Introduction
LGB+ Footnote 1 individuals have increasingly been able to free themselves from their social niche existence in recent decades, at least in many Western societies (Ayoub, Reference Ayoub2016). As a result, they have developed from an invisible electorate that could not be actively mobilized into a steadily more visible and mobilized segment among voters (Proctor, Reference Proctor2022). Nevertheless, LGB+ individuals are still a selectively visible social group and very often choose when to “unmask their sexuality which limits the ability of social scientists to study their attitudes and behaviours” (Turnbull-Dugarte, Reference Turnbull-Dugarte2020a, 518). Following this perspective, an increasing number of empirical studies have been devoted to the political behaviour, i.e. voting participation and voting behaviour, of LGB+ individuals and have demonstrated that this social group, which is still only marginally researched, might be and is capable of shaping politics (Guntermann and Beauvais, Reference Guntermann and Beauvais2022; Grahn, Reference Grahn2023; Hertzog, Reference Hertzog1996; Perrella et al., Reference Perrella, Brown and Kay2012; Turnbull-Dugarte, Reference Turnbull-Dugarte2020b; Wurthmann, Reference Wurthmann2023a). In contrast, the literature on LGB+ politicians and their political behaviour was considered “almost non-existent” (Everitt, Reference Everitt2015, 178) only a few years ago. Today, three main branches of research can be identified that are dedicated to the study of LGB+ political elites. These include the performance of LGB+ representatives in parliaments. Reynolds (Reference Reynolds2013, 259) finds, for example, “that the presence of even a small number of openly gay legislators is associated significantly with the future passage of enhanced gay rights”. Similarly, Bönisch (Reference Bönisch2022) and Haider-Markel (Reference Haider-Markel2010) find that LGB+ individuals act as mediators and advocates for the interests of sexual minorities. Worth mentioning is also the very recent research dedicated to the electoral success of LGB+ individuals and corresponding influential factors (e.g. Cravens, Reference Cravens2023; Everitt and Camp, Reference Everitt and Camp2014; Everitt and Horvath, Reference Everitt and Horvath2021; Magni and Reynolds, Reference Magni and Reynolds2018, Reference Magni and Reynolds2021). The third strand is explicitly dedicated to the public’s perception of LGB+ politicians and which indicators explain disapproval or approval (Albaugh and Baisley, Reference Albaugh and Baisley2023; Haider-Markel et al., Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017; Jones and Brewer, Reference Jones and Brewer2019).
These approaches do not, however, study whether the sexual identity of politicians affects which ideological positions or – more specifically – policy profiles on key issue dimensions they adopt. Analysis of population surveys suggests that the sexual identity of individuals influences the adoption of socio-economically left-wing and socio-culturally liberal attitudes (Turnbull-Dugarte, Reference Turnbull-Dugarte2020b). However, there are no findings supporting this effect for candidates or politicians from corresponding elite surveys reported yet. The present study aims to fill this gap and examines the effect of the sexual identity of parliamentary candidates on their policy preferences. Against the background of the finding that LGB+ candidates are perceived as more liberal than heterosexual candidates (Loepp and Redman, Reference Loepp and Redman2022; Magni and Reynolds, Reference Magni and Reynolds2021), and sexuality thus functions as a kind of cueing mechanism for voters (Magni and Reynolds, Reference Magni and Reynolds2018), this assumption could already prove to be too hasty. Preliminary findings provide evidence that parties of the radical right actively solicit votes from sexual minorities and elect members of this group to the highest party offices – instrumentally motivated to create sentiment against other minorities such as Muslims or trans* individuals (Magni and Reynolds, Reference Magni and Reynolds2023; Turnbull-Dugarte and López Ortega, Reference Turnbull-Dugarte and López Ortega2023). It is therefore all the more important to shed light on the question whether LGB+ politicians tend towards more progressive attitudes than heterosexual ones and if there is variation across parties with differing ideological profiles. Subsequently, the question arises whether a higher degree of descriptive representation, as described by Reynolds (Reference Reynolds2013), would necessarily lead to a higher degree of substantive representation (Pitkin, Reference Pitkin1967), which should be congruent with previous findings on attitudes of the LGB+ population (e.g. Turnbull-Dugarte, Reference Turnbull-Dugarte2020b).
In line with existing studies which find that MPs with a working-class background (e.g., Hemingway, Reference Hemingway2022), female legislators (e.g., Erzeel and Celis, Reference Erzeel and Celis2016) and MPs of specific ethnic origin (e.g., Ellis and Wilson, Reference Ellis and Wilson2013) represent the interests of their peer group in parliament, we expect parliamentary candidates who identify as LGB+ to represent the interests of this social group, so that LGB+ candidates for parliament should be more progressive on socio-cultural issues and more left-wing on socio-economic issues than heterosexual candidates. Empirically, we study in this contribution the determinants of policy preferences of candidates for a recent parliamentary election – the German Bundestag in 2021 – and analyse if parliamentary candidates who identify themselves as LGB+ adopt a distinct profile on key dimensions that reflect major conflict lines in modern societies.
In so doing, we take a closer look at the positions parliamentary candidates adopted on the socio-cultural dimension, thus differentiating between preferences for a pluralist position and a traditionalist position on the order of society and individual lifestyles on the one hand, and on a socio-economic dimension that distinguishes negative from positive views on fiscal conservatism on the other. We make use of the German Candidate Study 2021 (GLES, 2023) that covers a wide range of information on the attitudes, values and personal characteristics of politicians, including their self-reported sexual identity. We find that the self-identification as LGB+ results in a more progressive position on the socio-cultural dimension and a more state-interventionist position on a socio-economic dimension. Furthermore, individuals identifying as LGB+ tend to adopt less conventional stances on socio-cultural issues than what is typically associated with their political party. This suggests that having more LGB+ representatives in parliament enhances the influence of progressive perspectives and better represents the interests of the LGB+ community, even when controlling for the partisan affiliation of parliamentary candidates. To derive these findings, we review the existing literature in the following section and develop four hypotheses that guide the empirical analysis. Before presenting the results, we provide a more detailed description of the data and of the analytical strategy. The final section concludes by discussing limitations of the present study and by presenting ideas for future research.
The sexuality gap in political behaviour revisited
The voting behaviour and political attitudes of LGB+ individuals are still widely understudied as contributions focusing on this topic are still rare (Perrella et al., Reference Perrella, Brown and Kay2012; Mejdrich and Burge, Reference Mejdrich and Burge2018). Nevertheless, more and more studies address this issue and thus also deal with an old stereotype that “all the gays are liberal” (Worthen, Reference Worthen2020, 27). Indeed, LGB+ individuals do not only take liberal positions on issues that directly concern them, like, for instance, same-sex marriages or adoption rights for same-sex couples (Denise, Reference Denise2017). They also hold more progressive positions on abortion, climate and environmental policy, and the death penalty than heterosexuals do (Egan, Reference Egan2012; Worthen, Reference Worthen2020; Schnabel, Reference Schnabel2018). All in all, LGB+ individuals “are significantly more liberal in their attitudes than are heterosexuals with identical demographic backgrounds” (Hertzog, Reference Hertzog1996, 81). Differences in attitudes also manifest themselves in the fact that with their income rising, heterosexuals tend to be more hostile to redistribution efforts and are less likely to vote for left-wing parties, whereas this is less often the case for LGB+ citizens (Turnbull-Dugarte, Reference Turnbull-Dugarte2020b). Self-identifying as LGB+ is a strong predictor that shapes how individuals perceive the world (Schnabel, Reference Schnabel2018) and how they vote (Hertzog, Reference Hertzog1996). Indeed, the support of LGB+ individuals for social democratic or socially liberal parties is significantly above average as studies on voting behaviour and party preferences in the United States (Egan, Reference Egan2012; Hertzog, Reference Hertzog1996; Swank, Reference Swank2018), Canada (Perrella et al., Reference Perrella, Brown and Kay2012, Reference Perrella, Brown and Kay2019), Germany (Wurthmann, Reference Wurthmann2023a), and the UK (Turnbull-Dugarte, Reference Turnbull-Dugarte2022) indicate. This is explained by the longstanding support of LGB+ rights by these parties (Haider-Markel, Reference Haider-Markel2010; Turnbull-Dugarte, Reference Turnbull-Dugarte2020a; Wurthmann, Reference Wurthmann2023a), which reflects a congruence of parties and voters (Tremblay, Reference Tremblay and Tremblay2019). LGB+ individuals, however, turn out to be a particularly loyal group of voters: “Like women, LGBT citizens are geographically and ethnically dispersed, but unlike women, they do tend to vote for parties that are sympathetic to their group’s needs” (Reynolds, Reference Reynolds2013, 261). Interestingly, recent findings also provide evidence that bisexual and transgender individuals are about slightly less liberal than lesbians and gays (Jones, Reference Jones2021).
Haider-Markel et al. (Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017) show that openly gay and lesbian candidates are still rejected by about a third of all voters because of their sexuality. Some voters still penalize candidates belonging to a sexual or gender minority (Magni and Reynolds, Reference Magni and Reynolds2021). Nevertheless, LGB+ candidates perform as well as straight candidates on the ballots (Magni and Reynolds, Reference Magni and Reynolds2018) as long as they adopt heteronormative standards and are married and/or have children (Everitt and Horvath, Reference Everitt and Horvath2021). One explanation for this counterintuitive result is that “voters who would be inclined to vote against an LGBT candidate would probably not vote for most Democratic candidates” (Haider-Markel et al., Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017, 12), whereas gay Republicans enjoy a higher support than straight Republicans do because Democrats are “sympathetic to the sexuality cue as well as Republicans attracted to the partisan heuristic” (Loepp and Redman, Reference Loepp and Redman2022, 307). Generally speaking, gay candidates are perceived to be more liberal than their heterosexual counterparts (Magni and Reynolds, Reference Magni and Reynolds2021) and, in addition, they run more often for left-leaning or liberal parties than for conservatives (Everitt and Camp, Reference Everitt and Camp2014).
But what other forms of substantial representation can actually be expected when LGB+ candidates succeed in entering parliament remains unclear. We know that the social background and the identities of politicians originating from their gender, professional background, family structure or religious denomination can play an important role in the political process and can affect policy outputs and the final outcomes of the political process (Bailey, Reference Bailey1999; see also Baumann et al., Reference Baumann, Debus and Müller2015; Burden, Reference Burden2007). One can therefore expect that the election of LGB+ candidates into political offices leads to a greater degree of substantial representation of the interests of LGB+ individuals in the political process, so that policy outputs should stronger integrate and reflect the policy preferences of the LGB+ community (Bönisch, Reference Bönisch2022; Haider-Markel, Reference Haider-Markel2010; Magni and Reynolds, Reference Magni and Reynolds2021; Reynolds, Reference Reynolds2013).
Besides adopting more progressive positions on socio-cultural policies, we also know from existing research that LGB+ individuals tend to adopt more left-wing positions on economic issues than the heterosexual population (Turnbull-Dugarte, Reference Turnbull-Dugarte2020b). Kleiman et al. (Reference Kleiman, Spanierman and Smith2015) explain this pattern by referring to the experience of marginalization that many LGB+ individuals make, so that they are more likely to show solidarity with other, economically determined minorities. In addition, socialization and discrimination experiences of LGB+ individuals may have “altered their ideological predispositions to make them more supportive of leftist ideals, or that they are likely to be mobilized towards those who seek to cater to their aggregate policy preferences as a means of advancing their own welfare” (Turnbull-Dugarte, Reference Turnbull-Dugarte2020a, 531). For this reason, a straightforward expectation would be that candidates for public office who identify as LGB+ are also fiscally less conservative. Footnote 2
Following these existing findings on LGB+ individuals’ attitudes and policy preferences as well as their experience with discrimination in various policy domains, we argue that candidates for public office who identify themselves as LGB+ (and mention this accordingly in surveys) adopt not only policy positions similar to other LGB+ individuals. Moreover, candidates who compete for votes in elections should have an intrinsic motivation to change the status quo so that life conditions for individuals belonging to the same group should improve. One would therefore expect not only that candidates identifying as LGB+ adopt a distinct ideological profile in line with the one of their constituents, but also deviate from their party line on key policy issues that are considered to be relevant for these individuals. We know from studies on legislative behaviour that MPs with specific personal characteristics that can be related to a policy issue adopt distinct policy profiles in the process of legislative decision-making for vote-seeking reasons in general and for gaining a distinct policy profile within their party as well as among the public in particular (e.g., Burden, Reference Burden2007). For instance, religious denomination, gender, and parental status of elected politicians can matter for decision-making processes on moral policy issues, and elected politicians with a specific ethnic background are more active in legislative debates on citizen and minority rights (e.g., Bailer et al., Reference Bailer, Breunig, Giger and Wüst2022; Baumann et al., Reference Baumann, Debus and Müller2015; Dingler and Kroeber, Reference Dingler and Kroeber2023).
Drawing on these lines of reasoning, we expect not only that candidates for public office who identify themselves as LGB+ follow the ideological positions of LGB+ individuals on the socio-cultural and the socio-economic dimension, but also see themselves as more progressive and socio-economically less conservative than their respective party – because of an intrinsic motivation to change the status quo policy in favour of LGB+ individuals, in particular on diversity policy. On the basis of these considerations, we formulate the following hypotheses:
Hypothesis 1: LGB+ candidates adopt more liberal positions on the socio-cultural dimension than candidates who identify as heterosexual.
Hypothesis 2: LGB+ candidates favour a stronger welfare state than candidates who identify as heterosexual.
Hypothesis 3: Candidates identifying as LGB+ consider themselves as more liberal on the socio-cultural dimension than their respective party.
Hypothesis 4: Candidates identifying as LGB+ consider themselves as fiscally less conservative than their respective party.
Case selection and research design
We select the 2021 German Bundestag election to evaluate the hypotheses. Focusing on Germany has several advantages. First, diversity issues became – likewise in several other modern democracies – a highly salient issue in German party competition with the far-right populist Alternative for Germany (AfD) actively campaigning against anti-discrimination law proposals, whereas parties with societally progressive positions like the Alliance 90/The Greens (Greens), The Left, the Social Democrats (SPD) and the liberal Free Democrats (FDP) are in favour of policies that protect the rights of sexual minorities (Abou-Chadi et al., Reference Abou-Chadi, Breyer and Gessler2021; Jankowski et al., Reference Jankowski, Kurella, Stecker, Blätte, Bräuninger, Debus, Müller and Pickel2022; Wurthmann, Reference Wurthmann2023a). In addition, and as the descriptive results will show, the German case allows for analysing the policy profile of LGB+ candidates not only predominantly from liberal and progressive parties on socio-cultural issues, but also from parties which adopt moderate to explicit traditionalist profiles on the order of society like CDU, CSU and AfD.
Secondly, intra-party unity and cohesion among German parties is relatively high and deviating from the party line can be risky for individual politicians in terms of their chances to get (re-)nominated (Baumann et al., Reference Baumann, Debus and Klingelhöfer2017; Sieberer et al., Reference Sieberer, Thomas Saalfeld, Bergmann and Bailer2020), so that the focus on Bundestag candidates can be considered as a least likely case to find evidence for an impact of the sexual identity of candidates on their individual policy positions.
Third, a further advantage of the German case is the electoral system for the federal parliament and different processes of candidate selection that originate from the electoral system. 299 MPs were elected directly in districts, and the parties’ candidates for a direct candidacy in a district are selected in local party conventions and thus in a context where minority candidates in general, and LGB+ candidates in particular, tend to face higher hurdles to win a nomination (see, e.g., Debus and Himmelrath, Reference Debus and Himmelrath2024; Haider-Markel et al., Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017, Reference Haider-Markel, Gauding, Flores, Lewis, Miller, Taylor and Tadlock2019; Kulich et al., Reference Kulich, Ryan and Alexander Haslam2014). The remaining seats in the Bundestag are elected from party lists which cannot be changed by the voters. State-level party conventions, where the party leadership has more influence on the outcome of the candidate selection process than in decentralized meetings of the local party membership, decide on the composition of the list, and minority candidates have better chances to win list positions even in case of parties with a more traditionalist profile on socio-cultural issues since this signals diversity to the public when presenting the parties’ candidates (see, e.g., Mügge et al., Reference Mügge, van der Pas and van de Wardt2019). The empirical focus on the German Bundestag allows, therefore, to differentiate between different candidate modes and to evaluate whether the types of candidacies are related with the LGB+ background of candidates and their policy positions on the policy dimensions under study.
Fourth, the 2021 German Candidate Study (GLES, 2023) provides – besides data on the positions of candidates on key ideological dimensions, which form the dependent variable in our analysis – information on the sexual identity which the candidates for the 2021 Bundestag election identify with. All candidates of the parties that were represented in the Bundestag with parliamentary group status before the 2021 Bundestag elections were asked to participate in the study: the Christian Democratic Union (CDU), the Christian Social Union (CSU) Footnote 3 , the SPD, the AfD, the FDP, the Left and the Greens. In summary, 735 partial and full interviews were realized during the data collection process which took place from 19 October 2021 to 31 January 2022.
If we look at the percentages of the extent to which particular characteristics are over- or under-represented in the sample in direct comparison to the population of all candidates and elected MPs as a whole, it becomes obvious that Green candidates and elected Green MPs are clearly over-represented in the sample, while those of the AfD are clearly under-represented. Footnote 4 Further deviations with regard to other party groups, biological sex, age or type of candidacy are not of further concern (see Table A1 in the online Appendix).
Table 1 provides information on the number and share of candidates (and elected MPs) who identified as LGB+ in the candidate survey. In addition, Table 1 provides information on how many candidates refused to answer the sexual identity question, either by giving no response, stating explicitly not to answer a question on this topic, or by stopping the interview. Footnote 5 Taking these possibilities together, almost 11% of the candidates participating in the survey refused to answer the question on their sexual identity. This share does not vary much across parties. With the exception of the FDP candidates, of which 18% did not want to provide information on their sexual identity, between 8% and 11% of the respective parties’ candidates did not participate in answering this question. While the share of LGB+ candidates participating in the survey is highest among SPD, FDP, Greens and The Left and thus among those parties in Germany that adopt progressive positions in societal affairs in general (e.g., Jankowski et al., Reference Jankowski, Kurella, Stecker, Blätte, Bräuninger, Debus, Müller and Pickel2022) and on diversity issues in particular (Wurthmann, Reference Wurthmann2023a), the share of respondents with an LGB+ identity is clearly smaller among the candidates of the rather traditionalist CDU, CSU and the AfD. Nevertheless, there are LGB+ candidates among parties with a more traditionalist socio-cultural profile in the sample, which allows for evaluating whether LGB+ candidates are more progressive than other candidates and their party across the complete ideological party spectrum.
Table 1. Number of candidates and elected candidates covered in the analysis, by party
Source: German Candidate Study 2021 (GLES, 2023).
When turning to the share of elected LGB+ candidates, we find a very similar pattern for SPD, FDP and Green Bundestag members when comparing to their share of LGB+ candidates. Between approx. 12% and 15% of the elected candidates of these three parties who participated in the candidate survey identify themselves as LGB+, whereas the rate is significantly lower in case of the CDU (3.1%) and goes down to zero for CSU, AfD and The Left. These descriptive data already indicate that it is – besides theoretical reasons, because if LGB+ individuals are socio-culturally more progressive and fiscally less conservative, they should be more likely to run for parties to the left of the ideological centre – important to control for the party affiliation of the candidates in the multivariate analysis. Because of the missing representation of LGB+ candidates from AfD, CSU, and The Left who got elected to the parliament, we focus on the candidates only when evaluating the hypotheses. Separate regression models which are based on the elected candidates are presented in the appendix (see Table A2 in the online appendix).
Dependent variables
Our analysis aims at estimating the effect of LGB+ identity among parliamentary candidates on, first, their positions on a socio-cultural and on a socio-economic conflict dimension and, secondly, on the policy distance – individually perceived by the candidates – between the candidates and their respective party on both dimensions. In a first step, the Bundestag candidates who participated in the survey were confronted with a scale on which they were asked to position themselves. The socio-cultural dimension ranged from (1) priority of individual development opportunities to (11) priority of preserving traditions and lifestyles. In the socio-economic dimension, the conflict ranged originally between the extreme poles of (1) less taxes and contributions, even if this means fewer welfare state services, and (11) more welfare state services, even if this means more taxes and contributions. In order to have similar directions of the two scales, in which lower scores represent progressive or liberal policies and higher scores traditionalist or fiscally conservative positions, we reversed the direction of the socio-economic scale, so that lower scores represent preferences for an increasing scope of the welfare state, while higher scores indicate fiscally conservative preferences. In a second step, the respondents were asked to rank the position of their parties on these dimensions. To determine whether the candidates position themselves socio-culturally and socio-economically more to the left or to the right than their parties, we subtract the party position from the individual position. Footnote 6 Negative values then indicate that respondents position themselves to the left of the party, thus having more pluralist-progressive and fiscally less conservative positions, while positive values indicate that respondents position themselves to the right of the party’s assumed position, i.e. thinking that they are more traditionalist and fiscally more conservative than their party.
Focal independent variable
In order to find out the sexual identity of the interviewees, which is our focal explanatory variable, the survey respondents were given the opportunity to make this statement via the following selection of answer categories: (1) heterosexual, (2) homosexual, (3) bi- or pansexual, (4) asexual, (5) other orientation, namely [free field] and (6) the statement “I do not wish to comment on this topic”. We branded those respondents as individuals identifying as LGB+ who considered themselves as homosexual, bi- or pansexual or asexual as well as those respondents who entered statements in answer category 5 that could be classified as LGB+. Although distinct differences in policy preferences between lesbians, gays and bisexuals can be identified (Jones, Reference Jones2021), the number of respondents is too small for a more differentiated analysis in this case and future studies with a larger sample need to study if there are differences in the attitudes and policy positions between ‘sub-groups’ of sexual orientations and identities. Footnote 7 Accordingly, we created a new dichotomous variable ‘LGB+’ where candidates who identified as LGB+ individuals received a coding of (1), whereas all heterosexuals who were evident from choosing 1 as well as from the free-field statements from category 5 were coded with (0).
Not all members who consider themselves as LGB+ may be willing to reveal their sexual orientation. Simultaneously, we are explicitly referring to individuals who define their own sexual identity, using it as a manifestation of their self-identity construction (Hertzog, Reference Hertzog1996; Guyan, Reference Guyan2022). Given the absence of more preferable alternatives for self-disclosure, this practice continues to be widely employed (e.g., Turnbull-Dugarte, Reference Turnbull-Dugarte2022; Wurthmann, Reference Wurthmann2023a; Albaugh and Baisley, Reference Albaugh and Baisley2023). Nevertheless, there could be an additional self-selection bias, as some politicians may opt not to acknowledge their affiliation with the LGB+ community. Considering that the survey was conducted through respondents autonomously completing digital questionnaires, thus eliminating interviewer effects observed elsewhere (e.g., Kühne et al., Reference Kühne, Kroh and Richter2019), and given the stringent confidentiality measures provided to the participants regarding the information they provide on their sexual identity, this potential self-selection effect can be considered minimized but not completely eliminated due to the lack of alternative data.
Controls
Parties occupy the most diverse positions in political space, and individuals support or join parties because of a decisive overlap in the positions on key policy issues. Thus, LGB+ individuals should be more likely to support or join parties with, for instance, a liberal profile on the socio-cultural dimension. There is empirical evidence that LGB+ candidates are significantly more likely to run for parties that are more left-leaning (Everitt and Camp, Reference Everitt and Camp2014). However, it has been pointed out that members of sexual minorities are not a political monolith (Jones, Reference Jones2021; Turnbull-Dugarte, Reference Turnbull-Dugarte2022), not so much concerning political representatives, but rather to those politically represented. This is all the truer when, as in the present case of the Federal Republic of Germany, one of the leading figures of the radical right-wing AfD is a woman self-identifying as lesbian. Therefore – and because of the disproportionate representation of candidates from the left and liberal spectrum – we control for the party affiliation of the candidates in the following.
In addition, the age of candidates and the squared age (to control for a non-linear relationship) and a candidate’s sex serve as control variables in the statistical models. Furthermore, the inclusion of a differentiated measurement of the respondent’s place of residence is an essential control variable. Historically, LGB+ individuals have frequently relocated to urban areas due to persistent stigmatization in rural areas (Aldrich, Reference Aldrich2004). It is not for nothing that “the lesbian and gay rights movement has made its most dramatic political impact in the cities” (Bailey, Reference Bailey1999, 3). To date, it is evident that individuals from rural areas are less supportive of LGB+ individuals (Thompson, Reference Thompson2023). It is therefore all the more understandable that LGB+ candidates are also more likely to run in urban areas (Everitt and Camp, Reference Everitt and Camp2014; Haider-Markel, Reference Haider-Markel2010), and vote-seeking candidates should adopt more progressive positions on diversity issues, regardless of the candidates’ sexual orientation. Using rural areas and villages as a reference category, we, therefore, include ‘living in a small city’, ‘living in the suburb of a big city’ and ‘living in a big city’ in the form of dummies as further control variables. Because of the possible effects of the German mixed-member electoral system on candidate selection outcomes and the policy preferences of nominated candidates, we include a binary coded variable in the empirical models that identify those candidates who aimed at winning one of the 299 district seats in the Bundestag, which were directly elected by plurality rule. Candidates who appeared only on party lists or ‘dual candidacies’ who aimed at winning parliamentary representation by running on a party list and competing for district votes are coded (0). Footnote 8
For explaining the positions of the candidates and individually perceived distance between candidates and their party on the two ideological dimensions under study, we apply simple OLS regression models to test the set of hypotheses.
Results
Descriptive analyses
Before turning to the evaluation of the hypotheses with multivariate methods, we provide a descriptive overview on the characteristics of the dependent variables and how they vary across the sexual identity of candidates and their party affiliation. Figure 1 shows the positions that candidates of CDU, CSU, SPD, FDP, Greens, The Left and of the AfD adopted on the societal dimension (upper panel) and the economic dimension (lower panel). Furthermore, Figure 1 differentiates between the sexual identity of the candidates and thus between politicians with heterosexual orientation and those who consider themselves as LGB+. We refer to boxplots as a way to present the intra-party variation in the positions the candidates adopt. The boxplots in Figure 1 provide information on the distribution of the policy positions of the parties’ candidates on the socio-cultural and socio-economic policy dimensions, respectively. The line within the box shows the median position, the box provides information on the 25th and 75th percentile of the distribution of policy positions among the candidates of the parties.
Figure 1. Positions of candidates on the socio-cultural dimension (upper panel) and the socio-economic dimension (lower panel), differentiated by sexual orientation and party affiliation.
Source: German Candidate Study 2021 (GLES, 2023), own calculations.
As Figure 1 indicates, there is clear variation in the self-placement of candidates on both dimensions between and within parties. On average, the parties’ candidates reflect the standard patterns of party competition on both dimensions: while – in line with the analysis of spatial party competition in Germany – Greens, FDP, the Left and SPD adopt more progressive positions on socio-cultural issues, Christian Democrats and in particular candidates from CSU and the AfD take rather traditionalist positions. Yet, there is – when focusing on heterosexual candidates only – a remarkable intra-party heterogeneity on that dimension. For instance, there are some SPD candidates with heterosexual orientation who are more traditional on the socio-cultural dimension than their party colleagues and some CDU candidates with more progressive positions. However, when shifting the perspective to LGB+ candidates only, they are not only more or at least similarly progressive on that dimension. Moreover, there are only very few LGB+ candidates – all from CDU and CSU – across all seven parties, who adopt a rather traditionalist position on this policy domain. This descriptive finding speaks already in favour of hypothesis 1. We cannot observe a similar pattern for the economic policy positions of the candidates. While, as expected, the candidates of parties to the left of the centre (SPD, Greens and The Left) are more in favour of a stronger welfare state than for cutting taxes and the contrary is the case for parties of the centre and on the right of it (CDU, CSU, FDP and AfD), LGB+ candidates from the parties of the centre or right tend to adopt similar positions on the socio-economic dimension than their parties and do not differ that strongly from heterosexual politicians.
When shifting the perspective to the individually perceived distances between the self-placement of a candidate on the two dimensions under study and the position of the respective candidate attached to their party, we find that most candidates do not deviate strongly from the party line, regardless if we look at the socio-cultural or the socio-economic dimensions (see Figure 2). However, there is – according to the boxplots in Figure 2 – a slight trend that LGB+ candidates consider themselves a bit more in favour of progressive policies on the socio-cultural dimension than their party, indicated by negative scores in the upper panel of Figure 2. We can observe this pattern in particular for the candidates of SPD, Greens, FDP and – interestingly – for the radical right and populist AfD. There is a similar, albeit weaker pattern for the socio-economic dimension as the lower panel of Figure 2 indicates. Whether these descriptive patterns remain stable when controlling for several other important factors that have an impact on the candidates’ positions is evaluated in the following subsection.
Figure 2. Distances between candidates and their parties on the socio-cultural and the socio-economic dimensions, differentiated by sexual orientation and party affiliation.
Source: German Candidate Study 2021 (GLES, 2023). Comment: Negative scores indicate that candidates consider themselves more in favour of progressive policies on the socio-cultural dimension, while positive scores indicate that candidates take more traditional positions than their party. In the socio-economic policy dimension, negative scores indicate that candidates favour a stronger welfare state more than their party, while positive scores signal that candidates are more in favour of fiscal conservative economic policies than their party.
Multivariate analyses
The results of the OLS regression models in Table 2 below allow for evaluating hypotheses 1 and 2. The dependent variable is the self-reported position of a candidate on the socio-cultural dimension. In model 1, low scores represent a progressive position and high scores reflect a traditional position, while model 2 refers to the socio-economic dimension that distinguishes between preferences for an increasing government spending (low scores) and fiscally conservative positions (high scores). Our main variable of interest is the self-reported sexual identity of a survey respondent, that is, whether the candidate identifies as heterosexual or as LGB+. Model 1 indicates support for our first hypothesis. Even when controlling for important further personal characteristics, contextual features like the area of living and the party affiliation, and the mode of candidacy, we find that LGB+ candidates for the 2021 Bundestag election are significantly more progressive on the socio-cultural dimension.
Table 2. Determinants of the policy positions of candidates in the 2021 German Bundestag election
Note: The dependent variable is the position of a respondent on a social-cultural policy dimension (model 1) and on a socio-economic policy dimension (model 2). SPD candidates form the reference category. Numbers in parentheses are standard errors. Significance levels: + = p < 0.1; * = p < 0.05; ** = p < 0.01.
It is worthwhile mentioning that the effect of the variable covering information on the sexual identity of candidates remains stable and significant when we control for the party affiliation of the candidates. The variables covering information on the party affiliation of candidates take the theoretical perspective of substantial representation into account: individuals are more likely to join and run for a political party that shares their policy views, represents the candidates’ interests and the interests of the social group the respective candidates aim to represent. Candidates who consider themselves as LGB+ (and want to stand for the interests of LGB+ individuals) should be more likely to join and run for parties with more progressive or pluralist positions on socio-cultural issues. Indeed, candidates adopt more traditionalist positions if they were nominated by the CDU and in particular by CSU and AfD (and thus by parties with a conservative profile on the socio-cultural policy dimension), when compared to candidates of the SPD, which form the reference category in the regression models. By contrast, candidates of the Greens and the FDP are more progressive than SPD candidates. Candidates of the socialist Left do not differ significantly from representatives of the Social Democratic party.
The fact that the variable identifying LGB+ candidates helps to explain the positions of candidates on the socio-cultural dimension – despite controlling for the theoretically important party affiliation of candidates – suggests that an increasing share of LGB+ candidates would help to shape the status quo in socio-cultural policies in a more progressive direction. Indeed, when restricting the analysis to those candidates who won a seat in the Bundestag in 2021, they are significantly more progressive and thus in favour of pluralist forms of individuals’ way of life if they identify as LGB+, regardless of their party affiliation and further control variables (see Table A2 in the online appendix).
Model 2 in Table 2 tests the second hypothesis. We find – as expected – that candidates with LGB+ identity are fiscally less conservative than heterosexual candidates, even when controlling for the party membership and further structural characteristics of candidates. Again, and as expected, the party affiliation of candidates plays an important role and explains the position of candidates, with Christian democratic, liberal FDP and far-right AfD representatives being more in favour of fiscal conservatism than SPD candidates, while candidates of the socialist Left prefer a strong welfare state to a higher degree than Social Democrats.
Figure 3 shows the estimated position of the candidates on the socio-cultural dimension and on the socio-economic policy dimension. The estimates are based on models 1 and 2 from Table 2. Figure 3 demonstrates that LGB+ candidates are slightly but to a statistically significant degree more progressive than heterosexual candidates and tend to be more in favour of a stronger welfare state than candidates who consider themselves heterosexual. This substantive effect for both dimensions under study indicates that the sexual orientation of parliamentary candidates – in addition to the party affiliation of the respective politicians – matters for the policy positions candidates adopt on two key policy dimensions.
Figure 3. Estimated policy positions of candidates on the socio-cultural and the socio-economic policy dimensions, differentiated between heterosexual and LGB+ identity of candidates.
Comment: The predicted positions are based on models 1 and 2 from Table 2. Bars indicate 90% confidence intervals.
In a second step, we evaluate hypotheses 3 and 4 and thus the expectation that candidates who identify as LGB+ consider themselves more progressive on socio-cultural affairs and fiscally less conservative on the socio-economic policy dimension. By referring to studies on substantive representation and the literature on legislative behaviour, we argued that politicians with specific personal characteristics that can be related to a policy issue develop distinct ideological profiles for vote-seeking behaviour in general and for gaining a distinct profile within their party and among the public in particular. The results of OLS regression analyses presented in Table 3 provide evidence that LGB+ candidates consider themselves significantly more progressive on the socio-cultural dimension than their respective party (see model 1 in Table 3). There is, by contrast, no such effect when shifting the perspective to the socio-economic dimension. An LGB+ background does not imply that a candidate of parties represented in the Bundestag elected in 2021 considers themself fiscally less conservative than the party that nominated the respective candidate. Having a closer look on the party affiliation of the candidates reveals also some interesting results. The results presented in Table 2 indicate that candidates of the CDU and AfD assume a more progressive orientation on socio-cultural issues than their respective party, whereas Green party candidates and those of the socialist Left show a significantly less progressive position on that dimension than the party which nominated them. When shifting the perspective to the socio-economic dimension, all candidates adopt – with the exception of the ones of the Green party – more fiscally conservative positions than their respective parties when compared to SPD candidates who form the reference group in the regression models.
Table 3. Determinants of the perceived policy distance between the positions of candidates and their parties in the 2021 German Bundestag election
Note: The dependent variable is the distance between the position of a candidate and her party on a socio-cultural (model 1) and a socio-economic policy dimension (model 2). SPD candidates form the reference category. Numbers in parentheses are standard errors. Significance levels: + = p < 0.1; * = p < 0.05; ** = p < 0.01.
Figure 4 shows the substantive effect of the variable identifying candidates with LGB+ identity. While candidates with a heterosexual identity consider themselves as neither more progressive nor more traditionalist than their parties, LGB+ candidates position themselves as significantly more progressive than the party that nominated them. There is, by contrast, no similar effect when shifting the perspective to the socio-economic dimension and the individually perceived distance between a candidate and their party. Thus, there is only robust empirical evidence for hypothesis 3, but not for our fourth hypothesis: LGB+ candidates consider themselves more progressive than their party, but not fiscally less conservative than the party that nominated them. Likewise with the results of the analyses of the candidates’ positions on both dimensions, the findings on the individually perceived distance between the candidates and their parties indicate again that an increasing representation of candidates who consider themselves as LGB+ is likely to produce stronger shifts of the status quo in socio-cultural issues.
Figure 4. Estimated difference between the position of candidates and their parties on a socio-cultural and a socio-economic dimension, differentiated between heterosexual and LGB+ identity.
Comment: The predicted positions are based on models 1 and 2 from Table 3. Bars indicate 90% confidence intervals. Negative scores on the y-axis indicate that candidates consider themselves as more progressive/more in favour of increasing welfare benefits than their party, while positive scores indicate a more traditional/fiscally conservative position than the respective party of the candidates.
To check the robustness of these findings and if the estimated effects are of random nature, we differentiate between heterosexual respondents and those candidates who rejected to answer the question on their sexual identity or by stopping the interview. Footnote 9 The results show no significant effects, that is, candidates who refused to answer the question on their sexual identity do not differ from heterosexual candidates neither in their positions on the socio-cultural and the socio-economic dimensions nor in the distance towards their party on both dimensions (see Tables A4 and A5 in the online appendix).
Conclusion
We asked in this contribution whether representatives of political parties who consider themselves as LGB+ are socio-culturally and socio-economically more progressive and more in favour of a welfare state expansion than heterosexual politicians. We evaluated our hypotheses by using the candidate study of the 2021 German Longitudinal Election Study (GLES, 2023). We found that the variation in the candidates’ positions on the socio-cultural dimension and on the socio-economic dimension can be explained by information on the self-reported sexual identity of candidates, even when we control for important further explanatory variables like, for instance, the party affiliation of candidates, and, thus, the respective parties’ ideological orientation. In addition, the results of the analyses suggest that LGB+ candidates adopt significantly more progressive positions than their own parties on socio-cultural issues such as diversity policy, according to the self-placement of candidates and the placements of the parties by the candidates in the survey.
The findings of this contribution thus indicate that (increasing) representation of LGB+ individuals in parliaments and legislatures should make it more likely that more progressive positions on societal issues can be expressed in the parliamentary process. This is not only because of the descriptive presence of this social group among parliamentary candidates (and those who get elected), but also in terms of substantive representation since LGB+ candidates consider themselves significantly more progressive on socio-cultural issues than their parties. These patterns might also increase the likelihood that LGB+ friendly policies get implemented in the legislative process once more candidates with a LGB+ identity get elected, so that the status quo on socio-cultural issues in general, and on diversity policy in particular, is likely to change substantially in a progressive direction. Of course, future research should test the implications of an increasing presence of LGB+ politicians in parliament on policy outputs and policy outcomes in more detail. For instance, one would expect that laws and law proposals increasingly consider the interests of LGB+ people on issues related to, for instance, tax or family policy not only when societally progressive parties are more strongly represented in parliament, but also when politicians who identify themselves as LGB+ are more often represented in parliament compared to current or previous legislative periods. Furthermore, and with reference to recent studies on the effects of the parliamentary presence of women on interruptions in debates in the US Congress (Miller and Sutherland, Reference Miller and Sutherland2023), it would also be useful to investigate whether a more visible LGB+ presence in parliament also changes the content and style of parliamentary debates, and if so, in which form and which direction.
The findings presented here also point into the direction that the prejudice that “all the gays are liberal” (Worthen, Reference Worthen2020, 27) cannot be confirmed, but that such a basic tendency is just as true at the level of political elites as at the level of the population. Future studies must therefore focus more on systematic differences with regard to more specific issue preferences. We have, yet, to consider that the results presented here are based on a small subset of all candidates for only one parliamentary election at one point in time. Similarly, it was not possible to differentiate between different sexual identities. It would be desirable to consider this in future studies. However, this requires sufficient and more fine-grained data, which was not available in the present case. Future research should try to collect information on the sexual identity and further personal characteristics of candidates and elected politicians by means of sophisticated surveys in a comparative research design, so that, for instance, one could study processes of transnational LGB+ mobilization on the elite level (Ayoub, Reference Ayoub2013). For example, an increasing share of LGB+ legislators in a political system with a lower degree of party discipline than in the German Bundestag – for instance, in presidential systems where the government does not depend on a majority in the legislature – should result – depending on the societal policy preferences of the members of the executive – in more policies that favour and strengthen the rights of sexual minorities.
Moreover, in the current situation, unresolved inquiries warrant additional investigation. Owing to the limited sample size, it was not possible to explore age-related impacts on candidates. Over time, cohort effects may surface, suggesting that younger candidates exhibit less progressive thinking than their older counterparts, as the latter were raised in environments more heavily influenced by discrimination. Another issue left unanswered for future exploration is whether the finding by Haider-Markel (Reference Haider-Markel2010) that a growing number of LGB+ legislators correlates with an increase in bills favouring sexual minorities and a decrease in hostile bills can be replicated when not only taking the sexual identity of MPs into account, but also their policy preferences on a dimension that covers diversity policy.
Furthermore, the currently available data does not permit additional distinctions concerning attitudes within the LGB+ parliamentary candidates. However, such a differentiation would be advantageous, particularly considering potentially different interests and attitudes among the various subgroups behind the LGB+ acronym (e.g., Jones Reference Jones2021; Hertzog Reference Hertzog1996). As highlighted by Murib (Reference Murib, Brettschneider, Burgess and Keating2017), there is a pressing need for this differentiation, as it could shed light on disparities in treatment within the LGB+ community. This is especially pertinent in the context of the evolving societal relevance of the treatment of trans* individuals (Haider-Markel et al., Reference Haider-Markel, Miller, Flores, Lewis, Tadlock and Taylor2017; Jones and Brewer, Reference Jones and Brewer2019; Magni and Reynolds, Reference Magni and Reynolds2021; Wurthmann Reference Wurthmann2023b). Consequently, there is untapped potential for further research in the analysis of LGB+ political elites and their impact on the political process and its outcomes.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1755773924000018.
Acknowledgments
Authors are listed in alphabetical order. Both authors contributed equally to all work. We thank the anonymous reviewers and the European Political Science Review (EPSR) editors for valuable comments and suggestions. Support for this research was provided by the Deutsche Forschungsgemeinschaft/German Research Foundation (DE1667/4-3).
Competing interests
All authors declare that they have no conflicts of interest or competing interests.
Open access
