6.94 (East),6.18 (West) foreigners. This certainly also contributes to the differences in nativist attitudes among younger generations. Additionally, the power of parents to influence their children's attitudes through socialization should not be underestimated, even if this paper cannot account for it.

Figure 1: Populist attitudes among East and West Germans
Turning to the second nativist variable that measures the respondents' views on the assimilation of immigrants into German culture, the regional differences are less pronounced than in the first variable (see Figure 3). Nevertheless, the difference between East and West Germans is statistically significant (see Table 4, appendix). The proportion of older East Germans who strongly agreed with this statement was 10% higher than that of their West German counterparts. Older East and West Germans differ significantly in their nativist attitudes. However, the younger generations in both regions appear more similar in their nativist attitudes which makes the argument of similar socialization experiences plausible. This is further supported by the non-significant differences in means between younger East and West Germans. Nevertheless, this creates a great disparity to Pesthy et al.'s (2021: 81) result since they find significant differences in their analysis with the nativism index. The East-West differences, though, are not significant in younger respondents' ethnic national identity, which is one of their dimensions of nativism. Therefore, the different findings may be caused by different operationalization of nativism or the age groups.

Figure 3: Nativist attitudes among East and West Germans: assimilation
-Notes: relative frequencies; the respondents were asked for their opinion on the following statement: "Immigrants should be obliged to assimilate into the German culture." (GLES 2021a(GLES : 41, 2022a strongly disagree (1) -strongly agree (5); for specific codings of all variables see  Table 4, appendix) Source: 2021 GLES pre-and post-election survey (GLES 2021b(GLES , 2022b; own depiction Overall, the analysis of the respondents' nativist attitudes indicates a significant difference between East and West Germans with East Germans showing a higher degree of nativism. The argument that those living in the GDR were socialized to become more nativist than their Western counterparts seems reasonable, especially in the light of different generations. To gain a better understanding of the impact of populist and nativist attitudes on the regionally different AfD-success, the variables are added to the regression. Table 2 tests H3 by including the original version of the populism index (not the coded version for Figure 1) to investigate if the populist socialization in the GDR decreases regional In order to test H4, Model 7 incorporates both nativist variables with the previous models, resulting in a further decrease of the regional AME on AfD-voting. When controlling for socio-economic and socio-demographic variables as well as populist and nativist attitudes, residing in East Germany raises the average probability of voting for the AfD by 3.2% as compared to living in West

Model 6 in
Germany. As a result, the fourth hypothesis, which proposes the existence of an impact of nativist socialization in the GDR, can be confirmed.
In summary, this paper provides empirical support for Pesthy et al.'s (2021) argument about the socialization experience of East Germans in the former GDR and arrives at similar findings based on data from 2021. To enable some comparison with Arzheimer's work (2021: 70), satisfaction with democracy is added in Model 8. This leads to a rather small decrease in the regional AME, with living in East Germany increasing the average probability of voting for the AfD significantly by 2.8%, compared to residing in West Germany. Despite dissatisfaction being a crucial element in the debate of East Germany (Arzheimer 2021: 69; Struck 2017), the impact of controlling for satisfaction with democracy is surprisingly low. One explanation for this finding could be the interaction between some variables. Although multicollinearity is non-existent for this regression, it is essential not to overlook the theoretical coherence of the variables. For instance, it is plausible that respondents in precarious economic conditions attribute their circumstances to the current democratic system in Germany. This should especially be the case for those who assess their current or future economic conditions negatively. Such individuals may question the democratic system when they are concerned about their financial prospects.
Despite considering all those variables, living in East Germany still increases the average probability of AfD-voting significantly by 2.8% compared to living in West Germany, which can be interpreted as an East bonus for the AfD. This contradicts Arzheimer's (2021: 78) findings, as his analysis did not reveal a remaining East bonus for the AfD, despite including satisfaction with democracy, position on taxes and welfare state, and immigration, as well as age, gender, and education (Arzheimer 2021: 70). With Model 8, position on immigration, satisfaction with democracy and all three control variables are covered in this analysis, with the exception of the position on taxation and welfare state which is not relevant for AfD-voting, according to Arzheimer's (2021: 75) findings.
Thus, even though all variables are taken into account for the 2021 analysis, a remaining east bonus is still observed.

Conclusion
This paper aimed to explore explanations of the regional divide in AfD-voting. While the research question is not completely answered, the results suggest that neither different socioeconomic and socio-demographic conditions, dissatisfaction with democracy, nor populist and nativist attitudes can explain the regional impact entirely. The AfD benefits from an East bonus independent of all control variables this paper used. However, this paper highlights the importance of the subjective economic situation, religious denomination, populist, and nativist attitudes in decreasing the statistical East effect to a certain extent. The analysis of younger generations also suggests a potential decrease in the regional divide in the future. Nevertheless, the current success of the AfD in East Germany remains a concern and cannot be attributed to commonly assumed predictors. Therefore, increasing the wealth of losers of modernization alone is unlikely to eradicate the AfD's East bonus in the next election and mainstream parties need to directly address East Germans and their unique experiences to gather votes. To develop such strategies, policymakers require more specific information about reasons for the AfD's success. Further research using interviews or survey data could help to provide these answers.
The present study has several limitations that need more discussion. First, the operationalization of control variables is limited and needs to be refined in future research. This includes addressing the overlap between distinct approaches, which is a general problem in the literature on right-wing populist voting. Even the two main approaches -economic and culturalcorrelate to some extent (Ziblatt et al. 2022: 5