Political parties, issue salience, and the appointment of women cabinet members

Abstract

Do parties relegate female ministers to portfolios that are politically less important for them? This research note contributes to this debate and examines whether the issue salience of parties for specific policy areas has an effect on the nomination of a female minister. Previous theoretical work assumes that party leaders will be more likely to select men for those portfolios that are highly salient for the party. To test this assumption empirically, the paper analyzes the appointment of women cabinet members in the German states between 2006 and 2021. Notably, the findings contradict the theoretical expectations as well as previous empirical results from a cross-national study: On the German sub-national level the nomination of a female minister is more likely if the respective portfolio is highly salient for the governing party. Parties and their policy-preferences seem to be an important factor in explaining the share of women in sub-national cabinets.

Introduction

While the share of women in the executive branch of government has steadily increased over the last decades and the recent formation of ‘parity cabinets’ in Canada (Trudeau) and Spain (Sanchez) has attracted much attention in the media, this positive trend masks an important problem: The selection of women to cabinet positions is not gender neutral and, consequently, a higher number of female ministers does not reveal anything about the types of portfolios that women receive (Baumann et al., 2019; Goddard, 2019, 2021). Previous research has, for example, shown that women are disproportionally relegated to less prestigious ministries as well as to portfolios responsible for ‘soft’ or ‘feminine’ policy areas like education, culture or women’s affairs (Escobar-Lemmon and Taylor-Robinson, 2005; Krook and O’Brien, 2012; Kroeber and Hüffelmann, 2022). Thus, despite the growing number of female ministers, women are mostly excluded from the most powerful and prominent cabinet posts within the ‘inner cabinet’ (Davis, 1997) that would provide them with more influence on important policy decisions of the government as well as with more resources and public visibility.¹ However, this pattern has started to change and, while still not being the norm, women’s appointments to prestigious portfolios have increased substantially over the last three decades. According to Barnes and Taylor-Robinson (2018), as of 2016, 131 women received appointments to the finance ministry and almost 200 women held the foreign affairs portfolio. Barnes and O’Brien (2018) find that women even gain increasing access to the defense ministry, which in the past has been characterized as one of the most ‘masculine’ cabinet positions.

The overall prestige of a ministry, however, is only one factor in the gendered selection of cabinet members. By mainly focusing on system-level explanations, previous studies tend to neglect one of the most important actors in the appointment of ministers: political parties. Despite the positive effect of left-wing parties, we still know very little about the influence of party-level factors on the recruitment of women to cabinet (Annesley, 2015; Goddard, 2021).

Therefore, this study moves beyond the influence of the general prestige of a ministry and analyzes whether the importance that individual parties attach to specific ministries affects their likelihood to appoint a female minister. Previous research has shown that some portfolios are particularly salient for certain parties because these specific policy areas are highly important for them and their voters (Bräuninger et al., 2020). For example, the environmental portfolio is highly salient for Green Parties, even though it is generally not considered as one of the most prestigious or powerful ministries. If we want to assess the opportunities of women to influence public policy, it is therefore not enough to solely look at their mere presence in cabinet or at the overall prestige of the portfolios they receive. Beyond that, it is important to analyze whether women are also appointed to those portfolios that are particularly salient for their respective parties. Thus, taking these party-specific factors into consideration provides an important extension to previous studies.

A recent study by Goddard (2019) provides a first step in addressing this gap. She analyzes the appointment of women cabinet members across 29 European states between the late 1980s and 2014. On a theoretical level, Goddard expects that women are often excluded from close party networks that form the main pool of potential candidates for the most powerful and important political positions. Thus, party leaders will be more likely to select men for those portfolios that are highly salient for the party. In line with these expectations, her results show that women are significantly less likely to head a ministry if the associated policy area is salient to the party.

This research note contributes to this debate and analyzes the appointment of women cabinet members on the sub-national level in the German states between 2006 and 2021. Quite surprisingly, the findings contradict the theoretical expectations as well as Goddard’s (2019) empirical results: The analysis shows that the nomination of a female minister is more likely if the respective portfolio is highly salient for the governing party. Thus, the value-added of this note is twofold. First, the explicit analysis of parties and issue salience makes an important contribution to the growing body of literature on the political and institutional factors affecting the appointment of women cabinet ministers. Second, the analysis of the sub-national level in Germany expands the geographical scope to the sub-national level and provides alternative and unexpected empirical findings. The results present a novel perspective on existing research, add an interesting contribution to the scholarly debate about women in cabinets, and hopefully encourage further studies and debate about the places and conditions affecting women’s appointments to highly-salient cabinet posts.

How issue salience affects the appointment of women cabinet members

New institutionalist approaches assume that processes within distinct political arenas are structured by a set of formal rules, informal rules, and ideational norms (e.g. Hall and Taylor, 1996). More recently, gender scholars have developed a feminist variant of institutionalism which highlights how these ‘rules of the game’ can affect men and women differently and, as a consequence, how power in political systems can be unevenly distributed between the two sexes (Krook and Mackay, 2011; Chappell and Waylen, 2013). This also applies to the selection and appointment of ministers. Since the selection of ministers typically takes place behind closed doors, is made by a small set of the party elite and often lacks formal rules or criteria about who qualifies as a potential minister, feminist institutionalism has pointed to the importance of informal practices and selection criteria.

Previous research highlights that these informal rules and practices are likely to benefit the advancement of male potential ministers (Annesley, 2015; Bjarnegård and Kenny, 2016; Verge and Claveria, 2016): To minimize the risk of adverse selection and agency loss of future ministers, party elites tend to select candidates that are members of their close networks and who have demonstrated trustworthiness and loyalty to the party in the past (Bjarnegård and Kenny, 2016; Annesley et al., 2019). Due to their historical exclusion from politics and parties, women, however, only have limited access to these networks and are often seen as ‘outsiders’ by the party elites that heavily influence the recruitment process (Verge and Claveria, 2016).

To support this point, Bjarnegård and Kenny (2016: 385) point out that ‘trust is often homosocial, in the sense that we often perceive that we can trust people of the same sex as ourselves. We tend to trust people whom we perceive we can predict, and prediction is easier when we think we see ourselves in other people’. This ‘homosocial reproduction’ (Bjarnegård 2013) implies that elites tend to reproduce themselves and that (mostly) male party elites are more likely to nominate other men for a ministerial post (Cheng and Tavits, 2011). These informal networks of trust become particularly important for parties when appointing ministers to their highly salient portfolios. Because these offices are so important for the parties, they want to minimize the risk of an adverse selection and will therefore allocate these posts to a trusted member of the inner circle of the party. Due to the limited access of women to these close-knit networks of the party, they will be less likely to be appointed as minister for a highly salient ministry and, rather, be relegated to portfolios that are of less importance to the party (Goddard, 2019).

These assumptions are supported by the psychological literature on gender and leadership. According to Eagly and Karau’s (2002) ‘role congruity theory’, stereotypes and informal rules about the appropriate and acceptable gendered behavior produce distinct roles for male and female politicians which then prescribe how women and men should behave in specific situations. Whereas men are typically seen as dominant, controlling and are expected to be ‘prone to act as a leader’, women are stereotypically characterized as helpful, nurturing and sensitive (Eagly and Karau, 2002: 574). Therefore, role congruity theory predicts that women in leadership positions will be confronted with prejudice and hostility due to the dissimilarities between the typical expectations that people have about women and their ideal type of a leader (Baumann et al., 2019). Thus, parties are more likely to choose men than women as the ‘appropriate’ leaders of their most salient ministries and relegate women to those policy fields that are less important for the party.

Hypothesis: If the salience for a portfolio increases, the likelihood that parties appoint a female minister decreases.

Methods and data: new evidence on the appointment of women to salient ministries

The empirical part of the paper analyzes the appointment of women to sub-national cabinets in the German federal states between 2006 and 2021.² In comparison to Goddard’s (2019) cross-national study, analyzing female ministers on the sub-national level is a promising approach for at least two reasons. First, due to a high similarity in their political systems and institutional structures, the German states largely correspond to a most-similar systems design (see Raabe and Linhart, 2014).³ Thus, we can analyze the effect of issue salience while keeping the biasing influence of many potential omitted variables to a minimum. Second, the study adds new and interesting empirical evidence to the growing literature on the appointment of female ministers which has, until now, only very rarely been studied on the sub-national level (see Höhmann, 2017; Barnes et al., 2019).

The unit of analysis is an individual ministry in a single legislative period. The dependent variable of the analysis is the sex of the cabinet minister and takes on the value 1 if the minister is a woman, and 0 in case of a man.⁴ The coding relies on the minister’s first name and, in ambiguous cases, was verified through a background and image search of the respective person. Information on the composition of the cabinets could be obtained from the official websites of the state governments or upon request from the state chancelleries.

The primary independent variable of this study is issue salience. It is measured using quantitative content analyses of parties’ election manifestos and captures the extent to which governing parties emphasize specific policy areas in their manifestos. Data for this variable come from Raabe and Linhart (2014) and have been updated to include state cabinets until 2021. To capture issue salience, different sections of the party manifestos are manually assigned to one of ten policy areas. Because the ten policy areas used by Raabe and Linhart (2014) do not perfectly map onto the jurisdictions of the state ministries, saliences attached to the different areas are weighted by the substantive composition of the ministries. Hence, for each observation in the final data set, the variable issue salience indicates the percentage of the party platform that the governing party in control of the ministry relates to the respective portfolio’s jurisdiction.

The analysis controls for potential confounders that affect both the policy salience of ministries and the proclivity to appoint female ministers. First, I control for the general prestige of a ministry because previous studies have shown that women are more likely to be appointed to less prestigious ministries (see section 2) and that, at the same time, prestigious ministries can positively affect the salience of parties for these portfolios. Data for this variable are derived from Raabe and Linhart (2014) who conducted a survey of parties in the German states and asked party leaders to rate all existing ministries according to their prestige on a scale from 0 to 100 (the Prime Minister (PM) is used as a reference which received the score of 100).

Second, the analysis accounts for the gendered nature of different cabinet portfolios. Previous studies show that where women are included in the cabinet, they are frequently appointed to ministries that are stereotypically labeled as ‘feminine’ or ‘soft’ (Krook and O’Brien, 2012). Following the definition by Krook and O’Brien (2012: 844ff.), I therefore control for feminine ministries (dummy variable) that are associated to the private sphere or that are typically linked closely to women. This comprises ministries responsible for children and youth, education, culture, elderly, health and care, and women’s affairs.

Third, I control for party ideology. Previous research shows that parties on the left generally appoint more women cabinet members than right-wing parties (Siaroff, 2000) and that, simultaneously, left-wing parties emphasize different issues in their manifestos than their counterparts on the right (Bräuninger et al., 2020). Data on party ideologies are derived from Bräuninger et al. (2020) who locate parties on a single left-right dimension based on Wordscore-analyses of party manifestos for state elections.

Fourth, the analysis controls for the presence of a female PM and a female party leader. O’Brien et al. (2015) theoretically expect that women in leadership positions might ‘let down the ladder’ to other women and select more female ministers than their male counterparts. At the same time the presence of a female PM or party leader might bring more attention to women’s interests, thereby influencing the importance that parties devote to these topics. Female Prime Minister and Female party leader are two dummy variables coded 1 if the PM or, respectively, the leader of the governing party in control of the ministry is a woman.

Based on similar considerations, I also control for the share of female MPs in the party faction of the governing party in the federal state parliament. Women in parliament serve as one of the primary recruitment pools for a ministerial office and, thus, increase the intra-party supply of potential female ministers (Krook and O’Brien, 2012). More women in the party factions can also exert pressure on the party to represent issues that have been neglected in the past. Since parties on the state-level do not have official quota regulations for women cabinet members, the share of female MPs also serves as a proxy for general gender equality norms within a party. Data on the share of women in party factions was provided from the Federal Returning Officer and from the respective offices for Parliamentary Documentation of the State Parliaments.

The analysis includes an additional dummy variable to control for cabinets that have been formed in East Germany since previous research shows that the proportion of women in state governments in East Germany is significantly lower compared to their counterparts in West Germany (Höhmann, 2017).

Due to the binary structure of the dependent variable, I use logistic regression models to estimate the effect of issue salience on the likelihood of appointing a female minister. All models are estimated with robust standard errors clustered by cabinet to account for potential dependences within cabinets due to the limited number of seats at the cabinet table that are given to women. I also include party fixed effects to control for the possibility that – beyond ideology – some parties might generally be more inclined to appoint female ministers. Lastly, all models include a linear time trend (measured by the year in which the government has formed) to control for general increases in the attention that parties pay to women’s representation.

Results: the effect of issue salience on the appointment of women cabinet members

Descriptively, the results show that women were considerably under-represented in the executives of the German federal states between 2006 and 2021 (c.f. Höhmann, 2017). Across all states, only 39% of the ministries were headed by a female minister. The highest descriptive representation of women in the executive can be found in Rhineland-Palatinate where, on average, 58% of the ministries have been led by a woman. The states with the lowest average shares of women in executive cabinets include Saxony (31%) and Saxony-Anhalt (30%). More detailed descriptive evidence on the proportion of female ministers across states and ministries can be found in the Appendices A2 and A3.

To test the effect of issue salience on the appointment of a female minister, I estimate four different logit regressions. The results are shown in Table 1. Model 1 contains the basic specification discussed above. Model 2 is generally similar to Model 1, but includes additional fixed effects for the states. Models 3 and 4 additionally control for the seat share of a party in parliament and for the total number of ministries that a party receives.⁵ All coefficients are presented as log-odds with robust standard errors clustered by governments.

Table 1. The effect of issue salience on the appointment of a female minister

Notes: Logit regression. Dependent variable: Dummy variable coded 1 if a portfolio is headed by a woman. Standard errors (in parentheses) are clustered by governments.

* P < 0.10;

** P < 0.05;

*** P < 0.001.

The results from Model 1 show that – against the theoretical expectations – issue salience has a significant positive effect on the appointment of a female minister. This means that parties in the German states are more likely to nominate a woman for a ministerial post if the respective policy area is highly salient for them.⁶

Since a substantial interpretation of the log-odds is not very intuitive, I estimate the predicted probabilities to assess the actual effect size of issue salience.⁷ The predictions in Figure 1 show that the probability of appointing a female minister is slightly below 20% if a ministry is not salient at all for a party. If the salience for a specific portfolio rises, the likelihood of observing a woman at the top of this ministry increases steadily: If parties spend 15% of their manifesto on a specific policy issue (which corresponds to the mean salience score),⁸ the predicted probability for a female minister reaches 40% and further increases to 79% until issue salience reaches its maximum value that can be observed in the dataset (43%). The additional model specifications in Table 1 show that this strong positive effect remains unaltered if I include fixed effects for the states as well as additional control variables into the analysis.

Figure 1. The effect of issue salience on the appointment of a female minister, predicted probabilities (with 95% CIs).

Note: Estimates based on Model 1 from Table 1. All other variables enter the model with their empirically observed values.

As a further robustness check, I test whether the positive effect might be driven by a correlation between salience and stereotypically feminine ministries. In fact, as can be seen in Figure A4 in the Appendix, government parties on the German sub-national level particularly emphasize social welfare policies in their manifestos. The most salient policy areas are ‘Social Policy’, including the portfolios for women’s affairs, senior citizens, youth, children, and family as well as ‘Labor’. Since these portfolios are also often considered to be stereotypically feminine, this might be the main reason why women enjoy greater access to highly salient ministries on the sub-national level. Therefore, I re-run the main models of the analysis and exclude all ministries responsible for social and labor policies. The results in Table A5 in the Appendix demonstrate that the findings remain unchanged. Salience still has a significant and substantially important positive effect on the appointment of a female minister.

Conclusion

Do parties relegate female ministers to portfolios that are politically less important for them? This research note contributes to this debate and analyzes whether the issue salience for a specific policy area affects the likelihood of parties to appoint a female minister. For the national level, a recent study by Goddard (2019) finds a negative effect, meaning that appointed ministers are less likely to be female if the portfolio is of high salience to the party. This note presents new empirical evidence and studies the appointment of female ministers on the sub-national level in Germany. Quite surprisingly, the results show that parties are actually getting more likely to appoint a woman as minister if the policy field of a portfolio is especially relevant for the party. This finding is quite remarkable since it clearly contradicts the theoretical assumptions from previous studies as well as the negative empirical results from Goddard (2019). The analysis demonstrates that in the German states women are not relegated to those ministries that are less important for their parties, thereby corroborating the optimistic claim that women are more and more able to erode the old patterns of male dominance in the executive arena, reduce stereotypical beliefs about gender and leadership roles in politics, and to gain access to the most influential and powerful positions in the cabinet (Barnes and O’Brien, 2018).

To what extent are these positive developments generalizable to other countries or settings? Most of the German states are rather progressive in terms of the general representation of women in politics and many regional parties are committed to promote gender equality with their political recruitments (c.f. Annesley et al., 2019). Thus, the German regional level presents a most-likely case to detect early improvements in women’s access to salient ministries. Nevertheless, the positive findings should also travel to other sub-national governments in countries with comparable levels of gender equality norms.

Relatedly, the present analysis of parties on the sub-national level implies that the results are not immediately generalizable to the national level. In the following, I therefore discuss possible explanations why the results in the German states differ so starkly from the negative effect that was found by Goddard (2019) for cabinets on the national level.

First, the sub-national states in Germany often serve as an ‘experimental laboratory’ where parties test new policies or other political innovations (Freitag and Vatter, 2008). The appointment of women to highly salient posts in the state cabinets might be a ‘test run’ for the political parties which might or might not be transferred to the Federal government. The results of the present analysis may therefore foreshadow a long-term development within parties at the national level, characterized by a higher attentiveness towards feminist values as well as an equal representation of women in the most important ministerial positions within the cabinet.

Second, some of the portfolios that previous studies typically classify as ‘masculine’ do not exist on the sub-national level. This applies in particular to the defense ministry and the ministry of foreign affairs. The positive findings might therefore be driven by these characteristics of the German states.

Nevertheless, the results presented here are encouraging in that they provide evidence that the ‘old-boys networks’ within parties might begin to erode and that highly salient posts in the cabinet are now increasingly accessible to women.