Gender Quotas and Democratic Participation: Recruiting Candidates for Elective Offices in Germany

392 62 3MB

English Pages [195]

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Gender Quotas and Democratic Participation: Recruiting Candidates for Elective Offices in Germany

  • Author / Uploaded
  • coll.

Citation preview

Page viii → Page ix →

Acknowledgments If it takes a village to raise a child, then it took the equivalent of a large city to write this book—and almost the same amount of time as raising a child. What began as an attempt to explain regional variance in gender quota fulfillment grew into a book-length discussion of the political recruitment process as a whole. Along the way I benefited from a number of sources of help—financial, logistical, intellectual, and emotional—and these acknowledgements are the very least I can do to thank all those who have contributed to my endeavors. Funding for the Candidate Interest Survey came from the Alexander von Humboldt Foundation, the Miami European Union Center, and the University of Miami Department of Political Science. Thanks to the Zentrum Gender Studies at the University of Bremen, including Konstanze Plett and Ulrike Liebert, for their hospitality while I prepared the survey and conducted interviews. Thanks to Sandra Ahrens, Michael Davidson-Schmich, Sven Diekmann, Petra KrГјmpfer, Christina Rinja, Markus Thiel, and Barbara Wulff for their assistance translating and developing the survey questionnaire. Thanks to Dora and Josef Schmich, the Zentrum Gender Studies, and to the University of Miami Department of Political Science work study students for their logistical help in mailing the survey and to James Rutherford for data entry. Just as important as the money to collect data and conduct interviews was the uninterrupted time to write up the results that I received while on a University of Miami College of Arts and Sciences book leave. If the many members of the German candidate pool who I interviewed Page x →and surveyed had not responded to my queries about their experiences and insights there would be no book to read today. Instead, high percentages of those surveyed returned their questionnaires quickly and interviewees were not only generous with their own time but helpful in connecting me with others who could also provide information. These women and men’s contributions will become obvious in the pages to come. Even though I cannot mention their names here, I would like to extend my sincerest thanks to all who agreed to participate in my research and wish all participants the best of luck in their political careers, should they choose to pursue one. This book owes an intellectual debt to the many talented scholars who study gender quotas, gender and political recruitment, and German politics. Pioneering studies of gender quotas by Drude Dahlerup, Susan Franceschet, Mark Jones, Mona Lena Krook, Richard Matland, and Jennifer Piscopo got me thinking about the impact of quotas in Germany; Jennifer Lawless and Richard Fox’s work on gender and political ambition in the United States was the initial inspiration for using a similar research design in the European case. My work builds on an outstanding tradition of scholarship focusing on women in German politics written by Eva Kolinsky, Beate Hoecker, Joyce Marie Mushaben, Myra Marx Ferree, Lee Ann Banaszak, Sabine Lang, Brigitte Geissel, Angelika von Wahl, Kathrin Zippel, Sarah Wiliarty, Joanna McKay, Isabelle KГјrschner, Jessica Fortin-Rittberger, and Christina Xydias. Versions of this book were presented at meetings of the American Political Science Association, the Midwest Political Science Association, the German Studies Association, and the Council for European Studies as well as in talks at the Jean Monnet Centre for European Studies, at the University of Bremen, to gatherings of the Christian Democratic women’s auxiliary in Bremen, and to the University of Miami Department of Political Science Faculty Workshop. I am especially grateful to Mona Lena Krook for including me in her “Beyond the Numbers: The Impact of Electoral Quotas” Workshop at Washington University in St. Louis. Participants in all of these settings provided invaluable input into this project, and it would be far poorer without their observations. Thanks to Shouraseni Roy for help making the map, and a special thanks to my colleagues (past and present) at the University of Miami for their help with the quantitative aspects of this research. All interpretations and errors are, of course, my own. Thanks to Melody Herr, Michael Laver, Kevin Rennells, and the University of Michigan Press for their faith in this project. Thank you also to the anonymous reviewers whose comments on this manuscript made it a far stronger piece.

Page xi →Completing a book while simultaneously trying to be a teacher, a university administrator, a wife and mother of two young children, a member of community organizations, and a sprint triathlete—while also remaining a halfway sane person—is a tough job. Fortunately, I have been able to rely on a lot of people’s help over the years it has taken to write this book. In addition to all those mentioned above, I have had the support and friendship of a number of additional people in the profession, in cyberspace, and in “real life” here in Miami. Going to conferences and sharing this research has not only been intellectually interesting, it’s also been fun and a great source of support to know there are others out there juggling similar academic lives. Friends (including but not limited to) Merike Blofield, Barbara Donovan, Farida Jalalzai, Melanie Kintz, Jonathan Olsen, Melody Ellis Valdini, and Jennifer Yoder consistently remind me why I love what I do. Thanks to my Facebook friends who noticed my posts and self-imposed deadlines and asked how my writing was proceeding. Knowing that nonacademics including Simone Schmich, Andrea Scherer, Lou Tash, and Yvonne Visser-Guerrero were going to ask about my progress kept me writing. Yvonne, you will finish your own book one day too. My thanks also extend to the loving teachers at the University of Miami Canterbury Preschool and in the Miami-Dade County Public School system. They are the village that is raising my children, giving me time to write, and without their assistance I could not have completed this book. Thanks also to Linda Smith for cleaning my house while I work. Finally, thanks are due to my extended family. My parents Louise C. and Alan J. Davidson were always supportive of my choice to pursue an academic career, which I have come to appreciate more than ever as I advise students with far less understanding parents. Without my husband, Michael, I could not have become a tenured professor or completed this book. His consistent willingness to share in child care, follow me to research sites, serve as a single parent while I am out of town, and live in a city far removed from his cultural home have enabled me to follow my passions and I am deeply in his debt. Finally my two children, Klara and Nathan, are an enormous source of joy and pride for me and the best reasons why this book has taken a long time to write. I am sure Klara will become an author herself one day and Nathan a reader. While I love and thank all my family very much, it’s now Nathan’s turn for a book dedication.

Page xii → Page xiii →

List of Abbreviations ASF Social Democratic women’s auxiliary organization (Arbeitsgemeinschaft Sozialdemokratische Frauen) CDU Christian Democratic Union (political party) CIS Candidate Interest Survey CSU Christian Social Union (political party) EP European Parliament EU European Union FDP Free Democratic Party FU Christian Democratic women’s auxiliary organization (Frauen Union) MdB Member of the Bundestag (Mitglied des deutschen Bundestages) MEP Member of the European Parliament MP Member of Parliament NGO nongovernmental organization OECD Organization for Economic Cooperation and Development PDS Party of Democratic Socialism PPG parliamentary party group PR proportional representation (electoral system) PTA Parent Teacher Association SPD Social Democratic Party US United States VFD volunteer fire department WPA Women’s Policy Agency Page xiv →

List of German Terms and English Translations Arbeitsgemeinschaft Working group (in a political party) Bekanntheitsgrad Degree to which someone is known, name recognition Bezirk District level of government

Bezirkstag District council Bundestag The Federal Republic’s national parliament BГјndnis 90/die GrГјnen Alliance 90/the Greens (political party) BГјrgermeister Mayor die Linke The Left (political party) Deutscher StГ¤dtetag German federation of cities (lobbying group) Feierabendpolitiker After-work politician/amateur politician Frauenlisten All-women’s nonpartisan electoral lists Frauen Union The Christian Democratic women’s auxiliary organization Fundis The fundamentalist wing of the Green party Gemeinderat Local council (in towns) JuSos Social Democratic youth organization Junge Union Christian Democratic youth organization Karteileiche “Party card corpse,” an inactive party member kept on the membership rolls Kreistag County council Land/LГ¤nder One (or more) of the sixteen states in the Federal Republic Landtag State parliament Landrat County mayor Liberale Frauen The Free Democratic women’s auxiliary organization MinisterprГ¤sident Minister president, position akin to a U.S. governor Ochsentour “Ox Tour,” a slogging up the ranks of a political party Parteiengesetz The Federal Republic’s law governing political parties Parteiverdrossenheit Public loss of faith in political parties Realos The pragmatic wing of the Green party Sargspringerin Literally a “coffin jumper,” a woman who obtains a seat in parliament after a man higher up on the party list dies Page xv → Seiteneinsteiger A lateral-entry candidate who did not make her way up the ranks of a party Stadtrat Local council (in cities) Vorstand Executive board (of a political party) Volkskammer The German Democratic Republic’s national legislature

Wählergruppen Nonpartisan groups running candidates for elective office Wutbürger Middle-class citizens fed up with politics as usual

Page xvi → Page 1 →

Introduction Gender Quotas and Women’s Descriptive Representation—Missing Mechanisms The past decades have witnessed a marked rise of women into elective offices across all branches of government, but especially into national legislatures; worldwide, the percentage of female members of parliament (MPs) has tripled since 1970. Women’s presence in national parliaments has come to be considered an important component of democracy by international organizations, including the United Nations, and political equality is now viewed as requiring participation from both sexes (Dahlerup 2006; Krook 2009; Franceschet and Piscopo 2013). Moreover, because female political leaders are thought to serve as role models who can awaken other women’s and girls’ interest in public affairs, and because female representatives are believed to represent women’s interests, women’s inclusion in decision-making bodies is now regarded as improving the quality of democracy. Despite these changing global norms and the increase in women’s descriptive representation, however, men still remain overrepresented in almost every country. Although women make up half the global population, in 2014 fewer than one quarter of MPs worldwide were female (Interparliamentary Union 2014). As a result of a growing concern about women’s political underrepresentation, policy makers in countries as diverse as France, Argentina, Taiwan, and Rwanda have adopted gender quotas for elective office as a way to facilitate women’s political participation. Quotas involve setting percentages or numbers for the political representation of specific groups, in this case women and, at times, men. Quotas take various shapes including the creation of special seats reserved for women, formal electoral laws regarding the sex of candidates, and—the focus of this book—voluntaryPage 2 → commitments made by political party organizations. Political party candidate gender quotas initially began with some left-wing political parties in northern Europe in the 1970s, but have now spread across the political spectrum and the globe, requiring ever-higher percentages of women to appear on the ballot (Krook 2009; Thames and Williams 2013). Although at times very controversial, quotas have proliferated on the strength of several arguments including claims that they help achieve gender-equal participation in democracy and, with it, both role models for girls and better representation of women’s interests. Moreover, feminist scholars and practitioners have identified many discriminatory practices on the part of gatekeepers, or those who select candidates to appear on the ballot, and quotas were designed to overcome these barriers. An extensive body of research on cases from around the world confirms that if properly designed and implemented, even voluntary party quotas can indeed increase the percentages of women in elective office1 (e.g., Dahlerup and Freidenvall 2005; Kittilson 2005; Thames and Williams 2013; Tripp and Kang 2008). However, little empirical evidence exists of whether or not quotas’ goals of gender-equal participation in the democratic process, creating role models, and overcoming discrimination have actually been met. When quotas are successfully implemented, do male and female citizens participate equally throughout the political recruitment process used to winnow the millions of citizens legally eligible to run for office down to the few who actually become candidates? That is, do men and women obtain the qualifications necessary to run for office at equal rates? Are qualified men and women equally prone to aspiring to elected positions? Quotas have increased the percentages of women who appear on the ticket, but have they truly leveled the playing field, giving aspiring male and female politicians equal chances of being groomed as candidates by their parties and appearing on the ballot in winning places? While we do know that quotas can lead to more women in elective office, we do not know the answers to the above questions. The causal mechanisms through which quotas increase women’s presence in elective offices are poorly understood, making it impossible to evaluate whether quotas have achieved all that their proponents have hoped (see Dahlerup and Freidenvall 2010). Because there are far fewer positions in national parliaments

than there are citizens eligible to hold office, it is possible that women’s numeric representation in powerful legislatures can rise while women remain underrepresented among the citizens who are qualified for, aspire to, and are selected as candidates for elective office. If this were the case, quotas’ goal of improving women’s participation in the democratic Page 3 →process would only partially have been reached. This state of affairs could negatively impact women’s substantive representation as well, offering female citizens a narrower range of individuals from which to select a representative than male citizens enjoy. Further, because historically very few women have held elected office, it is possible that the numbers of women elected could increase while party leaders remain more likely to select male than female candidates. In such instances, quotas may lead to increased percentages of women in elective office but politically ambitious women would still face gender discrimination. Thus in order to determine whether quotas’ goals have been fully met, we must examine all steps on the road to becoming an elected official when quotas are in place, examining who obtains the qualifications needed to run for elective office in a given democracy, which qualified individuals in turn aspire to elective office, which of these politically ambitious citizens are selected to appear on the ballot, and who among them is ultimately elected by the voters. To date, however, most investigations of quotas begin at the candidate or elected official stage, rather than earlier in the political recruitment process. As a result, while we do know that quotas can increase the numbers of women in elective office, we do not understand how this increase occurs. If qualified women remain less likely than men to aspire to elective office or to be selected to appear on the ballot, additional measures to achieve gender equal participation in democracy are needed and quotas alone are not enough. Understanding whether this is the case or not is especially important because in recent years gender quotas have diffused, not only to parties across the ideological spectrum but also to different groups such as racial and ethnic minorities (Hughes 2011; Geissel 2013; BjarnegГҐrd and Zetterberg 2014; Krook and Zetterberg 2014) and from the political to the corporate world (Engelstad and Teigen 2012; Franceschet and Piscopo 2013). This book therefore departs from conventional practice and introduces a new method of studying the impact of quotas by systematically examining every phase of candidate recruitment in a setting where some—but not all—political parties employ gender quotas to select their candidates. The overall argument here is that, although quotas have indeed led to marked increases in women’s presence in legislatures, they have only had mixed success at attaining their other goals. Quotas have prompted parties to promote their female members to positions of inner-party leadership more often than their male members, qualifying women to run for elective office, and, in parties with parity or near-parity gender quotas, quotas have rendered female party members more likely than their male counterpartsPage 4 → to be asked by the party to run for elective office, to appear on the ballot, and to win elections. However, these results occur in part because quotas have been unable to spur equal numbers of women and men to join political parties and to aspire to elective office in the first place, creating favorable opportunity structures for the few women who do want to enter politics. In short, I argue that quotas have succeeded in some, but not all, of their goals and additional changes are needed to more fully achieve gender-equal participation in democracy. In this introductory chapter I first examine the initial rationale for adopting quotas and discuss why investigating the political recruitment process as a whole is required in order to determine whether quotas have achieved their intended aims. Second, I delineate the challenges of isolating quotas’ effects on political recruitment and then develop a method of doing so. Third, I preview the hypotheses that are tested empirically in the chapters to come. Finally, I conclude by discussing the sources of evidence used in this book and outline the remaining chapters.

Why Study Party Quotas and the Political Recruitment Process? Candidate gender quotas are at times quite controversial, but they have been adopted with increasing frequency around the world on the strength of several arguments including simple fairness, quotas’ symbolic potential, and their role in improving women’s substantive representation (Phillips 1995; Mansbridge 1999; SchwindtBayer and Mishler 2005; Franceschet, Krook, and Piscopo 2009; Dahlerup and Freidenvall 2010). While increasing numbers of women holding elective positions may seem at first glance to be consistent with these

goals, considering the political recruitment process makes clear that the two are not synonymous. One of the primary ways in which quotas have been defended is through appeals to fairness, employing what Anne Phillips calls the “principle of justice” (1995, 62). It is simply not equitable, this argument goes, that one sex dominates representation in a democratic system predicated on political equality, and, in order to correct this imbalance in political participation, quotas are required. As the United Nation’s Beijing Platform for Action put it, “Achieving the goal of equal participation of women and men in decision-making will provide a balance that more accurately reflects the composition of society and is needed in order to strengthen democracy and promote its proper functioning” (quoted in Dahlerup 2006, 16–17). Certainly,Page 5 → rising percentages of women in national legislatures go far to rectify the injustice political philosophers such as Phillips identify. However, given that there are only a handful of top elective offices available in any particular country, even half of these positions being held by women offers only limited evidence that female citizens are participating as fully throughout the democratic process as male citizens. Although women and men may formally enjoy the same right to run for elective office and while quotas have been found to increase the percentages of women candidates and elected officials, we know little about how quotas impact women prior to their becoming candidates for public office. Because virtually all existing research on quotas begins at the candidate stage or with elected officials themselves, it cannot shed light on whether or not male and female citizens obtain the qualifications for elective office at equal rates, whether qualified men and women have equal chances of developing political aspirations, being selected as a candidate, or elected to a public post. Yet if quotas are justified on the basis of fairness and giving women an equal chance to participate politically, it is important to know whether or not they actually render women and men equally prone to taking part in the various stages of political recruitment. If they are not, quotas can be considered only a partial success in achieving “justice,” even if they do increase the percentages of women in top elective offices. A second motivation for quotas is the hope that such affirmative action measures will have symbolic effects (Mansbridge 1999; Schwindt-Bayer and Mishler 2005; Dahlerup and Freidenvall 2010; Franceschet, Krook, and Piscopo 2012). As Jane Mansbridge argues, “Low percentages of Black and women representatives, for example, create the meaning that Blacks and women cannot rule, or are not suitable for rule.В .В .В . [T]he increased descriptive representation of women in the legislatures would undermine the perceptions that politics is a вЂmale domain’” (1999, 649). By increasing the numbers of women elected, it is expected, quotas will in turn inspire women and young girls to become eligibles, aspirants, and candidates just as their role models have done. Extensive research has established a connection between female elected officials and women and girls’ mass attitudes toward politics and political participation. When large numbers of women or very visible women take part in politics, some women and girls do know more about politics (Burns, Schlozman, and Verba 2001; Reingold and Harrell 2010), talk more about politics (Atkeson 2003; Wolbrecht and Campbell 2007), and express a greater desire to participate politically—for example, by running for elective office (Wolbrecht and Campbell 2007). However, far less attention has Page 6 →been given to whether women actually do participate more in these settings, and the studies that have been conducted unearth little connection between women in elective office and the average woman’s propensity to engage in other forms of political life such as contacting elected officials, attending political meetings, donating money to political causes, or working on election campaigns (Lawless 2004; Zetterberg 2009). To determine whether quotas’ symbolic effects extend to running for elective office, scholars must first identify whether the gender breakdown of the candidate pool changes after quotas are introduced and then compare male and female eligibles’ political ambitions. If, in practice rather than in opinion surveys, women exhibit a lower propensity to become eligibles and/or aspirants than men, quotas can be considered to have only limited symbolic effects. Thus, here too, in order to assess quotas’ full impact, all phases of candidate recruitment need to be studied. A third justification for quotas is that they improve women’s substantive representation; that is, quotas will

lead to women’s political interests being better served by elected bodies (Dahlerup 2006; Dahlerup and Freidenvall 2010; Franceschet, Krook, Piscopo 2012; Mansbridge 1999; Phillips 1995). As the Interparliamentary Union argued, “The concept of democracy will only assume true and dynamic significance whenВ .В .В . national legislation [is] decided upon jointly by men and women with equitable regard for the interestsВ .В .В . of both halves of the population” (quoted in Dahlerup 2006, 16). Studies of women’s substantive representation have established that the greater the numbers of women included in political decision making, the wider the range of women’s concerns that are taken up in political deliberations (Celis 2006, 2009; Weldon 2002). If women are less likely to be qualified for, and aspire to, elective office than men, a narrower spectrum of women’s than men’s issues will likely be brought to the table. In this case, female citizens would enjoy fewer options to represent themselves than male citizens. Moreover, many oppose quotas because they fear women elected via quotas will be mere tokens, unable to accomplish much for women (Dahlerup 2006; Dahlerup and Freidenvall 2010; see also Mansbridge 1999). If “quota women” are indeed grudgingly selected to fill affirmative action requirements, they may fear a “labeling” effect and wish to avoid speaking up on behalf of other women when they assume office (Franceschet and Piscopo 2008), reducing quotas’ ability to improve women’s substantive representation. Understanding the relationship between quota rules and political recruitment will help scholars better predict whether “quota women” are likely to be trivialized and fear being stigmatized, or whether Page 7 →quotas lead gatekeepers to groom women for elective office in ways similar to their male counterparts, endowing female officials with both legitimacy and a “mandate” to provide substantive representation (Franceschet and Piscopo 2008). Where similar ratios of qualified men and women aspire to available ballot slots, and where party leaders promote promising female candidates, achieving a nomination to run for office will be equally competitive for men and women and the latter can hardly be derided as tokens. Where few women are qualified for, and interested in, running for elective office, however, or where gatekeepers discriminate against women, female elected officials run a higher risk of being labeled tokens or fearing being labeled as such. Studying all phases of political recruitment where quotas are in effect will provide scholars with the evidence needed to determine which of these conditions is met. Finally, considering the effect of quotas on political recruitment is useful for the same reason political recruitment has been studied for years in a nongendered fashion: the process of winnowing out potential candidates inevitably contains biases in addition to sex including, for example, class, race, and ethnicity (e.g., Norris and Lovenduski 1995; Norris 1997). Just as certain types of men have been more likely to be selected as elected officials, so too will be certain types of women. Better understanding how the many eligible citizens are narrowed into a few candidates will shed light on what type of women quotas send to elective office and, in turn, the kinds of women who are likely to be sitting at the table when public policies are made and the sorts of “women’s interests” they may articulate.

Quotas and Political Recruitment: What Do We Know? An extensive literature documents both women’s continual rise among the ranks of legislators and the diffusion of party quotas worldwide (e.g., Dahlerup 2006; Krook 2009; Thames and Williams 2013). See figures 1 and 2. Countries where parties employ gender quotas feature, on average, more women in their national legislatures than states where parties have not committed to affirmative action.2 What is poorly understood, however, is how quotas interact with each phase of the political recruitment process used to select representatives of the population, leaving the mechanisms through which quotas escalate the percentages of women MPs hidden in a black box. In order to open this box, it is first necessary to identify the steps involved in becoming an elected official (see figure 3). In any given democracyPage 9 → elected officials represent the outcome of an extensive winnowing process that is termed political recruitment (Matland and Montgomery, 2003, 21). This trajectory begins with all the citizens who may legally run for elective office—in most democracies this includes a large pool of people who meet certain basic criteria such as a minimum age. Since women’s suffrage was obtained, roughly equal percentages of women and men may, in theory, become elected officials. Obviously, however, the vast majority of

citizens in long-term democracies never consider running for elective office; as Robert Dahl famously put it, “in liberal societies, politics is a sideshow in the great circus of life” ([1961] 2005, 305). Relevant for the study of quotas and political recruitment, then, is that small subset of citizens who are interested in politics and choose to obtain the qualifications that would make them a suitable elected official. Page 8 → Fig. 1. Women in National Legislatures Worldwide. (Sources: World: Interparliamentary Union (IPU); Party Quotas: Thames and Williams 2013 and IPU, author’s calculations; “1995” worldwide figures represent data from 1997.) Fig. 2. Number of Parties Worldwide Employing Voluntary Quotas. (Source: Thames and Williams, 2013, 85; quotaproject.org.) Fig. 3. The Political Recruitment Process. (Source: Based on Matland and Montgomery, 2003, 21.) These individuals are the eligibles, or those from whom the ranks of elected officials are drawn in practice; together eligibles make up the candidate pool. Even among those who are considered eligible to become an elected official, not all are interested in doing so. Eligibles may decline to run for office for myriad reasons, such as concerns a candidacy may impede their professional careers, a desire to achieve political goals through other means, or because of ill health. Those who do wish to seek elective office are considered aspirants and present themselves to the gatekeepers, or those who are responsible for allotting ballot positions. These gatekeepers thus have the power to further reduce the field of citizens able to participate politically at the highest level by selecting who becomes a candidate. In democracies, voters have the final say, selecting elected officials from among the ranks of candidates. Most research on gender quotas’ impact begins at the candidacy stage (e.g., Htun and Jones 2002; Jones 2004; Matland 2006; Meier 2004) or after elections have been held (Escobar-Lemmon and Taylor-Robinson 2008; Franceschet, Piscopo, and Krook 2012; Thames and Williams 2013).3 In order to fully evaluate gender quotas’ effectiveness in reaching their stated Page 10 →goals, though, all stages of the political recruitment process, not just penultimate and final phases, must be examined. While it is relatively easy to identify the elected men and women (a very small and very public subset), and to investigate the men and women who appear on the ballot (a larger but still well-documented group), it is much more difficult to determine who was eligible but not nominated or even interested in the first place. Many such individuals exist, but no public records are kept of those who did not seek or receive their party’s nomination. Identifying and studying the candidate pool is not impossible, however, and large-N surveys of eligibles have indeed been utilized to investigate gender and political recruitment at every stage of the process (Norris and Lovenduski 1995; Maisel and Stone 1997; Maestas, Maisel, and Stone 2005; Fulton et al. 2006; Evans 2008; Lawless and Fox 2005, 2010; Lawless 2012). Because, to my knowledge, such investigations have been conducted only where quotas are not in place, this work sheds no light on how successful quotas impact the candidate pool.4

A New Method for Studying Quotas and Political Recruitment This book, therefore, is one of the first to systematically compare men and women’s experiences across all phases of political recruitment in a context where various affirmative action policies are in place: no quotas, enforceable quotas requiring that approximately half of the candidates are women, and a less-binding, 33 percent “quorum” for women. I investigate whether quotas have symbolic effects, increasing the percentages of women among eligibles and aspirants, ask whether quotas change gatekeepers’ propensity to actively recruit female candidates, and check how quotas shape female aspirants’ odds of becoming a candidate and getting elected vis-à -vis male members of the candidate pool. In short, I bring the study of quotas into what Gallagher and Marsh (1988) call the “secret garden” of politics: the phases of political recruitment before candidates are selected to run for office. To do so, I develop a research design capable of identifying the causal mechanisms linking quotas to an increase in women in elective office. This method can be employed in other, similar cases around the world to further increase our understanding of quotas’ effect on political recruitment. Here I discuss the design of my study and then draw up arguments about the relationship between quotas and the various phases of political recruitment.

Page 11 →The Pathway Case Research Design Given the great cross-national variance in both gender quota rules (Dahlerup 2007; Krook 2009) and in methods for selecting candidates (Hazan and Rahat 2010; Siavelis and Morgenstern 2008)—combined with cross-country differences in factors influencing the percentage of women in elective office including the level of economic development (Inglehart and Norris 2003), political culture and religion (Inglehart and Norris 2003; Ruedin 2012), gender norms (Paxton and Kunovich 2003), electoral rules (Matland 2006; Salmond 2006) and party systems (Kittilson 2006)—it would be very difficult in a multicountry study to isolate the causal mechanisms through which gender quotas shape political recruitment. Moreover, locating candidate pools in multiple countries would prove even more challenging. In order to best determine how well-designed and implemented party quotas increase women’s descriptive representation in decision making bodies, then, a single country research design is required. John Gerring terms this research design a pathway case and argues that it offers a “uniquely penetrating insight into causal mechanisms” (2007, 238–39). To select a pathway case, a situation must be found in which quotas’ positive influence on political recruitment can be isolated from other causes such as those mentioned above. Studying a single country where quotas have been adopted and increased the numbers of women in elective office can better meet this standard than a multicountry approach, where controlling for alternative explanations would be quite difficult. Focusing on a single country also makes the complicated task of identifying eligibles who never become aspirants or candidates somewhat easier. A country employing electoral law gender quotas—which apply to all parties in the system—would be unsuited to this design, however, as assessing quotas’ impact on political recruitment would require comparing recruitment for elections prior- and subsequent-to quota adoption. When doing so a researcher could not be certain whether quotas or other temporal factors, such as, for example, increases in domestic attitudinal support for gender equality or changing global norms about women’s political representation, were driving any observed changes in recruitment. In addition, since party quotas are often considered less effective than electoral law quotas because the latter have the force of law to sanction violators, voluntary quotas represent a difficult case in which to detect quotas’ effects. If quotas are found to shape political recruitment where parties have only Page 12 →made voluntary promises, they are likely to have even greater effects where parties are legally bound to implement affirmative action. Thus an appropriate pathway case would allow a scholar to investigate candidate recruitment at a single point in time in a single case where the relationship of interest, gender quotas, and their impact on political recruitment varies. A country where some parties use quotas to increase women’s descriptive representation and others do not would fit this bill. Since symbolic effects do not occur overnight, but rather over a generation as girls undergo political socialization, the case selected also would need to have utilized voluntary quotas long enough for such generational effects to have occurred; electoral law gender quotas are a newer phenomenon than voluntary party quotas, again making them less suited for study here.5 These criteria considerably restrict possible pathway cases (see Thames and Williams 2013 for a complete list of party quotas and their dates of adoption). A study limited to a single country that was an early adopter of a range of voluntary party gender quotas, then, allows a range of independent variables to be held constant in order to better isolate the causal impact of gender quotas on political recruitment. Not all states featuring early party quota adoption are equally suited to answer the questions at hand, however. Some within-country variables, other than quotas, may still drive the candidate recruitment process. Two potential problems that must be considered involve cross-party variance in candidate selection procedures and in ideology. In some countries each party employs a different method to select its candidates (Hazan and Rahat 2010, 4). For example, one party might utilize primaries while in another a single party leader might draw up the ballot; as a result, diverging internal candidate selection procedures, rather than quotas, could drive observed variance in political recruitment. Similarly, there is a potential endogeneity problem. Voluntary party quotas originated in New Left political parties with progressive gender ideologies (Kittilson 2006) and cross-nationally left-wing parties have been most prone to select female candidates (Caul 1999), so variance in women’s political recruitment at each stage could be driven by party ideology rather than by quotas in and of themselves.

Despite these challenges, a suitable pathway case exists that allows me to control for these possible sources of variance: contemporary Germany. Some German political parties began to adopt voluntary quotas in the 1980s, and between 1980 and 2010 the percentage of women in the Bundestag increased threefold.6 Today, most—but not all—German parties employ voluntary quotas. The Federal Republic thus meets the criteria of Page 13 →a single country where quotas increased women’s descriptive representation, where some parties were early quota adopters while others never used quotas, and where no electoral law quotas exist. Moreover, Germany allows me to control for in-country variables as well. Due to concerns about undemocratic party gatekeepers in the wake of the Nazi era, the founders of the Federal Republic created very explicit laws governing candidate selection. As a result, all parties in Germany are required to follow the same procedure for choosing ballot nominees, regardless of their other characteristics (Roberts 1988, 97).7 Moreover, this type of candidate selection—termed party loyalist (Siavelis and Morgenstern 2008)—has been shown in cross-national research to be the context in which voluntary quotas have proven most effective in increasing women’s representation (Escobar-Lemmon and Taylor-Robinson 2008).8 Thus, Germany is typical of cases where voluntary party quotas have been shown to succeed, and the mechanisms through which quotas have increased women’s presence in top German posts are likely at work elsewhere. Further, while German quota regulations indeed began in the 1980s when the postmaterialist Green party adopted a minimum 50 percent quota for women, quotas have diffused across the political spectrum and are now employed by parties with a range of gender ideologies, mitigating the endogeneity problem. Following the Greens’ introduction of a quota, the Social Democratic Party (SPD) also phased in a voluntary quota requiring a minimum of 40 percent and a maximum of 60 percent of each sex on the party list.9 The SPD did so less out of an ideological commitment to gender equality, and more in the hopes of stemming the loss of postmaterialist voters (Kolinsky 1993). In December 2011 the party increased this quota to 50 percent of each sex for European Parliament and Bundestag lists. As is typical in other cases (Matland and Studlar 1996; Kittilson 2006), quotas then diffused further right in the political spectrum and in 1996 the Christian Democratic Union (CDU) adopted a 33 percent “quorum” for female candidates on electoral lists. This change was pushed onto a reluctant rank and file by top party leaders concerned about the CDU’s growing inability to attract female voters (Wiliarty 2010). This process of quota contagion helps alleviate concerns that feminist party ideology, rather than quotas, drives my findings. Moreover, while these very dissimilar German parties have adopted quotas, other parties with gender ideologies similar to the CDU’s and the SPD’s have not done so, allowing me to exploit a natural experiment and compare political recruitment in ideologically similar parties with and without affirmative action policies. In Germany, proponents of Christian Page 14 →democracy are found in two separate party organizations: the Christian Democratic Union and the Christian Social Union (CSU). The former has pledged to select 33 percent women candidates for national electoral lists while the latter has not. As a result, the CDU consistently sends a higher percentage of women to the Bundestag than the CSU does (Wiliarty 2013). A similar pattern can be observed at the state level (see chapter 1). These differences in recruitment cannot be attributed to Christian Democratic views about gender. Similarly, the Social Democratic Party and Free Democratic Party (FDP) have platforms stressing the importance of promoting gender equality with a similar emphasis on equality of opportunity—in contrast to the Greens’ and Left Party’s more far-reaching weight placed on equality of outcome, and the Christian Democrats’ focus on women’s family roles (Xydias 2013). After the Green Party adopted its quota, the FDP unveiled a Plan for the Advancement of Women that included informal methods to promote women within the party but no formal quota (Kittilson 2006, 95). Today the SPD utilizes a gender quota and the FDP does not; the former consistently sends higher percentages of women to the Bundestag than the latter, although up until the 1980s the FDP had been the leader in women’s descriptive representation (Schindler 1999, 636). Again these differences in recruitment cannot be attributed to ideological beliefs about equality of opportunity for women and men.10 Thus, because ideology and quota use do not perfectly coincide in Germany, cross-party differences in political recruitment are likely not driven by ideology, helping avoid endogeneity issues. Another way to demonstrate

quotas’ independent impact on political recruitment is to compare women’s electoral fortunes in the two parts of Germany’s mixed electoral system. Even within the same party, women have far better chances of election through the proportional representation half of the ballot—where quotas are used—than through the first-past-the post component where quotas do not apply (Davidson-Schmich and KГјrschner 2011). Selecting Germany as a pathway case thus allows me to control for a number of factors and isolate the impact of gender quotas on each stage of political recruitment. This pathway case research method can be applied to other cases meeting the criteria outlined here. Western European and Latin American countries where effective voluntary quotas have long been employed and where candidate selection procedures are similar across parties would be the most suitable sites for future investigations. Based on my examination of the case at hand, I conclude that quotas increase the percentages of women in elective office not through a symbolicPage 15 → effect—spurring women to join political parties and to aspire to elective office at rates similar to men’s—but by elevating the women who do overcome obstacles to party membership into positions of eligibility, leading gatekeepers to identify, train, and recruit promising women for elective office, and where quotas of over 40 percent are in place, to more often select women for promising ballot slots than their male counterparts. I discuss these arguments in detail below and in the pages to come. Future research in this vein can assess the external validity of the findings presented here.

How Quotas Shape Political Recruitment Political recruitment does not take place in a vacuum (see figure 4). Instead, it is embedded in what Matland and Montgomery term the recruitment environment or a broader set of social norms and institutions (2003, 21)—of particular interest here are those relating to gender, such as expectations about who performs household labor, who cares for the young and the sick, and who pursues what type of paid work outside the home. Gender quotas do not attempt, and would be unlikely, to change this recruitment environment in any significant manner. In addition to the overall societal backdrop against which elected officials are chosen, Matland and Montgomery also note that candidates are selected through recruitment structures including political parties’ formal and informal practices (2003, 21). While quotas can quickly change parties’ formal requirements regarding which sex may be selected for which ballot slot, quotas do not directly apply to informal party norms or rituals, which may themselves be highly gendered. In other words, quotas may alter rules governing who may selected as a candidate, but they do not prohibit mostly male local branch organizations from, for example, holding prolonged meetings in smoke-filled bars in the evening when no childcare is available. Over time, quotas may chip away at these traditional recruitment structures, but such change does not occur overnight. As a result, I hypothesize that the more influence the recruitment environment and traditional, informal recruitment structures have on a given phase of political recruitment, the less likely quotas are to be successful in reaching their goals at that stage of the process. I expect that the recruitment environment will matter most early on in political recruitment—hindering women from becoming eligibles and limiting their subsequent propensity to become aspirants—but that its influence will subside later in Page 16 →the process. Quotas should exert their biggest impact on gatekeepers and at the candidate stage. Below I discuss each of these expectations in turn. Fig. 4. The Recruitment Environment and Recruitment Structures. (Source: Based on Matland and Montgomery 2003, 21.) Eligibles In most democracies around the world, political parties play a key role in recruiting and selecting candidates for elective office (Hazan and Rahat 2010), and in party-loyalist systems such as Germany’s (Siavelis and Morgenstern 2008) becoming a party member is the first step in becoming an eligible. To be a viable contender for a party’s nomination, however, more than simple membership is expected; instead, the leading characteristic sought by parties is a potential nominee’s track record of service to the party and to the electoral district (Gallagher 1988, 248; Matland 2005; Matland and Montgomery 2003, 24). Although quotas

require changes to the sex of the candidates that parties select, they do not mandate changes to the nature of party life (recruitment structures), which is heavily masculinized. Moreover, quotas do not alter other aspects of citizens’ everyday lives (the recruitment environment)—such as gender discrepancies in professional careers and in unpaid household labor—which result in women possessing fewer resources and less time to devote to parties than men enjoy. Thus quotas are expected to have little impact on the rates at which women join political parties. However, if gender quotas are used not only to select candidates but also inner-party officers, as is common where party quotas are in place,11 affirmative action measures can have a salutary effect in that they promote the limited women present to positions of authority within the party. This Page 17 →quota-driven “elevator effect” is expected to create eligibles and is likely to make women overrepresented among party leaders when compared to their percentage among rank and file party members. Aspirants As figure 3 notes, not all individuals eligible to run for elective office are interested in doing so. People may join political parties and become officers within them not because they want to become politicians, but to have an influence on policy or experience social solidarity with others in the group (Gallagher, Laver, and Mair 2011, 332). Aspirants, then, are a subset of eligibles—those who desire elected office and who have the resources to pursue it. Just as in the first phase of political recruitment, at this stage, too, quotas are likely to come up against a recruitment environment that results in women having fewer resources than men possess, rendering female party members, on average, less likely than male party members to desire political office. To become aspirants, eligibles must be endowed with what Norris refers to as “political capital” (1997, 13): they need to be qualified for the job, have confidence in their own qualifications, and enjoy the time to pursue political office (Matland and Montgomery 2003; Murray 2010b; Lawless 2012). While gender quotas that extend to inner-party office can increase women’s objective qualifications for elective office in a party-loyalist system of candidate recruitment, quotas cannot increase women’s endowments of the other aforementioned resources. Quotas are not expected to (directly) change a highly gendered recruiting environment that values different types of behaviors and experiences, deeming some more appropriate for women than men. While quotas may elevate women to positions of inner-party leadership, they are unlikely to change the masculinized nature of “qualifications” for elective office and, on average, women are expected to be less interested in elective office and to feel less confident about their own abilities to run for elective office than men are. Even if an eligible individual wants, and feels qualified, to run for elective office, she must also have the time to do so. Responsibility for childrearing and household tasks clash with the long and irregular hours associated with a political career. Because the vast majority of candidate pool members also pursue professional careers in addition to politics, agreeing to hold any amateur elective office—for example, taking on a position with a city council—would add a “triple burden” to anyone, male or female, Page 18 →responsible for household and care work. However, women are on average more likely to find themselves in such ambition-dampening personal circumstances than are men. Higher-level elective offices that are full-time positions, and thus require “only” two shifts, may be more appealing to women, but in Germany and elsewhere the prerequisites for such professional offices often include first holding amateur local office. At this stage of recruitment too, then, quotas are hypothesized to exert only a limited impact and women are hypothesized to be less likely than their male peers to become aspirants. Gatekeepers Extant candidate pool studies—most of which focus on situations where quotas are not in place—consistently depict women as less often asked to run for elective office than their male counterparts (Evans 2008; Lawless and Fox 2010; Lawless 2012; Niven 1998; Sanbonmatsu 2006; Shepherd-Robinson and Lovenduski 2002). Indeed, in

Germany gender quotas were initially adopted because feminist activists hoped they could overcome such gatekeeper bias (Kolinsky 1991). While quotas have increased the percentages of female elected officials, scholars to date have not been able to determine whether gatekeepers are now grudgingly selecting available aspirants for ballot slots or whether parties are deliberately identifying and recruiting promising female eligibles, as they have in the past promoted auspicious men within their ranks (Patzelt 1995). By studying male and female eligibles’ experiences, however, it is possible to discern which mechanism is actually at work. I hypothesize that since gatekeepers’ goal is to select winning candidates, their reactions to quotas will depend on whether or not they believe women (or at least some female eligibles) will be attractive to the electorate. If gatekeepers fear that female candidates will not garner votes for their party, they are unlikely to want to implement quotas and thus unlikely to either recruit or select women. Where quotas have “teeth” and must be implemented, in contrast, party gatekeepers skeptical of female candidates’ viability cannot resort to evasive strategies. In such cases party leaders pressed to select female candidates have an incentive to actively identify, recruit, and train women who can win votes for their party. There is little theoretical reason to expect that in the long run party gatekeepers would sit back and allow quotas to thrust unwanted candidates upon them. Similarly, when gatekeepers believe that women are—or can be—winning candidates, they possess the same decision calculus. Gatekeepers’ Page 19 →estimation of female candidates’ chances may be high due to reasons unrelated to quotas—such as the overall level of socioeconomic development and religiosity in a country (Inglehart and Norris 2003)—but where quotas are present there are additional mechanisms that raise the likelihood that women will be seen as desirable candidates. For example, quotas can give ammunition to feminist activists who support the selection of female candidates (Matland and Montgomery 2003, 33). If quotas initially lead to the success of some female candidates, other women’s candidacies would subsequently begin to seem more viable. Finally, comparative research has determined that when the number of female gatekeepers increases, as is the case when quotas apply to inner-party offices, the probability that women will be asked to run for elective office rises as well (Caul 1999; Cheng and Tavits 2011; Kunovich and Paxton 2005). Whatever their motivations for recruiting female candidates, gatekeepers desiring to locate women to run for office face a complicated task; almost everywhere there are fewer female than male aspirants available (Matland 2005; Lawless 2012). In contrast, male aspirants are plentiful and, on average, are expected to be less in need of encouragement to run. As a result, I hypothesize that gatekeepers are likely to ask the women within their ranks to run for elective office when quotas are implemented. Quotas may also alter the informal channels through which parties locate viable candidates. Because women often do not hold the traditional, gendered qualifications for elective office (Murray 2010b), quotas may force gatekeepers to develop other methods of identifying competent party members when they are faced with determining which female eligibles to select for the ballot. Moreover, candidates have traditionally been selected by male gatekeepers who occupy gendered social networks (KГјrschner 2009, 18; Sanbonmatsu 2006, 153). Because parties cannot change the recruitment environment and create women with traditionally male careers or circles of friends, gatekeepers seeking electable female candidates as a result of quotas are expected to begin casting a wider net in order to identify potential candidates. Simply asking women to run for office, however, will likely not be enough to generate the requisite number of candidacies because there are likely to be fewer women than men in party organizations and these women are likely to be, on average, less confident about their abilities than their equally competent male peers (Lawless 2012). In order to generate a sufficient number of aspirants, organizations required by a quota to locate a steady stream of female candidates are predicted to create new training and mentoring programs designed to make women more confident about Page 20 →their own abilities—efforts not previously required when parties primarily sought aspirants among their more-confident male members. In sum, quotas are expected to be very influential at this stage of political recruitment. Where gender quotas are implemented, party gatekeepers are likely not only to ask women to run for elective office but also to alter the ways in which they train and recruit candidates.

Candidates and Elected Officials These gatekeeper efforts to develop a cadre of promising female candidates should yield observable effects when it comes to the next phase of political recruitment: becoming a candidate. Here too quotas are expected to have a marked impact on political recruitment; however, exactly what this influence is hypothesized to vary with quota type in Germany. In parties with parity or near-parity quotas, mandating at least 40 percent female candidates, women are expected to enjoy numerical advantages in the candidate selection process because a gap exists between the percentage of women in the party and the (higher) percentage of women required by the quota. As a result, female party members are expected to not only be actively recruited by party gatekeepers but also, on average, even more likely than their male counterparts to become candidates. In contrast, where a 33 percent quorum or other below-parity quota is in effect, the gap between female party members and the percentage of female candidates required is not expected to be as large and, as a result, women’s political opportunities are expected to be more circumscribed. Such affirmative action policies require only one of every three (or more) ballot spots to be filled by women, leaving the higher-ranked ballot positions for men. As a result, on average, male party members are hypothesized to be significantly more likely than their female counterparts to have been selected as candidates. Quotas’ strongest impact is expected to come in the final phase of political recruitment: becoming an elected official. Women in parties with parity or near-parity quotas are anticipated to be more likely than their male counterparts to have been elected. Quotas should ensure that, although fewer women than men join such parties, and although these women, on average, find themselves in personal circumstances that depress political ambition, the women who do join will be promoted to positions of eligibility, encouraged to run, and enough will be placed in electable ballot slots to give women better odds than the men in their party of being elected. Where quotas require only 33 percent women, female eligibles—who are Page 21 →ensured only one of every three ballot nominations—are anticipated to be less likely to win elective office than their male counterparts. Summary Quotas are expected to have only limited effects early in the recruitment process, exerting but a partial sway on women’s propensities to become eligibles and aspirants. The mechanisms through which quotas increase the percentages of women in elective office are predicted to come later in the political recruitment process. Quotas are hypothesized to wield a far stronger impact on gatekeeper actions, increasing women’s odds vis-Г -vis their male colleagues of being asked to run for elective office and prompting gatekeepers to cast a wider net for candidates, redefine selection criteria, and implement formal programs to increase women’s confidence about their abilities. In addition, due to the above-mentioned elevator effect, quotas, particularly parity or near-parity quotas, should increase women’s odds of successfully pursuing candidacies and becoming elected officials when compared to their male peers. Thus, in the pathway case studied here, quotas, especially (near) parity regulations, are anticipated to be successful in shaping the later stages of political recruitment; however, more will likely need to be done in order for their goals to be reached earlier in the recruitment process, enabling female citizens to participate in democracies in equal rates to their male counterparts. Quotas are unlikely to have an impact early on in the political recruitment process until a gendered division of household labor is overcome and parties alter their day-to-day practices in ways more welcoming to women.

Sources of Data These expectations are tested empirically here using two original sources of data collected by the author: a mail survey sent to over one thousand members of the German candidate pool and 41 semistructured interviews conducted in person with potential candidates in Germany. The Candidate Interest Survey (CIS) targeted eligibles—political party members playing leadership roles at the grassroots level. In Germany, the lowest level of party organization is the precinct level (Ortsverein) and the most entry-level leadership posts available are positions on the local executive board (Vorstand): (co)chair, vice chair, secretary, treasurer, and board member.Page 22 → The German political party law12 requires these positions to be filled through democratic inner-party elections held at least every two years; winners are easily identifiable on party websites and in local media reports. These precinct-level boards are thus where the broadest possible sample of German eligibles can be

found. Precincts were randomly selected from within five representative German LГ¤nder: Baden-WГјrttemberg, Bayern, Bremen, Hamburg, and Nordrhein-Westfalen (see figure 5).13 These five states vary on a number of dimensions associated with political recruitment and quota enforcement, including religion, population density, regional political culture, and the dominant party in politics; local and state electoral rules also differ across these LГ¤nder (Davidson-Schmich 2006b). Bremen and Hamburg, located in the northern part of the country, are densely populated, largely Protestant city-states frequently governed by left-wing parties. Historically, they have been at the forefront of women’s descriptive representation in Germany. In contrast, Bayern and BadenWГјrttemberg are large, more rural, heavily Catholic states. They have traditionally been governed by socially conservative Christian Democrats and have exhibited low levels of women’s descriptive representation. Nordrhein-Westfalen is an intermediate case, located in the central part of the country and possessing a mixed record of women in elective office. It is roughly equally divided among Catholics and Protestants, and between urban cores dominated by the SPD and more bourgeois areas where the CDU is popular. Sampling from these states ensures that CIS respondents represent the diverse range of settings from which the parties studied here recruit their candidates.14 Within each precinct, a survey was sent to the highest-ranking male and the highest-ranking female member of the executive board of the Green, CDU or CSU, FDP, and SPD branch organizations. Of the 1,068 surveys sent, 465 were returned for a 45.2 percent cooperation rate.15 Respondents were almost evenly divided among the five states and among each type of affirmative action measure studied (no quota, 33 percent quorum, near-parity, and parity quotas). Respondents ranged in age from 16 to 80 with a mean age of 49 years; 48 percent were women. Approximately one-third each came from towns smaller than 20,000 residents, cities of over half a million, and locales of intermediate size. The average individual completing the survey had been a party member for 17 years, with a range of one to 52 years. Fifty-five percent had been elected at the local level but 45 percent had not. Thus the sample contains a wide range of variance in terms of the areas from which respondents hail, the quota rules constrainingPage 23 → their political careers, and their life experiences. The recipients of the CIS were overwhelmingly ethnically German, however; only 1 percent of those identified as local party leaders had names indicating a Turkish or other immigrant background. This low figure parallels the very small numbers of migrants in state and national party leadership posts (Donovan 2007) and among high-level elected leaders in Germany (de Fonseca 2011, 121; Donovan 2012, 35). Chapter 1 includes an extended discussion of how CIS respondents’ backgrounds compare both to party members as a whole in Germany and to the general population. Further details about the survey sample, the questions asked, and their descriptive statistics are outlined in the appendix. Fig. 5. Map of German LГ¤nder Personal, German-language interviews were conducted with interviewees whose political experiences ranged from never having joined a party or run for office despite an interest in politics to serving multiple terms in the Bundestag, Germany’s national parliament. All were (or could have easily become) eligibles for elective office by virtue of their active contributions to party or public life in Germany. Moreover, three-quarters of these interviewees were also gatekeepers for elective office because they held Page 24 →leading roles within their parties. Their ages ranged from their twenties to their seventies, and their places of residence varied from an Alpine village of 5,000 people to a major German city of well over 500,000 inhabitants. Although some interviewees were not members of any political party, most were affiliated with the parties surveyed. Interview questions focused on respondents’ political careers to date, their motivations for and against running for particular elective offices, their perceptions of the nomination process in their precinct, of campaigning, and of serving in government, as well as their future plans. Interviewees were also asked what they believed made a good candidate, which careers and family structures they deemed compatible with elective office, and what role certain characteristics such as gender played when their local party organization selected individuals to run for office. All interviewees were generous with their time and insights. In order to protect their privacy, I do not use their names here, but other identifying characteristics have not been altered. The appendix includes a list of interviewees’ backgrounds.

Finally, I supplement my original data with translations of German-language sources including academic studies, government and party documents, and media reports.

Plan of the Book I investigate the mechanisms linking gender quotas and women’s representation in elected bodies as follows. Chapter 1 provides some empirical background information about German parties’ voluntary gender quotas and women’s descriptive representation, the country’s recruiting environment, and its party-loyalist recruitment structures. Drawing on the Candidate Information Survey and German-language research, chapter 2 compares the percentages of men and women joining German political parties and ascending to the leadership positions that render them eligible to run for elective office. This chapter further takes an intersectional approach, comparing the demographic traits of the women and men who do join parties and become eligibles to the underlying German population. Chapter 3 then draws on personal interviews and the CIS to measure political ambition among eligibles, depicting males’ and females’ propensity to become aspirants for various elective offices. Using the same sources, chapter 4 delves into the behavior of party gatekeepers, studying both their inclination to ask women and men to run for elective office and the extent of their attempts to identify and prepare promising female party members for Page 25 →candidacies. Chapter 5 deploys CIS data to compare men’s and women’s experiences in seeking and receiving a ballot nomination and the relative frequency with which male and female eligibles are elected to public office. Chapter 6 summarizes the book’s empirical and methodological contributions and concludes by making both policy recommendations and suggestions for future research.

Page 26 →

One The German Political Recruitment Process I have only been able to have [a career in party politics] in part due to the fact that I am a woman and at times it was exactly a woman who was needed. —German chancellor Angela Merkel (quoted in Koelbl 2002, 53) Before I investigate the mechanisms linking German party quotas and women’s increased participation in elected bodies, some background information on the German political recruitment process is needed. This chapter first introduces the German parties and their quotas and then establishes that, since their introduction in the 1980s, these affirmative action policies have indeed increased the percentages of German women in elective office—especially at high levels of government. Moreover, the chapter documents a diffusion effect, whereby even parties without quotas have increased women’s descriptive representation, albeit to a lesser extent. The chapter then goes on to present each phase of the German political recruitment process including the elective offices available, the electoral systems used, the gatekeepers who select candidates, the traits selectors deem important to electoral success, and the individuals who are considered “eligible” to run for office. The remainder of the book then proceeds to examine how quotas have intersected with each stage of the above political recruitment process to produce the observed increase in women holding elective office.

Page 27 →Voluntary Party Quotas in Germany Germany’s party system features six salient parties with a range of positions on gender-related issues: three large and three smaller parties.1 The oldest party in Germany, and the largest on the left side of the political spectrum, is the Social Democratic Party (SPD). The party’s platform expresses support for gender equality and believes in the need for state intervention to assure job and educational opportunities for women, along with state-sponsored measures to reconcile home and family responsibilities (Xydias 2013, 8). Its counterweights to the right are the Christian Democratic Union and the Christian Social Union (CDU/CSU); the CDU contests elections in 15 of Germany’s 16 states while the CSU is present only in the state of Bayern. At the federal level these two parties combine to form a single parliamentary party group (PPG). The CDU/CSU has historically held very traditional views of women’s roles in society, although under the influence of the women’s auxiliary organization, the Frauen Union (Women’s Union), and the leadership of Angela Merkel, these views have modernized in the past decades, accepting for example that mothers might work outside the home (Wiliarty 2010). The Christian Democrats’ recent platforms continue to view women primarily through the lens of their role as mothers and family members, however (Xydias 2013, 8). Either the SPD or the CDU/CSU (or both) has been present in every state and federal Cabinet since the founding of the Federal Republic in 1948. As a general rule, the SPD is stronger in Protestant northwestern Germany and the CDU/CSU dominates the Catholic south (Green et al. 2008, 88). Founded with the creation of the Federal Republic, the libertarian Free Democratic Party (FDP) served as a “kingmaker” between these two larger parties in the latter half of the twentieth century, forming national coalitions with the CDU (on economic issues) and the SPD (on social issues); in 2013, for the first time since its creation, the FDP failed to surpass the 5 percent threshold necessary for inclusion in the Bundestag. The Free Democrats contest elections in all 16 states, although they have fallen below the 5 percent threshold on numerous occasions, especially in eastern German LГ¤nder. The FDP’s libertarian platform expresses support for gender equality, but, in contrast to the SPD’s position, rejects state-driven measures to achieve this goal, instead preferring to rely on private sector solutions or an end to gendered state regulations, such as a tax on dual-income families (Xydias 2013, 9). The Green Party emerged from 1970s social movements, including the Page 28 →feminist movement, and entered

the national stage in 1983. After merging with some eastern German citizen movements when Germany reunified in 1990, the party changed its name to Alliance 90/Greens. It governed Germany in coalition with the SPD from 1998 to 2005. The Greens are present in most western state legislatures but have achieved less electoral success in eastern Germany. Given its activist origins, the party assumes feminist stances on a range of gendered issues. In contrast to the parties discussed above, the Greens’ platform views women not just as family members or workers but also as individuals whose life chances are circumscribed by unequal gender roles in society. Alliance 90 calls for an end to women’s double burden of paid work outside the home and unpaid care work within it, focusing on equality of outcome, not simply equality of opportunity (Xydias 2013, 7). The newest party to enter the national parliament is The Left, which resulted from a merger of the heir to the East German communist party, the Party of Democratic Socialism (PDS), and dissident elements from the SPD. The Left is extremely successful in eastern German state and local elections (Green et al. 2008, 88)—essentially replacing the SPD as the largest party on the left in eastern LГ¤nder; however, it is a newcomer to state-level politics in western Germany and has yet to serve in government at the federal level. The Left party’s view of gender issues reflects its Marxist heritage; its platform stresses how neoliberal economic policies and patriarchy create unequal gender roles to the detriment of women and prescribes extensive state intervention to achieve both equality of opportunity and equality of outcome (Xydias 2013, 6). Since the 1980s, these six parties have adopted a range of policies regarding the sex of the individuals who may occupy both inner-party offices and ballot slots (see table 1.1). As has been common across western Europe, and in keeping with their stress on equality of outcome in gender relations, the postmaterialist Greens were the first of the German political parties to adopt a voluntary gender quota (Kittilson 2006).2 In 1986, three years after it initially entered the national parliament for the first time, with one-third of its PPG female, the Green Party adopted a formal parity quota, requiring women to occupy all odd-numbered list places; the only way a man can be assigned an odd-numbered ballot slot is with the permission of the majority of the female party members present when candidates are selected. Even-numbered positions can be contested by either male or female aspirants, permitting all-female electoral lists. This parity quota also applies to inner-party offices; top positions are to be held by male and female cochairs and all inner-party bodies are to be filled on a parity Page 29 →basis. The Greens’ attempts to promote gender equality go further than simply setting numerical targets, however. In order to counter male domination of discussions, male and female speakers must alternate at national party congresses and debates can be terminated at the point when no more women desire to speak. Moreover, in keeping with their stress on equality of outcome, the party statutes mandate that childcare be available at formal national party gatherings; lower-level branch organizations are urged to follow suit. Individuals holding top national positions in the party are provided with both childcare and eldercare subsidies to pursue their posts.3 As a result of their new quota, the Greens sent record numbers of women to the national parliament in 1987; following the election, 57 percent of their parliamentary party group was female. Prior to that, women had never made up more than 10 percent of the Bundestag as a whole, or more than 20 percent of any other party’s PPG (Schindler 1999, 636–37). As the Greens made strides toward women’s descriptive representation, more and more women began to vote for the party rather than either the SPD, the beneficiary of younger women’s votes during the 1970s, or the CDU/CSU, which had benefited from women’s support during the postwar period (Kittilson 2006; Kolinsky 1989, 1993). The next organization to adopt a gender quota was the Social Democratic Party. The SPD’s motive for adding a gender quota to its statutes was less ideologically driven than the Greens’—instead, strategic electoral considerations were at the forefront. Members of the SPD’s women’s auxiliary organization, the Arbeitsgemeinschaft Sozialdemokratischer Frauen (ASF), had been pressing their party to adopt a quota since 1977 but had met with little success. Between 1984 and 1986 the party agreed to promote women into leadership positions within the organization and give preference to women when ballot nominations were drawn up, but these measures were simply recommendations, not requirements. Only after the 1987 election, when the ASF could document the SPD’s electoral losses to the Greens, did the party agree to a binding quota for women (Kittilson Page 30 →2006; Kolinsky 1993). Between 1988 and 1994 the SPD phased in a near-parity quota, which ultimately required that no fewer than 40 percent, and no more than 60 percent, of their candidates be of a single

sex. This regulation was combined with a placement mandate stipulating that for every five places on the party’s list, the first and the third would go to one sex, the second and the fourth to the other, and the fifth place could be held by a candidate of either sex. In 2011 the SPD’s quota for the Bundestag and European Parliament lists was increased to a parity quota, alternating men and women down the entire list. The SPD’s statutes include no allowance to violate these quotas. These rules also apply to all inner-party offices and leadership positions. The party statutes do not specify any other policies, such as childcare or opportunities for women to take part in discussions, to promote inner-party gender equality.4 Table 1.1. German Parties and Their Quotas Party Quota Green Party (Bündnis 90/die Grünen) ≥ 50% women The Left Social Democratic Party Christian Democratic Union Christian Social Union Free Democratic Party

50% women 40%–60% women 33% women no quota no quota

After German unification in 1990, the PDS entered the German party system and began to contest democratic elections using a parity quota similar to the Greens’, except that women are to receive one of the first two places on the list, and then all the odd-numbered slots—as long as candidates can be found. Just as with the Green’s, the PDS’s quota did not guarantee men any nominations and all-women’s lists were permitted. The quota also applied to all inner-party posts and bodies; if a “woman’s” position in the party was left vacant or filled by a man, and subsequently a woman desired to occupy it, a new election could be held at any time. In 2007 the PDS merged with some breakaway SPD members to form the Left Party, which continued with the same parity quota.5 The PDS/Left’s motivation for quota adoption was similar to the Greens’ in that it was primarily ideologically driven, designed to undo the subordinate position of women in a capitalist, patriarchal society.6 To this end, the party’s statutes require free childcare at all federal-level party meetings and alternating opportunities for men and women to speak at these gatherings. As elsewhere in western Europe (Matland and Studlar 1996; Kittilson 2006), German quotas began to diffuse across the political spectrum. In the 1970s and early 1980s women in the CDU’s Frauen Union were not as vocal as their counterparts in other parties and the Christian Democrats did not face inner-party pressure for a quota as the SPD had. Only later in the 1980s did a generational change occur in the Women’s Union and an influx of professional women began to express interest in holding elective office (Kittilson 2006; Wiliarty 2010). In 1986 the CDU set voluntary targets for women’s representation, calling for women to occupy inner-party posts in proportion to their membership within the party and demanding that women appear on the ballot in electable positions. In 1988 Page 31 →the party agreed to consult with the Women’s Union when selecting candidates. By the early 1990s, however, it became clear that these measures were not increasing the percentages of women in elective office; after the 1994 national election the proportion of CDU women elected to parliament actually declined. Moreover, the Christian Democrats continued to lose female voters to parties employing quotas. As a result, the Frauen Union called for a binding 33 percent quota for women in both inner-party and elective offices. This measure was rejected by rank and file members at a 1995 party congress, much to the dismay of Chancellor Helmut Kohl who favored the policy in order to improve the CDU’s electoral fortunes. Driven by a desire to attract female voters, party leaders revamped the proposed quota into a nonbinding “quorum,” specifying that one of every three party offices and list places should be given to a woman. It did not specifically require that the other two slots be reserved for men, so in theory women could campaign for these positions too. However, all women’s lists seemed unlikely because, in contrast to the original quota proposal, the quorum could be violated if a majority of all party members voting on an officer or candidate slate—not just the female members—agreed to do so. This weakened form of affirmative action was ultimately passed at the 1996 CDU party conference following appeals from top leaders including Chancellor Kohl and Bundestag president Rita

SГјssmuth (Wiliarty 2010). Thus the quorum is both lower and easier to violate than the other parties’ (near) parity quotas. Just as the SPD did, the CDU adopted affirmative action measures for women primarily due to electoral—rather than ideological—concerns. At the time the Candidate Interest Survey was conducted, the CSU and the FDP did not employ candidate gender quotas. Like the SPD and CDU before it, however, the CSU performed poorly electorally without a quota and, in an attempt to modernize the party’s image, the party leadership pushed through a quota for inner-party offices in late 2010 (Issig and Vitzthum 2010). The Christian Social Union’s current quota is far more limited than the other parties’, however, requiring 40 percent women on inner-party executive boards at the regional level and above; there is no quota for candidates or for lower-level inner-party offices.7 As of 2015, the Free Democrats remain the sole quota-less organization among Germany’s major parties. In the late 1970s and the early 1980s, the FDP had the highest percentage of women in their Bundestag delegation of all parties, and at the same time as the Greens established their quota the Free Democrats took several informal steps to increase women’s presence within the party, including changing the timing of congresses to be more Page 32 →family friendly and establishing an annual report on the status of women in the FDP. The latter practice was adopted by other organizations and data from such reports helped justify the SPD’s and CDU’s quotas (Kittilson 2006; Kolinsky 1989, 1993). The FDP’s own statistics in the late 1980s and early 1990s indicated that women in the organization had made few inroads into either party or elective office. However, rather than adopting quotas, which was viewed as too bureaucratic a solution, in 1995 the party again attempted to use quota-less measures to improve women’s position within the organization. By the early 2000s the Free Democrats had still failed to increase women’s descriptive representation by much and, as a result, in 2003 the FDP launched a Women’s Campaign to attract more women to the party, on the assumption that more female members would lead to more female leaders (McKay 2004); this attempt was followed by a similar drive in 2006. In addition, in 2008 an official mentoring program was begun to assist women in scaling the ranks of the party. None of these efforts proved very successful, however, and, in 2010, the FDP’s women’s organization, the Liberale Frauen (Liberal Women), pledged to seek new ways of advancing women’s status within the party (Liberale Frauen 2010). At the 2012 party congress the Liberal Women submitted a proposal calling on the party to adopt binding measures to increase women’s descriptive representation. The proposal failed, however, and the FDP remains without a quota. German parties, in sum, feature an array of policies regarding women’s presence in inner-party bodies and on party lists, ranging from parity (Greens, Left), to near-parity (SPD), to 33 percent (CDU), and at the time research for this project was conducted, to no quotas (CSU and FDP). In contrast to countries where legislative gender quotas are in effect (Krook 2009), however, these affirmative action policies are simply voluntary promises, and no election authority checks lists to see whether quotas are implemented before accepting ballot nominations. Similarly, there are no state-sanctioned punishments for parties failing to fulfill their promises. The closest Germany comes to a legal mechanism for enforcing quotas is its Party Law (Parteiengesetz), which requires political parties’ internal operations to be democratic. Each party is legally required to draw up a statute and adhere to it, so parties that have formal quota policies are legally bound to abide by them. However, with one exception, all the quotas studied here contain loopholes.8 The Greens’ quota can be ignored if the women selecting inner-party officers or candidates approve; this occurred, for example, in the 2002 and 2005 federal elections when Green Page 33 →foreign minister Joschka Fischer was placed at the head of the party’s list rather than a woman (Walker 2012). The Left Party’s quota does not have to be implemented if a woman cannot be found for a particular spot, and the CDU’s quorum can also be ignored if a majority of electors choose to do so. Only the SPD’s statutes do not permit any exceptions. Their women’s auxiliary organization has indeed reported violations of the quota to the Social Democratic Party’s adjudication body; these have been corrected internally before courts of law had an opportunity to intervene (e.g., “Wolfgang Rose” 2007). Members of the ASF report using the threat of the Party Law to ensure quota compliance. Empirical research has found that, although they are voluntary, Germany’s party quotas are usually implemented in high-level elections. For example, in the 2013 European Parliament elections all parties adhered to their statutes when putting together party lists (Bundeswahlleiter 2013a). In the 2009 and 2013 German federal

elections, parties adhered to their quotas in over 95 percent of their list nominations (Davidson-Schmich 2010; Bundeswahlleiter 2013b). One study of variance in Land-level quota implementation during the 1990s found that quotas were implemented approximately two-thirds of the time (Davidson-Schmich 2006b). In addition, the women’s auxiliary organizations that monitor their own parties report that quotas are fulfilled for top-level inner-party posts as well (BГјndnis 90/GrГјne 2011; CDU 2010; Die Linke 2011; Ferner 2011). For lower-level elective elections, however, parties adhere less often to their own quotas. For example, in the largest cities in Germany, the Left Party, the SPD, and the CDU met their quotas in less than half the municipalities; the Greens complied with their own regulations in only 55 percent of the metropolises (Holtkamp; Wiechmann and Schnittke, 2009, 46; see also Eder and Fortin-Rittberger 2013). The parties’ own internal investigations also find weaker quota application for lower-level party posts (CDU 2010; Die Linke 2011; Ferner 2011). These results clash with the conventional wisdom that a “law of increasing disproportion” exists, whereby the more powerful the position, the fewer members of minority groups will be present (Putnam 1976).9 Despite these shortcomings in filling quotas at the local level, German parties’ voluntary quotas have indeed increased the numbers of women in elective office, as is the case elsewhere in the world (Thames and Williams 2013). The remainder of this book discusses the effects that these quotas have had on political recruitment in Germany and explains the lesser representation of women at lower levels of government. Before doing so, it is first Page 34 →necessary to describe the Federal Republic’s political recruitment process itself. Below I depict, in reverse order, each stage of the progression from eligible to elected official.

Female Elected Officials in Germany As might be expected from the above discussion on quota implementation, these affirmative action measures have been effective in increasing women’s descriptive representation over the past three decades, especially at high levels of government where relatively few female Members of Parliament (MPs) are required to fulfill a quota. At lower levels of government, where far more elected offices must be occupied by women in order to comply with quotas, however, the latter measures have been less effective in bringing women into public positions, although they have increased women’s numbers (see figure 1.1). The evidence presented in this section also documents a diffusion effect, indicating that parties without quotas now send more women to elective offices than they did before other parties adopted affirmative action measures. In order to establish that quotas have in fact increased women’s descriptive representation it is necessary to first depict the positions to which individuals can be elected in the Federal Republic. Unlike the reserved-seat quotas used in countries such as Taiwan and Uganda (Krook 2009), the voluntary quotas described above did not create special seats set aside for women but were designed to improve women’s chances of being elected to already-existing offices that can legally be occupied by either men or women. Germany is a federal system and member of the European Union (EU) and, as such, elects candidates to four levels of government, including local, state, federal, and European parliaments (see table 1.2). Levels of local government can include cities/towns, counties, and regions, depending on the state. Germany’s Party Law requires that parties are governed in an internally democratic fashion with opportunities for members to take part in decision making. In addition to electing a national executive board at least every two years, each organization must select a state level leadership and—with the exception of three city-states—each state-level party organization is obligated to form smaller units that also regularly choose slates of officers. Generally, these substate organizations parallel the elective bodies depicted above, so that for every council or parliament there is generally a parallel party organization. Page 35 → Fig. 1.1. Number of Positions and Women’s Descriptive Representation: 2010. (Source: Deutscher StГ¤dtetag 2012.) Table 1.2. Elected Offices and Electoral Systems in Germany

Level of Government Local

High

Type of Elective Electoral System Used Office Town/City PR (14 states; 12 have open Council lists) County Mixed (2 states) Council Regional Council

Do Quotas Apply? Yes, to original list Yes, to list portion

Professional Position? No

Mayor Plurality No Depends on community size State Mixed (13 states; 1 with open Yes, to original list portion Yes Member of list) Yes, to original list (13 of 16 States) Parliament PR (3 states; 2 with open list) portion (Landtag) Federal Mixed Yes, to list portion Yes Member of Parliament (Bundestag) European PR Yes Yes Member of European Parliament

Source: www.wahlrecht.de. Page 36 →One of the main traits sought in candidates for public offices in Germany is prior experience in a lower-level elective position (Conradt and Langenbacher 2013, 191; Kintz 2011), and as such these four tiers of government represent interconnected rungs on a career ladder. Local-level positions are primarily amateur posts occupied by what Germans call Feierabendpolitiker or “after-work politicians,” while being elected to a higher level of government constitutes full-time employment. Most members of the eligibility pool retain their day jobs and do not ascend the level of, as Max Weber put it, politics as a vocation ([1919] 2008). Those who do, however, must first establish a rГ©sumГ© in low-level posts. Below I portray each step of the career ladder in turn. Local-Level Elected Offices Each of the Federal Republic’s sixteen states, or LГ¤nder, features a range of substate elective offices. The lowest level of government includes elected local councils, usually called the Gemeinderat or Stadtrat, depending on whether the body is in a small town or a large city. The executive branch at this level includes a mayor (BГјrgermeister). Small-town mayors are often indirectly elected members of the council while larger cities’ majors are frequently elected directly (Wollmann 2004). With the exceptions of Germany’s three city-states (Berlin, Hamburg, and Bremen), the LГ¤nder also feature elected county-level councils (Kreistage). In some states, the county mayor (Landrat) is also directly elected. The state with the largest area, Bayern, elects regional councils as well (Bezirkstage). Given Germany’s multiple political parties and its relatively proportional electoral systems, local councils there have more seats on them than do typical city commissions in the United States. For example, the U.S. city South Miami, Florida, and the German city Dossenheim, Baden-WГјrttemberg, both have approximately 12,000 residents, but the South Miami City Commission is made up of only 4 members whereas Dossenheim’s Gemeinderat contains 22 seats. On average, there is one local council seat for every 1,800 inhabitants in Germany. Moreover, because multiple parties contest these many seats, an even larger number of candidates are required each electoral cycle. Thus there are many positions to which an aspirant can be elected, and there are even more

opportunities to appear on the ballot as a local-level candidate in Germany. This means that, in order to fill their quotas, parties must locate many women willing to run for office in every community. The total number of local-level elected officials—and the correspondingPage 37 → ratio of officials to the underlying population—varies from Land to Land depending how many levels of government are present. For example, the small, densely populated city-state of Hamburg does not elect county or regional councils. There are only 361 elected local officeholders in this state of almost two million residents, or one local-level elective office for every 4,988 inhabitants. In contrast, Germany’s geographically largest state, Bayern, features multiple levels of local government including town, county, and regional councils. There are 19,320 local-level elected positions in this state of almost 13 million, or one for every 646 residents. In light of this considerable state-tostate variation in political opportunity structure, later analyses control for an eligible’s state of residence. The ratio of elected officials to citizens also varies with each Land between densely populated urban centers and more sparsely settled rural areas. For example, in the state of Baden-WГјrttemberg, the large city of Stuttgart has one city council seat per 10,000 inhabitants and the small town of Biberach has one city council seat per 1,000 residents. This occurs because the councils are roughly the same size, but Stuttgart is almost twenty times as large in terms of population. Again, then, later analyses control for the size of the community in which an eligible resides. The introduction of party quotas had a positive, but limited, impact on women’s local-level descriptive representation in Germany (see figure 1.2). In 1985, before the Greens adopted their quota, women held 14.4 percent of local council seats in Germany (Hoecker 1987, 60). By 2010, when four parties utilized quotas, this figure almost doubled to an average of 25.2 percent. While parties employing quotas had an average of 27.1 percent women in their local council delegations in 2010, only 21.0 percent of the local councilors from parties without quotas were women (Deutscher StГ¤dtetag 2012).10 The latter figure is only 7 percent above that in the 1980s. There is also a clear correlation between women’s representation and community size. Germany’s smallest towns have an average of only 22.3 percent women on their local council whereas the largest urban areas’ city commissions contain on average 37 percent women (Deutscher StГ¤dtetag 2012). See table 1.3. High-Level Elected Offices This relationship continues up the career ladder; the fewer elected positions available, the higher the percentage of women elected and the more successful quotas have been in increasing women’s descriptive representation. In addition to the local-level positions depicted above, there are three Page 38 →other legislatures to which German citizens can be elected, referred to here as “high-level” elected offices. Fig. 1.2. Average Percentage of Local Council Seats Held by Women: 1985–2010. (Note: N Positions 2010: 45,700; includes only communities of 10,000+ citizens; no data on gender and party available prior to 2002; quotas/ no quotas figure excludes independent voter groups.) Table 1.3. Local Council Seats Filled by Women: 2010 Local Population Number of Positions to Fill % Filled by Women (2010) 10,000–20,000 19,692 22.3% 20,000–50,000 15,733 25.4% 50,000–100,000 4,796 27.5% 100,000–500,000 3,747 32.2% 500,000 + 1,205 37.1% Source: Deutscher StГ¤dtetag 2012. First, Germany is broken down into 16 states and each has its own parliament or Landtag.11 Together these legislatures contain slightly over 1,800 members (Mitglieder des Landtages or MdL). Here again, quotas have exerted a clear positive impact on women’s descriptive representation. While only 11.5 percent of state

legislators were women at the time the Greens adopted their quota (Hoecker 1987, 63–67), by 2010 this figure had almost tripled to 32.2 percent (see figure 1.3). Of the Landtag seats held Page 39 →by parties with quotas, 34.6 percent were occupied by women compared to 19.8 percent of the seats won by quota-less organizations. While this latter figure is clearly lower than that for parties with quotas, it is still almost double the percentage of women these parties sent to state legislatures in the prequota era, suggesting quotas have had a diffusion effect at this level. The ideology-transcending effects quotas have on women’s descriptive representation can be clearly seen when comparing the CDU’s and CSU’s track records of electing women to state legislatures (see figure 1.4). While both parties sent similar percentages of women to the Landtag prior to the CDU’s quorum, the latter steadily increased Christian Democratic women’s presence at the Land-level while CSU women enjoyed far fewer gains until their party adopted a 40 percent quota for inner-party offices in 2010 (and party leaders pushed for a candidate quota as well). Second, at the national level, Germans can be elected to the federal parliament called the Bundestag.12 This parliament’s composition varies depending on election results but it has at least 598 members (Mitglieder des Deutschen Bundestages or MdB). As at the state level, the introduction of quotas markedly changed the composition of the Bundestag (see figure 1.5). Prior to the Greens’ quota adoption, fewer than 10 percent of MdB were women, but by the time the CDU’s quorum went into effect this figure had tripled. After the 2009 election, 36 percent of the seats held by parties employing quotas were filled by women in contrast to only 23 percent of the mandates won by quota-less organizations. While below that of parties with quotas, this latter figure is more than double than that of the prequota era, suggesting that quotas have had a diffusion effect on other parties at the federal level as well. At this level, too, the impact of a quota on the two ideologically similar parties can be observed. In the 2013 Bundestag election, 50 percent of the MdB elected via proportional representation (the tier where quotas are employed) were women compared to only 21 percent of the deputies elected through the first-past-the-post tier (where quotas are not used) (Davidson-Schmich 2014). Similarly, since the adoption of its quota, the CDU has consistently sent more women to the Bundestag than the CSU (Wiliarty 2013). Third, because Germany is a member state of the European Union, it currently elects 96 Members of the European Parliament (MEPs).13 Here, too, quotas have had a positive impact on women’s descriptive representation (see figure 1.6). In the first direct European Parliament (EP) election in 1979 only 15 percent of Germany’s delegation to the EP was female; as early as 1989 this figure doubled to one-third, where it remains. At this highest tier of the German political system, quota diffusion can most be Page 40 →clearly seen; in 2014 38 percent of the EP seats held by parties employing quotas were occupied by women and 32 percent of those held by quota-less parties were as well. Thus, while quotas have clearly had a positive impact on women’s descriptive representation on all rungs of the career ladder, their effect has been most pronounced at the highest levels. Similarly, quota diffusion is much more marked at higher levels of elective office. Fig. 1.3. Percent of State Legislature Seats Held by Women: 1985–2010. (Note: N in 2010: 1859. Sources: 1985: Hoecker 1987, 63–67; 1990: Author’s calculations based on Hoecker 1995; 1995: Hoecker 1995, 117–33, excludes Berlin; 2000: author’s calculations based on Davidson-Schmich 2006b; 2005 and 2010: author’s calculations based on parliamentary websites.) Fig. 1.4. Percent of CDU and CSU State Legislature Seats Held by Women: 1985–2010. (Sources: See fig. 1.3.) Clearly, then, the adoption of gender quotas in Germany has increased women’s descriptive representation—as quotas have done in settings across Page 42 →the world. As argued in the introduction, to understand how quotas have produced this outcome—and to explain why quotas have been less successful at lower, less powerful levels of government—one must examine how quotas impact each phase of the political recruitment process. Forthcoming chapters devote themselves to this task. Before turning to this subject, however, I continue delineating Germany’s system of political recruitment, depicting how candidates become the elected officials just described. Quotas have changed some, but certainly not all, of this process. Page 41 → Fig. 1.5. Percent of Bundestag Seats Held by Women: 1985–2010. (Note: N Seats 2010: 620. Sources: 1985, 1990, 1995: Schindler, 1999, 635; 2000: Feldkamp 2005, 158; 2005 and 2010: Davidson-Schmich and Kuerschner 2011, 26.) Fig. 1.6. Percent of Germany’s Seats in the European Parliament Held by Women: 1985–2010. (Note: N seats 2010: 99. Source: 1985, 1990, 1995: Hoecker 1995, 150; 2000: Holzapfel

and Loeffler 2000; 2005 and 2010: Bundeswahlleiter 2013a.)

How German Candidates Become Elected Officials By definition, in order to ascend to elective office a citizen must be elected by her peers. Here I describe the electoral systems that are used to choose the representatives outlined above and discuss the gendered implications of these rules. While there are some variations in the electoral systems used across offices and states, political parties are of central importance in every instance. In fact, parties play such an important role that Germany’s political system has been termed the “party state” (Conradt and Langenbacher 2013, 135). This is the reason the Candidate Interest Survey targeted party leaders. The German Personalized Proportional Electoral System The most frequently employed electoral system in Germany is the personalized proportional, or mixed, electoral system, which is used, in slightly different versions, to select members of the Bundestag, 13 states’ legislatures, and local councils in two states (see table 1.2).14 In this system, citizens receive two votes; the first is awarded to a single candidate running for a constituency seat elected via plurality, and the second vote is given to a list of candidates selected using proportional representation (PR). The two tiers of this system are linked, meaning the overall makeup of elected bodies parallels the results of the second vote. In addition, dual candidacies are possible and quite common. The net effect of this electoral system is similar, but not identical to, that of proportional representation. Elections are largely depersonalized and vote choices are driven primarily by the voter’s view of political parties rather than individual candidates (Conradt and Langenbacher 2013, 190). With their first vote in the personalized proportional electoral system, citizens pick a single candidate running in the constituency in which they Page 43 →reside; the candidate receiving a plurality of the votes wins the direct seat for that district. Each party is allowed to nominate one candidate per district for this section of the ballot and the quotas described above do not apply to this part of the ballot. In fact, all but one of the major parties’ statutes mentions nothing about the sex of the candidates to be selected here. The exception is the CDU charter’s admonition that, when nominating directly elected candidates, the party should “work toward the sufficient participation of women” (quoted in Davidson-Schmich 2010, 158); however, the statute specifies neither how gatekeepers should go about doing this nor how many women would suffice. It is technically possible to appear on this part of the ballot without being nominated by a political party. All that is needed is 200 signatures of eligible voters in the constituency. Nonpartisan candidates may select their own keyword to appear on the ballot in place of a party name. However, were they to be elected, nonpartisan candidates would be largely impotent in the Bundestag as virtually all resources and powers in the parliament are awarded to parliamentary party groups (Davidson-Schmich 2006a, 54). In addition, German parties receive generous public funding to finance their campaigns (Green et al. 2008, 89) and enjoy “brand name” recognition, making it difficult for nonpartisan candidates to successfully contest an election (Davidson-Schmich 2006a, 26). As a result, all candidates ever elected to the Bundestag via the first vote have run with a party label. The so-called second, and more important, vote in the mixed electoral system is for a party that is running a closed list of candidates; the federal parliament is ultimately composed in proportion to the percentage of the second vote parties receive. The chancellor and her cabinet are in turn chosen by the party or parties that have the majority in the Bundestag. At the federal level, only political parties are allowed to place candidates on this part of the ballot and the above-mentioned quotas apply here. Thus to appear on the Bundestag ballot for the second vote, and to have a chance of winning via the first vote, a candidate must be nominated by a political party. Since the adoption of quotas, women’s candidacies on the first and second parts of the ballot have increased at similar rates (Fortin-Rittberger and Eder 2013); however, the two components of this mixed system are gendered in different ways and women are more commonly elected through the proportional part of the ballot (Iversen and Rosenbluth 2010). This is in part because the party quotas discussed above apply to PR lists (the second vote) but not to the constituency nominations (the first vote). As a result, far more women are elected from party lists than enter elective office through directly elected seats. For example, in the 2013 Bundestag elections,

Page 44 → 50 percent of those elected from party lists were women but only 21 percent of constituency seat winners were female (Davidson-Schmich 2014; see also Davidson-Schmich 2010). Here again, the positive effect of quotas on women’s descriptive representation can be observed. This also explains why in settings where mixed electoral systems are used—the Bundestag and many state legislatures—women’s descriptive representation has plateaued below the level of the Greens’, Left Party, and SPD’s quotas. Thirteen LГ¤nder use variations of this mixed system to elect their state legislatures, which in turn indirectly elect the state’s executive branch. While most versions are similar to the federal electoral system, Bayern utilizes open lists for the second vote and voters select one party member from the list rather than the list in its entirety. Here too more women enter the state legislature through the PR portion of the ballot (for which quotas apply) than through the plurality component (Hennl and Kaiser 2008). These state-level electoral systems are somewhat more open to nonpartisan candidates than the federal electoral system. Nonpartisan candidates may also appear on the first portion of the ballot—although they often must collect a certain number of valid signatures to do so, whereas parties face no signature requirement. In contrast to federal laws, some (but not all) states’ laws allow groups of independent voters (WГ¤hlergruppen) to place nonpartisan lists of candidates on the second portion of the ballot. Even where they are permitted, however, nonpartisan candidates or lists face hurdles to election similar to those described above: the state-level executive branch is chosen by the majority parliamentary party group(s) and independent deputies in the state parliament have virtually no power. Political parties receive government funding to contest electoral campaigns and enjoy brand-name recognition. Thus nonpartisan candidates and WГ¤hlergruppen are very rarely elected to state legislatures, meaning that to become a successful state-level candidate in Germany an individual must receive a political party’s nomination just as at the federal level (Davidson-Schmich 2006a). Proportional Representation in Germany The second most common type of electoral system used in Germany is proportional representation (PR), variations of which are used to elect Germany’s representatives to the European Parliament, members of three states’ legislatures, and local councils in 14 states. To elect its members of the EP, Germany uses closed-list PR.15 Each voter receives a single vote that she may cast for an unalterable list of candidates put forth by a political Page 45 →party. Germany’s 99 seats are distributed to the parties in proportion to the percentage of the votes that each receives; the seats go to the top-ranked candidates on the party’s list. Only political parties may put forward lists of candidates for European elections. This same electoral system is used to select members of Saarland’s state legislature. In the states of Hamburg and Bremen a form of open-list PR is used in statelevel elections.16 All substate electoral systems involve some form of proportional representation for electing town, county, and regional councils. Local electoral systems differ from state to state in the number of votes a citizen receives and whether these votes are cast for closed party lists or individuals on a party list. In Berlin and Saarland the local electoral system parallels the PR system used to elect members of the EP: voters can cast a single vote for a closed party list. In twelve other states open lists are utilized and, rather than receiving a single vote to cast for a closed list, voters receive either three votes, five votes, or as many votes as there are seats on the council. They may then distribute these among individual candidates in any way they choose. For example, a citizen could give all her votes to a single candidate, one to each candidate on a single party’s list, or several votes to her favorite candidates across parties. Parties then receive seats on councils in proportion to the percentage of the overall votes that went to the members of their lists, and the parties’ allocated seats are then distributed to the most popular candidates on the list. In almost all states, nonpartisan lists are permitted in local elections. Especially in southern and eastern Germany, such lists commonly do appear on the ballot and win seats in local governments. Each unique WГ¤hlergruppe has its own policy regarding gender and candidate selection. In some communities feminists have formed independent women’s lists (Frauenlisten). In such locales, these all-women’s lists have had an important influence on getting women elected at the local level and spurring other parties in the town to promote female candidates as

well (Holuscha 1999, 201–3). On balance, however, Germany’s WГ¤hlergruppen are very male dominated and, even though some electoral lists contain only women, across Germany as a whole independent voter groups send lower percentages of women to local councils than even parties without quotas (Deutscher StГ¤dtetag 2002).17 If an individual harbors political ambitions beyond the local level, affiliation with a nonpartisan group may be less attractive than beginning a political career with a political party that is also represented at the state, national, and European levels. Running as a member of a political party rather than as a nonpartisan candidate also offers a candidate more resources with which to finance a campaign because German parties, but Page 46 →not nonpartisan groups, enjoy federal funding. Indeed, even at the local level WГ¤hlergruppen hold only 13 percent of the seats (Deutscher StГ¤dtetag 2011, 107). Because nonpartisan lists do not play a significant role in the Federal Republic’s state, national, and European politics the remainder of this book focuses solely on parties. PR electoral systems such as those described above have been found in cross-national research to be the most conducive to women’s descriptive representation (Salmond 2006) and Germany is no exception, even beyond the Bundestag. States and localities using PR electoral systems exhibit higher percentages of women than do states and localities where elections are conducted using the mixed system (Holtkamp, Wiechmann, and Schnittke 2009). Similarly, Germany sends higher percentages of women to the European Parliament, elected through pure PR, than to the Bundestag, elected through a mixed electoral system. As a result of these varying Land electoral systems, state-level controls are included in the analysis conducted later. However, PR systems are no guarantee of women’s descriptive representation (Salmond 2006), as can be observed at the local level in Germany. Although all substate electoral systems involve some form of PR, on average women’s descriptive representation is lower in these bodies than in the Bundestag and the state legislatures elected through mixed electoral systems. In some contexts open lists such as those used in some German communities have been found to disadvantage women as sexist voters push women down the list in favor of men, but empirical research in Germany and other contexts of high gender equality finds this not to be the case (Valdini 2012). Instead, in communities above 5,000 people, on average female candidates are more likely than male candidates to be bumped up the list into more advantageous positions (Holtkamp, Wiechmann, and Schnittke 2009). It is thus unlikely that electoral laws are responsible for the variance in quota implementation between higher levels of government and local councils. Plurality in Germany The electoral system used does likely explain some of the low levels of women elected at the local level, however. In some communities a single-member-district plurality electoral system is employed to directly elect mayors (Wollmann 2004). In these contests party quotas do not apply, with predictable results: only 9.2 percent of Germany’s 1,306 elected mayors in 2010 were women (Deutscher StГ¤dtetag 2012). However, mayors were only 3 percent of local officials in Germany. Page 47 →While mayoral candidates may be nonpartisan, BГјrgermeister must work with a council made up of people elected via a party or WГ¤hlergruppe list, however, and thus are generally affiliated with one party or independent voters’ group. The former are better able to finance mayoral campaigns. Here again, then, if an individual harbors political ambitions beyond the local level, affiliation with a nonpartisan group may be less attractive than beginning a political career with a political party. In sum, the vast majority of German officials are elected to posts where quotas apply to some or all of the ballot nominations. The most striking aspect of the German political system from the perspective of a potential candidate is the importance that being selected as a candidate by a political party plays in making or breaking a political career; unless an aspirant is selected as a candidate by a political party, a political career beyond the local level in Germany is unthinkable.

German Gatekeepers and How They Select Candidates

Who Are the Gatekeepers? Given the above electoral system, then, German political parties determine who appears on the ballot under their banners and therefore are the gatekeepers in this system of political recruitment. Hence feminists interested in increasing women’s political representation in Germany pressed for party quotas as opposed to other measures (Kolinsky 1993; Kittilson 2006). The following section describes the formal process of candidate selection and then turns to the informal aspects, maintaining a focus on how political party gender quotas relate to candidate selection. While quotas require that gatekeepers select female candidates for certain positions, they do not change how candidates are selected. I draw here on both official laws governing candidate selection and personal interviews depicting how these laws are applied in practice. The Candidate Interest Survey provides additional data. In comparative perspective, German candidate selection is very bureaucratic and relatively decentralized. The former characteristic has been shown cross-nationally to be favorable to women (and other traditional political outsiders) as it allows aspirants to identify the steps they need to undertake to become a candidate. Decentralization in candidate recruitment, in contrast, has been argued to hinder women’ chances as it allows many possible veto points where women’s candidacies can be thwarted Page 48 →(Caul 1999; EscobarLemmon and Taylor-Robinson 2008; Siavelis and Morgenstern 2008; Hazan and Rahat 2010). Germany’s election laws clearly delineate nomination rules in an attempt to prevent the kind of undemocratic candidate selection prevalent prior to World War II; such detailed regulations are not typical of other western European countries (Roberts 1988, 97).18 The law stipulates that—for all levels of government—a party’s list candidates (whether for an open or a closed list) and its nominee for a direct seat must be selected via a secretballot vote taken during a meeting open to all party members eligible to vote for that list or seat in the upcoming election, or at a meeting of delegates who themselves were selected at a meeting open to all party members eligible to vote in the upcoming election. Delegates may be used to select candidates in any organization where there are more than 250 party members; quotas apply to the selection of delegates as well, ensuring women’s inclusion in this phase of candidate recruitment in parties where quotas are used.19 Thus local council candidates are selected by members of the precinct-level organization, county candidates by party members in the county or delegates thereof in larger counties, and state, federal, and European level list candidates are chosen at state-level meetings of delegates. Directly elected mayors, state, and federal candidates are selected by party members in the electoral district in which they stand. Technically, then, rank and file party members are gatekeepers as they have the final say in candidate selection, rendering it a relatively decentralized process. Any party member may suggest a candidate at the selection meeting and each candidate must be given adequate time to introduce herself and her ideas to those assembled. The only candidates the law rules out are those who are members of another political party or individuals who are not eligible to vote in the election. If a dual ballot is used, people may, and often do, appear on both sections of the ballot. Once the candidates have been chosen, the head of the party organization must submit nominees’ names to election officials along with evidence of the time and place of the selection meeting, the invitation to that meeting, the number of party members who attended, and the outcome of the votes taken. The party leader must sign the proposed ballot and be able to prove that she herself was democratically selected by her own party. Meetings to select the single candidate for the first section of the dual ballot, or for any directly elected mayoral post, then, formally involve the candidates for the nomination verbally presenting their qualifications to the members or delegates present followed by a secret ballot vote to choose the candidate. In practice, however, interviews indicate that much inner-partyPage 49 → politicking goes on before the selection meeting, with ambitious candidates making the rounds of party members in an attempt to win their support in advance. An interested candidate who does not anticipate winning the party’s nomination may ultimately decide not to attempt to pursue it at the selection meeting, ceding to the more popular choice. Deciding upon the party’s list, even an open one, for a proportional representation election is more complex. The heads of the relevant party executive board generally draw up their preferred electoral list, run it by the board as a whole for approval, and then present it to the broader membership/delegates at the candidate selection

meeting. Then there is an election for each slot on the list, beginning with the first. Anyone interested in this spot stands up and presents herself to the assemblage. A secret ballot vote is taken and the winner receives the first slot on the list. The process continues down the list until there are no more candidates willing to run. Here, where quota rules have been adopted they must either be adhered to or, outside the SPD, circumvented as described in the party’s statute—even when open lists are used and voters may ultimately change the list order. One interviewee described the process of creating an electoral list as follows: You must imagine it like this: And then every single position gets voted on. There’s always someone who stands up and says, “Here we have place number one on the list. Here we’ve envisioned our current mayor as the new lead candidate. He’s done great work, his administration’s record is wonderful, and we really need him to continue in office. Are there any comments from the floor?” Then maybe there’s someone who speaks against the mayor, and then a vote is taken. Every single position on the list will be worked through in this way. And you always have the opportunity to stand against the proposed candidate.В .В .В . [The list] usually winds up approximately as it was proposed, but there are indeed changes. It’s always in question. That’s why the three [party leaders who prepare the proposed list] have to do a good job at the start. If they put a weak candidate near the top of the list, for example, at place 15, and they know he is a controversial candidate, then it can turn out that already at slot 15 there are alterative candidates being proposed. Once the list gets opened up, once the genie is out of the bottle, then everything can change.В .В .В . That’s why they have to be careful when they prepare the list.В .В .В . One goes and talks with people and says, “listen, I know he’s controversial and you are one of his critics but we want him for this and this reason. Can you live with this?” Then you get the Page 50 →answer “no” or “well, OK, but I want, for example, that my preferred candidate also is well placed on the list.” So everything is discussed in advance, and it is already a very long balancing process before it comes [to the candidate selection meeting]. Other interviewees from large parties agreed that while the leadership’s preferred list was not always adopted exactly as proposed, radical changes to the list were rare. In contrast, members of the smaller parties reported somewhat less predictable outcomes in their candidate selection meetings. Interviewees from all parties agreed, however, that the very top slots on the list are often uncontested as party members defer to incumbents, the hopeless spots at the end of the list rarely provoke much controversy, and the slots toward the middle of the list are the most hotly contested as ambitious candidates fight to get on the list in an electable position.20 Similarly, there is a high degree of competition for direct seats a party seems likely to win, but less competition to be selected for a seat the party is unlikely to ever gain. The competition for a ballot nomination also varies depending on the level of office, the size of the party organization, and its popularity within a given community; the well-paid state, federal, and European parliament positions described above generate many aspirants whereas there are far fewer aspirants for volunteer positions in small communities where a party is weakly organized. These varying odds are discussed in more detail in chapter 2. Before moving on it is also important to note that although small parties including the Greens and the Free Democrats are unlikely to win directly elected seats via plurality, or as many seats as larger parties via PR, and although they have fewer members than the larger parties, small parties still place similar numbers of candidates on the ballot as larger ones to appear confident and give their supporters an option to vote for them. In the analyses to follow, I therefore include controls for party size. Technically, then, anyone who is a party member in Germany possesses gatekeeping powers to vote on candidates for the ballot. In practice, however, inner-party leaders at each level of party organization enjoy agenda-setting power for that level of office, putting together a suggested list and politicking before the selection meeting to gain support for this list. Thus the Candidate Interest Survey reached the lowest-level gatekeepers—the leaders who draw up lists for local councils and who make up part of the selectorate for directly elected seats for the Landtag (where relevant) and Bundestag. It is also possible that some CIS respondents simultaneously held leadership positions at the county, district, state, or federal level, and Page 51 →played a role in drawing up higher lists as

well. The role of local party leaders in crafting draft lists renders the German system of candidate selection more centralized in practice than it appears on paper, an advantage for traditional political “outsiders,” such as women, seeking to break in to politics. What Quotas Require of Gatekeepers Quotas are not employed for directly elected positions so there is no formal requirement that gatekeepers consider an aspirant’s sex when selecting a single-member district candidate. However, quotas do provide ammunition to feminist actors within the party interested in promoting female candidates for such seats; arguments that a woman should be nominated for a directly elected seat are indeed made, often by members of parties’ women’s auxiliary organizations (see also Davidson-Schmich 2014). Different quota policies require various types of attention to be paid to an aspirant’s sex when a party’s list is drawn up. For the Greens’ zipper quota, women must be granted the odd-numbered positions on the ballot and are also eligible to compete for the even-numbered ones. The Left Party’s quota functions in a similar manner: the first name on the ballot may be a man’s or a woman’s but after that the two sexes must alternate. The SPD’s quota requires alternating the sexes in positions one through four and leaves the fifth ballot spot open to either sex, continuing this pattern down the list. In these parties there are no, or very few, “battles of the sexes” over a given list position; instead, female aspirants compete with female aspirants for (almost) half the ballot slots and male aspirants with other men for their half of the list. By allotting women and men the same percentage of ballot positions, the Green and Left parties’ parity quota design downplays the idea that quotas are only about women (for the importance of this fact, see Murray 2014). The SPD’s nearparity quota design lets women and men compete against one another only for the last slot in a group of five, making it less likely that giving that list position to a woman would be perceived as tokenism. The Christian Democratic Union’s quorum, requiring one out of every three ballot spots to go to a woman, in contrast, could indeed lead to competition between men and women for a particular ballot slot. Such gendered competition is especially likely if women resist the party’s general practice of putting female candidates in the third of every three slots and demand a higher position. Only the CDU’s quota design, then, routinely pits women against men and overtly frames affirmative action as a “women’s” issue rather than a men’s issue. Page 52 →Where quotas are not employed, sex is not formally a criterion when drawing up a party list. However, quota contagion means that feminist activists within quota-less parties (and gatekeepers concerned about winning women’s votes) have a reason to press for women to be selected for promising ballot positions, so quota diffusion may also spark a “battle of the sexes” for good places on the list where affirmative action policies have not officially been adopted. Below I discuss the criteria, in addition to the candidate’s sex, that inner-party leaders utilize for selecting candidates. While quotas now require that gatekeepers include women on certain list positions, and quota contagion means that an aspirant’s sex is considered in other cases as well, affirmative action policies have not changed the other traits deemed important by gatekeepers. Gatekeeper Preferences: Party Loyalty Across German parties, states, and levels of government, one quality emerges as the most important for a nominee: party leaders, delegates, and rank-and-file members unsurprisingly look for candidates who they believe will best represent the interests of their party, regardless of sex (Holuscha 1999, 162). As one state-level party leader argued, it is important to look for a candidate whose “reliability you can trust”; the party doesn’t want candidates who are “going to break out and follow their own initiatives” instead of the party line. One mayor said he advocated for the selection of people who were going to be “team players” in his party’s delegation to the local council. Where closed lists are used, party loyalty becomes particularly important for democratic accountability as voters cannot reward or punish individual MPs.

If party members’ number-one criterion for a candidate is someone who is going to be a good representative of their interests, how do they determine who will represent them well? The most common answer is to select an incumbent who has already proved her loyalty when legislative votes were taken (Reiser 2014).21 While in the absence of quotas incumbency is a huge obstacle for female aspirants or other “outsiders,” when quotas are employed the incumbent for a woman’s slot on the ballot will already be a woman (Shair-Rosenfield 2012; Shair-Rosenfield and Hinojosa 2014). This elimination of male incumbency advantage is one of the major reasons why women are more successful in getting elected via list PR than via plurality to German parliaments such as the Bundestag (Davidson-Schmich 2014). Page 53 →Even female incumbents must have begun a political career at some point, however, so parties require a mechanism other than incumbency to determine which initial aspirants are likely to be loyal to the party once elected. Interviews suggest that gatekeepers generally prefer to pick someone from their own ranks; hence being active in one’s party and assuming a leadership position within the party organization is in practice the most important attribute of a candidate for elective office in Germany (Conradt and Langenbacher 2013, 191; Kintz 2011)—as it is in other party-loyalist systems of candidate recruitment (Siavelis and Morgenstern 2008; see also Matland 2005; Matland and Montgomery 2003, 24). The process of working one’s way up through the ranks of a political party is in German called an Ochsentour—literally an “ox tour” but better translated as “a slogging through the ranks” (Patzelt 1995; KГјrschner 2009). The higher the level of elective office one aspires to, the further within the inner-party leadership ranks of the party one must rise. So, for example, a candidate for county council is expected to be active in the party at the town and county levels; a state-level candidate will usually have held inner-party office at the town, county, regional, and state levels. As one Social Democratic interviewee put it, “You have to be anchored in the party; [for me to receive the nomination] it was important that people knew me, that my local party group was behind me.” Someone who just arrives at a selection meeting and “makes a big fuss” isn’t as likely to get elected as someone who has done “core work” for the party for a long time. His Christian Democratic counterpart agreed: “Before the public can elect you, the party must select you.” He pointed out that there are a number of people who would like to simply appear at a party meeting and say “[s]elect me” but “without having done their Ochsentour” this is unlikely to happen because ordinary party members who have worked hard on behalf of the party are the ones to select the candidates. A member of the FDP agreed that “having a party book is a prerequisite” for being a candidate, but also not enough. She observed that in order to become a candidate one must also take an active role in the party so that one is known and trusted by other party members. Even for members of the Green Party, which came into being in part as a reaction to the entrenched role that political parties play in Germany, party work is important. One member observed “lots of people want a good spot on our [state-level] list” and “it is a competitive game.” You have to “pay your dues” to get a good spot on the list; “you have to tender your services for a long time. You have to prepare for a year or two and then you have a chance. You have to deliberately head toward” a candidacy. Another Green interviewee summed up that there is Page 54 →a lot of “psychology” involved in candidate selection and it is “difficult” for an “unknown” quantity to get nominated. This lesson is not lost on aspiring politicians. One ambitious young party member interviewed explained the steps he was taking to lay the groundwork for a future political career: “It takes a hell of a lot of time” to establish a reputation within the party, one has to take the “Ochsentour, be industrious, [and do] in particular the work nobody else wants to do” including “hanging up campaign posters in all kinds of weather.” Another said he had taken the role of secretary in his local party group based on the slogan “wer schreibt, der bleibt” (he who writes, stays around). A state legislator compared his maintenance of close ties to his local party organization to the actions of someone with an affluent grandmother: “If you want to inherit, you are particularly nice to this grandmother. You call her a lot and visit her a lot and remember her birthday.” Gatekeeper Preferences: Previous Office Holding For elective office above the local level, assuming a leadership role within the party is not in and of itself sufficient. Instead, regardless of whether a quota is in place or not, having also held a lower-level elective office is usually a prerequisite for selection as a candidate for a high-level elected body (Conradt and Langenbacher 2013,

191; Davidson-Schmich 2006a). A candidate for a county council has usually served on a local council and proven herself effective, a candidate for the state legislature has often served on a county council and established a reputation there, and so forth. Thus the Ochsentour involves not only slogging one’s way through the party ranks but also working one’s way up the hierarchy of elective offices. As one party member put it, “If there are two equally skilled people [running for the state legislature]В .В .В . and one has been on a local council and the other hasn’t, the choice is clear.” Someone who has served in elective office has a known track record so party members can assess whether or not that person would make a good representative of their interests at a higher level. The use of quotas at the local level has helped women gain the qualifications needed to appear on the ballot for higher level offices—helping to increase the number of female candidates, even for the plurality component of the ballot (Davidson-Schmich 2014). Service to the party and previous office-holding are made important because of the mixed and PR electoral systems in place. Because quotas don’t change the electoral system, they do not change the importance of these characteristics. The vast majority of candidates selected in Germany, Page 55 →then, come from the ranks of party members, especially those active in the party, holding inner-party leadership positions. Since parties are highly decentralized and have a wide array of individuals on their local executive boards, however, additional factors are considered when selecting which active party member will be a candidate for a direct seat or fill an open-list position. The Candidate Interest Survey asked respondents to rate six candidate traits on a four-point scale from unimportant (0) to very important (3). The most highly rated traits, in order of importance, were public speaking skills (average rating 2.4), high public visibility (2.3), involvement in community voluntary organizations (2.2), and substantive knowledge of a particular policy area (2.1). Personal interviews with gatekeepers allow me to expand on why these particular traits are considered important. Below I discuss each of these characteristics in turn, focusing on why gatekeepers prioritize these attributes. Gatekeeper Preferences: Public Speaking Ability The party leaders interviewed frequently mentioned public speaking ability as an attractive trait in a potential candidate because the latter must be articulate enough to explain the party’s position to voters. Moreover, good public speaking skills are vital for winning the party’s nomination at the selection meeting. While gender quotas regulate the sex of candidates who may be selected, quotas do not change the method through which candidates are picked, and party members are unlikely to support the candidacy of an individual who cannot even effectively make her case to her own ranks. As one party leader put it, “One must be able to present oneself well. You can forget about [a nomination] if you give a bad speech” at the selection meeting. A successful candidate at the local level agreed, attributing his nomination to the fact that his professional training and job as a teacher had given him good public speaking skills. Similarly, a former member of the Bundestag believed he was a better candidate than his rivals because he possessed “a certain eloquence” that they did not share. Gatekeeper Preferences: High Public Visibility Simply delivering a charismatic speech at a nominating meeting is unlikely to be sufficient for achieving a competitive ballot position, however. Parallel to the Ochsentour of service to the party, and a history of lowerlevel elective offices for high posts, a successful German nominee must have established a personal reputation for herself outside the ranks of her party Page 56 →as well. When selecting an individual candidate for a singlemember district in Bundestag, state parliament, mayoral, and some local elections, gatekeepers pay considerable attention to whether a potential candidate has a widespread positive reputation among members of the electoral district’s voting public; being widely known in the community is important because voters are voting directly for a specific candidate, rather than party, on this portion of the ballot.22 On closed lists, only the top few candidates’ names appear on the ballot. When granting one of these very secure places at the top of the list, parties also consider how widely known a prospective candidate is. Similarly, if the list is an open one—as is utilized for some state and most local elections—successful list

candidates also need to have name recognition beyond their party organization. Because voters can select individuals from among list candidates, parties hope to attract voters by running candidates who citizens at large may personally know. In some states’ local elections voters receive as many votes as there are seats on the local council, and interviewees maintain that parties may even put locally popular citizens somewhere on their list—even if these individuals have absolutely no interest in actually serving in elective office. Since voters have many votes to distribute they may be willing to cast a few of these for someone they know personally even if they don’t necessarily support that person’s party (see also Roberts 1988, 110). Germans often refer to this characteristic as a person’s Bekanntheitsgrad—literally “the degree to which someone is known” or visible in the community. The concept of Bekanntheitsgrad is similar, but not identical, to the American idea of name recognition. In the U.S.’s media-driven campaigns, it is important the voters can recall the name of a candidate when they go to vote. In Germany, electoral districts are geographically much smaller than in the United States and there is an expectation on the part of many voters that their representative is known to them personally. One CDU member of a county council noted that her strongly conservative rural county included ten towns and each had its own town festival. At each festival, voters expected a high-level representative of each party to be there. She argued that if the CDU’s representative did not appear at her town’s festival, the voters would charge the party with having gotten arrogant and losing touch with voters. Similarly, a Green state legislator in an urban area argued that [as a state legislator] you also have official engagements that you must take part in because it is simply expected that, for example, if Page 57 →somebody somewhere is dedicating a new hospital or welcoming a new doctor, it is expected thatВ .В .В . you also come. People will look: “The SPD and the CDU were there; where were the Greens?” If you can’t make it, you have to at least send a nice card. There are number of ways to cultivate the visibility needed to be selected as a single-member district candidate, a top entry on a closed list, or as an open-list member. As mentioned above, being an incumbent (or holding another elective office) gives an individual a public platform with which to present herself to the wider community on a regular basis. An older party member noted that first-time candidates often had trouble receiving a large number of votes in her town’s open-list elections because they were not as well known in the community as incumbents; she noted that the number of votes she received went up with every election she had contested. In small towns across Germany, voluntary organizations play a key role in organizing social life. Every community, but especially smaller, rural ones, has a broad array of social clubs including, but certainly not limited to, sports clubs, parent-teacher associations (PTAs), church groups, choirs and musical groups, volunteer fire departments, animal breeding groups, and folklore societies. Each officially chartered club, like each political party, is required to have a democratically elected executive board. The leaders of these clubs are generally wellliked and well-respected members of the community who often appear in the local news when their teams win games or when they organize concerts, parades, festivals, and meals to which the general public is invited. Interviewees concur that, especially in local elections, parties are thrilled to have the leaders of these groups on their lists because these individuals are well known and have large social networks.23 One Social Democratic mayor from a small town attributed his successful start to a political career to the fact that his father was active in the volunteer fire department and people recognized his last name; he himself was also a member of a local bowling league and choir. Finally, if local party members have a choice, they generally prefer to select a candidate who has long-term local roots rather than someone who has recently moved to the area; the likelihood that the former is well known in the community is higher than the latter (see also Gallagher, Laver, and Mair 2011). As one new resident of a town pointed out, Germans prefer a candidate who has “Stahlgeruch”—a term used by horse breeders to refer to the fact that horses can recognize other horses from their barn because they can sniff out “the smell of their stable.” He and another interviewee were pessimistic about their electoral chances because they had transferred Page 58 →to their current residence for professional reasons and did not believe they stood a good chance against native candidates. A small-town mayor believed he was elected in part because “I was born here and I plan to

die here”; he said the citizenry trusted that he would avoid missteps as the mayor because he would have to continue to live with the results of his decisions for the rest of his life. The higher up the chain of elective offices one goes, clearly, the broader the geographic definition of “local” is. However, such a personal reputation need only be confined to the aspirant’s electoral district, even for a Landtag or Bundestag candidacy—in contrast, an ambitious party member needs a statewide reputation for an auspicious list position on the state, national, or European Parliament ballot. Gatekeeper Preferences: Involvement in Voluntary Organizations/Substantive Knowledge The candidate characteristics deemed the third and fourth most important by CIS respondents included involvement in voluntary organizations and substantive knowledge of a particular policy area (see also Conradt and Langenbacher 2013, 191; Holtkamp, Wiechmann, and Schnittke 2009, 52). The process of creating a party list in Germany is much like the procedure used to select a vice presidential candidate in the United States: balancing the ticket to appeal to as many potential voters is of paramount importance (Gallagher, Laver, and Mair 2011, 344).24 Further, the balancing of various interests also helps to achieve the consensus among the diverse party members needed to approve the list. Again, quotas have not changed the importance of these factors, which stem directly from the PR component of the German electoral system, its method of selecting candidates, and its parliamentary form of government. As mentioned above, being a leader in a club or association helps aspirants gain public visibility, but involvement in such groups is valued for additional reasons as well. Because the people elected from the list ultimately must constitute all or part of a party’s parliamentary party group, and in most instances its representatives to the executive branch, the party needs to ensure that it has members who are capable of addressing the various issues with which the legislative party group is confronted during and after the election, either as (shadow) ministers or members of Germany’s relatively powerful legislative committees (Gallagher, Laver, and Mair 2011, 56). Where the executive branch is indirectly elected, parties will be sure to place the potential chancellor, cabinet members, state-level Minister PrГ¤sident, or mayor at the top of their list, followed by future leaders of the Page 59 →PPG, such as the whip and the potential committee chairs. Moreover, having people on the list who stress different themes allows parties to reach out to a range of voters; gatekeepers interviewed worried that if their list were deficient in a particular area, competing parties would be quick to seize on this fact (see also Roberts 1988, 116). Thus, because the process of putting together a party list involves including multiple areas of expertise, having a particular “profile” (e.g., as an expert on education, the environment, or women’s issues) allows an aspiring candidate to gain the backers within the party needed to secure an electable spot on the party list drafted by gatekeepers. One way to establish expertise in a particular issue area is to be involved in an interest group or a social club related to the subject; leadership in such organizations is a common path to local office (Davidson-Schmich 2006a). One interviewee, a restaurant owner in an Alpine resort area, headed her village’s Tourism Association and credited her political success to her ability to speak on a subject vital to the community’s economic health. Another interviewee who was a mother of four and active in the PTA ultimately parlayed her knowledge of educational issues into a seat in the state legislature. An additional method through which expertise is demonstrated is one’s professional background. One statelevel party leader pointed out that if you’ve “studied accounting and then serve on the budget committee you are going to be better able to understand the budget [than another candidate] because of your background.” The need to include candidates with technical expertise rises with the level of office. As one local politician noted, almost anyone in his party could contribute to a town council debate about renaming a street, but helping the county council PPG decide whether or not it made medical or financial sense to close a county hospital required a deputy with more specialized knowledge. This northern Social Democrat’s comment was echoed almost verbatim by a CSU member residing near Germany’s southern border when she explained the advantage a local nurse enjoyed when her party’s county council list was drawn up because the council oversaw the running of the county hospital.

All parties are interested in putting together a list with a range of expertise, but they are particularly certain to include experts in the areas most important to their traditional voters—businesspeople in the CDU/CSU and FDP, labor experts in the SPD, environmental interests within the Greens (Roberts 1988, 108). As a Green Party member observed, “if one has, for example, done a lot on environmental or women’s policy issues, .В .В .В then it is clear that he automatically pulls a lot of weight” when the list Page 60 →is drawn up.25 In rural areas party leaders try to include farmers on their lists as well. In addition to leading a local NGO or having a professional background, aspirants often develop their “profile” in a certain area by leading the inner-party organization devoted to their interest (Conradt and Langenbacher 2013, 191). For example, the Christian Democratic Party contains a large array of inner-party organizations including those for workers and Protestants, the Social Democrats have many groups including those for gays and lesbians and one for small-business owners, the Greens have a group for union members, and The Left for the disabled. Where active, these groups can usually negotiate an auspicious list position for one of their members. Parties also attempt to balance ideological strands within the party when creating a list (Conradt and Langenbacher 2013, 191); for example, the Greens fill lists with both Realos (or pragmatic party members) and Fundis (the more ideologically pure Greens) and the Left Party balances between the communist platform and more moderate elements (Walker 2012).26 Thus, devotion to a particular strand of party ideology or issue expertise—demonstrated by ties to voluntary groups, professions, or inner-party groupings—helps establish a “profile” an aspiring politician can cultivate within her party in order to increase her chances of appearing in an auspicious list position. Gatekeeper Preferences: Other Demographic Characteristics In addition to balancing substantive areas of expertise, diverse personal characteristics are also needed to achieve a well-rounded list. Because such inner-party balancing went on throughout the postwar period, quota advocates were able to portray the adoption of quotas not as a new imposition on gatekeepers but rather as the logical extension of existing practices.27 In addition to sex and the factors outlined above, the demographic criteria that are also taken into consideration when putting together a party list in Germany include religion, ethnicity, age, and place of residence. The larger parties (CDU/CSU/SPD) take pains to stress their openness to citizens of all confessions, and they have traditions of ensuring that both Catholics and Protestants occupy top positions (Wiliarty 2008). Interviewees report, however, that in contrast to the United States, there is no expectation that elected officials be openly religious, for example, by attending church services or praying in public. This likely stems from the fact that, on average, 21 percent of party members are openly atheist (Klein 2011, 54) and voters are highly secular as well (Conradt and Langenbacher 2013, 73). Page 61 →German political parties’ members are overwhelmingly white, native German speakers, in contrast to the country’s increasingly heterogeneous population. While there are no explicit affirmative action policies for people of immigrant background—except in Berlin’s Social Democratic Party organization—the gatekeepers interviewed, especially on the left, expressed interest in recruiting candidates of Turkish or other nonGerman origin if they could locate such party members (see also Reiser 2014). These gatekeeper preferences can be seen in the steadily rising percentage of nonethnic Germans nominated for and elected to the Bundestag over the past two decades (WГјst 2011, 254; Fonseca 2011, 122); the numbers of migrants in state legislatures have also increased (compare Donovan 2007 and Donovan 2012). Even more than religion and ethnicity, balancing candidates of varying age is a strong consideration when drawing up a party list. Because of the importance of incumbency and the Ochsentour for candidate selection, there is an innate advantage in the system for older candidates. Interviewees report that the youth wings of all parties have been active in trying to counter this bias by demanding that certain spots on the list be reserved for young people. Here, however, formal age quotas have not been adopted at the national level, although local party organizations can—and do—adopt them in their charters. In local elections where the voting age in some states drops to 16, youth wings have argued, this rule increases the attractiveness of the list to the electorate. Further, having a diversity of ages is also an asset for political party organizations as a whole as it ensures that they have

access to a steady stream of individuals who can assume posts as older members phase out their activity. In addition, one state party leader observed, having young members on the ballot each time an election is held leads to “permanent renewal,” which helps the PPG get “new ideas, follow trends, and avoid becoming calcified.” There was an implicit assumption underlying many interviewees’ comments that indicated there is a strong expectation that when someone turns 65 she will no longer run for elective office. As the German population has aged, however, senior citizens in the German parties have also been active in founding their own inner-party organizations and parties also try to include an elderly candidate on their lists. For higher level elective offices, where lists of candidates can include party members from large geographic areas, parties often work to balance regional interests as well (Gallagher, Laver, and Mair 2011, 344; Reiser 2014). Making a list with regional balance ensures that party leaders will be able to secure support for their list from the members/delegates of all the Page 62 →party groups within the district. In addition, having representatives from across the electoral district will allow parties to better keep in touch with their constituents between elections. Especially where PR is used, having parliamentarians from all regions ensures that citizens will easily be able to find “their” representative should they require assistance. Thus, while quotas require gatekeepers to consider candidate sex, they have not diminished the importance of these other demographic criteria. The Role of Money in Campaigns Before turning to the aspirant and eligible phases of political recruitment, it is instructive to note which characteristics are not mentioned in the above discussion, because the fact that these attributes are not sought after also is an important component of candidate selection in Germany. In stark contrast to the U.S. political system, fundraising ability and financial resources are not priorities in the Federal Republic. German parties receive generous subsidies from the government (Green et al. 2008, 88–89), and elected officials are expected to donate part of their salary or honorarium to their party as well. Thus, German candidates are relieved from having to do extensive fundraising or bring their own financial resources to the campaign, although they are free to spend their own money to supplement party funds if they so choose. Even if they were to do so, however, campaign expenses are not high. For example, while the average U.S. Senate candidate raised almost $1.5 million in 2011 (Open Secrets 2013), one report on campaign expenditures for the 2009 Bundestag election found the winning candidate to have spent approximately $26,000, most of it saved from her parliamentary salary (KrГјger-LeiГџner 2009). The most frequently mentioned campaign expenditures in interviews included printing up personal flyers in addition to the party brochures, purchasing “give away” trinkets including candy and ice scrapers with a person’s name on them, or buying beer to thank their campaign volunteers. Generally, interviewees attest, campaigns are brief—six weeks or so—and most primarily involve hanging posters produced by the party, canvassing neighborhoods, staffing “Info Stands” in the town marketplace to distribute the party’s printed brochures and paraphernalia, appearing at social gatherings involving beer, and attending any local events to which the general public is invited (see also Conradt and Langenbacher 2013, 193). Radio and TV ads are more rare, and publicly financed when they are used (Conradt and Langenbacher 2013, 192). In sum, gender quotas have reserved certain ballot positions for women Page 63 →(and at times men), but the other traits considered desirable in a nominee—party loyalty, previous experience in elective office, public speaking skills, high public visibility, ties to voluntary organizations, substantive knowledge, religion, ethnicity, age, and place of residence—have not been altered by quotas.

Factors Deterring German Eligibles from Becoming Aspirants Not all individuals who possess the aforementioned traits are actually interested in running for elective office and serving in an elected capacity, however. Those people who are qualified for, and interested in, doing so are considered aspirants for elective office. A subsequent chapter will investigate gendered reasons why eligible women may not aspire to elective office and ask whether the adoption of quotas has exerted a symbolic effect, changing women’s decision calculus. Here I discuss nongendered reasons why Germans who are eligible for

elective offices may decide not to pursue them; the main deterrents to eligibles’ developing political aspirations include the time-intensive nature of elective office and its incompatibility with career development. The Candidate Interest Survey presented eligibles with a list of 17 reasons that might deter them from running for elective office and asked them to select factors that had (or might) dampen their political ambition. Few significant gender differences emerged, with both sexes identifying the same six factors as the strongest deterrents to running for elective office (see table 1.4). Both men and women expressed concern about the time commitment involved in running for elective office, agreeing that the primary deterrent to running would be less time to spend with their families. For women, concerns about time away from friends was the second most frequently mentioned reason for not running, with worries about a lack of free time in sixth place. For men, concerns about a lack of free time to pursue hobbies and a social life were the second and third most frequently mentioned hurdles to developing political aspirations. Eligibles’ reservations about their ability to balance political, personal, and professional pursuits are well founded, considering that running for and holding an elective office in Germany are indeed time-consuming pursuits. One study found that, on average, local mayors work 68 hours a week (Holuscha 1999, 297). While county and some large city mayors are paid professionals, the remaining mayoral positions and local council seats—positions that are the prerequisites for higher level elective office—are all Page 64 →amateur posts whose occupants generally receive only token honoraria. As most members are otherwise employed during the day, council, committee, and related party meetings take place during the evening hours or on weekends, which could otherwise be spent with family, friends, or hobbies. While both women and men expressed concerns about losing free time if they were to run for office, women were significantly less likely than men to hold this reservation; chapter 3 explores this difference in more detail. For women the second, and for men the third, most frequent deterrent to running for office involved the difficulties of balancing a professional career and elective office. Half of the individuals interviewed brought up this concern as well. Local officeholders need to be free evenings and weekends and occasionally during the day, which may clash with work schedules as well as opportunities to pursue career-development-related activities. High-level elective offices in Germany are well compensated, but assuming them requires an individual to step away from other employment for an uncertain period time and to commute to the state capital, Berlin, or the other EU member states in which the European Parliament is headquartered. Serving in 13 of Germany’s state legislatures is considered a full-time position and pays on average €6,200 per month plus expenses; as a frame of reference, the median monthly income in Germany is €3,311 (Statistisches Bundesamt 2013).28 Working as a Member of the Bundestag is similarly a full-time position earning an MdB over €8,000 per month, plus a considerable expense allowance. Membership in the European Parliament is also a lucrative full-time position with a pretax salary of over €7000 per month plus many untaxed benefits. However, MEPs must shuttle between their constituency in Germany and the EP’s offices in Luxembourg, Belgium, and France. Pursuing high-level elective office in Germany, therefore, is equivalent to selecting politics as one’s profession, albeit for an indeterminate period of time as the possibility of losing an election always exists. Page 65 →Eligibles in professions that require continuous service or updating of skills report being unwilling or unable to step away from their place of employment to pursue elective office. Similarly, those who are concerned about their ability to return to their previous post, or obtain a new position following a stint in politics, will be reluctant to run for elective office. Table 1.4. Factors Deterring Political Aspirations Percentage of respondents agreeing with the statement: Women Men Running would take time away from family life 43% 46% Running would take time away from social life with friends 26% 25% An incumbent holds the seat I would be interested in 25%** 18% Running would hurt professional career 24% 25% I see no deterrents to running 19% 20%

Running would take time away from hobbies/free time

18%*** 27%

Source: Candidate Interest Survey (N = 430). ** Female/male difference of means significant at p ≤ 0.05; *** Female/male difference of means significant at p ≤ 0.01. Interestingly, for the women surveyed, the fifth, and for the men surveyed the sixth, most common response when asked what would deter them from running for office was that they could see no deterrents.29 Finally, as has been observed in the literature on political ambition (Fulton et al. 2006), both sexes, but more so women, were sensitive to political opportunity structures, expressing concerns that they would be unable to unseat an incumbent politician as a deterrent to running for elective office. Subsequent chapters delve into this topic in more detail. In sum, while gender quotas require parties to select female candidates, they neither change either the timeconsuming nature of elective office nor protect the future career prospects of women who choose to run for elective office—both critical influences on political ambition.

Eligibles for Elective Office in Germany This final section turns to the very first phase of political recruitment—becoming an eligible. While quotas require gatekeepers to select women for certain ballot positions, they do not change what constitutes “eligibility” for elective office. Formally, all German citizens eligible to vote in a given election are also permitted to stand as a candidate in that election; for most contests, this includes German citizens who are at least 18. For some state and local offices, those who are at least 16 may also vote and run for office. In European Parliament elections and local elections, citizens of other European Union member states living in the constituency have been granted suffrage. In contrast to the United States, where a citizen must register to be eligible to vote, there is no specific election registration process in Germany, although Germans must have lived at their current address for at least three months to vote and run in the electoral district where they currently reside. German law also includes provisions barring from voting the mentally disabled, those institutionalized in homes for the criminally insane, and incarcerated criminals whose voting rights have been removed by a judge. In theory, then, there are approximately 60 million eligibles in Germany.30 In practice, however, given political parties’ monopoly on ballot nominationsPage 66 → and the aboveoutlined preferences of gatekeepers, eligibles are first and foremost loyal party members. Loyalty to the party can initially be demonstrated in many ways but the primary method through which this is done is by becoming active in one’s local party organization and assuming an inner-party leadership post (Conradt and Langenbacher 2013, 191; Davidson-Schmich 2006a; Kintz 2011). Thus, virtually all German elected officials are long-time party members (Kintz 2011). Despite this fact, or perhaps because of it, the German population has a generally negative opinion of political parties (Schulze 2007). As a result, parties are routinely pressured by the media and disgruntled voters to nominate candidates from outside their ranks. Such individuals are referred to as “Seiteneinsteiger” or people making a lateral entry into politics (KГјrschner 2009, 44). These candidates are generally individuals sympathetic to the party but who do not have the time to become active within it because they focus their energies on another area, usually in professional associations. Sometimes such candidates are nominated to mobilize voters traditionally close to a political party. For example, the SPD occasionally selects union activists, the CDU and FDP businesspeople, and the Greens members of NGOs to run on their ticket. At other times, lateral entry candidates are chosen because they may help a party reach out to a constituency that does not traditionally vote for it, or because the party lacks expertise in that person’s area of strength; for example the Social Democrats may select a lateral entry candidate with a business background. While these nominations may be popular with voters, gatekeepers argue that Seiteneinsteiger make less effective elected officials than long-term party members as they are easily frustrated by the protracted, compromise-filled negotiations required to make public policy in Germany and, for this reason, gatekeepers say they are loath to nominate them. As a result, the number of lateral entry candidates in Germany is limited (KГјrschner 2009, 44) and interviewees report that when such candidates are

indeed elected, they often only serve one term. In order to trace the process of political recruitment in Germany, then, one must begin by examining loyal party members. While quotas do apply to the inner-party leadership positions required to demonstrate party loyalty, they do not apply to who can become a party member. Although membership in a political party is open to virtually all voters—male and female—who can pay minimal dues, only a small percentage of citizens can realistically be expected to join a party. The German General Social Survey found that only 10 percent of Germans report a very strong interest Page 67 →in politics (GESIS 2011) and another public opinion survey found only 14 percent of the population claiming they could imagine joining a political party (Klein 2006, 43). Thus the potential eligibility pool is far smaller than the number of citizens formally eligible to run for elective office. The statistics above suggest that there are only 4.7 million individuals in Germany are very strongly interested in politics. In 2010, however, Germany’s six major parties reported a total of only 1.3 million members (Niedermayer 2011)—or a mere 2.2 percent of the electorate. In 1980 the comparable figure for West Germany was 4.5 percent, indicating that in Germany, as elsewhere in Europe, party membership has been plummeting (van Biezen, Mair, and Poguntke 2012). The statistics above probably overrepresent the number of actual members since these are self-reported figures and parties have an incentive to count everyone whoever joined as a member as long as they have not dropped out, even if that person is not actually active; these Karteileichen (literally “party card corpses”) beef up membership figures and make parties look more popular than they are. Even these likely inflated numbers represent only about 30 percent of those Germans who report a strong interest in politics. One reason for this low percentage of politically interested citizens—male or female—joining parties is the poor reputation political parties have in Germany, as elsewhere in Europe. Across western Europe, the phenomenon of Parteiverdrosseneheit (dissatisfaction with or distrust of political parties) emerged with the rise of postmaterialism and the end of traditional societal milieus (Schulze 2007). In eastern Germany, like other eastern European nations, distrust of political parties is even more pronounced, given the forty-year dictatorship by the Communist Party there (Yoder 1999). Many people politically active in the nonpartisan citizen movements that brought an end to the communist regime disliked the concept of political parties so much that they abandoned their involvement in politics when confronted with the need to run on a party label after German unification (Davidson-Schmich 2006a, 166–67). Roughly one-third of Germans responding to a national survey did not believe any German party was the most capable of solving what the respondent believed to be the most pressing problem facing the country (Rattinger 2000, 218). Recent political events in Germany, including widespread public protests in Stuttgart against development plans there, the rise of the antiestablishment Pirate Party, demonstrations across the country against nuclear weapons and nuclear waste transport, declining levels of party identification (Dassonneville, Hooghe, and Vanhoutte 2012), and Page 68 →high electoral volatility, indicate the growing public dissatisfaction with party politics as usual in Germany. Across party lines, politically active individuals who I interviewed noted declining interest in party membership among citizens in their communities. A young Christian Democrat said that while he thought his generation was interested in politics, they were not interested in parties. One Green Party leader observed that because of what he termed “a disdain for” or a “negative view of” politics, “people aren’t standing in line” to join his party. As one frustrated FDP leader argued, “Many whose [involvement] would be welcome won’t have anything to do with a party in the first place becauseВ .В .В . the image of parties in general is bad.” One state legislator projected, I think that [people] don’t go into politics because they would find very little recognition among the general public. I mean, it’s the same for us. When we go out and say we are members of theВ .В .В . state legislature then people say, “oh, you guys, what are you up to again?” and people just rail at you. In principle you are seen as holding the worst job in the world. You could almost say that if you are, for example, a custodian somewhere, people say, “OK, that’s great, come sit down with us.” But if you are a politician, people say “ugh”; that’s just the way it is.

The small number of Germans joining and becoming active in political parties is due to more factors than simple dislike of parties, however. Even among those politically interested individuals who may not have an objection to partisan organizations per se, some may not fit ideologically into the German party system and may be unable to find a group with which they would want to affiliate. Several interviewees who were active in German interest groups specifically cited an unwillingness to represent a party with which they did not agree on all issues as the main reason they had decided against running for political office—even when asked to run by a party traditionally associated with their voluntary group. The same was true of some staffers working for German parties. Indeed, even some party members who had been elected at the local level, where issues are relatively pragmatic rather than ideological, were reluctant to run for higher office because they feared having to toe the party line with which they at times disagreed. Thus, aside from a basic lack of interest in politics, there are many nongendered reasons why a politically aware German might not join a political party and become an eligible. Party quotas do not directly address these aspects of the political recruitment process and have required gatekeepersPage 69 → to find increasing numbers of female candidates against a backdrop of a shrinking eligibility pool. Moreover, there are several important and highly gendered reasons why women may, on average, be less likely to join a political party than the average man. These reasons lie outside the purview of quotas, in the recruiting environment and in informal recruitment structures, such as the nature of party life itself. It is to this topic that the next chapter turns.

Page 70 →

Two Eligibles I thought to myself, no way! —A young mother describing her reaction to getting involved in a local party group after attending their meeting and encountering only older men The way [party] meetings are run, and the times they are held, the length and intensity with which people are really only yakking on and on.В .В .В . When this eternal blabber begins, I start to think to myself, вЂMy god, you could be at home doing a load of wash. You could be doing this or that.’ The time that is wasted is appalling. —Female party member (quoted in KГјrschner 2009, 223; author’s translation) Proponents of gender quotas often argue that these affirmative action measures are important because increasing the numbers of women in elective office (women’s descriptive representation) will lead to women’s symbolic representation. That is, quotas and the ensuing cadre of female politicians will send a signal to female citizens that they are capable of ruling, that their political participation is desired, and that political institutions are responsive to their concerns. One expected manifestation of symbolic representation is that women and girls will in turn become more interested, and active, in politics. Germany provides an ideal testing ground for this expectation. As depicted in the previous chapter, quotas have been in effect since the mid- to late 1980s and top-level elected bodies in the Federal Republic have Page 71 →contained a critical mass of women for more than a decade. Moreover, since 2005 Angela Merkel has served as the very visible chancellor of the country accompanied by a cabinet containing one-third women. In addition, the country’s population has one of the highest levels of gender egalitarian attitudes in the world, a ranking based on, among other factors, citizens’ beliefs about whether or not women make better political leaders than men (Inglehart and Norris 2003). Of all countries worldwide, Germany ranks in the top ten most gender-equal rated by the United Nation’s Gender Inequality Index (UNDP 2013). If quotas were to inspire women to become politicians, Germany would be among the most likely cases for this to occur. This chapter investigates whether, almost three decades after the Greens first adopted their quota, German women and men now become eligible to run for elective office at equal rates. As depicted in chapter 1, the vast majority of German candidates for elective office are active members of political parties. If German women were inspired by quotas or the presence of female politicians to pursue elective office, the first steps on this route would be to join a party and then to assume leadership positions within it. Do women and men join political parties at similar rates? Are they equally likely to rise to positions of leadership within their party organizations? The answers to these questions are important, for they will tell us whether quotas have been successful in achieving Anne Phillips’ concept of “justice” (1995, 62), increasing women’s political participation in the very earliest stages of the political recruitment process. The answers also have ramifications for women’s descriptive representation—whether or not parties will be numerically able to fill their own quotas—and hint at quotas’ impact on substantive representation as well. Taking an intersectional approach and investigating which types of women join parties and become leaders in them will shed light on how potential female representatives may conceive of other women’s interests. To address these issues, I draw on two types of data. First, as with the other chapters of this book I rely on my Candidate Interest Survey and personal interviews with German eligibles. Second, I draw on several Germanlanguage sources about mass political behavior, including research on the gender breakdown of party members (Niedermayer 2011), the largest existing survey of rank-and-file party members in the Federal Republic (Spier et

al. 2011), parties’ self-studies of their membership and leadership, public opinion data from the German General Social Survey, and statistics from federal and Land-level government agencies. This chapter proceeds as follows. I begin by discussing quotas and Page 72 →symbolic representation and then I delineate aspects of the recruitment environment and informal recruitment structures that restrict quotas’ symbolic effects. I subsequently provide quantitative data depicting the gender breakdown of German parties and other voluntary groups, finding that—a generation after quotas were adopted—parties remain male-dominated organizations. Here I also present qualitative evidence of gendered hurdles to party membership created by the recruiting environment and informal recruiting structures. Next I examine party leaders in Germany and demonstrate that quotas have an “elevator effect,” creating a relatively gender-equal eligibility pool in parties employing parity and near-parity quotas. I conclude by discussing the implications of these findings for both the descriptive and the substantive representation of women.

Gender Quotas and Symbolic Representation Proponents of gender quotas often argue that these measures are important because increasing the numbers of women in elective office (women’s descriptive representation) will lead to women’s symbolic representation (Mansbridge 1999; Schwindt-Bayer and Mishler 2005; Franceschet, Krook, and Piscopo 2012). As Jane Mansbridge argues, Low percentages of Black and women representatives, for example, create the meaning that Blacks and women cannot rule, or are not suitable for rule.В .В .В . [T]he increased descriptive representation of women in the legislatures would undermine the perceptions that politics is a “male domain.” (1999, 649) That is, quotas and the resulting cadre of female politicians will send a signal to female citizens not only that they are capable of ruling but also that their political participation is desirable. Women elected to public office are thought to become role models who will in turn awaken other women’s and girls’ interest in politics and inspire them to get involved. One observable implication of this form of symbolic representation would be that men and women become eligibles at similar rates. These expectations are often tested using public opinion surveys. Extant empirical evidence suggests that when large numbers of women or very visible women take part in politics, women and girls do know more about politics (Burns, Schlozman, and Verba 2001; Konig 2009; Reingold and Harrell 2010), talk more about politics (Atkeson 2003; Wolbrecht and Page 73 →Campbell 2007), and express a greater desire to participate politically—for example, by running for elective office (Wolbrecht and Campbell 2007). Fig. 2.1. German Men’s and Women’s Mean Political Interest: 1980–2010. (Source: GESIS 2011.) Indeed, since the adoption of quotas in Germany, the gap in women’s and men’s political interest has narrowed. Since 1980, the German General Social Survey has questioned citizens about their degree of interest in politics, scaled here from 0 (a complete lack of interest) to 1 (a very strong interest). Figure 2.1 presents the difference between men’s and women’s mean score over the 30-year period during which quotas were phased in. Over these three decades, as expected by those who argue quotas will have symbolic effects, the gender gap in political interest was reduced by approximately one-third. Additional studies indicate that if more nuanced questions are asked—for example about interest in specific issues such as social policy or environmental politics—this gap disappears altogether (Banducci and Semetko 2002; see also Dolan 2011). Quotas have certainly had some symbolic effects. However, the above studies of symbolic representation involve relatively low-commitment activities: answering a question on a survey or engaging in casual conversation. When it comes to more intensive commitments, scholars have failed to locate robust empirical evidence that women’s descriptive representation increases women’s deeper involvement with politics. Lawless (2004) finds having a female representative to Congress does not predict whether a woman will attend a political meeting, donate to a candidate, or volunteer for a party or

campaign. Similarly, Page 74 →Zetterberg’s (2009) case study of Mexico argues that quotas do not increase women’s propensity to contact their elected representatives, political parties, or other public officials. Nor did he observe quotas inspiring women to get involved with political campaigns. This disconnect can be explained by examining the social backdrop against which descriptive representation occurs. Political party quotas may alter some aspects of the recruitment environment (i.e., social norms about whether women can or should succeed as politicians), they do not change the overall gendered division of labor in society. Moreover, while quotas do change formal political recruitment structures (i.e., by changing party rules governing who can be nominated for certain ballot slots or inner-party posts), they have a much more limited impact on informal party norms. The overall political recruitment environment and informal party norms, even in a highly gender egalitarian country such as Germany, exert gendered effects that make it difficult for even politically interested women to join and become active in parties.1 As a result, I expect that—despite the use of voluntary quotas for over a generation—the German parties still contain more men than women. In other words, quotas’ symbolic effects are limited when it comes to changing women’s actions rather than their attitudes. Below I delineate the gendered hurdles to party membership in Germany and then go on to empirically investigate the eligibility pool.

Gender Quotas and the Recruitment Environment The previous chapter outlined hurdles to party membership in Germany including a generalized distrust of parties and an inability to locate an organization close to one’s policy preferences. While these problems likely apply equally to men and women, several other facets of the recruitment environment hinder primarily women’s participation in political parties—even when gender quotas are in place. Comparative research finds citizens propelled to join political parties for ideological, social, and career reasons (Gallagher, Laver, and Mair 2011, 332). Germany’s recruitment environment renders women, on average, less likely to possess all three motivations than men. Moreover, even women who do wish to join a political party may be unable to find the time to do so. Despite the fact that the political interest gap between men and women in Germany has narrowed over time, as of 2010 a 10 percent gap in men’s and women’s mean level of political interest still remained. While 14.1 percent of the men surveyed expressed a strong interest in politics this Page 75 →figure fell to 6.3 percent for women (GESIS 2011). This gap is replicated across many different surveys and question formations, and persists even when controlling for education, age, and other variables that may shape an individual’s interest in politics (Westle 2009); it is also present in other long-term democracies (Dow 2009). Women’s greater average propensity to care about certain policy areas, such as the environment or social issues, does not seem to translate into a more generalized interest in politics as a whole—the focus of party groups. Because German women remain less likely than men to be “very interested” in politics, they are less likely to hold the strong ideological beliefs that propel individuals to join parties.2 Another route to party membership comes as individuals perceive career-related advantages to joining a political party. Germany’s post–War War II conservative welfare state (Esping-Andersen 1999), however, was strongly oriented toward a male breadwinner model. The country’s taxation system continues to provide considerable incentives for single earner couples (European Commission 2013) and one-third of German women ages 15 to 64 are not in the paid labor force—about 10 percentage points below men’s employment rates. Of these women, almost half are in part-time positions and, as a result, the gender wage gap in Germany is higher than in almost all OECD countries (European Commission 2013). Thus, German women are less likely than German men to pursue careers that might be enhanced by joining a political party. In addition, they are less likely than men to possess the personal disposable income to pay the dues needed to join. The gender gaps in political interest and employment have a spillover effect in that, as in other countries, men and women come to occupy sex-segregated work, volunteer, and social networks (Sanbonmatsu 2006, 153). Because they often interact with other politically dispassionate individuals, women are less likely than men to find themselves in social situations that awaken their political interest and involvement. Men, in contrast, are more

likely to have the peer-to-peer interactions have been found to lead to increased political participation, as friends recruit friends for partisan activities (Klofstad 2011). While German women are on average less likely than German men to be found in full-time paid employment, they are more likely to engage in unpaid work in the home. Germany’s conservative welfare state relies on (female) family members to provide care for children, the elderly, and the sick. The creation of service-sector jobs has been particularly slow (Esping-Andersen 1999), rendering housekeeping, childcare, home healthcare, and other services either nonexistent or prohibitively expensive. As a result, in Page 76 →Germany, even more so than in many other OECD countries, women bear an unequal share of household work and childrearing tasks (Iversen and Rosenbluth 2006, 11). This in turn means that the German women who are interested in politics and inspired to join a party may still find themselves unable to become eligibles (Geissel 2013). While on paper all that is needed to join a party is an application form and dues payment, party membership in practice involves a considerable time commitment. A majority of the members of German parties engage in some party-related activity each month and over half of them spend more, at times far more, than five hours each month on party work (Heinrich, LГјbker, and Biehl 2002, 27–29). Because parties have by and large been organized by men employed outside the home with stayat-home wives, local party group activities take place after working hours, usually in the evenings or on weekends. Routine meetings are held during which members are expected to contribute to discussions of political issues and help plan activities. Members are often asked to represent the party at local events (for example if there is a street festival, the party might have an information table that needs to be staffed), to attend programs sponsored by the party (for example a public talk about a pressing political problem), or to meet with citizens concerned about a particular issue (for example, parents concerned about conditions in the local school). Parties often sponsor training sessions to educate their members about political issues or to train them in political skills such as how to successfully organize an event. Some members are chosen as delegates to represent their precinct-level organization at higher-level party conventions and others get involved in the inner-party bodies such as those for women or students. During campaign season members hang up posters, pass out brochures, and go door to door. These activities generally take place outside of working hours. In addition to these activities undertaken by rank and file party members, there are also a number of internal leadership positions within local party organizations that Germany’s Party Law requires be filled on a regular basis. Often local party groups contain only one or two dozen members, but they are all required to elect at least three members to form a board of directors (Vorstand). In addition to a chair and one or two vice chairs, local party groups often elect other officers including a secretary, a treasurer, and additional advisory board members. Occupying these leadership positions within a party requires a time commitment over and above that expected of rank and file members. People in the party Vorstand are the ones who organize the events described above, manage the local group’s finances, Page 77 →and keep records of its activities. They are likely among the 10 percent of German party members who spend over 20 hours a month on party related activity (Heinrich, LГјbker, and Biehl 2002, 29). As delineated in the previous chapter, such activities are necessary for an individual to be considered a contender for elective office. While quotas may have reduced the gender gap in political interest, they cannot free up women’s evenings and weekends to devote to party work. For working women, this is the time that can be utilized for household tasks for which they are responsible, and for women with children these are hours that are given over to childcare. Thus, for most politically interested German women to join a political party and become active in it would mean assuming a triple burden: work outside the home, domestic work, and party work. Men who wish to take part in party life are, given Germany’s gendered division of household labor, on average more likely than women to have someone at home to lighten their load. Studies of other OECD countries have found housework and childcare to be the primary drivers of women’s lower level of political involvement (Mestre and Marin 2012). Because quotas are unable to change the gendered division of household labor and free women to become active party members, the recruiting environment is expected to limit quotas’ ability to increase the percentages of women within German parties.

Gender Quotas and Informal Recruitment Structures I expect that this hurdle to the average woman’s political participation will in turn create a vicious cycle that will repel even those women who are politically interested and who do have the free time to join a party. While Germany has a highly visible woman chancellor, high-profile female cabinet ministers, and a critical mass of women in its European, national, and state-level parliaments, the first point of contact for women desiring to join a political party is their local precinct group. Regardless of whether or not a quota is in place for selecting candidates, these local organizations are expected to be heavily male dominated, meaning that becoming an eligible in Germany will be tantamount to joining a club made up mostly of men. The fact that they do not physically resemble other party members may have a deterrent effect on some potential female eligibles. Moreover, the strong descriptive representation of men is likely to have an additional substantive impact in terms of how everyday party life is conducted (Geissel 2013). Male-dominated party organizations worldwidePage 78 → have been found to possess informal norms that are highly masculinized (Shvedova 2005, 35–36; Lawless 2012, 58; Kenny 2013; BjarnegГҐrd 2013). As Marianne Githens writes, even when parties are “amenable to women’s entrance”—as is the case when they adopt candidate gender quotas—political organizations possess “normative standards for behavior whichВ .В .В . assumeВ .В .В . male values as universally valid” (2003, 43). Jennifer Lawless concurs, “political organizationsВ .В .В . that have always been controlled by menВ .В .В . operate with a gendered lens that promotes men’s participationВ .В .В . and does not sufficiently encourage women” (2012, 58). For this reason, women cross-nationally participate more often in informal social movements or activities such as boycotts than in formal political organizations such as parties (CoffГ© and Bolzendahl 2010; Sauer and WГ¶hl 2012, 346). While quotas change formal recruitment structures such as candidate nomination procedures, they do not generally address the everyday ways in which party members interact with one another, creating a masculinized ethos that is expected to deter the politically interested women with time to devote to voluntary groups from joining political parties—even organizations with quotas. While this problem is expected to exist everywhere in Germany, it is likely to increase as the size of the party organization decreases. In thinly settled rural areas, many political party groups—especially those that are not particularly electorally successful—possess very few members and are weakly organized. In such contexts, a party group that is one-third female may literally only contain two or three women. In such settings it is easy for newcomers to feel very isolated. Inner-party life in such small groups is rarely conducted according to formal rules and procedures and is instead dominated by informal (masculinized) practices. In contrast, in densely populated urban areas precinct-level party organizations tend to be much larger. Even if women make up only a third of the group, they are greater in sheer numbers than in small communities. It thus becomes less likely that a woman attending her first party get-together will encounter no, or only one or two, other women. The likelihood of a woman, or women, who can serve as a “critical actor” effectively challenging male norms also increases (Childs and Krook 2006). Moreover, the larger the size of the party organization, the more organized inner-party life becomes and the more difficult it is for an “old boys’ network” to get away with making decisions through informal channels.3 Studies in other contexts have demonstrated a close relationship between urbanization and women’s political representation (Carbert 2009; Eder and Fortin-Rittberger 2013). For these Page 79 →reasons, I expect women’s underrepresentation in political parties to be particularly pronounced in thinly settled areas, regardless of whether a quota is present. To become eligible for elective office in Germany, however, an individual must not simply join a political party; he or she must also become active within it, assuming positions of inner-party leadership. Here, gendered hurdles to women’s participation are expected to be somewhat lower and quotas’ impact stronger. A woman who is not deterred from party membership by masculinized associational norms may also not be dissuaded from additional involvement. Similarly, a woman who can find enough time for her party to make dues payment worthwhile may possess the time needed to play a leadership role. Moreover, quotas enhance the political opportunity structure for women. In all parties with quotas—but especially those employing parity or near-parity quotas—the percentage of inner-party leadership posts designated for women is likely to be higher than the percentage of female party members. Even if women and men do not join parties in equal numbers, the Greens,

the Left Party, and the SPD’s quotas require roughly equal percentages of each sex on party boards, among delegates to conventions, and in other inner-party bodies. To the degree that women remain underrepresented in the rank and file, quotas are expected to have an “elevator effect,” creating higher percentages of female leaders (eligibles) than party members. In short, Germany’s recruitment environment and its informal recruitment structures, even after over a generation’s experience with quotas, are expected to reduce women’s propensity to become party members, especially in rural parts of the country. However, quotas—especially (near) parity quotas—can level these discrepancies by elevating the women who do join parties to inner-party posts, creating a more genderbalanced eligibility pool. The following sections present empirical evidence supporting these expectations.

Women’s Membership in German Political Parties In the mid-1980s, when gender quotas began to be introduced in West Germany, 23 percent of all political party members were women (Kolinsky 1989). In the 25 years following quota adoption, the percentage of German party members who are women did rise, in both parties with and without quotas (see figure 2.2). Parties with quotas now have a higher percentage of female members (33 percent on average) than do parties without quotas Page 80 →(21 percent) (author’s calculations based on Niedermayer 2011, 16). The Social Democrats’ female membership increased the most, rising 6 percent to 31 percent; the initially quota-less CSU’s rose 5 percent to 19 percent. The Greens’ female membership grew 4 percent to 37 percent and the Christian Democratic Union’s women members edged upward 4 percent to 26 percent. In contrast, the quota-less Free Democrats (FDP) saw their female percentage decline 1 percent to only 23 percent of party members. Today’s Left Party is comprised of 37 percent female members; no comparable figures are available from the 1980s as the party did not exist in its current democratic form then. Fig. 2.2. Male and Female Party Members: 1985–2010. (Sources: Kolinsky 1989; Niedermayer 2011, 16.) On the whole, however, these increases are minimal. In 2010, over a generation after quotas were first adopted, women remained a minority in every major German party—even those with quotas. Sixty-three percent of Green and Left Party members are men. The figure rises to 69 percent and 74 percent for the SPD and CDU, respectively. Moreover, because German parties have shed members at a rapid rate since the 1980s, the overall number of women in German parties has actually declined by almost 100,000 since the adoption of quotas—even though all of these parties expanded their geographical scope after German unification in 1990. The current low percentage of women is not an artifact of parties’ male dominated, prequota pasts either. Women comprised a mere 30 percent of parties’ new members between 1999 and 2009 (Klein 2011, 56) and membershipPage 81 → in party youth organizations, such as the CDU’s Junge Union and the Social Democrat’s JuSos, remains male dominated (CDU Deutschland 2010, table 21; JuSo Bundeskongress 2009). Thus it is not possible to conclude that quotas have had a strong symbolic effect, spurring women to become party members and hence eligibles for elective office. In fact, it is likely the reverse: a rise in female members in the 1970s inspired leftist German parties to adopt quotas in the first place (Kolinsky 1991). Similarly, after the Bavarian Christian Social Union experienced the upswing in female members depicted in figure 2.2, it too adopted a quota for certain offices.

Germany’s Recruitment Environment Why is it that women’s increased political interest has not translated into a sharper increase in party members? Surveys of the German population indicate that, as expected above, much of this trend can be attributed to the recruitment environment in Germany. Repeated studies of women’s political participation indicate that female citizens in the Federal Republic face time constraints that hinder their public involvement. When it comes to civic engagement requiring little time commitment, few male-female differences can be found. Both men and women are equally likely to vote in national elections and to sign petitions (Hess-Meining 2005, 368 and 400; GESIS 2011).

However, when it comes to civic engagement that takes a greater time commitment, gender differences begin to emerge. The German Ministry for Women reports that while 39 percent of male citizens claim to participate in voluntary activities, this figure falls to 32 percent for female citizens (Hess-Meining 2005, 393–94). When asked exactly how much time they devote to such pursuits, men and women’s responses diverge even more sharply (see figure 2.3). On average, men are much more likely than women to report being able to make a sustained time commitment to voluntary groups—the kind of commitment necessary to become eligible to run for elective office in the Federal Republic. In contrast, women on average mention considerably more sporadic involvement. Instead, time use surveys establish, German women spend more hours on household chores than do German men (Iversen and Rosenbluth 2006). Given that there are a finite number of hours in a day, these two findings are likely highly correlated: German women simply have less time on average than German men to devote to voluntary activities such as party membership. Fig. 2.3. German Citizens’ Time Spent Per Week on Voluntary Activities. (Source: Hess-Meining 2005, 397.) Page 82 →The Candidate Interest Survey clearly illustrates the gendered nature of Germany’s recruitment environment. The women polled, who all had assumed leadership positions in their parties, were significantly more likely to be primarily responsible for housework and childcare than their male counterparts—regardless of whether their party employed a quota or not (see figure 2.4). These striking differences in domestic responsibilities represent a best-case scenario, as the women surveyed actually did manage to join a party and became active. The results indicate that over half of these individuals have some type of domestic assistance—either supportive family members or hired household help—advantages from which not all German women benefit. The low percentage of eligibles having primary responsibility for childcare depicted in figure 2.4 illustrates how difficult it is for Germans with young children to become active in a political party that requires their presence during the evening and on the weekend. Roughly two-thirds of the eligibles responding to the CIS did not have minor children living at home. This percentage does not differ from households across Germany; however, what does vary is the proportion of single parents. While 7 percent of German households are headed by single parents, only about 1 percent of party leaders were trying to combine raising children on their own with party work (Statistisches Bundesamt 2008, 46). The respondents who did have children had relatively few—1.7 on average—and the children were older—the mean age was 10 years. Regardless of whether their Page 83 →party utilized a quota, there were no differences in the propensity of male and female leaders to have children, nor were there significant differences in the ages of their children. Fig. 2.4. Eligibles Primarily Responsible for Household Tasks. (Source: Candidate Interest Survey [N = 447]. All differences of means significant at p ≤ 0.00.) The fact that one-third of eligibles do have young children indicates that while difficult, it is not impossible, to combine employment, childrearing, and political activity in Germany.4 For parents who wish to become involved in politics, it is possible to rely on other family members to help with childrearing. Several politically active fathers observed that their party work would have been impossible if their wives had not assumed the lion’s share of childrearing and housework. When one father of three who successfully ascended the ranks of his party was asked about the possibility of combining politics and family life, he replied, “you can survive it [but] you must pay the price somewhere.” He claimed one needed to have a wife who had to act essentially as “a single mother.” In contrast, none of the women interviewed mentioned a husband who acted as a single father. Instead, politically active women mentioned relying on grandmothers or hiring nannies and babysitters to care for children, as well as housekeepers to deal with household tasks. One CDU member, a mother of four, argued, “if one wants to [get involved in a party], one can”; another mother active in politics agreed, “everything can be organized.” To do so, however, does require considerable financial resources not available to all Germans; I return to this point below. Page 84 →In contrast, parents who wish to, or must, care for their children on their own—regardless of their

sex—find it difficult to combine party life with childrearing and can easily be deterred both from joining a political party and becoming active while their children are young. One single father stressed that it was a “balancing act” for him to combine the two; he had turned down higher-ranking positions in his party because he wanted to respect “the human rights of the children” and did not want to subject them to the many party meetings he would have to take them to. Another man with two small children said he was putting off getting more involved in party work than he already was until the kids were older because “I want to be the one to raise my children” and more activity in the party would “cause me to be away from them too oftenВ .В .В . when I should be home putting them to bed.” Since women are more likely than men to be responsible for childcare, however, this problem disproportionately impacts mothers. Over and over again, interviewees repeated verbatim that mothers need to wait “until the kids are out of the worst of it” (“aus den GrГ¶bsten raus”) to become active in party work (see also KГјrschner 2009, 118). Interviews make clear that while parties can relatively easily recruit young women and older women, it is difficult to find female members in their childbearing years. As one party member in her fifties summarized: We try again and again to get young women excited.В .В .В . We have a youth organization, the Green Youth, that is relatively activeВ .В .В . and there are a lot of good people, also people with very good public speaking skills, many good young women. But then at some point they break away.В .В .В . A lot happens in the evenings and on weekends, even the meetings of party members. These kinds of events can only take place in the evenings because, of course, the other members work all day. So most things take place at 7 p.m. or 6 p.m.В .В .В . Even if you are capable of managing your time well, there are certain things you can’t manage because you don’t know how long they will take.В .В .В . You are constantly working with unknowns. That is naturally difficult if you have hired someone, or asked someone, to look after your children and you consistently come an hour late.В .В .В . that makes both you and the other person dissatisfied. Who wants to always live with something like this? It’s stressful. She then turned to her own experience: Fig. 2.5. Age Distribution of Eligibles. (Source: Candidate Interest Survey [N = 456].) Page 85 →I was always employed professionally at the clinic. I used to be involved with the union and as an ombudsman, and such things, but I didn’t have time for anything else. I worked full time and had two kids. After [the kids were older] I had a little more time, I joined the Greens.В .В .В . Now my kids are grown so I don’t have any problems with childcare. My Candidate Interest Survey illustrates the net effect of such individual decisions (see figure 2.5). The women who do join parties and assume positions of leadership, eligibles, are significantly older than their male counterparts.5 The mean age for female party leaders surveyed was 50.5 years whereas the mean age for men was 48 years. Only 18 percent of female respondents were under the age of 40 compared to 31 percent of men, indicating that women do indeed wait until their children were “out of the worst of it” before taking on leadership roles within a party. This pattern held in both parties with, and those without, quotas. In contrast, male party leaders were more evenly distributed across age groups. As male party leaders age they appear to be replaced by younger men but female party leaders are able to hold on to leadership positions longer, likely due to less competition from younger entrants. In sum, having a spouse at home who will aid in household tasks clearly Page 86 →facilitates party work and, in the current German recruitment environment, this spouse is generally the wife. As one woman observed, having a family “backs up” men so they can participate in parties, but having a family is a “burden on the backs” of women desiring partisan involvement. While the Greens and the Left Party recognize this fact and attempt in their quotas to overcome it, by for example providing childcare at national party congresses and granting national officeholders child- and eldercare subsidies, these efforts are very little, very late in the

recruitment process and benefit only those who have already found the resources to join and become active enough in the party to work their up way to a high-level post. Thus, as expected above, gender quotas in Germany have not reduced women’s unequal share of domestic responsibilities, making it difficult for women—especially those with children—to enjoy the necessary time to join and take part in party life.

Masculinized Recruitment Structures A closer investigation of German men’s and women’s participation in voluntary groups indicates that, as expected above, women’s underrepresentation in political parties is not only a matter of time constraints. While women on average devote less time to voluntary activities than men, they also spend their available time in different organizations then men (see figure 2.6). The most popular groups for Germans to join, regardless of sex, are sports teams, social clubs, and music or cultural groups; higher percentages of men than women report involvement in both sports and social groups while equal percentages of both sexes are members of cultural groups. Women’s fourth and fifth most frequent groups to join include the PTA and charitable groups and higher percentages of women than men maintain they are active in these type of bodies. Instead, twice as many men as women report membership in a professional organization. Figure 2.6 makes clear that, as suggested in the previous chapter, political parties are not a particularly popular type of organization for any citizen to join. However, they also seem to be particularly unattractive to women. While, for men, parties were the sixth most frequently joined group on a list of fourteen, for women parties ranked twelfth. The only organizations that women were less likely report affiliations with were the volunteer fire department and crime-fighting groups (of which only 1 percent of men and women were members). The 6 percent gap between men’s and women’sPage 87 → membership rates in parties was second only to the 6.5 percent difference in professional association membership, which can be explained by women’s significantly lower full-time labor force participation in Germany. Women’s less frequent presence in parties than men’s does not appear to be entirely explainable by differences in political interest either—on average only 3 percent fewer women than men reported involvement in nonpartisan citizen initiatives (Hess-Meining 2005, 395). Multivariate analysis controlling for a host of factors shown to shape political participation still finds that sex exerts an independent effect on whether German citizens join a party, with men more likely to become members than women (Klein 2006, 47; Glatte and de Vries 2014). Clearly, something about political parties make them particularly unattractive to even politically interested, civically engaged German women. Fig. 2.6. Citizen Involvement in Voluntary Associations, by Type. (Source: Hess-Meining 2005, 395.) Interviews with eligibles and potential eligibles in Germany, along with the results of my Candidate Interest Survey, confirm that, despite quotas, German parties have retained their male-normed rituals (see also Geissel 2013). These qualitative sources allow me to identify four mechanisms through which informal recruitment structures deter women from joining or participating more fully in political parties: local groups continue to be made up mostly of men, these organizations’ meetings maintain a masculinized ethos, the substantive discourse at such gatherings is at odds with women’s preferences, and party activities remain unnecessarily time consuming. Below I depict each of these mechanisms in turn. Page 88 →Party Groups Contain Mostly Men Respondents to the Candidate Interest Survey in parties without quotas indicated that, on average, their party group contained 73 percent men; while the figure for parties using quotas was indeed significantly lower,6 the mean party group still featured 65 percent men. Thus quotas have not altered the fact that, in Germany, joining a party means signing up for a club containing mostly men. This fact is not lost on politically interested women who consider becoming party members. Female interviewees mentioned a male-dominated atmosphere as a reason they were initially reluctant to join a party—either with or without a quota. One Social Democrat recalled that she had been active in politics as a student, but after she got married and moved to her husband’s rural hometown, she went to a meeting of the

local SPD and found only older men present. Then, in her words, “I thought to myself, no way,” and she did not return to the party for years. A 2009 study observed similar reactions to Left Party organizations in eastern Germany (Holtkamp, Wiechmann, and Schnittke, 54). A state legislator I interviewed remembered the first time she went to a meeting of the Christian Democrat’s youth wing, “there were only young men there in pinstripe suits” leading the gathering; another CDU member described her fellow party members as “men with beer bellies.” A female Free Democrat in her thirties recalled going to an FDP working group meeting in her city to encounter only older men; she then tried to join her party’s women’s auxiliary organization but found it barely functional. One CSU woman reported, “And the first thing that actually shocked me when I went to my first party meetingВ .В .В . was that the members seemed like really old men to me.В .В .В . I basically saw no other women.” Another CSU member confessed, [Y]es, you go into such a county-level party group and at first you’re shocked. I was with my girlfriend from school and together we were the only two women. If I’m completely honest I’ll admit I’d never have been able to stand it alone. (quoted in KГјrschner 2009, 135) A Masculinized Ethos Even if men are not overtly hostile to women’s presence, women entering a male-dominated domain may feel out of place. One recalled the reception she received when walking into the room for her first party meeting, the “very old men” at the gathering “looked at me really strangely and at Page 89 →first they thought I had surely made a mistake when entering the room” (quoted in KГјrschner 2009, 135). This feeling may emerge because other, informal, aspects of party life beyond the scope of candidate gender quotas remain highly masculinized. One such example involves when party groups choose to meet. Party meetings generally occur on the evenings and weekends, times when women pursue domestic responsibilities. For women who work part time, daytime meetings when children are in school may be preferable. One leader of the Social Democratic Party’s women’s auxiliary organization noted: In the younger years it’s the same [for girls] as boys in school and at the universityВ .В .В . but [later] it becomes difficult to activate them and get them involved.В .В .В . In general it is so that in the biography of women politics is unattractive because the political parties conduct their activities in the classic old times in the evenings. That often creates a big hindrance for those in their middle years to take part in politics. Because the party is a hindrance with its activities, we’ve got to try to create attractive possibilitiesВ .В .В . for women who aren’t members. And, as to what form this would take, we have not yet come up with a coherent idea yet. Moreover, where parties choose to meet may also serve as a deterrent to some women. One woman interviewed said her teenage daughter had decided against joining the local youth wing of the Social Democrats because the meetings were too “smoky and macho.” Isabelle KГјrschner’s 2009 study of women in the CSU found repeated references to meetings held in “backrooms” (135), “smoke filled” locales (135), or over beer (204). The above SPD leader praised a plan to hold breakfast meetings but lamented: That’s the kind of thing that men reject quickly; they say, you’ll just get together and eat, but that’s not true. Women would really like to have a different atmosphere. That doesn’t mean that they aren’t just as political or not getting down to business, just that the atmosphere really makes a difference. Even if official party meetings take place in party offices or during daytime hours, informal—and more binding—decisions are often made by a few party elders over informal beers drunk after the official meeting had ended. One woman observed: Page 90 →You have to be able to tolerate alcohol. You have to know how politics works. Political decisions are also made at night at the bar.В .В .В . you have to be there at night, you have to discuss things then and also drink something. Because that’s the way men conduct political business.

And whether you think that’s good or not, if you’re not there, your position won’t be recognized. (quoted in Kürschner 2009, 204)

This type of politicking is not limited to quota-less parties. Politically interested women, from different regions of the country, who were not active party members suggested that being shut out of such informal networks had deterred them from further involvement with the SPD as well. Type of Discourse and Substantive Focus In addition to off-putting meeting locations, many female interviewees objected to the substantive content of party meetings. Some women interviewed reported outright sexism in their male-dominated party organizations. A Christian Democrat recalled being one of two women at a meeting of the party’s youth wing when officers were chosen; when the other woman, a dental hygienist, volunteered to become the organization’s treasurer, the men told her she should “stick to drilling teeth” because women didn’t know about accounting. A female Free Democrat in her thirties recounted going to an FDP economics working group meeting and having older men ask her what a “little miss” (FrГ¤ulein) like herself was doing at a meeting about economic issues. A Social Democratic state legislator found that male party members often talked among themselves when she was speaking, rather than listening to what she had to say; they were more likely to interrupt her than one another and often overlooked calling on her if she raised her hand to ask a question. A Green state legislator noticed that, because she was quite petite, men often simply talked over her head at each other, ignoring her altogether. Members of an established all-male party group may become uneasy when women enter their ranks. One Christian Democrat recalled that after she was elected to the board of her local CDU youth group the other male board members were very uncomfortable: “They said вЂnow we will have to clean up our act and not tell dirty jokes.’ I thought that was just silly because I grew up with brothers so I thought of the dirtiest joke I knew and told it. This broke the ice.” Many female interviewees noted that male-dominated discussions in Page 91 →party gatherings often had a different substantive focus than they would prefer. One Green state legislator argued that while women viewed such occasions as a time to discuss policy substance (Sachfragen) men considered meetings an opportunity for self-presentation (see also KГјrschner 2009; Holtkamp, Weichmann, and Schnittke 2009, 56). One interviewee concluded that this is why women prefer interest groups and social movements to political parties, as the former tackle problems directly. A Christian Democratic woman argued that women in her party simply wanted to make good policies, but men “want to become this [position] or that [position].” A woman who ultimately decided not to join the Social Democrats recalled a party meeting she attended where environmental policy was being discussed. One female member made an impassioned speech about local environmental problems and urged her fellow party members to take up these concerns; some men present argued that the environment had become a passГ© topic politically and that the party should focus on other issues more likely to win the party votes. When she disagreed, a male colleague argued, “Do you want to be right or do you want to win?” She maintained that women “had a different relationship to power” than men and were more concerned about being substantively “right,” whereas men were more concerned with winning and “self-affirmation.” She concluded that events like these had deterred not only herself but other women as well. Other Christian and Social Democrats across Germany reported similar incidents. Female interviewees also lamented that if women desired to get ahead in politics, and in turn make the substantive policy changes they favored, they were required to engage in the same jockeying for power and self-promotion, even if they were uncomfortable with it.7 Otherwise, as one male party leader observed, even a hard-working, talented woman will “get lost in the shuffle.” The Recurring Problem of Time This type of discourse creates externalities that further depress women’s interest in becoming involved with

political parties. As each individual party member endeavors to call attention to him or herself, meetings drag on and become redundant. As one CSU member put it, “What one quickly discovers is that men in the party talk very, very much. This taking of stances, this self-presentation through large, long speeches, according to the slogan вЂeverything has been said, but not by everybody’—that’s the way party meetings are run” (quoted in KГјrschner 2009, 222; author’s translation);Page 92 → this comment was repeated almost verbatim in interviews done elsewhere in Germany (see Holtkamp, Wiechmann, and Schnittke 2009, 56). One Christian Democratic woman who was a doctor before entering a full-time state legislative position mourned: You have to be called to politics and like playing power games, either that or you have to be crazy or a masochist. When I worked as a physician, I saw thirty patients a day and accomplished something for each of them. [As a state legislator] I can sit all day in a meeting and nothing will have gotten decided. Sometimes on the way home I have a crisis; in my profession I could accomplish much more in one day than I can in politics. While of course not all men enjoy sitting through long-winded speeches either, because of the gendered nature of the recruiting environment, which places more domestic responsibilities on women, women simply have less time available to devote to such long meetings. As one Green state legislator put it, “sometimes you also have to make sure that there’s something in the fridge or that maybe you do a load of wash” and it would be nice “if you didn’t get home again at 11:30 p.m.” These hurdles create a vicious cycle: because women on average have less time than men, they are less likely to join political parties than men, and because parties are male dominated they continue their male ethos, which keeps meetings long and masculinized, further deterring women and perpetuating the cycle. Can Quotas Change Informal Recruitment Structures at All? To the degree that quotas successfully require women to assume top positions of party leadership, they may be able to propel women into posts where they can alter the time, location, or tenor of party meetings to be more welcoming of other women’s participation. Several female party leaders interviewed argued that as soon as they became head of their precinct party organization they had tried to make their group more female-friendly; such efforts may indeed inspire more women to join that particular branch of the party. One politically ambitious college student described her selection criteria for choosing a party: “#1: What does the party stand for? Where do I fit best [ideologically] and #2: What’s this party like in my town? What do they offer? How is the party group composed?” She ultimately decided to join the SPD’s branch organization, led by another woman. Older women reported that instances of blatant sexism had Page 93 →declined over the years and indeed, female members of parties with quotas currently in their late twenties and early thirties recounted fewer negative experiences with overtly sexist party organizations than older women or women in parties without quotas. Quotas like those of the Greens, which mandate that women are equal cochairs of local party groups, are best suited to placing women in positions of power where they may alter informal recruiting structures and, indeed, Green interviewees expressed concern about male-normed party practices less often than members of other parties. Both the Green and the Left parties’ statutes recognize the problem of masculinized discourse and require that national-level party meetings end when no further female speakers can be identified. These two parties do have the highest percentage of women members in Germany. However, both parties still contain a majority of men. The above-mentioned rules regarding discussions only apply to formal, high-level gatherings, which occur infrequently. They do not apply to the low-level, day-to-day interactions that shape whether or not politically interested individuals decide to become involved in a party in their own community. Of the respondents to the Candidate Interest Survey, only 10 percent reported that their precinct-level party group contained 50 percent or more women. Half of these groups were indeed Green Party organizations, but still even 80 percent of Green CIS respondents found themselves a party group containing over half men. Even parties employing quotas thus remain male dominated, and female interviewees were still conscious of being in the minority. On balance, then, quotas have exerted only a very limited impact on informal recruitment structures.

Population Density and Party Membership

Also as expected, there was a clear correlation between population density and women’s participation in political parties. Figure 2.7 compares the average percentage of female members in Germany’s three most thickly settled LГ¤nder (the city-states of Berlin, Hamburg, and Bremen) and its three most rural states (Mecklenburg-Vorpommern, Brandenburg, and Sachsen-Anhalt).8 In all parties except the FDP—which has few women in either rural or urban areas—higher percentages of women are present in the densely populated citystates than in the more rural LГ¤nder. The gap—10 percent—is most apparent in the columns for the Bayernbased CSU because its results are based on district- rather than state-level memberships, which can more finely distinguish between urban and rural areas. Page 94 →The Candidate Interest Survey also found that, for all parties except the Greens, respondents from towns smaller than 20,000 residents had lower percentages of women in their precinct organizations than did respondents hailing from cities with over 500,000 inhabitants. Fig. 2.7. Female Party Membership and Population Density: 2010. (Sources: Greens: Niedermayer 2011, Die Linke 2011; SPD: Ferner 2011, 17; CDU: CDU Deutschland 2010, Table 3; CSU: KГјrschner 2009, 102; FDP: Liberale Frauen 2010.) Interviewees offered various explanations for these differences. Many argued that gender roles are less traditional in large cities, making sexist party leaders less likely (see also Holtkamp, Wiechmann, and Schnittke 2009).9 Others observed it took more time to get involved in a rural party organization because going to meetings often involved a long drive to another area, whereas in cities party meetings were always held nearby, rendering participation less burdensome for time-strapped women (see also Eder and Fortin-Rittberger 2013). Finally, one urban interviewee maintained: I think it’s the structure of the cityВ .В .В . there’s always been the interaction between densely populated spaces and women who were in involved in many different areas and who built networks; these women are much more likely to find other like-minded women here than in the countryside.В .В .В . I hear it time and again from students, Page 95 →from young women [in the party], that they can’t imagine living in a rural area, and I hope that they appreciate the allies that they’ve been able to find here. Women therefore remain a minority in all political parties, particularly in rural areas, and this helps account for why parties do not always fulfill their quotas at the local level, especially in small communities.

Gender Quotas and Eligibles However, as established in the previous chapter, in Germany eligibles for elective office are not simply party members. Instead, gatekeepers select those who have demonstrated loyalty by assuming inner-party leadership posts. Studies of the German parties’ mass membership find only minimal sex differences in terms of who becomes active within a party; in 1998 female party members were slightly more active than male party members and in 2009 men nominally more so than women (Spier 2011, 100–101). It is in this regard that quotas exert their first noticeable impact on the political recruitment process: they turn the small number of female party members into a much larger percentage of female eligibles. In all parties with quotas, the affirmative action measures employed require higher percentages of female party officers than the proportion of women in the party. While women make up only 37 percent of the Greens’ and Left Party’s members, (at least) 50 percent of the leadership posts are reserved for them. Women comprise 31 percent of the SPD’s rank and file but between 40 percent and 60 percent of inner-party offices are required to be filled by women. The CDU’s 33 percent quorum is closer to, but still above, the actual percentage of female party members (26 percent). Just as with the implementation of the quotas for women in elective office, these inner-party quotas are generally met, especially in high-level positions, where smaller numbers of women are required to fill the quota than for more localized positions. Figure 2.8 compares those parties employing quotas and those without quotas in terms of their percentages of women among rank-and-file members, their county board members in the five states surveyed by the Candidate Interest Survey, all their Land-level governing bodies, and their federal executives.10 This figure demonstrates quotas’ elevator effect. While parties with quotas average 30 percent female members, 37

percent of their county-level inner-party leaders, 44 percent Page 96 →of their Land, and 42 percent of their national boards are composed of women. All parties with binding (near) parity quotas for inner-party offices met or exceeded their quotas at the state and national levels and the CDU came close (31 percent) to meeting its 33 percent quorum at the state level, although it failed to do so at the national level. While no party was able to meet its own quota at the county level, the average percentage of female inner-party leaders was higher than the percentage of women among rank-and-file party members. Fig. 2.8. The Elevator Effect. (Sources: Author’s calculations based on party organizations’ websites in 2013; Left Party and SPD County Leader data from Ferner 2011, 20 and Die Linke 2011. For membership figures, see fig. 2.2.) Comprehensive data about local-level party executive boards is not available for most parties, but the SPD’s internal documents reveal a similar, elevator-like pattern. Although only 31 percent of SPD members are women, in the most densely populated states women make up almost 40 percent of local party boards and for the five LГ¤nder surveyed by the CIS the figure is 34 percent (Ferner 2011, 21). Internal Left Party, SPD, and CDU studies also establish that, compared to their percentage in the rank and file, women are for the most part overrepresented in positions such as delegates to nominating commissions, the leadership of most inner-party Page 97 →organizations, and in party working groups (Ferner 2011; Die Linke 2011; CDU Deutschland 2010).11 In contrast, as depicted by figure 2.8, where quotas are not in place, the national board contained a lower proportion of women than the party’s already very low percentage of female rank-and-file members. Parties without quotas did promote slightly higher percentages of women to county and state leadership positions than they had within their ranks, but this elevation occurred to a far lesser extent than where affirmative action was employed. Moreover, the overall percentage of women leaders was far lower in quota-less organizations. One study of the CSU found only 9 percent of local party leaders to be women (KГјrschner 2009, 101), a lower percentage than the ideologically similar CDU had in every other state (CDU Deutschland, 2010, table 8). The FDP does not keep detailed statistics about the gender breakdown of local party boards, delegates, or working group members, but given the low percentage of women in the party and at top levels of leadership, it appears unlikely that women are overrepresented in these positions. Thus, while quotas have not increased parties’ female membership by much, they have accomplished an increase in the percentage of female party leaders, especially in the parties with parity and near-parity quotas. Today almost half of Green and Left party eligibiles, and 40 percent of the SPD’s, are women. The fact that women occupy board positions is essential for political recruitment because these are the peak positions from which top elected officials are drawn (Kintz 2011). These patterns are also in sharp contrast to the prequota era, when women were markedly underrepresented among party leaders (Hoecker 1995). Parties without quotas continue to have far lower percentages of female eligibles. In sum, despite a generation of quotas in Germany, women remain underrepresented in all parties, even those with parity affirmative action policies. The highly gendered recruitment environment, combined with masculinized recruitment structures, particularly prominent in rural areas, render female citizens, on average, less likely to become party members than their male counterparts. Quotas have not achieved “justice” or the equal participation of men and women in this first phase of the democratic process. Parity and near-parity quotas do improve women’s participation at the eligible stage, however, by creating almost equal numbers of male and female eligibles. These potential elected representatives are drawn from a narrow subset of the citizenry, however. The remainder of the chapter delves into the implications of these findings for women’s substantive and descriptive representation.

Page 98 →Implications for “Justice” and Women’s Descriptive Representation While women are underrepresented among party members, they are not so underrepresented that women’s descriptive representation cannot be achieved in most communities. As of 2010, citizens of the Federal Republic elected fewer than 50,000 representatives and German parties contained well over 300,000 female members. The

actual number of women required to fill party quotas is far lower than even 25,000 because the SPD’s locallevel quota and CDU’s quorum are below 50 percent and quotas do not apply to directly elected seats. Moreover, an individual may hold more than one elected position at once. Thus, not only are there enough women in parties to fill half of all elected offices in the country, but also enough to occupy every single elected position with a woman several times over. Table 2.1 depicts the ratios of each party’s female members to the number of women that would have been necessary to fulfill its minimum quota given actual election results.12 Thus there are plenty of female party members in every organization from among whom state, federal, and European Parliament candidates can be selected. While less pronounced than at higher levels of government, there are also multiple female party members available to fill every local office. However, here the ratios of party members to available seats are much smaller and, as indicated above, female members are not distributedPage 99 → equally across Germany. Because women are less likely to be found in rural or weakly organized party organizations, the precise odds will differ from community to community. In very small communities it may be difficult for some parties to find enough female members to fulfill their quota promises.13 Multiple studies of quota implementation in Germany have found positive relationships both between the population of the community, the degree of organization of political parties, and parties’ ability to meet their self-imposed targets for women in elective office (Davidson-Schmich 2006b; Holtkamp, Wiechmann, and Schnittke 2009; Eder and Fortin-Rittberger 2013). As a result, in later analyses I control for community size. Table 2.1. Ratio of “Women’s” Elected Positions to Female Party Members in Five Representative LГ¤nder Local-Level Offices High-Level Offices SPD 1 seat for every 29 female members 1 seat for every 500 female members 1 seat for every 38 male members 1 seat for every 740 male members CDU 1 seat for every 28 female members 1 seat for every 460 female members 1 seat for every 39 male members 1 seat for every 678 male members Green 1 seat for every 8 female members 1 seat for every 202 female members 1 seat for every 14 male members 1 seat for every 334 male members CSU 1 seat for every 16 members 1 seat for every 932 members (3 of whom are women) (177 of whom are women) FDP 1 seat for every 14 members 1 seat for every 354 members (3 of whom are women) (81 of whom are women) Source: Author’s calculations, based on party members: Bundeszentrale fГјr Politische Bildung; elected positions: state statistical offices of Baden-WГјrttemberg, Bayern, Bremen, Hamburg, and Nordrhein-Westfalen. See chapter 1 for a discussion of which offices are considered “local.” Outside of small communities, however, German women’s low level of participation in political parties does not have much of a negative impact on these organizations’ ability to fulfill their gender quotas and achieve women’s descriptive representation, especially at higher levels of government where there are large numbers of female eligibles for each available seat in parliament. This finding is in keeping with research done elsewhere that found no “supply side” problems in locating candidates to fill electoral law quotas (Casas-Arce and Saiz 2011; see also Ashe 2015). Concerns that women elected from parties with quotas are mere tokens seem unfounded; indeed, studies of women’s substantive representation in the Bundestag clearly establish that female MdB do not hesitate to speak up on behalf of women (Xydias 2008). Table 2.1 also demonstrates that, while there is considerable competition for “women’s” ballot spots, the level of competition is indeed lower than for the “male” positions in parties with quotas. This disparity may even be greater given that it is not party members, but leaders, who are tapped to run for elective office. Since women have better odds of obtaining a leadership position than their male counterparts where quotas are employed, female members are more likely than males to be in the subset of members selected to run for these

elective seats. In addition, these figures are calculated based on the minimum quotas for women. Because membership figures are self-reported, they also very likely overestimate the number of actual party members. The percentage of active party members is also far smaller than the overall number of members, further reducing the number of women (and men) who are actually eligible to run for elective office (Spier 2011, 100–101). To summarize: while there are enough women to fill German parties’ quotas and to create competition for “women’s” list positions, there is less competition for a candidacy among women than among men where quotas Page 100 →are used. Where quotas are not employed, men and women possess (at least formally) equal odds of selection, but there are very few women present.

Implications for Women’s Substantive Representation The lower number of women in political parties will likely have a much stronger impact on women’s substantive representation than it will on descriptive representation. Because there are far fewer women than men in German political parties, from the very start of the political recruitment process Germans’ female elected representatives are drawn from a much smaller pool than their male politicians. Empirical research has shown that the larger the number of women who take part in policy debates, the wider the range of women’s issues and interests brought to the table (Celis 2006 and 2009). Moreover, the ways in which representatives frame “women’s” interests is closely correlated with partisanship (Xydias 2008), and to the degree that women are more underrepresented in certain parties than in others these perspectives will in turn receive less attention in elected bodies. The limited number of women in the German candidate pool thus translates into a more constrained choice of female representatives than male representatives. This outcome is especially pronounced where parties do not have quotas or where quotas require low percentages of women in leadership posts. Who Are Eligibles? Even in terms of men, however, the choices are narrow to begin with. Not only are Germany’s political recruitment structures and environment gendered in their effects, they also make it easier for some types of women (and men) to join parties than it is for others. As a result, the women who do join German parties and rise to positions of local leadership are not representative of the underlying population; neither are male eligibles. Instead, these organizations’ members have four main differences to nonparty members: they have different social networks than most Germans, they are better educated than the general population, have a higher household income and different employment patterns than the citizenry at large, and are likely to be ethnically German rather than possess an immigrant background. Those who go on to become party leaders—those eligible for elective office—are even less representative of German society; they Page 101 →are more educated, affluent, and atypically employed than rank-and-file party members. It is also likely that their social networks are even more politicized than the average party member’s. Thus not only do Germany’s recruitment environment and recruitment structures limit the number of women who join parties, they also ensure that the women who do are not typical German women. (Of course, neither are male party leaders typical German men!) I discuss the key differences below. Regardless of whether they are male or female, or whether their party has a quota or not, eligibles in Germany are more likely to come from households where members are more prone to talk about politics, join parties, or run for office than are other Germans.14 Of the party leaders surveyed by the Candidate Interest Survey, 64.7 percent said that they often talked about politics in their home when growing up. Interviewees often brought up these family experiences as reasons why they had chosen to become party members. In addition, 35 percent of them reported having a family member who had run for elective office and 7.9 percent listed two family members. Of those respondents who were married or who had a domestic partner, 50.2 percent said that their spouse was also a member of a political party. Given that only 2.2 percent of the German electorate joins a party (Niedermayer 2011), and fewer than that actually run for office, it is clear that members of the eligibility pool do not hail from typical social networks.

Similarly, members of German political parties—regardless of whether they are male or female or whether their party employs a quota or not—have much higher levels of education than the general population; those who rise to the leadership of their local party organization had even more schooling. Given that interest in, and knowledge of, politics are highly correlated with educational attainment, this finding is not surprising (Verba, Schlozman, and Brady 1995, 350). While only 12 percent of the German population had a completed university degree or a PhD (Statistisches Bundesamt 2008, 131), 39 percent of rank-and-file party members (Klein 2011, 47) and 63 percent of the party leaders surveyed did.15 Regardless of their sex or whether their party features a gender quota, eligibles’ household income falls above the median. While the median household in Germany brought home between €1,700–2,600 per month (Statistisches Bundesamt 2008, 549), the median household income category of party leaders surveyed was between €3,200 and €4,500 per month.16 This finding helps explain how female eligibles were able to afford the household help depicted earlier in this chapter. Page 102 →Employment Patterns Both political party members in Germany and party leaders are, on average, more likely to work outside the home than other Germans and their employment patterns are far from typical. Instead, members and leaders of German political parties are disproportionately employed in jobs that allow them the flexibility to combine paid work with volunteer political activity—or they are not currently in the workforce at all (see figure 2.9). Half of the individuals interviewed mentioned the ability of a politically interested person to combine paid employment with volunteer party work as one of the most important factors that could either help or hinder someone from becoming an active party member or elected official. Here I discuss party members’ and party leaders’ employment patterns in more detail. As is the case in the German population as a whole (European Commission 2013), male eligibles were more likely than female party leaders to be employed and, as will be discussed below, there were slight differences between women from parties with and parties without quotas. For the most part, however, men and women in the candidate pool—regardless of what party they are in—more resemble each other than they do rank-and-file party members or the citizenry at large. Working party members are most commonly employed in the public sector; in Germany civil servants who pursue political activity are allowed to adjust their work schedule to accommodate any official party event they must take part in and, should they get elected to full-time elective office, they are allowed to leave their post and are guaranteed an equivalent position if they should subsequently lose an election or voluntarily leave politics. While only 7 percent of the German workforce was employed in the public sector, 35 percent of rank-and-file party members worked in public posts (Klein 2011, 50). Similarly, 36 percent of the male and 28 percent of the female eligibles surveyed by the CIS held public sector positions. A Social Democrat employed as an administrator at a public university reported that it was “very easy” for her to combine her job and her political activities; when she had to leave her job to attend a political function it was “self-evident” to her boss that she could do so and she believed that it was “viewed positively” in her workplace that she was so involved in public affairs. She claimed that “no one has ever complained to me” about her having to miss work. A public school teacher concurred that all he had to do was show the principal the invitation to a partyrelated event he had to attend and a substitute teacher would be arranged. In stark contrast, 83 percent of employed Germans work in the private Page 103 →sector. This figure fell to only 40 percent for rank-and-file party members (Klein 2011, 50). Even fewer party leaders reported employment in the private sector: the figure was 26 percent for male leaders and a mere 21 percent for female eligibles. Private sector jobs render it especially difficult for people to become active in their local party organization and to assume leadership positions. A Christian Democrat interviewed who had worked his way up to be a chief executive officer of an insurance company said that while he would have liked to have become more involved in politics, his profession had not allowed it. He had moved seven times over the course of his career and after each move he had to reestablish himself in a new party organization. A young female employee of a mercantile company feared that devoting too much time to political activity would limit her professional opportunities, and a party member from a quickly developing high-tech industry feared that her political work would take time away from keeping up with developments in her field. Private employers, in contrast to public sector workplaces, are under no legal obligation

to release employees for political work or hold their positions for them while they assume elective office. Fig. 2.9. Employment Patterns among Eligibles, Party Members, and the General Public. (Source: Eligibles: CIS; Party Members/General Public: Klein 2011, 49–50.) The self-employed, however, do enjoy the freedom to set their own hours, which enables them to combine party membership with employment. Less than 1 percent of the German workforce is self-employed but 6 percent of rank-and-file party members are (Klein 2011, 50). In stark contrast, 23 percent of male and 21 percent of female party leaders worked for themselves. One active member of the CSU owned a restaurant and maintained that her freedom to close her establishment the day of party-related Page 104 →functions was what had enabled her to assume inner-party leadership posts; an owner of a flooring and wall coverings shop also shut his business when his presence at party activities was required. Others who performed freelance work, for example two journalists interviewed, also enjoyed the flexibility to set their own schedule, which allowed them to take part in party work when they needed to. Depending on whether or not their party utilized a quota, female party leaders did have some cross-party differences in employment. Female eligibles in parties without quotas were significantly more likely to be selfemployed than women in parties with quotas (p ≤ 0.000). In contrast, female party leaders from parties with quotas were significantly more likely to be employed in the public sector compared to female leaders from parties without quotas (p ≤ 0.10). However, the same pattern occurs with male party leaders as well and is likely driven by the fact that the quota-less FDP is strongly supportive of entrepreneurs. Both of these employment patterns are atypical of the underlying population, however. Other party members and party leaders were not currently active in the workforce at all. Forty-three percent of rank and file party members were retirees compared to 29 percent of the overall population (Klein 2011, 49). However, while many of these older members may continue to pay dues, they are no longer active in the party. This can be seen by the fact that only 8 percent of male and 13 percent of female party leaders were retirees and the average age of party leaders was the same as that of the German population as a whole—49 years (Statistisches Bundesamt 2008). Interview respondents reported strong norms encouraging those over 65 to step aside from elective office or inner-party posts to create opportunities for younger party members. At the other end of the age spectrum, among party members as a whole and among party leaders, the percentage of students or trainees (3 percent) paralleled that of the underlying population (4 percent) (Klein 2011, 49). The unemployed were underrepresented in comparison to the general population; given that the unemployed are likely to be consumed with looking for work, their underrepresentation in voluntary leadership is perhaps unsurprising, however. Finally, the CIS asked respondents to self-identify as “homemakers” or “other.” Just as in the general population (European Commission 2013), female party leaders were far more likely (12 percent) than their male counterparts (3 percent) to fall into this category. The female party leaders surveyed were significantly less likely to be engaged in paid employment than male party leaders. Moreover, while women across parties were equally likely to be unemployed, students, or housewives, women in parties Page 105 →with quotas were significantly more likely than women in parties without quotas to be retirees.17 By reserving inner-party positions for women, parties with quotas ensure that older women are not pushed aside by younger men wishing to begin their Ochsentour; these senior women face less competition from younger women who, on average, are likely to be occupied by building families and careers and have little time to devote to party leadership. Thus German party members and especially party leaders have different employment patterns than the German population as a whole. They are more likely to work in the public sector or be self-employed than the general public and less likely to be employed in the private sector, retired, or unemployed. Only students were represented in political parties’ memberships and leaderships in proportion to their underlying percentage in the German population. Ethnicity

Finally, party members and leaders—regardless of their sex or whether their party employs a gender quota—differ from the overall German population in an additional important respect: they are less likely to have what can be described as a migrant background. In Germany, 15 million individuals are identified as “migrants”—these are people who (a) were not born in Germany or (b) are not German citizens or (c) had a parent who fell into one of the first two categories; this amounts to almost 20 percent of the German population. Of these, only eight million are German citizens (Die Beauftragte der Bundesregierung 2012, 27; see also WГјst 2011, 256). This migrant population is far from homogeneous. Many migrants are ethnic Germans from eastern Europe who moved to Germany in the early 1990s with the fall of the Iron Curtain. Another large group includes Turks who came to western Germany as “guest workers” and their descendents. While eastern European ethnic Germans automatically received German citizenship, Turks have faced high barriers to becoming German citizens. Iranians are the second largest non-ethnic-German group among naturalized citizens (Donovan 2007, 459–60). Of adult German citizens (i.e., eligible voters and those eligible to run for elective office) only 10 percent are of immigrant background (WГјst 2011, 256). Whether they are citizens or not, migrants may join the major German parties—except for the CDU, which limits its members to German and EU citizens. Political parties do not keep formal records of the number of migrants in their ranks. However, interviews with active party members—evenPage 106 → in areas with high immigrant populations—make clear that the rank-and-file membership of all the major German parties does not come close to the 10 percent migrant mark in the electorate, let alone the almost 20 percent in the underlying population. When asked about migrant members of their local party organization, most rural interviewees indicated that there were none at all or could list only one or two people. In larger cities with higher immigrant populations, especially in parties of the Left, interviewees were more likely to report having migrant members in their precinct organization, but they also underscored that there were disproportionately few compared to the numbers of migrants in their neighborhoods. Reasons given for this state of affairs ranged from migrants’ low levels of education and high levels of unemployment, interest in home country rather than host country politics, language barriers, a lack of effort on the part of the parties to recruit migrants, and discrimination by German party members. Survey research has also found migrants to be less interested in and knowledgeable about politics than nonmigrants; polls have found that migrants were less able to name the major German political parties and less likely to identify with any of them than nonmigrants (WГјst 2004). Other studies of minority groups in Germany have also found only weakly institutionalized ties between migrants and politically powerful entities like parties (Brady 2003); this holds true elsewhere in Europe as well (Erlanger 2008). Of all the groups in society, Turkish women were the least likely to be engaged in voluntary activities of any sort (Hess-Meining 2005). While statistics on migrant membership in political parties are difficult to obtain, what is easier to document is that naturalized German citizens are underrepresented among the ranks of party leaders. Of the 1,068 CIS recipients, less than 1 percent had names that suggested they were not ethnic Germans.18 While Turks are the second largest immigrant group to Germany, and while most Turks are Muslims, only one (female) party leader surveyed was a Muslim (0.2 percent of all respondents). Because of the dearth of ethnic minorities in the CIS sample, subsequent quantitative analysis is unable to provide further intersectional analysis in terms of national origin. The lack of ethnic diversity among grassroots party leaders extends upwards. One study of all political parties’ federal-level leadership found a total of only three executive board members who were nonethnic Germans—while each party’s executive board contained dozens of members. At the state level, seven states had no nonethnic Germans on any party’s executive board. In those nine states that did have non-German members on at least one party’s executive board, the average number of Page 107 →non-German board members was a mere 1.7 per state while, again, parties’ executives routinely contained one to two dozen members (Donovan 2007). Thus the leadership of the major German parties is far less likely to contain members of migrant groups compared to Germany’s population as a whole. As a result, very few candidates or elected officials in Germany are of migrant descent. In the 2005 Bundestag election, only 4 percent of candidates had a non-German background (de Fonseca 2011, 121) and in 2013 only 16 Members of the Bundestag, 2.6 percent of the total, were of migrant descent and only five of those were women

(Donovan 2012, 35).19 While in Berlin’s state legislature 9 percent of MPs were migrants, in some of Germany’s state legislatures no migrants are represented at all. In Berlin, two-thirds of the migrant legislators were women, likely attributable to the fact that the Berlin SPD also has a quota for migrants and can “kill two birds with one stone” by nominating minority women (Donovan 2012, 35; Geissel 2013; Reiser 2014). Thus a double quota seems to create an elevator effect for minority women (Hughes 2011; see also Kintz 2011). In sum, then, the women who join parties and become eligibles for elective office—just as the men who do—are not typical of the German population as a whole or even rank-and-file party members. They are more educated, come from more affluent households and politicized social networks, are more likely to be ethnic Germans, and tend to cluster in occupations including public service and self-employment. Although “quota women” are more likely to be retired than nonquota women, both groups are very atypical, when compared to women in their own party and to women in the population as a whole. Intersectionality and Substantive Representation These differences are important when it comes to women’s (and men’s) substantive representation. The women who do join parties and especially those who are eligible to become elected officials will likely have certain conceptions of other women’s needs that are colored by their own experiences. Party members’ high levels of education and household income will likely shape their conceptions of which economic issues are important. For example, under a female chancellor and with a Bundestag containing almost one-third women, Germany has made considerable reforms to its reconciliation policy (policies designed to allow parents to combine career and family). These reforms, however, have enabled professional women to more easily pursue careers (for example, by hiring low-wage domestic Page 108 →help) while not increasing the social safety net for poorer women (von Wahl 2011). Interviewees also report that voters often complain to them that parties made up of public sector workers do not adequately grasp the challenges globalization poses for private business. Moreover, the small numbers of unemployed in the top ranks of political parties likely limit politicians’ firsthand experiences with programs such as the Harz IV welfare reforms, and the tiny percentage of single parents leading German parties suggests that the challenges faced by this (predominantly female) group may not be appreciated. While a majority of low-wage workers in Germany are women, few if any women elected at the national level have experienced such employment, and progress toward a legal minimum wage in Germany was slow in coming (Henninger 2013). The homogeneous nature of German parties may limit politicians’ interaction with, and attention to, the concerns of migrants, poorly equipping them to develop public policies to improve immigrants’ lives. Ethnic German feminists and their allies in political parties often fail to identify female migrants as “women,” categorizing them as “Muslims” instead (Rottman and Ferree 2008). As a result, even with a critical mass of women in government, Germany only passed a very weak antidiscrimination law and has failed to vigorously enforce it (von Wahl 2011). Although the Bundestag passed an immigration law in 2005, designed to help integrate immigrant women into German society, the law initially failed to make much of an impact because it did not address the issues of primary concern to female migrants. As result, two years later the law had to be significantly amended to better achieve its aims (Donovan 2013). Finally, given that party members come from highly partisan social networks they also may be out of touch with the average citizen’s dissatisfaction with party politics. Women (and men) who do feel comfortable in Germany’s parties may be ineffective ambassadors to recruit other citizens, male or female, for political life. Observers of German politics lament the increasing disconnect between partisan politicians and ordinary citizens and are becoming increasingly vocal in their calls for reform of German political recruitment structures (Dassonneville, Hooghe, and Vanhoutte 2012). This Parteiverdrossenheit has manifested itself of late in many forms including the rise of the antiestablishment Pirate and Alternative fГјr Deutschland (AfD) parties and tens of thousands of WutbГјrger (irate citizens) taking to the streets of Stuttgart to protest development plans. The small numbers of white, affluent, educated, socially active German women present in parties also translates into fewer options to representPage 109 → citizens. Debates over the desirability of legislated quotas for women

on corporate boards in Germany illustrate how even women with similar socioeconomic backgrounds from the same party—in this case the CDU—may hold very different perspectives. In this debate Labor Minister Ursula von der Leyen vehemently favored quotas, Women’s Minister Kristina SchrГ¶der equally adamantly opposed legal quotas, and Chancellor Angela Merkel ultimately brokered a compromise to postpone making a decision on the issue (until after her own reelection campaign). Limited numbers of potential female eligibles—even those from similar backgrounds—translate into fewer choices of female representatives for citizens. In addition to shaping substantive representation in this way, the atypical background of women who do rise through the ranks of political parties provides a certain kind of symbolic representation as well. Studies in other contexts have shown that when members of racial minority groups are elected to visible positions, citizens from these groups have higher levels of political interest (Banducci, Donovan, and Karp 2004). Women with low levels of education, low household income, or of migrant background do not see themselves re-presented in the way that more educated and affluent ethnic German women do, and may thus remain less interested in politics, further limiting the demographic diversity present in parties. In sum, gender quotas and the resulting increase in women’s descriptive representation in Germany cannot be said to have offered symbolic representation, at least when it comes to inspiring women to join political parties. A generation after the adoption of gender quotas, all German parties remain male dominated. Moreover, fewer women in total are members of political parties today than when quotas began to be adopted. Given the gendered division of labor in Germany, women—especially those with young children—have difficulty finding the time to join and become involved in political parties unless they enjoy domestic assistance. Moreover, informal recruitment structures remain highly gendered, based on a male life course and masculine norms, further depressing women’s participation in this first phase of the political recruitment process. Just as their male counterparts, the women who do join political parties in Germany are far from typical women. Quotas do exert one noticeable effect at this stage, however, as they alter the process for selecting inner-party leaders and guarantee that the small numbers of women who do join political parties with quotas enjoy a favorable political opportunity structure. Through this “elevator effect,” quotas have aided women in ascending the ranks of parties with quotas, especially Page 110 →parity and near-parity quotas, so that women are well represented in the ultimate eligibility pool. In the quota-less parties and the CDU, women remain underrepresented among eligibles. Not all inner-party leaders may desire to become elected officials, however, and the next chapter turns to the question of whether the women who do manage to overcome hurdles and become inner-party leaders aspire to elective office at similar rates as their male peers.

Page 111 →

Three Aspirants If a 30-year-old woman with two kids and a full-time job came to me and asked whether she too should serve in local government, I’d say keep your hands off this position. Or are you prepared to work seven days a week over the course of years? I have to admit, my conscience won’t allow me to encourage another woman, because I’ve learned from experience what it is like. —Female SPD local council member (quoted in Holtkamp, Wiechmann, and Schnittke 2009, 55; author’s translation) I’ll wait and see what happens. I don’t even know if I could do [the job]. There are so many other people in the party who do want to run. I don’t want to take anyone else’s spot away. —Response of a 27-year-old female law student, vice president of an SPD precinct organization, and leader of the state-level Young Socialist organization, when asked whether she would ever run for her city council The previous chapter reveals that candidate gender quotas have done very little to change the political recruitment environment and informal recruitment structures in Germany, making it difficult for women to join and become active members of political parties. Despite their low percentages of female members, however, several parties have adopted parity or near-parity quotas for inner-party offices, elevating the percentages of women in the eligibility pool. The same occurs, to a lesser extent, in the CDU, which employs a 33 percent “quorum.” In these parties, female party members enjoy better numerical odds than their male counterparts of being selected Page 112 →for posts subject to a quota, including both inner-party offices and electoral list positions. Because sex is not a consideration in the remaining parties, the few women present among party leaders benefit from no such numerical advantages and men have no such disadvantages. This chapter investigates quotas’ impact on women’s propensities to develop aspirations for elective office. As the model of political recruitment presented earlier notes, not all individuals eligible to run for elective office are interested in doing so. People may join parties and become officers within them not because they want to become politicians, but so they can have an influence on policy or experience social solidarity with other members (Gallagher, Laver, and Mair 2011, 332); some may be drafted to a party board not due to political ambition but because they possess bookkeeping or clerical skills, making them a suitable treasurer or secretary. Moreover, German law requires that all local party organizations elect a slate of officers; in small communities where a particular party is not very strong, individuals may be roped into service simply because an organization requires someone to occupy a legally mandated post.1 Here I examine the second step in the political recruitment process: becoming an aspirant. Aspirants, or individuals with political ambition, are a subset of eligibles—those qualified people who desire elected office and who have the resources to pursue it. Research conducted where quotas are not in place consistently finds female members of the candidate pool less ambitious than their male counterparts (Norris and Lovenduski 1993; Thomas, Herrick, and Braunstein 2002; Lawless and Fox 2005, 2010; Fulton et al. 2006; Maestas et al. 2006; Lawless and Theriault 2006; Lawless 2012). In contrast, political ambition has rarely been systematically investigated where quotas are employed. Instead, studies of gender quotas’ impact generally begin at the candidacy stage (e.g., Htun and Jones 2002; Jones 2004; Matland 2006; Meier 2004) or after elections have been held (Escobar-Lemmon and Taylor-Robinson 2008; Franceschet, Krook, and Piscopo 2012; Thames and Williams 2013). This chapter extends the study of gender and political ambition beyond the quota-less Anglo-Saxon context and expands the study of quotas beyond the final phases of political recruitment. I investigate a second possible symbolic effect of gender quotas, and ask whether, where quotas are used, women in the candidate pool are as likely as men to develop political ambitions. Quotas have not resulted in women flocking to political parties, but

can they awaken the political ambition of those women who do join? I find that quotas have been successful in creating female eligibles who have at least considered a run for elective office. However, quotas have Page 113 →been less efficacious in prompting these women to say they would, after thinking it over, actually accept a nomination to run for office at the local level. Despite the use of quotas, women are still on average more likely to doubt their own suitability for elective office than men and to find themselves in domestic circumstances that depress political ambition. Here again, although quotas can change formal recruitment structures such as who may occupy a certain inner-party post or ballot spot, they cannot alter the backdrop against which political recruitment occurs—neither the gendered recruitment environment nor the informal party norms that shape what constitutes a “qualified” candidate. I develop these conclusions as follows. First, I discuss the importance of political ambition in a democracy. Second, I turn to the theoretical relationships among candidate gender quotas, the recruitment environment, informal party practices, and political ambition. Third, I empirically investigate levels of ambition among German eligibles, and, using interviews and data from my Candidate Interest Survey, I depict the mechanisms that shape political ambition in Germany. Finally, I discuss the ramifications of my findings.

The Importance of Political Ambition One of the first political scientists to systematically study political ambition was Joseph Schlesinger; he expected that a certain proportion of the population in any given democracy would harbor the desire to become a politician, and, once elected, these politicians would likely aspire to be reelected or to move up to an even higher political office (1966; for a review of this literature and contributions to it, see Lawless 2012). This constant supply of ambitious politicians is vital for ensuring democratic accountability. If no candidates come forward to challenge officeholders, voters will be unable to “throw the bums out” should elected officials not perform well. Similarly, if elected officials are unconcerned about their own reelection, they have little incentive to serve constituents during the period they are in office. In a context where women, or other groups, remain descriptively underrepresented, a consistent supply of politically ambitious women, or minorities, assumes particular importance for democratic accountability. If low numbers of women occupy elected positions, female voters may be lulled into believing that the few representatives who share their descriptive characteristics will also substantively represent them and they may fail Page 114 →to check up on their descriptive representative’s behavior in office (Mansbridge 1999, 640). Alternatively, even if members of a minority realize “their” member of parliament is failing them, they may remain loath to challenge her if no alternative can be found. Thus, the presence of other ambitious minorities who desire an elective post offers an important corrective, as these up-and-coming politicians possess incentives to detect and publicize flaws in their group’s existing representative. Thus, where women have not yet reached parity in descriptive representation, women’s political ambition plays an additional important role in achieving democratic accountability. The existing literature on quotas does not offer much empirical evidence about gender quotas and political ambition. While political ambition is often studied where “entrepreneurial” recruitment processes are in place—that is, where parties are relatively weak, quotas are not used, and candidates must come forward on their own (Norris 1997, 12)—political ambition is hardly ever examined in party-loyalist systems of candidate selection where party quotas are used (Siavelis and Morgenstern 2008).2 The few studies of women in political parties with quotas that do touch on gender and political ambition do not survey men, so no meaningful gender comparisons can be made (e.g., Geissel 2000; HorstkГ¶tter 1990). Most of those examining the impact of quotas begin their investigations at the candidacy stage (e.g., Jones 2004; Htun and Jones 2002; Matland 2006; Meier 2004) or after elections have been held (Escobar-Lemmon and Taylor-Robinson 2008; Franceschet, Piscopo, and Krook 2012; Thames and Williams 2013) rather than studying eligibles. This research indicates that quotas are indeed effective at increasing the numbers of women who appear on the ballot, and, as a result, some scholars have questioned “supply side” explanations of women’s

underrepresentation, arguing that a lack of aspirants cannot explain low numbers of female MPs in top legislatures (Krook 2010; Murray 2010b; Cases-Arce and Saiz 2011). However, while this research reveals that enough female candidates can be located to fill top ballot spots, it cannot tell us if such politically ambitious women are the norm or the exception, making it impossible to assess whether quotas have a symbolic effect on early stages of the political recruitment process by increasing women’s political ambitions and rendering them equally as likely as their male counterparts to aspire to elective office. Understanding whether or not this is the case is vital, not just for democratic accountability but also for evaluating whether the original goals of quotas—increasing women’s participation in democracies, creating role Page 115 →models, and providing substantive representation for women—have been met. Because elected bodies contain far fewer positions than individuals eligible to hold office, it is possible that women’s numeric representation in powerful elective bodies can rise while lower percentages of female than male eligibles develop political aspirations. If this were the case, quotas’ goal of obtaining the equal participation of men and women in this second phase of the democratic process would not have been met. If female eligibles remain less likely to become aspirants than male eligibles, it may negatively impact women’s substantive representation as well, offering female citizens a narrower range of individuals from which to select a representative than male citizens enjoy. Similarly if, despite over a generation’s use of quotas, a female chancellor, and a critical mass of women in high-level legislative bodies, German women are less likely than German men to develop political aspirations, quotas’ symbolic impact can be considered limited. Without studying all phases of political recruitment, then, we cannot determine how quotas function and whether their goals have been fully met. Moreover, many oppose quotas because they fear women elected via affirmative action will be mere tokens, unable to accomplish much for women (Dahlerup 2006; see also Mansbridge 1999). If “quota women” are indeed selected among only very few aspirants, and men are competitively chosen, when women assume office they may fear a “labeling effect” and shy away from speaking up on behalf of other women (Franceschet and Piscopo 2008), reducing quotas’ ability to improve women’s substantive representation. Understanding the relationship between quota rules and political ambition will help scholars better predict whether “quota women” are likely to be trivialized and fear being stigmatized, or whether women face inner-party competition for ballot nominations in ways similar to their male counterparts, endowing female officials with both legitimacy and a “mandate” to provide substantive representation (Franceschet and Piscopo 2008). Where similar ratios of male and female aspirants to available ballot slots exist, achieving a nomination to run for office will be equally competitive for men and women and the latter can hardly be derided as tokens. Where few female aspirants are present, however, female elected officials run a higher risk of being labeled tokens or fearing being labeled as such. Studying political ambition where quotas are in effect will provide scholars with the evidence needed to determine which of these conditions is met. Research that begins by examining who ultimately is placed on the ballot when quotas are used, rather than digging deeper and comparing male and female aspirants, thus provides only a limited view of quotas’ impact. Page 116 →Earlier phases of political recruitment must be studied in order to evaluate women’s and men’s propensities to develop political ambitions.

Gender Quotas and Political Ambition As delineated in chapter 1, not all those eligible to run for elective office will desire to do so (see also Matland and Montgomery 2003); both men and women are deterred by the toll that running for office would take on their families, friendships, careers, and hobbies. Instead, only a subset of party members hopes to become politicians one day. Many studies of political ambition tend to stress the impact of political opportunity structures on political aspirations, while other work, especially on gender and political ambition, has argued that political aspiration is also a function of other resources, such as time and self-confidence. Each of these schools of thought helps us to understand the interplay between quotas and ambition. While quotas can easily improve some political opportunity structures, the highly gendered recruitment environment dampens women’s political ambitions,

even where quotas create a favorable political opportunity structure. Below I discuss each of these influences on ambition in turn. Political Ambition and Political Opportunity Structures Most academic literature on political ambition has focused on the United States (Schlesinger 1966; Fowler and McClure 1989; Maisel and Stone 1997; Stone and Maisel 2003) and stressed the importance of political openings on individuals’ aspirations—unsurprising given the importance of individual initiative in the entrepreneurial form of candidate selection used in the United States (Siavelis and Morgenstern 2008). Schlesinger argued that “ambition for office, like other ambitions, develops with a specific situation, that it is a response to the possibilities which lie before a person” (1966, 8). Subsequent literature has found that political opportunity structures, such as whether a person’s party is consistent with the partisan composition of the electoral district, the presence or absence of a popular incumbent, or the legislative salary a politician would earn, strongly condition political ambition (Maisel and Stone 1997; Fulton et al. 2006). As detailed in chapter 1, in Germany such opportunities are expected to include how popular an eligible’s party is in his or her electoral district, the number of seats on the city council per city residents, and whether the individual hails from a large party capable of winning a plurality seat or is a member of one of the Page 117 →smaller parties that rarely win directly elected mandates. The aforementioned opportunities (or lack thereof) are unrelated to an eligible’s gender and shape both men’s and women’s chances of being elected. Students of comparative politics rarely study individual-level variables such as political ambition but do have a thorough understanding of how country-level variables create political opportunity structures that help or hinder women’s presence in national legislatures. This research indicates that, viewed comparatively, the German political system ranks as conducive to women’s political opportunities. The PR aspects of its electoral system have been found to have a positive effect on women’s representation (Kittilson 2006). Moreover, since the adoption of quotas women’s performance in the plurality component has markedly improved (FortinRittberger and Eder 2013) and female Bundestag candidates from parties using quotas are now more likely than male candidates to be elected (Manow and Flemming 2012; Hennl and Kaiser 2008). Germany’s comparatively high level of economic development and secularization create voters who have highly gender egalitarian attitudes (Inglehart and Norris 2003). Further, multiparty systems—especially ones containing Green or New Left parties like Germany’s—have been associated with comparatively high percentages of women in elective office (Matland and Studlar 1996; Kaiser 2001; Kittilson 2006). Finally, in contrast to other settings where party leaders have often been reported to be biased against or to actively discourage potential women candidates (Fox 2000, 245–46; Niven 1998, 2006; Sanbonmatsu 2006), or where parties approach fewer women than men to run (Fox and Lawless 2010; Lawless 2012), German gatekeepers across parties are more likely to encourage female than male party members to run for elective office.3 Thus in terms of formal political opportunity structures there is little reason to believe that German women should be any less politically ambitious than men in the Federal Republic. In fact, in addition to these favorable political opportunity structures for women, quotas create further advantages for female eligibles. First, German quotas are voluntary promises made by parties—rather than legally mandated regulations—and by adopting quotas parties have sent a strong symbolic signal to women that their candidacies are desired. Second, quotas have created a critical mass of elected female officials who can serve as role models. Third, and likely most important, quotas exert an “elevator effect” that continues beyond promotion to positions of inner-party leadership. Because women make up lower percentages of the rank and file than quotas require for inner-party leadership bodies and electoral lists, female party members have better numerical odds than men of receivingPage 118 → their party’s nomination for a ballot slot. As chapter 2 demonstrated, women in parties with quotas enjoy better political opportunity structures than do men. Research conducted in other cases implies that these good odds are likely to exert a strong positive impact on women’s political ambition, as women on average have been found to be particularly sensitive to political opportunity structures (Fulton et al. 2006). If political ambition were indeed solely a function of the opportunities open to an individual, therefore, we would

expect to find that female eligibles in parties employing quotas were more ambitious than men in their own party. Because sex is not a consideration when selecting candidates in parties without quotas, women in such parties enjoy no such numerical advantages and men no such numerical disadvantages; hence there would be no reason to expect sex differences in political ambition in parties where quotas are not employed. Political Ambition and Personal Circumstances However, political opportunity structures are not the only determinants of whether an eligible will become an aspirant. In order to do so, a person also needs to have the ability to take advantage of the favorable opportunities present. Just as in the previous phase of political recruitment, even when quotas are in place the gendered recruitment environment and informal recruitment structures in Germany are expected to render women less able to act on auspicious political opportunities than their male counterparts. As a result of personal circumstances, the average female eligible is expected to be less politically ambitious than the average male eligible. To become aspirants, eligibles must be endowed with what Norris refers to as “political capital” (1997, 13): they need to be qualified for the job, have confidence in their own qualifications, hold the financial resources necessary to pursue politics as a vocation, and enjoy the time to pursue political office (Matland and Montgomery 2003; Murray 2010b; Lawless 2012). While gender quotas that extend to inner-party office can increase women’s qualifications for elective office in a party-loyalist system of candidate recruitment, quotas cannot increase women’s endowments of the other aforementioned resources. Indeed, as discussed in the previous chapter, quotas’ failure to induce more women to join political parties in Germany suggests that, on average, German women have lower levels of political capital than German men. Some personal hurdles to political ambition would seem to apply equally to men and women. The main reasons CIS respondents give for not running for office involve the personal cost of a political career in terms of Page 119 →time away from family and friends (see table 1.3). There is little theoretical reason to expect that there would be gender differences in valuing relatives or friendships. Similarly, for an eligible to become an elected official she needs to be able to afford to do so;4 a fear that running would jeopardize one’s professional career was the other main reason eligibles gave for a lack of ambition. Again, there is little reason to expect that either men or women would be less sensitive to jeopardizing their employment status.5 Self-Perceived Qualifications However, some personal circumstances are highly gendered, creating hurdles to women’s political ambition, regardless of whether a quota is in place. Objectively, all of the individuals queried by the Candidate Interest Survey are qualified to run for elective office by virtue of their inner-party leadership role, a valued commodity in Germany’s party-loyalist system of candidate recruitment.6 For an eligible to become an aspirant, however, she must also perceive herself as qualified to run for elective office. Research conducted in a range of contexts consistently reveals that women are less likely than men to view themselves as competent—even when there are no objective sex differences or when women actually score higher than men on unbiased instruments. This finding holds for self-assessments of mathematical ability (Wigfield, Eccles, and Pintrich 1996), verbal skills (Pajares 2002), intelligence (Furnham and Rawles 1995), and value as a highly skilled employee (Bowles, Babcock, and McGinn 2005). This pattern may stem from the fact that women with qualifications identical to men’s are indeed perceived by others, including those in positions of power, as being less qualified than their male peers (Moss-Racusin et al. 2012). Against this societal backdrop—well beyond the purview of quotas—female eligibles are likely to be less confident in themselves than male eligibles. In their work on the United States, Jennifer Lawless and Richard Fox establish that on average women in the American candidate pool feel less qualified to hold elective office than equally qualified men (Lawless and Fox 2012; Lawless 2012). Moreover, this lack of confidence is especially detrimental to women’s political ambition because, on average, women have been found more likely than men to believe that qualifications are an important factor in determining whether one should run for office (Fox and Lawless 2004; Fox, Lawless, and Feeley 2001).

Proponents of quotas often argue that by placing women in visible positions of power, affirmative action measures will exert a symbolic effect, building other female citizens’ confidence by providing role modelsPage 120 → who demonstrate that women are capable of ruling (Mansbridge 1999; Schwindt-Bayer and Mishler 2005; Franceschet, Krook, and Piscopo 2012). However, even when women fill top elective positions, these offices remain highly masculinized (Jalalzai 2013; Runyan and Peterson 2014). Character traits commonly associated with politicians—such as leadership ability and decisiveness—are also attributes frequently connected to men rather than to women. Some have credited Angela Merkel’s political success in part to her ability to portray herself as an “honorary man” (Mushaben 2006) who is “professionally masculine” (Kutch 2012). Regardless of quotas, then, female eligibles’ mental images of what constitutes a qualified politician may be so gendered that they fail to see themselves fitting the bill (Githens 2003). Moreover, contrary to the expectation that highly visible women in politics will have a positive role model effect, Jennifer Lawless found Hilary Clinton’s experiences with a sexist media while running for president to have exerted a negative role model effect, scaring off female eligibles in the United States (Lawless 2012). After witnessing Angela Merkel’s treatment in the press—including criticisms that her hair and makeup were too dowdy, caricatures of her cleavage when she did wear a low-cut gown to an opera, and sexualized references to her relationship with French president Nicolas Sarkozy—many German women may have become disenchanted with running for office as a result of, not despite, their female chancellor (Mushaben 2006; Yoder 2011). In addition, “objective” qualifications not overtly related to sex may also be more frequently or easily held by men than by women due to the gendered recruitment environment, further masculinizing what constitutes suitability for elective office (Iversen and Rosenbluth 2010, chap. 6; Murray 2010b, chap. 3). While quotas require women to be selected for inner-party offices and certain ballot slots, they do not change what is considered a qualification for elective office. The latter informal recruitment structures are beyond the scope of quotas. For example, in Germany, following party loyalty, gatekeepers most prize an eligible’s public speaking ability. An extensive literature on gender and political deliberation finds that in groups dominated by men, such as local party organizations in Germany, women tend to speak less frequently than men and are viewed as having less authority than men when they do contribute to discussions (Karpowitz and Mendelberg 2014). As a result of these social dynamics, men may be more likely than women to believe they possess the rhetorical skills needed to be a politician. This problem is likely compounded by the content of the speeches that Page 121 →successful aspirants must deliver. In order to secure a ballot nomination, an aspirant must make a verbal presentation in front of other party members arguing why she should be selected over others for a particular ballot slot. This type of self-promotion is at odds with modesty, which is considered a socially appropriate “feminine” trait; those who pursue gender untypical behavior are in turn often met with a backlash in social interactions. Women and girls have been found to internalize such norms and adjust their behavior accordingly, often having difficulty speaking up on their own behalf (for a review of this literature, see Amanatullah and Morris 2010). Men, in contrast, are socially rewarded for self-promoting behavior and thus are more likely to engage in it. These aspects of the recruitment environment lie far beyond the scope of quotas and are hence unlikely to be altered by it. Thus, even women with high levels of education and professional experience may not feel qualified, or comfortable, engaging in the type of public speaking required of a successful candidate for elective office. Other traits considered qualifications for elective office may also be similarly gendered. German gatekeepers also seek high public visibility, involvement in voluntary organizations, and substantive knowledge in key policy areas. As a result of gendered socialization, over the course of a lifetime men and women collect different education, work, and volunteer experiences (Dow 2009), meaning that while they may indeed be well known, active members of their communities and professions, men and women are likely to have achieved this prominence through different channels. For example, men are more likely to have studied economics, worked in business ventures, and volunteered with the chamber of commerce, while women are more likely to have studied nursing, worked in care-related professions, and volunteered with health-related charities. Both sets of experiences would be highly helpful when serving on a county council charged with overseeing a publicly run hospital. However, given the social prestige of the former “masculine” activities, and the lower status of the latter

“feminine” ones, on average, it is more likely that a male eligible with the above rГ©sumГ© would perceive his experiences as qualifying him for elective office than would a woman with the above CV. Thus, while quotas do have a positive impact on helping eligibles demonstrate party loyalty, as they are used to select inner-party leaders and candidates for low-level offices, they do not (directly) change a highly gendered recruiting environment that values different types of behaviors and experiences, deeming some more appropriate for women than men. Thus, while quotas do elevate women to positions of inner-party leadership and reserve ballot spots for them, they do not change the masculinized nature Page 122 →of “qualifications” for elective office, and, on average, women are expected to be less confident about their abilities to run for elective office than men. Domestic Responsibilities Even if an eligible individual wants, and feels qualified, to run for elective office, however, she must also have the time to do so. Responsibility for childrearing and household tasks clash with the long and irregular hours associated with a political career. Because the vast majority of candidate pool members also pursue professional careers in addition to politics, agreeing to hold elective elected office at the local level—a volunteer post—would add a “triple burden” to anyone, male or female, responsible for household and care work.7 Since women perform the bulk of unpaid household labor in Germany (Iversen and Rosenbluth 2010, 73), women are on average more likely to find themselves in these ambition-dampening personal circumstances than men. Studies in contexts where quotas are not employed have found women with children at home to have lower levels of political ambition than other groups (Flammang 1997, 162–67; Fox and Lawless 2003; Fulton et al. 2006; but see also Lawless 2012). One study of the United States established that women were more likely than men to leave political office if it clashed with their family responsibilities (Thomas, Herrick, and Braunstein 2002, 414). It is unlikely that even quotas’ favorable political opportunity structures will be able to undo this state of affairs. In sum, while quotas can relatively easily change political opportunity structures, they are less able to alter the recruitment environment in a given society or the informal norms as to what constitutes a qualified candidate. As a result, I expect that on average even eligible women will be less politically ambitious than their male counterparts because, given Germany’s gendered recruitment environment and its informal recruitment structures, they possess less time and confidence to run for office. Quotas, Ambition, and the Eligibility Pool This underlying gender difference in political ambition is expected to manifest itself differently across parties, however, depending upon whether or not sex is a consideration in selecting party leaders. While not all inner-party leaders in Germany will harbor political ambitions, all those party members who do will attempt to become local party officers because these positions are prerequisites for elective office. In most settings, however, Page 123 →there will be far more local-level board (Vorstand) seats to fill than there are ambitious party members to assume them. These remaining posts will be staffed by active party members who do not necessarily desire to become elective officials. Where sex is not a consideration in choosing inner-party leaders, both politically ambitious men and women are expected to rise to top board positions. Because women make up a minority of party members and because women are expected to be, on average, less ambitious than men, however, there will be far fewer such women than men. Less-ambitious party members will then receive the remaining Vorstand places. Given the constraints both on female party members’ time and self-confidence, the tendency to undervalue women’s qualifications, and the overall underrepresentation of women in political parties—especially those that do not employ quotas—it is unlikely many of these remaining seats will be occupied by unambitious women. Instead, men are more likely to be tapped for the vacancies. The resulting board is thus expected to be comprised of a few ambitious women and a larger pool of men with mixed ambitions.

When quotas are implemented, in contrast, the outcome is expected to differ. Because quotas require a lower percentage of men in party leadership bodies than among the rank-and-file membership, opportunities for men will be limited. Ambitious male party members will be sure to endeavor to obtain leadership posts; where local party groups are large, there may be more male aspirants than there are board positions. The chances of unambitious men being tapped to fill a vacancy are reduced in comparison to parties without quotas, because there are simply fewer such spots available. In contrast, quotas require far higher percentages of women in board positions than there are among rank-and-file party members; to the degree that women are less ambitious than men, an even more favorable ratio of leadership spots to ambitious party members exists when quotas are in place. Politically ambitious women will certainly strive for these leadership positions, but after they assume top spots, vacancies will remain. Quotas require parties to tap less ambitious women for the remaining Vorstand places—who would otherwise easily be overlooked in organizations comprised mainly of more ambitious men. Thus parties employing quotas likely will not only have more female leaders than parties without quotas, they will also enlist women with a broader range of political ambitions. As a result, where quotas are employed, boards are expected to be filled by a relatively ambitious group of men and a broader pool of women with mixed levels of political ambition. In sum, I expect that underlying sex differences in political ambition will be magnified among the leaders of parties with quotas and muted in Page 124 →the candidate pool of parties without. The following sections test these expectations empirically, using several measures of political ambition.

Empirical Evidence: Gender, Quotas, and Political Ambition There are three main ways in which political ambition has been measured. First, researchers have simply asked respondents whether they would ever like to run for elective office in the abstract (Constantini 1990; Lawless and Fox 2005, 2010; Maestas et al. 2006; Lawless 2012). Second, scholars have investigated individuals’ actual behavior, studying whether respondents ever in practice considered running for political office (Lawless and Fox 2005, 2010; Lawless 2012). Third, political scientists have measured respondents’ attitudes toward running for office in a specific, but future, scenario (Maisel and Stone 1997; Fulton et al. 2006; Maestas et al. 2006). The Candidate Interest Survey utilized each of these measures and below I report the results of each in turn.8 Political Ambition in Theory The CIS asked eligibles whether they agreed with the statement “I have always wanted to be a politician.” Given that the individuals surveyed are all members of the German candidate pool, a surprisingly low overall percentage agreed with this statement; a mere 6 percent confessed to always having desired political office. This paucity of political ambition is likely due at least in part to social desirability. As discussed in chapter 1, politicians are generally held in low regard in Germany and people may be loath to admit to admit to a history of political aspiration. As one state party leader put it, “there’s a widespread contempt of politics” in Germany and a “negative view” of politicians; “people aren’t standing in line to get into the state legislature. Its reputation is not so good and the pay is not so good. This is not a very good job description—who would want to take it?” In keeping with the above hypotheses, significant gender differences in ambition emerged in parties with quotas, where 11 percent of the men sampled and only 2 percent of the women maintained that they had always aspired to elective office (see figure 3.1). Even though quotas not only create excellent political opportunities for women in these parties but also bring about many female political role models, a gender gap in abstract political ambition remains among German eligibles in parties employing quotas. In contrast, no significant gender difference was found among Page 125 →eligibles in parties without gender quotas; there approximately 4 percent of the local party leaders surveyed had always wanted to be politicians, regardless of sex. It is important to keep in mind, however, that there are very few women leaders in these parties. These results indicate that unambitious men less often achieve inner-party leadership positions in parties employing quotas, while in the absence of quotas unambitious women are less frequently selected party leaders. The men from parties with quotas were significantly more ambitious than their male counterparts in parties not employing affirmative action, whereas women leading parties without quotas were on average more ambitious than their counterparts in parties with

quotas.9 Fig. 3.1. I Have Always Wanted to Be a Politician. (Source: Candidate Interest Survey [N = 433]. ***p ≤ 0.01.) Table 3.1 examines the effects of respondents’ sex and their parties’ rules for selecting inner-party leaders on their likelihood of always wanting to be a politician. As expected, women in parties with quotas are significantly less likely, 7.7 percent on average, than other respondents to claim they have always wanted to be a politician. In and of itself, sex had no independent effect on this kind of abstract ambition. Eligibles from parties employing quotas are approximately 4 percent more likely than eligibles in parties without quotas to harbor political aspirations. Admittedly, this model explains extremely little of the variance present. However, “nascent political ambition”—the desire to become a politician, rather than, for Page 126 →example, a doctor, actress, or athlete—differs considerably from a strategic decision to enter a particular race and is less easily explainable by political variables (Lawless 2012). Political Ambition in Practice Of greater interest here is not what respondents have “always” wanted to do with their lives, but what they actually do once they gain an inner-party post and become eligible to run for office. Do the less ambitious women tapped by quotas to serve in inner-party offices develop political ambitions as a result of their quota-driven political opportunities? In addition to asking what a respondent had always wanted to do, therefore, the CIS also inquired about whether a respondent had ever actually considered running for certain elective offices, both at the local and high levels. Figure 3.2 depicts the percentages of women and men in parties with and without quotas who have ever considered running for office either at the city/town, county, or district level. The results stand in sharp contrast to the question about always having wanted to be a politician. Although the vast majority of eligibles maintain they had not always wanted to be an elected official, over 90 percent of them had at least considered running for a local-level office. There was very little variance in this regard; both women and men in parties with and without quotas responded affirmatively.10 Serving as a local party leader in Germany places an individual in a position where he or she is prone to considering a run for elective office, even if he or she may not have always desired a political career. Thus one positive effect of Page 127 →quotas is to elevate women who do not initially harbor political aspirations into situations that nurture ambition; where quotas are not employed such women are less likely to gain such experiences. Table 3.1. Determinants of Always Wanting to Be a Politician: Logistic Regression Results Change in Probabilitya Variable Coefficient (Standard Error) (Change from 0 to 1) Female Sex 0.10 (1.02) not significant Respondent’s Party Employs Quota 1.18 (0.76)* Female Respondent in a Party with a Quota в€’1.99 (1.20)* Constant в€’3.26 (.72)*** N 433 0.071 Pseudo R2 0.003 Prob > chi2

+4.1% в€’7.7%

Source: Candidate Interest Survey. aBased

on Stata command “prchange.”

* p ≤ 0.1, *** p < 0.01. Fig. 3.2. Eligibles Ever Having Considered Running for Local-Level Elective Office. (Source: Candidate Interest Survey [N = 430]; differences of means not significant.)

Personal interviews demonstrate the mechanisms through which political ambitions develop over time as a result of holding inner-party office. A woman who was trained as a doctor but ultimately became a state cabinet minister described this process. She claimed, “I never wanted to go into politics,” but once she began to get involved in her party she discovered “it was fun” so she decided to run for the her city council, then worked her way up to head of the council’s Christian Democratic Party group, and served as an adviser to her state legislator. She recalled, “politics is like an addiction that eats you up. [Political] power is a drug and once I had access to some, it acted upon me quickly and I was addicted.” When the state legislator with whom she worked retired, she viewed herself as “his logical successor and pursued this goal.” When asked why he decided to run for the state legislature, one man responded, “I fell into this position, or, you could also say I climbed here. At first I just fell into positions within the party as people asked me to do various things. After a while, one begins to want some things in exchange for all his hard work, some influence.” He added that “that’s when I began climbing” the ranks and working toward a professional career in politics. One state legislator argued, “I didn’t catapult into this position. I built up to it and grew into it organically.” By holding various inner-party posts, he explained, “I learned different aspects Page 128 →of politics and gained a foundation from which to speak.” Eventually he observed what he perceived as a poor job being done by other members of his own party in the state legislature and said to himself, “They are doing such a bad job, and you’ve been around for so long, you can do it better.” Individuals active in local politics discussed how their interest in running for higher-level elective offices grew the more experience they gained with politics. One Green member of a county council, employed in the public service, recalled that upon her election to her town council she was “astounded and shocked” at how little the other amateur members of the council knew about the administrative issues about which they were required to render decisions. She maintained that “before [being elected] I had no idea that local politicians were so weak[ly]” prepared for their jobs; in her opinion, they had “no clue.” These firsthand encounters with other town council members gave her the confidence to run for the county council. The Candidate Interest Survey, therefore, asked eligibles not just whether they had considered a run for local-level elective offices but also for more powerful positions. Figure 3.3 portrays the percentages of eligibles who had in practice considered running for either the Landtag, Bundestag, or European Parliament. Here both for men and women, and in parties with and without quotas, the percentages are considerably lower than for local-level offices. Fifty-seven percent of all respondents reported having mulled over a top elective office. That fewer respondents had considered becoming a candidate for such posts than for local-level positions reflects the considerably different nature of elective office at this level. Serving in the state legislature, Bundestag, or European Parliament is a full-time job that in most cases requires travel away from home and stepping away from one’s chosen career for an indefinite period of time. Moreover, such positions are highly sought after and the odds of receiving one low. These personal circumstances and political opportunities clearly shape individuals’ aspirations (or lack thereof) for a career at this level. One interviewee described the decision to leave his low-level amateur council seat and day job to become a professional state-level parliamentarian as “crossing the Rubicon.” As expected above, in parties employing quotas for inner-party office, significantly fewer women than men had considered running for a professional political position. While 59 percent of men claimed to have thought of doing so, only 50 percent of women had done so. Although gender differences remain, half of the female inner-party leaders surveyed had indeed considered competing for a Landtag, Bundestag, or European Parliament mandate. Quotas succeed in placing many women in inner-party posts where one out of every two at least considers becoming a professionalPage 129 → politician. In parties without quotas the difference between men’s and women’s means was not statistically significant at conventional levels, but a higher percentage of women (64 percent) than men (60 percent) had thought about throwing their hats into the ring for one of Germany’s highest elected positions. These female party leaders are only a small minority of the eligibles in parties without quotas, however. Fig. 3.3. Eligibles Ever Having Considered Running for High-Level Elective Office. (Source: Candidate Interest Survey [N = 340]; *p ≤ 0.10.)

Willingness to Accept a Hypothetical Candidacy Thinking about running for elective office in and of itself does not render someone an aspirant, however, as an aspirant is someone who has thought it over and then decided that he or she does want to run for office. Thus, Candidate Interest Survey recipients were also asked whether or not they would be willing to accept a hypothetical nomination from their party to run for elective office in their community or for the state, federal, or European parliaments. Local-level offices included either a seat on the city or town council, the county council, the regional council, or a directly elected mayoral position at one of these levels. Here the percentage of positive responses falls approximately 20 percent when compared to just thinking about running; on average over 70 percent of those surveyed indicated they would be willing to accept a nomination to run for a local-level elective position. As expected, and as indicated in figure 3.4, significant gender differences emerge in parties selecting their officers using gender quotas. Women were less ambitious than their male counterparts; 65 percent of Page 130 →women eligibles in parties with quotas claimed they would accept such a nomination compared to 73 percent of the men. Where sex was not a consideration in selecting eligibles, equal percentages of men and women expressed a willingness to run for local-level office. Fig. 3.4. Eligibles Willing to Accept a Hypothetical Nomination for Local-Level Elective Office. (Source: Candidate Interest Survey [N = 441]; ** p ≤ 0.05.) Similar results occur when asking respondents about their willingness to accept a nomination for a high-level elective office including the state, federal, or European parliament (see figure 3.5). On average, far fewer eligibles indicated interest in these full-time political posts than in amateur local ones; only 36 percent of the candidate pool members surveyed said they would accept a hypothetical nomination compared to the approximately 60 percent who had thought about it. Here again, as expected, significant gender differences emerge in parties using quotas to select their leaders. While 38 percent of men in parties employing quotas asserted they would agree to appear on the ballot for the Landtag, Bundestag, or European Parliament, this figure fell to 30 percent for their female counterparts. The difference between men and women’s means in parties without quotas shrunk to only 2 percent with 44 percent of male and 42 percent of female party leaders maintaining they would agree to be nominated for these high-level elective offices. Thus, although quotas create many women leaders who consider running for local-level office as often as their male counterparts do, gender gaps in considering running for high-level office and willingness to accept hypothetical nominations, both at the local and at the high levels, emerge—despite the favorable political opportunity structures created by Page 131 →quotas. In other words, over a generation after parties adopted quotas in Germany, lower percentages of female eligibles become aspirants than do male eligibles. Fig. 3.5. Eligibles Willing to Accept a Hypothetical Nomination for High-Level Elective Office. (Source: Candidate Interest Survey [N = 440]; * p ≤ 0.10.)

Explaining Gender Variance in Political Ambition What prompts these gender differences in parties with quotas? As hypothesized above, both situational factors (political opportunity structures) and personal circumstances (such as domestic responsibilities and selfperceptions) are likely to drive this variance. The following section explores why women in parties with quotas are on average less willing than men in their parties to consider running for high-level office or become a candidate either for local- or high-level elective office.11 Tables 3.2 and 3.3 present the results of multivariate analyses investigating the causes of the above sex differences in eligibles’ hypothetical willingness to consider running for a high-level elective office (Model 1), as well as to become a candidate for local-level (Model 2) or high-level (Model 3) elective offices. These logistic regression models analyze respondents from parties employing quotas and include a number of independent variables,Page 132 → both for an individual’s personal circumstances and for the political opportunity structures she faces, as well as controls for sex, age, party size, and state of residence. A respondent’s self-perceived qualifications

for elective office are measured along a four-point scale of not at all qualified, partially qualified, qualified, and very qualified. An individual’s domestic responsibilities are measured utilizing two separate variables. One, an indicator variable, registers whether or not the party leader is primarily responsible for childcare. The second is an additive index of the number of household tasks for which the eligible is primarily responsible: cooking, cleaning, laundry, and shopping. The model also includes a categorical variable that checks whether individuals believe running for elective office would harm their careers. The final personal circumstance tested is the degree of importance respondents place on spending time with their families along a four-point scale: not important, rather important, important, or very important. The models also include a categorical measure of party strength—whether or not a respondent reports that her party holds a directly elected seat in her electoral district—and a measure of the district’s population along a seven-point scale.12 As argued above, women in parties employing quotas are expected to be less willing than their male counterparts to consider or accept a hypothetical nomination for elective office. Respondents who doubt their qualifications for elective office, or who believe that running would harm their career, are predicted to be unlikely to consider running for office or to accept a hypothetical nomination. Similarly, those who value time with their family or are primarily responsible for childcare or household labor are expected to be less willing to run for office than those unburdened by domestic responsibilities. Favorable political opportunity structures are hypothesized to stoke political ambitions. Therefore respondents in thinly populated areas are expected to be more ambitious than those in densely populated cities because they face less competition for elective office; similarly, those whose parties are locally popular should also perceive good opportunities and be more willing to accept a nomination than those from organizations facing poor electoral chances. In contrast, members of small parties including the Greens and FDP are expected to be less ambitious than those in larger parties because directly elected positions, such as mayors or members of state legislatures and the Bundestag elected through the first vote, are rarely if ever won by small parties. State-level controls are included as well to tap into the varying political opportunity structures across the LГ¤nder. Finally, older citizens are expected to be less likely than younger citizens to want to serve in elective office in the future (Fulton et Page 133 →al. 2006); conversely, given that they have had more time to think about it, older people are expected to be more likely to have ever considered running for office than younger eligibles. Table 3.2 depicts the determinants of whether a respondent in a party with quotas had ever considered running for a high-level elective office (Model 1). Here, sex does not play an independent role. Instead the biggest determinants of thinking over becoming a candidate at this level were self-perceived qualifications, the impact of running on one’s career, and domestic responsibilities. While both men and women’s positive assessment of their ability to run for office increased their likelihood of considering doing so, the impact for women (+16.1 percent for every one unit change in self-perceived qualifications) was greater than that for men (+13.7 percent). See table 3.3. Women who felt their career would be harmed by holding office at this level were 8.8 percent less likely than other women to have considered seeking a nomination for the Landtag, Bundestag, or European Parliament, holding all other independent variables at their means; under the same circumstances, the chances for men decreased by 7.4 percent. Household tasks also deterred individuals from considering a run for top-level elective office, reducing women’s chances by 3.7 percent per additional chore (cooking, cleaning, laundry, or shopping) and men’s by 3.2 percent. Being primarily responsible for childcare, valuing time with family, and unfavorable political opportunity structures did not deter eligibles from at least considering a run for high-level elective office. The control variables were not statistically significant here; the only exception was residence in the state of Nordrhein-Westfalen where a comparatively large number of party members contend for a relatively low number of seats. Model 2 in table 3.2 depicts the determinants of an eligible in a party employing a quota’s willingness to accept a hypothetical nomination for local-level political office. Here, as expected, the female eligibles in parties with quotas are significantly less likely than their male counterparts to be willing to appear on the ballot in their town, county, or district—even controlling for a range of personal and structural characteristics known to shape political ambition. Holding all other variables at their means, women in parties with quotas are 11.2 percent less likely than their male counterparts to express a willingness to run for office at this level (see table 3.3). Moreover,

responsibility for household tasks exerts a separate, negative impact on political ambition for local-level office, unsurprising as this represents a “third shift” in addition to professional and political work. For each additional household task for which a female eligible is primarily responsible—cooking, cleaning, laundry, or shopping—ambition is Page 136 →reduced by 5.3 percent. Domestic work dampens men’s ambition as well, diminishing the willingness to accept a nomination by 4.0 percent per task; at this level, though, primary responsibility for childcare did not predict political ambition. Also as hypothesized, as respondents became older, they became less willing to express willingness to run for a future elective office. Members of smaller parties, less likely to win directly elected seats, were also less willing to become candidates. Page 134 → Table 3.2. Determinants of Political Ambition in Parties with Quotas: Logistic Regression Results (1) (2) Have Would actually accept a considered (3) hypothetical running Would accept a hypothetical nomination for high-level elective nomination for highofficec for locallevel level elective elective officeb officea Coefficient Coefficient Variable (Standard (Standard Coefficient (Standard Error) Error) Error) Personal Factors Sex 0.24 (0.41) в€’0.57 в€’0.10 (0.35) (0.34)* Feel Qualified 1.14 0.10 (0.22) 1.07 (0.25)*** (0.30)*** Responsible в€’0.27 в€’0.24 в€’0.06 (0.15) for Household (0.19)* (0.15)* Tasks Responsible в€’0.59 в€’0.22 в€’1.11 (0.68)* for Childcare (0.89) (0.58) Running в€’0.70 в€’0.16 в€’0.90 (0.35)*** Would Hurt (0.44)* (0.33) Career Value Time with Family

в€’0.33 (0.30)

Party Strength в€’0.59 (0.46) Population в€’0.10 (0.18) Age Small Party

в€’0.25 (0.25) 0.84 (0.39)** в€’0.23 (0.10)**

0.01 (0.02) в€’0.03 (0.01)*** в€’0.11 в€’0.43 (0.43) (0.33)*

в€’0.54 (0.25)** Political Opportunity Structures 1.28 (0.36)*** 0.06 (0.10) Controls в€’0.07 (0.01)*** в€’0.29 (0.35)

State: Baden- в€’0.58 0.67 (0.57) WГјrttemberg (0.63) State: Bremen 0.97 (1.00) в€’0.52 (0.66) State: 0.71 (1.01) в€’0.12 Hamburg (0.66)

в€’0.50 (0.51)

State: NordrheinWestfalen

в€’2.88 (1.11)***

в€’0.39 (0.50)

Constant

1.69 (1.56) 3.88 (1.30)***

1.96 (1.55)*

N

234 0.215 0.0000

298 0.23 0.0000

Pseudo R2 Prob > chi2

0.07 (0.48)

299 0.15 0.0000

в€’0.09 (0.68) 0.38 (0.67)

Source: Candidate Interest Survey. a

Categorical dependent variable: 1 = respondent had thought about running for at least one of the local-level elective offices described in chapter 1; 0 = not willing or uncertain. b

Categorical dependent variable: 1 = respondent would be definitely willing to accept his or her party’s hypothetical nomination for at least one of the local-level elective offices described in chapter 1; 0 = not willing or uncertain. c

Categorical dependent variable: 1 = respondent would be definitely willing to accept his or her party’s hypothetical nomination for at least one of the state, federal, or European parliaments; 0 = not willing or uncertain. * p ≤ 0.1; ** p < 0.05;*** p < 0.01. Page 135 → Table 3.3. The Relative Impact of Independent Variables on Political Ambition in Parties with Quotas (1) (2) (3) Considered Running for Would Accept Hypothetical Would Accept Hypothetical High-Level Office Local-Level Nomination High-Level Nomination Women Men Women Men Women Men Sex not significant not в€’11.2% not significant not (Change significant significant Male в€’ Female) Feel +16.1% +13.7% not significant not +21.6% +22.4% Qualified significant (1 unit change, 4point scale) Responsible в€’3.7% в€’3.2% в€’5.3% в€’4.0% not significant not for significant Household Tasks (1 unit change, 4point scale)

Responsible not significant for Childcare (0 to 1 change) Running в€’8.8% would hurt Career (0 to 1 change)

not not significant significant

not в€’17.6% significant

в€’18.7%

в€’7.4% not significant

not в€’16.4% significant

в€’17.2%

Value Time with Family (1 unit change, 4point scale) Party Strength (0 to 1 change) Population of Community (1 unit change, 7point scale) Age (1-year change) Small Party (0 to 1 change)

not significant

not not significant significant

not в€’11.0% significant

в€’11.4%

not significant

not +17.0% significant

+12.3%

+29.6%

Not significant

not в€’5.1% significant

в€’3.9% not significant

not significant

not significant

not в€’0.7% significant

в€’0.6% в€’1.4%

в€’2.2%

not significant

not в€’10.0% significant

в€’7.6% not significant

not significant

+ 28.9%

Source: Candidate Interest Survey. Note: Results based on Models 1–3 in table 3.2, using Stata command “prchange,” varying sex and holding all other variables at their means. Model 2 in table 3.2 also finds that, as expected, political opportunity structures condition political ambition, with women more sensitive to these considerations than men. Holding all other variables at their means, female respondents whose parties are strong in their electoral district, winning a key directly elected seat, were 17 percent more likely to claim they would accept a local-level nomination than if their party were weak; being a member of a popular party increased men’s aspirations by 12.5 percent (see table 3.3). Similarly, as the population of their community rises, increasing the competition for ballot slots, women become even less likely than men to be willing to accept a nomination. For women, a one unit change in community population decreases willingness to be nominated by 5.1 percent and for men the figure is only 3.9 percent. Similarly, while men from small parties were 7.6 percent less likely than men from other, more competitive parties to be willing to run for office, the figure for women rose to 10 percent. As in other cases cross-nationally, German women appear to be particularly sensitive to political opportunity structures. Other independent variables did not achieve conventional levels of statistical significance. Model 3 in table 3.2 depicts the logistic regression results for whether or not a respondent in a party with a quota would agree to be a candidate for the state legislature, the Bundestag, or the European Parliament. Contrary to expectations, sex did not exert an independent effect on this type of political ambition. A range of other personal

circumstances did exercise considerable impact on an individual’s political aspirations, however. Of these the most influential was an eligible’s sense of his or her own qualifications for elective office (see table 3.3). For every one unit increase in self-perceived qualification on a four-point scale, women’s willingness to run for elective office increased by 21.6 percent and men’s willingness by 22.4 percent. Also, as expected, primary responsibility for childcare decreased a willingness to accept a hypothetical Landtag, Bundestag, or European Parliament nomination by 17.6 percent for women and 18.7 percent for men; the number of household tasks for which a respondent was responsible did not impact ambition at this stage, however. Those respondents who placed greater value on spending time with their families were less willingPage 137 → to express an interest in a highlevel nomination than others; for every one unit increase in valuing family time along a four-point scale, women’s and men’s political ambitions were reduced by approximately 11 percent. Finally, as expected, the ability to balance political and professional life exerted a significant impact on eligibles’ willingness to pursue a full-time political career. Men who believed that running for office would harm their careers were 17.2 percent less likely to want to accept a nomination to run for high-level office than men who did not perceive this tension; for women the figure was 16.4 percent. Also as expected older respondents were less likely to express a willingness to embark on a future high-level political career. Notably, women’s ambitions dimmed less with age (в€’1.4 percent per year) than men’s (в€’2.2 percent per year) perhaps because they face fewer challenges from younger competitors than do men (see figure 2.5). Political opportunity structures strongly shaped political ambition for high-level office as well; women’s and men’s willingness to campaign for a high-level office was increased by 28.9 percent and 29.6 percent, respectively, in electoral districts where the respondent’s party had a successful track record. Population at this level was not significant, likely because the odds of winning a top post are low, regardless of community size, and because state and federal legislative districts are more proportionally drawn than local ones. Controls for state of residence and party size also did not prove significant.

Mechanisms Reducing Women’s Political Ambition Thus, once a number of other independent variables have been controlled for, gender is not an independent predictor of whether an eligible in a party with a quota had considered running for an elective office or expressed a willingness to accept a hypothetical nomination to run for state, national, or European elective offices. This finding represents a significant achievement of quotas. Although women are less likely than men to join a party or to have always wanted to be a politician, quotas lift the women who do join parties into positions of eligibility where at least half of them have considered running for elective offices and, ceteris paribus, are no less likely than men to want to accept a ballot nomination for a powerful parliament. The same was true in parties without quotas, where on average men and women did not differ in their mean propensity to express political aspirations and sex did not exert an independent effect on eligibles’ propensity to consider running for office or to accept a nomination at any level. Page 138 →However, in contrast to parties in which quotas are used, female eligibles in parties without quotas are few and far between and, on average, more ambitious to begin with. Despite quotas’ achievements, however, gendered hurdles to political ambition remain as some of the independent variables that predict political ambition are themselves highly gendered. While political opportunity structures such as party strength and district population apply equally to men and women in a given party organization, other key determinants of ambition such as self-confidence, domestic responsibilities, and age are unequally distributed among the eligibles. Moreover, in parties with quotas the female sex retains an independent, dampening effect on the willingness to accept a nomination for local-level elective office—the first rung on the German career ladder. The following section depicts the results of personal interviews to demonstrate the mechanisms through which the above factors shape political ambition. Nongendered Influences on Political Ambition Both male and female interviewees reported that their political careers had been shaped by political opportunity

structures such as their party’s chances in their electoral district. For example, a Christian Democratic member of the Bundestag said that he had originally pondered running for the European Parliament as he had a keen interest in international affairs, but decided against it because the CDU was not strong enough to win an EP mandate in his area. A Social Democrat who decided against running for office argued, “To campaign costs a lot of timeВ .В .В . and psychological energy. Why get into all this if you have no chance of winning?” One Green state representative argued “there are a bunch of reasons why” an individual would not run for office and proceeded to enumerate six reasons; first on her list was that the Green Party candidate might not be able to win. Another Green woman agreed that no one wants the “torture” of running for “an unwinnable [directly elected] seat. But when it really counts, when it comes to [safe] places on the list, there is lot of bickering. Lots of people want this and it’s a competitive game.” A CDU member in a Left-leaning district agreed that when his precinct organization needed a candidate for their area’s directly elected Bundestag seat, “we must search for volunteers” because the mandate was sure to go to the SPD. Similarly, both male and female interviewees noted the greater competition for nominations in densely populated urban areas compared with rural ones. A leader in the CDU’s Women’s Union, responsible for recruitingPage 139 → candidates in her town of 8,000, claimed it was “tremendously difficult” to find enough candidates in her area, whereas her urban Frauen Union counterparts reported “giant fights” for promising nominations. A Green described what she viewed as a “glaring urban to rural drop-off” in the ratio of ballot spots to party members; in her town of 20,000 she maintained “it’s only a matter of time” until active party members, male or female, run for office because the candidate pool was so small. In interviews men and women frequently mentioned the ability to combine elective office with their professional aspirations as a major boost to political ambition. The Candidate Interest Survey found men and women equally likely to express concern that their careers would be hindered by a run for elective office; this finding held in parties with and without quotas. One female state legislator said she had been willing to take the risk of becoming a full-time politician because she was not worried about finding a job should she lose her seat: “I am resourceful; I will find something,” she claimed. Those employed in the public sector appreciated their legal right to take time off work to pursue political activities. One civil servant attributed her pursuit of local- and county-level posts to her employer: “What’s decisive is that I can professionally afford it. Not everyone can afford to be an amateur politician; in most professions it’s not at all possible. Some people have to work from morning вЂtil night and others can’t get any time off” to attend local functions. A local business owner explained that he enjoyed the flexibility to serve on the city council because he could close his small business during the periods he needed to attend political functions. He observed that for party members who were employees, rather than self-employed, this is more difficult to do. “The boss will say, вЂWho do you want to work for? Me or politics?’” Indeed, one of the most frequent reasons interviewees gave for not aspiring to particular posts was the incompatibility of such offices with their careers. The above-mentioned CDU member continued that he would be unwilling to run for a full-time political position such as mayor or a high-level parliament because these offices would require him to shutter his business altogether. A self-employed woman in the CSU expressed similar concerns. A Christian Democratic Party leader charged with creating a party list for elections in her community maintained that local merchants often declined to run for office when asked, fearing that they would lose customers who supported other parties as a result; Free Democratic leaders, whose party members tend to be disproportionately self-employed, often expressed a similar problem. Regardless of sex, when interviewees believed that an elective office Page 140 →would detract from their home life they were unlikely to aspire to it. Male and female CIS respondents valued time with their families equally. One woman who had been asked to run for the Bundestag and claimed to have had a good chance of being elected, declined to do so. She maintained, “my family life would have been in pieces. I would have had to write off my family,” spending the week in Berlin and only shuttling home on the weekend. A man who had an opportunity to become a state-level cabinet minister turned it down, arguing “this was a conscious decision made for the well-being of my children”; in his view a powerful political position and active fatherhood were “not compatible.”

Gendered Influences on Political Ambition: Domestic Responsibilities While male and female CIS respondents were equally likely to report valuing time with their family, female eligibles were far and away more likely to report being primarily responsible for an array of household tasks. As depicted in figure 2.4, only 14 percent of male inner-party leaders claimed they were responsible for doing laundry compared to 74 percent of female eligibles. Sixty-three percent of the women surveyed had primary responsibility for cooking meals compared to only 23 percent of men. Shopping for household needs followed a similar pattern: while 53 percent of women were mainly in charge of this task, only 27 percent of men were. When it came to housecleaning, 47 percent of female party leaders were accountable compared with only 14 percent of male eligibles. Of those who had minor children, only 3 percent of men were the main caregiver; the figure for women was 48 percent. This may explain why female CIS respondents were significantly less likely than their male counterparts to mention a loss of free time for hobbies as a deterrent to running for elective office—many women have very little free time to begin with. These different private sphere burdens clearly manifested themselves in personal interviews. Of all the men interviewed for this book, only one (a single father) mentioned a concrete example of how domestic responsibilities impacted his political work, talking about racing to drop off his daughter at school before a morning meeting, having to bring her to appointments held outside school hours, and taking time to help her with homework. In contrast, female interviewees routinely made references to the day-to-day logistics for which they were responsible. A state legislator and mother of four claimed that, while both male and female members of her parliament spent considerable time away from their families, it was Page 141 →women who were “handicapped” by domestic responsibilities. “The only problem men might have is a nagging wife who asks what time they are finally going to come home, whereas a woman has this issue and then has to go out and buy shoes for the kids.” An older female state legislator said her position would be difficult for a woman in her “family phase” to hold; it is not just that women are responsible for childcare, she argued, but for the social life of the entire family. “Who is it that remembers birthdays?” she asked rhetorically. Another woman mentioned struggling to combine grocery shopping, laundry, and her responsibilities as a state legislator; she had put off running for elective office until her children were old enough to care for themselves. These commitments in turn reduced women’s aspirations for elective office. A mother of a three year old employed as a nurse and active in the CDU concluded that “this is enough” without holding elected office as well. A working mother of three was approached about running for her local council after her husband resigned his seat because it clashed with his work (not household) responsibilities; she turned down the opportunity, citing childcare and other domestic burdens as the main reason. A female state legislator asked herself before running for office, “Do I really need this in addition to my home and family responsibilities?” Although she had a husband who could help her at home, she lamented “it’s up to us [women] to organize everything.” A party leader responsible for putting together party lists in her community maintained it was difficult to find women willing to run, claiming the first reason they give is “we can’t do this to our children.” Another party leader agreed that when she approached qualified working women with children about running for office—the kind of women who would make excellent candidates due to their high visibility in the community because of their volunteer commitments in the PTA, church, or clubs—they often turned her overtures down, claiming “it’s too much” to also run for office. She cited two very capable women in her party who had recently resigned from her town’s council for this reason. In contrast, neither the men interviewed, nor party leaders’ accounts of attempting to recruit men for elective office, mentioned double and triple burdens as a hindrance to male ambition. One state legislator concluded: We come up yet again against the issue of work/family balance. This includes not just work and family, but also free time. When gender mainstreaming accomplishes what we imagine it can—with equal rights in all spheres—it should be no problem for women to fully realize their interests in their free time. But that means that Page 142 →we have to include men. We have to bring men to the point where they also sometimes stay home and don’t go play soccer, that they don’t go indulge their own hobbies, but rather say [to their female partner], “OK you now have this hobby and you see the need to become active in this hobby, so we’ll share work in the family. Go for it!”

Because if that happens, when one starts in politics, it becomes a virtuous cycle. A woman can’t really then say, “But I can’t continue in this [political role] because I now have to stay home and take care of the kids.” The whole family has to adjust. That’s clear. That’s what equal rights are about.

Gendered Influences on Political Ambition: Self-Perceptions Self-perceived qualifications also played a vital role in encouraging individuals to run for elective office. One state legislator recalled that she had originally run for office “out of anger”; she had attended a local council meeting about youth centers and other child-related issues in her community only to find that the mostly male city council had “no idea” about the issues and that the quality of their discussion was very poor. She, a mother of two boys, and a female friend in the audience that evening decided to run for office to bring their knowledge of youth issues to the council. A Social Democrat who founded an antinuclear group in the wake of the Chernobyl meltdown tried to lobby members of her town council’s environmental committee to improve her community’s nuclear disaster plans. After dealing with the committee, she quickly realized, “they have no idea [about nuclear disasters]. You have to do it.” She ultimately became her town’s vice mayor. A member of the Bundestag firmly asserted that he had pursued and received his party’s nomination because “I was better” than the other potential candidates, explaining, “[I had] political knowledge. As a candidate, or a potential candidate, you must take part in many debates and people notice whether you tell them dumb things or whether you are sensible.” A Social Democrat took on the role of deputy mayor in his town because “I thought I could do it”; a Christian Democrat offered a chance to run for his town council decided, “why not? I can try.” Conversely, respondents who doubted their ability to serve as an elected official refused to, or were very reluctant to, run for elective office. Occasionally male interviewees expressed that they were unqualified for a particular position. For example, one Social Democratic town council member asserted he would never consider running for a county mayor position (Landrat) because to do the job well, he believed, one must have studied Page 143 →law and he had not. However, in interviews, and in party leaders’ accounts of recruiting candidates for elective office, women were far more likely to express, or to be reported as having articulated, concerns about their qualifications for elective office. When asked about running for a high-level office, a female local councilor maintained, “if it happens that I receive a nomination it would be great, but I am not going to push my way forward. There are others who have more abilities and can do the job better.” A Russian immigrant, she worried that her (objectively quite good) German-language skills would not be sufficient. A leader in the CDU Women’s Union, whose colleagues maintained would be an excellent candidate, but who repeatedly turned down suggestions that she run, simply laughed off their praise, sarcastically arguing, “What am I supposed to run for? Mayor? I can never remember anyone’s name. How can I run for office?” One Christian Democratic Party leader charged with drawing up her local-level list observed that, in addition to citing children as a reason they could not run for elective office, women often claimed they were incapable of doing the job. She argued, “Women are too honest. If a man is asked whether he can do a job he’ll say вЂyes, sure’ right away,” while women actually think through whether they are suited or not and often come to doubt their abilities. The head of a women’s policy agency and active Social Democrat concurred, “Women have very high expectations of themselves.В .В .В . They ask, вЂCan I really do a good job at all aspects of the job, from A to Z?’ whereas men are easier on themselves and say, вЂthings will take care of themselves.’” A leader of the FDP’s Liberal Women claimed recruiting female candidates was difficult because many women she spoke with were not comfortable having to stand up in front of a group and argue for the party’s endorsement. Indeed, female interviewees repeatedly expressed hesitations about their ability to run for office based on the fact that they felt “uncomfortable” (unangenehm) with the type of activities a potential candidate had to undertake. One Christian Democrat argued that to develop the public visibility required for a successful nomination, one had to frequently appear in the local newspaper, which required running to the front of any group

about to have its picture taken. While she claimed her male counterparts were quite willing to stand front and center, even if they had a “beer belly sticking out. I do not like to do this; it’s uncomfortable to be in the foreground but it comes with the territory.” One state legislator had pondered calling all her friends in the party and asking them to come to the nominating meeting during which candidates were to be selected but she decided against it, not wanting to appear to be a “solicitor.” A Green who Page 144 →actively worked for the party as an adviser but not as a candidate argued, “if a woman decides to run for office then she will definitely get caught up in power plays (Machtspielen) whether she wants to or not. And that is simply uncomfortable.” A Social Democratic state legislator agreed: My predecessor decided not to run again and then the county party organization began to look for a successor. At the time it didn’t occur to me to run, but then others came to me and said, “you could do it.” At first I thought, no, I can’t do that.В .В .В . At first, I was a little scared of getting into a competition [for the nomination] with people with whom I was friends.В .В .В . That seemed like it would be uncomfortable and I went back and forth [about running]. Finally, her daughter convinced her to throw her hat into the ring by arguing “don’t act like this. You’re always preaching something completely different: that women should have self-confidence.” She went on to gain her party’s nomination, but described the process of visiting all the local party groups in her electoral district to shore up support for her candidacy as akin to “streetwalking.” A Green state legislator agreed and criticized the public way in which prospective candidates were voted on by other party members at a nominating meeting, arguing “running for office isn’t funny.” Especially for the state legislature, where the nominating convention received media attention, “the press is there and they use terms such as вЂincompetent’ or вЂfailed’ to describe you and you have to read about yourself in the newspaper. And not on the back pages either, but on page one, and if you are unlucky, with a picture too.” While some female interviewees did mention women role models who inspired them to run for elective office, other interviewees echoed the above concerns about the way the media would portray them based on their observation about other female politicians’ media treatment. One Christian Democrat fumed: It’s a fact that 52 percent of the population is female and so it should be fully clear that a similarly high percentage of women get elected. Unfortunately this is not so.В .В .В . Look at the situation, there were several interviews about Angela Merkel where [commentators] spoke not about her competence as a politician but rather about her ugly hairdo, her old-fashioned clothing, and her bad makeup. But these things don’t even tell us anything about a person’s competence. Politics shouldn’t depend on a person’s appearance but, Page 145 →rather, what does a person stand for? Can she accomplish something positive for Germany? But these [gendered comments] are things that we observe again and again and they make it fundamentally difficult [for female candidates]. Fig. 3.6. Eligibles’ Self-Perceived Qualifications for Elective Office. (Source: Candidate Interest Survey [N = 445]; *p = 0.1, **p ≤ 0.05.) Of course not all male eligibles feel comfortable campaigning either. One 12-year veteran of local politics posited, “Others are laid back about it but I have some problems [with the process of obtaining a ballot nomination]. It’s never very nice. You never get used to it. You learn who your true friends are and come to recognize your enemies. This can lead to personal disappointment and hurt.” However, explicit questioning of one’s own qualifications and concerns about the discomfort of running were far more often expressed by female than by male eligibles. The Candidate Interest Survey corroborates these qualitative findings. Although by virtue of their inner-party leadership position, all CIS respondents fulfilled gatekeepers’ primary qualification—loyal service to the party—women and men in parties with quotas had significantly different views of their suitability for elective office (see figure 3.6). On a scale from 0 (not at all qualified) to 3 (very qualified), women in parties employing

quotas gave themselves a mean score of 1.77 while men’s mean rating was 1.91. In contrast, women in parties without quotas were significantly more confident in their own abilities when compared to their male counterparts—women on average scored themselves a 1.94 compared to men’s 1.78. The Page 146 →few women who do manage to rise to the top of male-dominated parties without quotas are quite self-confident. Where quotas are used, a broader range of women emerge as party leaders, replacing some less ambitious men. Fig. 3.7. I Would Not/ Have Not Run for Office because I Am Not Qualified. (Source: Candidate Interest Survey [N = 430]; p ≤ 0.05.) Similarly, when asked why they had not ever run for elective office or would in the future be deterred from running, 8.3 percent of female eligibles in parties employing quotas maintained a lack of qualifications had held, or would hold, them back from seeking a nomination (see figure 3.7). The figure for men was significantly lower—only 3.5 percent. In parties without quotas significant sex differences were again not present with 4.4 percent of women and 3.6 percent of men citing a lack of qualifications as a reason for not seeking elective office. Gendered Influences on Political Ambition: Age Interviews also made clear the close connection between age and political ambition. A town mayor prophesized that when his term ended he would be 64 and a half and ready to retire. A businessman turned down an offer from his party to run for the state legislature when he was in his late fifties, arguing it would take years of preparation to achieve the visibility needed to win a seat and then several terms in the Landtag to move from the back benches to a position of power and influence. By this time he would have been in his seventies and he thought that “it would be better to let a young person” have the nomination. He suggested a 26-year-old woman from his precinct organization instead. A 60-year-old breast cancer survivor was offered a chance to run for the state legislature but turned the opportunity Page 147 →down, maintaining, “I don’t know how many years I have left—I want to enjoy life and remain unencumbered” by political responsibilities. The broader range of women in leadership positions in parties employing quotas also reflects itself in eligibles’ ages. In parties employing quotas, female eligibles are significantly older than their male counterparts with a mean age of 51 years compared to men’s 47. As willingness to accept a nomination decreases each year, here too women’s ambitions are depressed in parties employing quotas. In contrast, the women who emerge as leaders in parties without affirmative action measures are, on average, significantly younger (48 years) than their female counterparts in parties with quotas and closer in age to the average man in their own party (50 years).

Implications: Have Quotas’ Goals Been Fully Met? In sum, gender quotas have pulled many German women into elective office. If ambition is measured in terms of actually having considered running for local-level elective office, no significant gender differences emerge in the German candidate pool. In parties without quotas, a slightly higher percentage of female than male leaders had considered pursing a local-level elected post than their male counterparts—however, there are few such women in parties without quotas. When ambition is measured in terms of individuals’ willingness to accept a hypothetical candidacy, however, a gender gap emerges at the local level in parties employing quotas; female eligibles are significantly less likely than their male counterparts to express a desire to run for local office. While sex does not exert an independent effect in terms of considering running for or accepting a ballot nomination for high-level elective office, many of the detriments to political aspirations—including a lack of self-confidence, domestic responsibilities, and age—apply more frequently to women in parties with quotas than to men. As a result, on average, women in parties with quotas surveyed by the CIS were less likely to have considered running for high-level elective office and expressed less willingness to do so should they be asked to run for such a post. In sum, while quotas elevate women to positions of leadership within political parties and place them in a position where they at least consider elective office, women are less likely than men to become aspirants for local-level

offices and still face hurdles to ambition in terms of high-level elective office. Thus quotas have been only partially successful in achieving Mansbridge’s concept of justice. Quotas do bring a broader range of women Page 148 →to inner-party positions, and hence eligibility for elective office, than are present when quotas are not employed. Rather than being passed over in favor of more politically ambitious men, women who are not initially politically ambitious are tapped to fill party boards and find themselves in circumstances conducive to developing political aspirations. In parties without quotas women with little political ambition do not often rise to positions of inner-party leadership, curtailing the overall number of female eligibles when quotas are not employed. Nonetheless, quotas’ symbolic effects are limited. Even the women who are interested enough in politics to join parties and accept positions on local party boards when a quota is in place, remain on average less likely than their male counterparts to have always wanted to be a politician and less willing to accept a local-level nomination. The latter is a vital first step on the career ladder in Germany. Women are also more likely than men to find themselves in personal circumstances that depress ambition. Moreover, women still are less likely than men to view themselves as having the (normatively male) qualifications needed for elective office. As a result, then, while quotas ensure that citizens have a greater number of female aspirants from which their representatives can be chosen, quotas do not result in equal levels of political ambition among German men and women. The unequal distribution of household labor and the informal norms about what constitutes candidate quality—factors beyond the scope of quotas—dampen women’s political ambitions and again leave women underrepresented at this stage of the democratic process. The conclusion discusses policy recommendations to overcome these problems. These mixed effects likely impact women’s substantive representation. On the one hand, quotas succeed in bringing a larger number and broader range of women to inner-party office than in cases in which quotas are not utilized. Thus affirmative action measures do increase the scope of potential candidates to represent women’s interests, improving the odds that a fuller picture of women’s needs may be painted (Weldon 2002). On the other hand, however, because the women in parties with quotas remain less likely than their male colleagues to be willing to set foot on the career ladder, citizens are again left with fewer choices of female than male representatives. Moreover, the women who do aspire to elective office are more likely to be those who have higher levels of self-confidence and fewer encumbering family responsibilities than other women. The same, of course, is true for men, but such men are more typical of men as a group than women with few household responsibilities are typical of other women. The findings presented here do have positive implications for female Page 149 →elected officials’ ability to speak on behalf of other women once elected. Despite gender differences, well over half of both male and female eligibles, in parties with and without quotas, claim they would accept a hypothetical nomination for local-level elective office, and over 30 percent maintain they would consent to running for high-level elective office. As portrayed in table 2.1, the numbers of candidates actually elected to public office at the state, federal, and European levels are far, far smaller. Even for Green women, who enjoy the most favorable political opportunity structure, there are over 200 female party members for every winning “women’s” ballot slot. Women’s rates of aspiration, even if lower than men’s, seem sufficient for successful female nominees to avoid charges of tokenism. I now turn to the next phase of political recruitment and examine what party gatekeepers do when faced with both an eligibility pool with a gendered political ambition gap and quotas that call for higher percentages of women on the ballot than they have in their rank-and-file membership.

Page 150 →

Four Gatekeepers It wasn’t my intention [to run for the city council and later the state legislature]. I’m self-confident, but I wouldn’t have trusted myself to take on an elective office. But [the head of the local party organization] supported me very much. I had a lot of reservations and I needed someone who believed in me. —33-year-old woman from a party with a quota The previous two chapters established that quotas have only a limited ability to overcome gendered barriers in the political recruitment environment, resulting in fewer women than men joining political parties and indicating a willingness to campaign for local-level elective offices. However, by changing recruitment structures and requiring parties to select among their few female members for inner-party officers, quotas’ elevator effect has created a pool of female eligibles in Germany and, on average, these individuals do at least consider running for local posts as often as their male counterparts. Moreover, although there are low percentages of women in political parties, there are even fewer elective offices available, so that the number of women in parties still exceeds the number of offices designated for women at the European, national, state, and many local levels, allowing parties to implement their quotas if they choose to do so. Here I examine the next phase in the political recruitment process—the role played by party gatekeepers in identifying, training, and encouraging promising eligibles within their ranks to run for elective office. This process of recruiting auspicious eligibles occurs well before the nomination Page 151 →phase when candidates are selected to appear on the ballot. This chapter finds evidence that while quotas have not altered the recruitment environment, they have been able to shape the practices of party gatekeepers by modifying some political recruitment structures—not only by changing the formal criteria for drawing up lists of candidates but also by altering the informal process through which gatekeepers develop a cadre of electable candidates. This chapter investigates three questions. First, have quotas led parties to ask female and male eligibles to run for elective office at the same rate? Second, have quotas exerted a diffusion effect, also changing the behavior of gatekeepers in parties without official affirmative action policies? If so, through what mechanisms does this diffusion occur? Third, have quotas changed the ways in which parties identify and train promising candidates for elective office? Whether quotas lead women to be asked to run for elective office or not is important because recruitment by party gatekeepers has proven vital around the world in developing potential officeholders’ political ambitions and helping them become successful candidates (Kunovich and Paxton 2005; Lawless 2012; Norris 1993, 1997). Indeed, recruitment of candidates is one of the primary functions of political parties in a democracy (Gallagher, Laver, and Mair 2011). Moreover, while being encouraged to run for office is important for all candidates, being asked is particularly important for women because they are considered more sensitive than men to political opportunity structures; the expectation of a favorable response from party gatekeepers boosts a woman’s propensity of developing political ambitions more than it does a man’s (Fulton et al. 2006; see also chap. 3). In fact, where gender gaps in political ambition exist, they can be closed when influential individuals recruit women to run for elective office (Lawless 2012, 155). Finally, when parties make concentrated efforts to groom women for elective office, female candidates are less prone to accusations of tokenism and more liable to be considered serious politicians. If quotas lead parties to recruit their female members, rather than simply select them to appear on the ballot, then, they will not only enhance inner-party democracy, they will also help women to develop political aspirations, become candidates, and be taken seriously by others. In addition, the questions asked here are important because they contribute to a debate about the qualifications of individuals elected via quotas. Worldwide, considerable opposition to quotas exists because some believe

affirmative action measures will force gatekeepers to nominate unqualified women for elective office (e.g., Dahlerup 2006). The chapter makes both a Page 152 →theoretical and an empirical argument that, where binding candidate gender quotas are in place, gatekeepers have rational incentives to develop quality female candidates rather than simply make due with unsuitable women. This chapter represents one of the first efforts to investigate the microfoundations of the relationship between quotas and gatekeepers’ actions. The pathway case study research design used here allows me to assess the impact of various affirmative action measures on gatekeeper actions vis-Г -vis male and female eligibles, while controlling for confounding factors such as political culture, economic development, electoral institutions, and party ideology. This research design also enables me to document quotas’ impact across the political spectrum. I can investigate the mechanisms through which other parties’ quotas pressure gatekeepers in quota-less parties to recruit women to run for elective office. This chapter proceeds as follows. I begin by discussing the importance of parties asking their members to run for elective office and then develop arguments about the relationship between quotas and party leaders’ recruitment attempts—both in parties with quotas and in those without. I review existing evidence from the prequota era in Germany and then contrast this information to contemporary gatekeeper behavior. These latter findings are derived from the Candidate Interest Survey. Then, based on personal interviews as well as Germanlanguage sources, including media reports and academic studies, I depict the mechanisms that prompt gatekeepers both to ask women to run for elective office and to modify their traditional recruitment structures in order to identify and train auspicious women within their ranks. Finally, the chapter concludes by discussing both the ramifications of my findings and some questions they raise.

The Importance of Being Asked In party-loyalist candidate recruitment systems like Germany’s, being asked to run for elective office by one’s party is a vital precondition to securing a ballot nomination (Kintz 2011, 25–26; KГјrschner 2009, 17; Manow and Flemming 2012; Siavelis and Morgenstern 2008, 19–21). While being recruited to run is obviously not the same as winning elective office, not being asked to run is a virtual guarantee of not being elected. Even when entrepreneurial candidate selection processes are used, and parties are much weaker, however, recruitment efforts by gatekeepers such as party leaders exert a significant impact both on an individual’s political aspirations and whether she ultimately becomes a candidate (Lawless 2012; 180; Sanbonmatsu 2006). Thus, regardless of the mechanisms used to select Page 153 →candidates, party gatekeepers play a critical role in democracies worldwide (Gallagher and Marsh 1988; Hazan and Rahat 2010; Matland 2005; Norris 1997; Rai 2011; Siavelis and Morgenstern 2008). In all democracies—whether or not quotas are present—voters desire candidates who will represent their interests if elected (Iversen and Rosenbluth 2010, 137), and gatekeepers’ main goal is to select the candidates they believe will help their party win votes (Matland 2005, 97). While there is little evidence that voters discriminate on the basis of sex in long-term democracies like Germany’s (Casas-Arce and Saiz 2011; Fox 2010; Holtkamp, Wiechmann, and Schnittke 2009), scholars have put forth multiple explanations of why gatekeepers are often more likely to view men rather than women as winning candidates, including the gendered nature of what are perceived to be qualifications for elective office (Holuscha 1999; Matland and Montgomery 2003, 24; Murray 2010b), women’s inability to commit to uninterrupted political careers due to familial responsibilities (Iversen and Rosenbluth 2010), the gendered social networks in which gatekeepers find themselves (KГјrschner 2009, 18; Sanbonmatsu 2006, 153), the fact that most gatekeepers are men (Cheng and Tavits 2011), and misplaced concerns that voters object to female candidates (BjarnegГҐrd and Zetterberg 2011, 189). Interdisciplinary work indicates that human biology might play a role as well (Klofstad, Anderson, and Peters 2012). With so much working against female eligibles, it is perhaps not surprising that studies conducted where quotas are not utilized consistently depict women as less often asked than their male counterparts to run for elective office (Evans 2008; Fox and Lawless 2010; Kenny 2013; Lawless 2012; Niven 1998; Sanbonmatsu 2006; Shepherd-Robinson and Lovenduski 2002). Indeed, in many cases gender quotas were initially adopted because feminist activists hoped they could overcome such bias (Grolle and Bake 1995, 142; Kittilson 2006; Kolinsky

1991; Holuscha 1999, 207). To date, however, little evidence exists as to whether quotas actually change the behavior of gatekeepers or not, because, outside of the quota-less United States and United Kingdom, there are few systematic studies of candidate pools, making it difficult to ascertain on the basis of existing evidence how gender quotas influence the odds of parties asking their eligible female and male members to throw their hats into the ring. Some qualitative research conducted where quotas are in place has examined party leaders’ actions vis-Г -vis women, but these case studies have not made paired comparisons with men’s experiences, rendering it impossible to measure the gender gap in recruitment (for the German case, see Geissel 2000; Holuscha 1999; Horstkotter 1990). Instead, most investigations of quotas begin by examining women’s ballot placement or electoral success, withoutPage 154 → inquiring as to how candidates came to appear on their parties’ tickets (Casas-Arce and Saiz 2011; Escobar-Lemon and Taylor-Robinson 2008; Htun and Jones 2002; Jones 2004; Matland 2006; Meier 2004).1 While increased percentages of female candidates or legislators indicate that gatekeepers select female candidates, these numbers cannot tell us whether gatekeepers deliberately recruit women. This difference is normatively and substantively important. Gatekeepers may be forced by quotas to grudgingly place persistent women on the ballot after actively discouraging their aspirations (simply selecting women) or they may actively seek out and groom future female candidates (recruiting them). Clearly the latter is more consistent with quotas’ normative goal of increasing women’s participation in the democratic process than the former outcome. Moreover, if gatekeepers tend to prime quality male candidates for elective office and overlook identifying and promoting equally promising female candidates, women’s substantive representation may suffer as well—a result diametrically opposed to quotas’ original purpose (Mansbridge 1999; Phillips 1995). Indeed, much of the opposition to quotas stems from those who predict that affirmative action measures will lead to the selection of candidates who parties might otherwise not recruit (Mansbridge 1999; Dahlerup and Freidenvall 2010; Krook 2009; KГјrschner 2010). In many cases, gatekeepers express concern about being forced to accept poorly qualified female candidates to fulfill a quota and some women fear being labeled a “quota woman” or a token because of this practice (Berger, von Bothmer, and Schuchardt 1976, 13 and 52; Dahlerup 2006; Franceschet and Piscopo 2008; Holuscha 1999, 78; McKay 2004, 70; Rai 2011). Without systematically studying the experiences of both male and female members of the candidate pool, however, it is not possible to discern which mechanism is actually at work—are parties simply selecting any available woman (if women are selected at all) or are they strategically recruiting female eligibles? By interviewing and surveying both male and female members of the German candidate pool about their experiences I am able to empirically answer this question here.

Gender Quotas and Party Gatekeepers in Theory: Why Ask Women to Run? Theoretically, there is little reason to expect that in the long run party gatekeepers would sit back and allow quotas to thrust unwanted candidates upon them. Let us take a look at the situation from their perspective: since Page 155 →gatekeepers’ goal is to select winning candidates, their reactions to quotas will depend on whether or not they believe women (or at least some female eligibles) will be attractive to the electorate. If gatekeepers fear that female candidates will not garner votes for their party, they are unlikely to want to implement quotas and thus unlikely to endeavor to either recruit or select women.2 Empirical evidence abounds of parties failing to comply with (the spirit of) quota regulations, ranging from placing women mainly in unelectable positions (Htun and Jones 2002; Jones 2004; Murray 2004)—for example as backup candidates should the main candidate be unable to serve (Baldez 2007)—implementing quotas inconsistently (Meier 2004), paying fines rather than implementing quotas (Murray 2010b), ignoring quotas altogether (Krook 2009, 41), intimidating female candidates (Rai 2011, 205), or even going so far as resorting to electoral fraud (Krook 2009, 41).3 Where quotas have “teeth” and must be implemented, in contrast, party gatekeepers skeptical of female candidates’ viability cannot resort to such evasive strategies; for example, in countries with electoral law

quotas containing placement mandates, parties failing to put female candidates in winnable list places will not appear on the ballot at all, guaranteeing electoral loss. In such cases party leaders pressed to select female candidates have an incentive to actively identify, recruit, and develop the female candidates who can win votes for their party. One qualitative study of company leaders following the adoption of gender quotas for corporate boards in Belgium summarized their reaction to the law as trying to “make the best out of itВ .В .В . in such a way that the advantages [of having to select female board members] outweigh the disadvantages” (Roos and Zanoni 2015, 8). Similarly, when gatekeepers believe that women are—or can be—winning candidates, they possess the same decision calculus. Gatekeepers’ estimation of female candidates’ chances may be high due to reasons unrelated to quotas—the overall level of socioeconomic development and religiosity in a country (Inglehart and Norris 2003), a leftist ideology (Caul 1999), or the electoral system (Iversen and Rosenbluth 2010)—but where quotas are present there are additional mechanisms that raise the likelihood of women being seen as desirable candidates. First, quotas give ammunition to campaigns in support of quota implementation undertaken by feminist activists (Matland and Montgomery 2003, 33). Such “watchdogs” may be found among voters, NGOs, women’s policy agencies, sympathetic journalists, and within the ranks of the party itself, especially in women’s auxiliary organizations. Similarly, once a party is subject to a quota, its rivals may scrutinize whether or not it complies with quota rules in the hope of identifying violations that can be exploited Page 156 →by charging the party with hostility toward women. Second, if quotas initially lead to the success of some female politicians, other women’s candidacies subsequently begin to seem more viable. This has occurred even in sociocultural contexts with low levels of gender egalitarian attitudes (Beaman et al. 2009; Bhavnani 2009). Third, comparative research has determined that when the number of female gatekeepers increases, as is the case when quotas apply to inner-party offices, the probability that women will be asked to run for elective office rises as well (Caul 1999; Cheng and Tavits 2011; Kunovich and Paxton 2005). Whatever their motivations for recruiting female candidates, gatekeepers desiring to locate women to run for office face a complicated task; almost everywhere there are fewer female than male aspirants available (Matland 2005; Lawless 2012; Ashe 2015). In Germany, not only are there fewer female party members, those in parties employing quotas exhibit, on average, less willingness to accept a nomination to run for local-level elective office and, on average, possess lower levels of confidence about their own qualifications for holding elective office (see chapter 3). In contrast, male aspirants are more common and, on average, more confident about their ability to serve as an elected official. The higher the quota a party must fulfill (or the larger the gap between the percentage of female aspirants and percentage of female candidates required), the harder gatekeepers must look to find the quality female candidates they believe will aid their party’s electoral fortunes.4 As German parties’ quotas are set far higher than the percentage of women within their rank and file, especially in small localities, the Federal Republic fits this description. As a result, German gatekeepers in parties implementing affirmative action measures will have to ask more women than men to run for office in order to fill their quota. As the gap between the percentage of women required by the quota and the percentage of female eligibles increases, so too should the efforts of gatekeepers to ask women to run for elective office. Moreover, where quotas are binding, gatekeepers have less leeway to circumvent nominating women. Thus, the Greens’ and SPD’s higher, more binding quotas should prompt greater gatekeeper efforts to recruit women than the CDU’s lower, easy-to-circumvent quorum. Even gatekeepers in parties without quotas, however, may come to the conclusion that women can win if they observe parties employing quotas successfully contending elections with female candidates. Moreover, they may be subject to pressure from rivals or feminist activists among voters, NGOs, party members, and the media calling for more female candidates. This type of quota “contagion” has been observed as quotas have spread across parties, countries, minority groups, and public offices (Matland and Page 157 →Studlar 1996; Caul 1999; Engelstad and Teigen 2012; Thames and Williams 2013; BjarnegГҐrd and Zetterberg 2014; Krook and Zetterberg 2014) and are likely to similarly influence recruitment practices. Quotas in Germany therefore are expected to exert a diffusion effect on organizations without affirmative action measures. While these parties may not be more

likely to ask women than men to run for elective office—no quota requires them to do so and, as indicated in chapter 3, women who rise to the top of quota-less parties are already at least as politically ambitious as their male colleagues—they at least have incentives not to discriminate against women.5

Gender Quotas and Party Gatekeepers in Theory: Changing Recruitment Structures Quotas should not only shape gatekeepers’ propensity to ask their female members to run for elective office, they also can change recruitment structures, or the methods party leaders use to tap viable eligibles. These changes can take three different forms: redefining what constitutes a qualification for elective office, seeking female candidates in nontraditional places, and offering women mentoring and training programs to improve their selfconfidence. Because traditional candidate selection criteria have been more difficult for women than men to obtain (Matland and Montgomery 2003, 24; Murray 2010b), gatekeepers forced by a quota to select female candidates may need to rethink or refine selection criteria in order to locate enough women who they believe will be successful candidates. Comparative research indicates that potential candidates’ most widely valued characteristic is their track record of service to their party and electoral district (Gallagher 1988, 248; Matland 2005; Matland and Montgomery 2003, 24). This is especially true in parliamentary democracies featuring proportional representation; in such cases, voters prefer to vote for parties whose MPs reliably toe the party’s line once elected (Iversen and Rosenbluth 2010, 137). In the party-loyalist system of recruitment, then, a qualified candidate is one who has served her party well and can be expected to do so in the future (Siavelis and Morgenstern 2008, 19–21). When quotas are deployed in this context, party gatekeepers will need to identify female aspirants who they believe will be loyal to the party. As a result, gender quotas in such systems are often set not only for ballot slots but also for inner-party offices, creating opportunities for women to prove their fealty (see European Parliament 2008 for examples). In additionPage 158 → to holding inner-party leadership posts, party loyalty has traditionally been demonstrated by longevity of service to the organization. However, because women are often unable to have a long-term party career due to interruptions for care work (Iversen and Rosenbluth 2010), quotas may prompt gatekeepers to develop other methods of identifying faithful members when they are faced with determining which female eligibles may be counted on to toe the party line. Similarly, candidates have historically been selected by male gatekeepers who occupy gendered social networks (KГјrschner 2009, 18 and 204; Sanbonmatsu 2006, 153). Because parties cannot change the recruitment environment and create women with traditionally male careers or circles of friends, gatekeepers also will have to cast a wider net to find female candidates and attempt to identify nontraditional (read: female) ways of achieving visibility and a reputation within an electoral district. This broadening of selection criteria represents the second effect of quotas on recruitment structures. Indeed, cross-national evidence finds that “quota women” are not less qualified than their male colleagues in legislatures, but they are at times differently qualified (see contributions to Franceschet, Krook, and Piscopo 2012); for example, in Argentina both male and female members of Congress are highly educated, but women are more often doctors and teachers than men, who are more prone to possess degrees in engineering, accounting, and the social sciences (Franceschet and Piscopo 2012). Finally, where gatekeepers decide to recruit females, simply asking women to run for office may not be enough to generate sufficient candidacies. Because there are fewer women than men in political parties, and women are on average less confident about their abilities than their equally competent male peers (Lawless 2012; see chapter 3), organizations seeking female aspirants may have to create new training or mentoring programs designed to make women more confident about their own abilities—efforts not previously required when parties primarily sought aspirants among their, on average, more confident and more plentiful male members. Such candidate development programs have been implemented when parties were required to comply with quotas in countries as diverse as the Nordic states (Freidenvall, Dahlerup, and Skjeie 2006, 64), the Balkans (AntiД‡ and Lokar 2006, 159), and Indonesia (BylesjГ¶ and Seda 2006, 261–62). Thus, where gender quotas are enforceable and/or seen as desirable, party gatekeepers are expected to not only

ask women to run for elective office but also to alter the ways in which they recruit candidates. In Germany, the latter efforts include using characteristics other than simple longevity of party service to identify loyalty, expanding the social networks from which candidates are Page 159 →sought, and implementing candidate training or mentoring programs for women. The greater the gap between the percentages of women in a party organization and the percentage of female candidates mandated by a quota, the more pronounced these efforts will be. In contrast, parties not compelled to select female candidates due to a quota have much less incentive to undertake targeted measures to develop female candidates or revise selection criteria and thus are not expected to do so. While such parties may be equally likely to ask male and female eligibles to run for elective office, the women asked are likely to be more similar to their male counterparts than women in parties deploying quotas. The following section draws on the CIS and personal interviews to investigate these hypotheses in the German case. Alternative Explanations Before doing so, however, it is important to reiterate that the research design used here allows alternative explanations to be ruled out. For example, party ideology can be excluded as a cause of varying recruitment efforts: in Germany the two main Christian Democratic parties (CDU and CSU) have different rules for selecting candidates and any differences in their recruitment practices cannot be attributed to ideology. One might also argue that if a high percentage of female eligibles are asked to run for office in Germany, this outcome may be due not to gender quotas, but rather to the electoral system or the fact that Germans are simply likely to view women as winning candidates, that is, a result of the recruitment environment. It is the case that proportional representation electoral systems are considered more favorable to women’s candidacies (Matland and Studlar 1996), and Germany uses a mixed electoral system with a PR component; similarly, Inglehart and Norris’s (2003) analysis of the World Value Study data does show that Germany ranks among the top 10 countries worldwide in terms of progressive attitudes toward gender equality, a measure strongly influenced by citizens’ tendency to disagree with the statement “men make better political leaders than women.” Certainly, the electoral system and a gender egalitarian political culture have played a role in the success of the German Green party, the adoption and diffusion of gender quotas, and the strong propensity of German parties to recruit their female eligibles. However, political culture and the electoral system are constants, and any differences found in the recruitment efforts of parties with high, enforceable quotas, weaker quotas, and those organizations with no quotas will demonstrate the independent effect that quotas exert on gatekeepers. Page 160 →Moreover, evidence from other cases also suggests that quotas change gatekeeper behavior even in settings with low levels of gender egalitarian beliefs. For example, as Balkan countries began to adopt quotas, social democratic parties there turned to the Socialist International for assistance in building the capacity of female candidates in the region (AntiД‡ and Lokar 2006, 159; see also BylesjГ¶ and Seda 2006, 261–62). Finally, while mass attitudes toward gender equality shift only slowly as a country modernizes, in many cases gatekeepers’ actions vis-Г -vis female aspirants appear to have altered rapidly when quotas were introduced (Brown 2001; Lovenduski 2005), even in PR electoral systems (Cases-Arce and Saiz 2011). I now turn to the German case to explore how quotas have shaped gatekeeper behavior there.

Gender Quotas and Party Gatekeepers in Germany: Evidence from the Prequota Era Germany is a quintessential case of party-loyalist candidate selection; election laws grant parties exclusive control to select their candidates for the ballot and parties in turn choose these contenders from within their own ranks (Conradt and Langenbacher 2013, 191). As described in chapter 1, gatekeepers’ primary selection criteria was traditionally the completion of years of service to the party, the so-called Ochsentour. This recruitment process has made political careers very difficult for German women because it is at odds with the female biography (Kolinsky 1989; Holuscha 1999). Many politically interested women have been unable to pursue an uninterrupted party career, steadily moving up the ranks over time, because doing so is incompatible with childrearing and other domestic care work that may crop up over the course of a lifetime. While German women have often been very active in party politics as pupils and students, and later when their children have left

home, they have often dropped out of party life during their childbearing years and thus often had a thinner party rГ©sumГ© than men of their age. Even women with inner-party track records similar to that of their male peers may have been discriminated against by gatekeepers suspicious that their personal circumstances might change in the future. Compounding this situation, historically many gatekeepers believed that female candidates would be rejected by voters (Berger, von Bothmer, and Schuchardt 1976, 21, 44–45), and thus possessed little incentive to prime women in their party for electivePage 161 → office by promoting them within the party’s ranks, even if they were available to serve. Because completing an Ochsentour was more difficult for women than men, then, they were less attractive candidates for selection than their male counterparts. Experiences gained while moving up the party ranks have also been found to heighten German politicians’ political ambitions by teaching them the skills needed to hold political office and developing their confidence (Patzelt 1995, 266–68; chapter 3); excluding women from the Ochsentour likely not only made women unqualified for elective office, it likely depressed their interest in running as well. While no large-scale surveys of female and male party members exist from the prequota era, the extant anecdotal evidence strongly suggests that women were not often asked by their parties to run for elective office at the time (Berger, von Bothmer, and Schuchardt 1976; Kahlweit 1994; Schmalz-Jacobsen 1981). In 1981–82, members of the major German parties in one state were surveyed and 39 percent of the female party members expressed a willingness to hold an elective office; at the same time, however, only 18 percent of state legislators were women. The report’s author concluded that political parties were not tapping into their reserves of potential candidates (Hoecker 1986, 72; see also Hoecker 1994, 558). Mentoring and training programs for female candidates did not exist at that time. Instead, female party members’ participation was relegated to the women’s auxiliary wings—the Christian Democratic Women’s Union (Frauen Union), the Social Democratic Party’s Working Group for Social Democratic Women (ASF), and the Free Democratic Party’s Liberal Women (Liberale Frauen)—groups geared not toward elective office but toward social activities (Kolinsky 1993). Local party leaders spent less time communicating with new female party members than they did with new male members (Hoecker 1986, 71). When women were asked to run by male gatekeepers, they were assigned token positions low on the party list (Holuscha 1999, 54). For example, sometimes a young, attractive woman would be placed at the bottom of the ballot so the party could pose her front and center on their campaign poster in order to attract visual attention to their advertising (Berger, von Bothmer, and Schuchardt 1976, 21; Holuscha 1999, 177).6 Women who ultimately entered parliament were often dubbed Sargspringerinnen (“coffin jumpers”) because they received their seats midterm as a result of vacancies due to the death of a man ranked higher up on the electoral list (Berger, von Bothmer, Schuchardt 1976, 47; see also Grolle and Bake 1995). Page 162 →Against this backdrop, female activists within many German parties in the 1970s began to openly criticize both the Ochsentour criteria for candidate selection (Kittilson 2006; Kolinsky 1993) as well as sex discrimination on the part of male party leaders (Kolinsky 1991). The Liberal Women convinced the FDP to begin a “Women’s Campaign” in the hopes of recruiting more female party members (McKay 2004). The ASF called on the Social Democrats to adopt gender quotas to rectify these problems (Kittilson 2006). Such activism led to the diffusion of voluntary party quotas across the political spectrum: the Greens, and following unification the Left Party, adopted a minimum 50 percent quota for women on electoral lists, the SPD implemented rules requiring between 40 percent and 60 percent female candidates on their lists, and the CDU committed to a 33 percent quorum for women. Each quota is higher than the percentage of female party members and each contains a placement mandate. (See chapter 1 for more details of these measures). At the time the Candidate Interest Survey was taken, neither the CSU nor the FDP employed quotas for inner-party offices or party lists. This variation in affirmative action policy allows me to investigate various party quotas’ impact on contemporary gatekeepers and their propensity to recruit male and female eligibles for elective office.

Gender Quotas and Contemporary Party Gatekeepers in Practice: Who Is Asked to Run? The Candidate Interest Survey indicates that, a generation after quotas’ introduction, women in parties with

quotas are more likely to be asked by party gatekeepers to run for elective office than their male counterparts; significant sex differences in recruitment do not exist in the parties without quotas. The size of the gap between the percentage of women in a given party and the percentage of female candidates mandated by its quota does not consistently influence the size of the gender gap efforts to groom female candidates—although some indicators do suggest that the Green’s and SPD’s higher, more binding quotas lead to even more women being asked than the CDU’s weaker quorum. The CIS included both direct and indirect measures of gatekeeper actions. On the one hand, respondents were asked about their own personal interactions with party leaders; on the other, they were queried about their perceptions of gatekeepers’ preferences. Here I discuss each of these measures in turn. Page 163 →Direct Measures of Gatekeeper Actions CIS respondents were asked whether they had ever been approached by their party to run for any elective office. Those who answered yes were coded 1 and those who answered no were coded 0; this indicator variable is the dependent variable used in the multivariate analysis below. The survey clearly reached the targets of gatekeepers’ recruiting efforts; 81 percent of the respondents had been asked by their party organization to run for elective office. As expected, the difference between the means for men and women in political parties employing gender quotas was statistically significant. While 83.5 percent of the women surveyed had been asked to run for elective office, this figure fell to 77 percent for men (see figure 4.1). This statistically significant gender difference held in both parties with (near) parity quotas and the weaker quorum as well. However, contrary to expectations, the gender gap in being asked to run was greater in the CDU, the party employing the 33 percent quorum, than it was in parties with higher, more binding quotas. Parties not utilizing affirmative action measures evince no statistically significant differences in gatekeepers’ propensity to ask male and female eligibles to run for elective office. Notably, however, contrary to evidence from the United States and the United Kingdom where quotas are not used, women were not significantly less likely to be recruited than men in these parties either. The positive relationship between the female sex, quota type, and party leaders’ propensity to ask an eligible to run for elective office holds even when controlling for additional traits of interest to gatekeepers. A number of other variables besides quotas likely systematically shape parties’ considerations about who to ask to run for elective office. These factors are included in the models presented in table 4.1. Given the importance of the Ochsentour in Germany, I expect that the longer a person has been a party member, the more likely he or she is to be asked by the party to run for office; thus I include a measure of the number of years a person has been in his or her party in the analysis below. I also expect that party members who possess traits underrepresented in political parties will be more likely to be asked to run than people from overrepresented groups. While German parties purport a desire to represent a wide range of societal interests (Davidson-Schmich 2006b, 28–29), their members are not a typical cross section of society (Spier et al. 2011), making it difficult in practice for gatekeepers to develop a balanced slate of candidates. As detailed in chapter 1, underrepresented groups include, Page 164 →in addition to women, racial and ethnic minorities; those with low levels of education or income, or both; and individuals employed in the private sector. Combining employment in a private company with elective office in Germany is quite difficult, whereas the selfemployed have greater flexibility, making them gatekeepers’ prime targets according to interview respondents (see also Kintz 2011, 23). Hence I include an indicator variable measuring whether a respondent is self-employed. Because the CIS uncovered so few members of ethnic minorities in the candidate pool, there is no meaningful variation with which this hypothesis can be tested here. The model below does investigate the impact of household income and education level on an aspirant’s chances of being asked to run for elective office. Fig. 4.1. Eligibles Having Been Asked by Their Party to Run for Elective Office. (Source: Candidate Interest Survey [N = 444]; *p ≤ 0.1.) Finally, when I test these hypotheses below I also control for political opportunity structures such as the size of a respondent’s party and community, her state of residence, and her party’s strength in her electoral

district.7 Smaller parties have fewer members from which to draw their candidates and may thus be more likely to ask their members to run for office. This is particularly true because in addition to the few seats they win, small parties also feel compelled to nominate candidates for offices they are not at all likely to gain. In Germany, local councils are roughly the same size in sparsely populated rural districts and in densely settled urban areas, meaning that gatekeepers in small communities have access to far fewer aspirants per seat than do those in larger cities and hence may have to recruit more heavily to fill all available ballot slots. This is especially true for “women’s” slots on the ballot as parties have even fewer female members in thinly populated areas than in urban ones. Similarly, locally popular precinct organizations likely to win many seats need to ensure that their Page 165 →electoral lists are long enough and may also have to approach more party members about running than would gatekeepers in a precinct where the party is unlikely to gain many seats. Lastly, I control for state of residence as there is considerable state-to-state variation in Germany in terms of political culture and stateor local-electoral systems; moreover, variance in German states’ populations create different ratios of party members to available seats across the LГ¤nder. Further details about the coding of these variables and descriptive statistics are available in the appendix. Clearly, in addition to the variables outlined here, a range of other less-measurable factors may also come into play when gatekeepers consider who to recruit, such as a party member’s personality, public speaking ability, degree of public visibility, or the presence of a scandalous past. These variables are difficult to reliably capture in a survey and cannot be taken into account here, reducing the explanatory power of the models presented. Quotas’ systematic impact could be more clearly demonstrated if it were possible to control for such idiosyncratic factors. Table 4.1 depicts the results of this multivariate analysis. As expected, in parties employing quotas (Model 1), sex is a significant predictor of being asked to run for elective office, even controlling for personal traits and political opportunity structures.8 In parties without affirmative action rules, women were not any more likely than their male colleagues to be asked to run for elective office by party leaders (Model 2). However, it was also the case that women in these organizations were not systematically overlooked by gatekeepers either; an eligible’s sex played no statistically significant role in an eligible having been asked by the party to run for elective office. In both iterations of logit analysis, the longer a respondent had been a party member, the more likely he or she was to have been asked to run for office by the party, as would be expected in a party-loyalist system. In parties with quotas, the self-employed were targets of gatekeepers’ requests to run as well. Living in a highly populated area reduced the likelihood an individual would be asked to run for office and residing where the party was popular increased the chances. These variables did not significantly predict recruitment in parties without quotas; high education did, however. Contrary to expectations, household income had no consistent bearing on whether an individual had been asked to run for elective office and controls for party size as well as for state of residence were not significant. Table 4.2 compares the relative impact of the variables found significant in Model 1. In parties with quotas, women were 7.4 percent more likely to be asked to run for elective office than their male peers, holding all other variables at their means. To put the magnitude of this change into perspective,Page 167 → it is roughly equivalent to having an additional decade of experience in a party and is far greater than moving from a suburb to an urban center or from an electoral district where one’s party has mixed electoral fortunes to a constituency where the party dominates politics. Page 166 → Table 4.1. Predictors of an Eligible Being Asked by the Party to Run for Elective Office: Logistic Regression Results Model 1: Model 2: Parties with Quotas Parties without Quotas Variable Coefficient (Standard Error) Coefficient (Standard Error)

Candidate Traits Sex (1 = Female) Years in Party Respondent Self-Employed Educational Attainment

0.48 (0.31)* 0.03 (0.01)** 1.14 (0.51)** 0.02 (0.09)

1.02 (0.91) 0.09 (0.05)* 1.02 (1.06) 0.32 (0.23)вЂ

в€’0.08 (0.10) в€’0.12 (0.27) Political Opportunity Structures/Controls Population of Electoral District в€’0.04 (0.01)* 0.48 (0.45) Household Income

Party Strength State: Hamburg

0.36 (0.28)†−0.51 (0.46)

0.00 (1.04) в€’17.1 (4786)

State: Bayern State: Nordrhein-Westfalen State: Baden-WГјrttemberg Small Party Constant N

в€’0.08 (0.72) 0.20 (0.61) 0.20 (0.71) 0.25 (0.37) 0.52 (1.03) 282 0.067 0.076

0.98 (5434) в€’16.2 (4786) в€’15.1 (4786) 16.0 (2573) в€’2.9 (5434) 81 0.2733 0.218

Pseudo R2 Prob > chi2

Source: Candidate Interest Survey. Note: Dependent Variable: Respondent Was Asked by Party to Run for Office (0–1). †p < 0.20, * p ≤ 0.1, ** p ≤ 0.05, *** p ≤ 0.01. Table 4.2. The Relative Impact of Independent Variables on Being Asked to Run in Parties with Quotas Men Women Change from man to woman NA +7.4% 1 standard deviation change in years in party +7.2% +5.4% (SD Men = 12 years; SD Women = 10 years) Change from non-self-employed to self-employed +16.2% +11.6% 1 standard deviation change in population of electoral district (7-point scale) −2.0%−1.5% 1 standard deviation change in popularity of party in electoral district (3-point scale) +3.7% +2.8% Source: Candidate Interest Survey. Note: Results based on Model 1 in table 4.1 using Stata command “prchange,” holding all variables (except sex) at their mean. In sum, as expected, direct measures of gatekeeper activities indicate that women in parties employing candidate gender quotas are more likely to have been asked to run for elective office by their parties than are their male peers; this effect is greater than an additional decade of experience in the party. Higher quotas were not consistently found to cause greater gender gaps, however. Women and men in parties without affirmative action policies were equally likely to have been asked by gatekeepers to become a candidate. Indirect Measures of Gatekeeper Actions In addition to these direct measures of gatekeeper propensities, the CIS also contained some indirect measures,

querying survey respondents about their perceptions of gatekeepers’ actions. Most eligibles believed that their party would react favorably (rather than neutrally or negatively) if they personally declared their political aspirations; in parties employing candidate gender quotas, as predicted, no significant sex differences in this perception emerged (see figure 4.2). In contrast, women in parties without quotas were significantly less likely than their male counterparts to believe their party would react favorably to their political aspirations. Disaggregating by quota type indicates that “quota women’s” optimism about gatekeepers’ responses to their candidacies is primarily driven by SPD and Green respondents’ views. Women in the organization with the weak women’s quorum—the CDU—were significantly less likely to expect positive responses than their male counterparts; in contrast, women and men in parties with (near) parity gender quotas were equally likely to expect a favorable response from gatekeepers. These varied views of party gatekeepers’ preferences continued when respondents were asked which sex would be more likely to be nominated by their local party organization for a directly elected post (see figure 4.3). In this figure, a high score would indicate that party members believe women have greater chances than men of being nominated while a low score implies the opposite. On average, eligibles in parties utilizing gender quotas believed men and women enjoyed equal chances in this regard, while individuals whose parties did not employ gender quotas were significantly more likely to predict a man would have an advantage over a woman. Fig. 4.2. How Would Your Party React if You Declared a Desire to Run for Elective Office? (Source: Candidate Interest Survey (N = 420]; *p ≤ 0.10, **p ≤ 0.05.) Page 168 →Not all quotas are created equally, however, and evidence of this can be ascertained by breaking down Candidate Interest Survey responses by gender quota type (see figure 4.3). Eligibles in parties deploying parity or near-parity quotas paint a picture of gatekeepers who are more enthusiastic about female candidates than do those individuals hailing from the organization with only a weak women’s quorum. Christian Democrats were more likely than their Social Democratic and Green counterparts to believe that a man had a better chance of being selected as a candidate than did a woman. Given that the quorum is interpreted as granting men two seats for every one allotted to women, this result is unsurprising. This indirect evidence offers stronger support than did the direct measures that the size of the gap between the percentage of women present in a party organization and the percentage of women required by the quota exert an impact on gatekeeper actions. As expected, members of the candidate pool in Germany perceive different gatekeeper attitudes toward women’s candidacies depending upon whether their political party employs a gender quota or not. Indirect measures also indicate that party members in the organizations with the largest gaps between the percentage of female eligibles and quota-mandated female candidates believe gatekeepers to be more likely to Page 169 →promote female aspirants than do respondents from the party with the less-enforceable quorum. Fig. 4.3. Who in Your Party Group Is More Likely to Be Nominated for a Direct Seat? A Man or a Woman? (Source: Candidate Interest Survey [N = 459]; ***p ≤ 0.01.)

Mechanisms: Why Do Quotas Lead Gatekeepers to Ask Women to Run for Office? Qualitative evidence, including interviews and media reports, establishes several mechanisms through which quotas have led to women being asked to run for elective office in Germany.9 First, as parties experienced electoral success using quotas, gatekeepers became convinced that women were indeed viable candidates. Second, quotas have given ammunition to watchdog groups concerned with securing women’s political representation. Finally, quotas have created a class of female gatekeepers who can recruit other women and promote their candidacies. Below, I discuss each of these mechanisms in turn. Perceptions of Female Candidates’ Viability The Green Party’s use of a parity quota went hand in hand with its increasing electoral success in the late 1980s. As a result, gatekeepers in other parties began to view nominating women not as a liability, but as an asset. Page 170 →One study of Bundestag members revealed that almost half of the non-Green respondents believed the

rise of the Greens had had an impact on their own party by prompting it to more actively recruit female candidates (Kaiser 2001, 108–9). Indeed, German media reports regularly depict parties as believing female candidates are electoral assets (e.g., “Frauen sind” 2007; Hoffmann 2008). The head of one Land-level women’s policy agency attributed the gains made by women in politics in her state over the past several decades to the advent of the Greens and their quota; the success of their female candidates in her state, she argued, proved to other party organizations in the region that women were willing and able politicians. A state legislator attributed the rising number of women on her city council to her own efforts in the 1990s to start a women-only candidate list to diversify the all-male city council. After two women entered the council from the women’s list, other parties began ensuring that they could boast of female nominees too; by the early 2000s her city’s council was one woman short of parity and the FDP had recruited her to run for the Landtag. In fact, in the over 40 interviews conducted for this project, only two conversation partners expressed concern that women could not win elections, and both of these younger, female interviewees attributed negative perceptions of female candidates not to the electorate as a whole but specifically to older, conservative men. Quota “Watchdogs” Given the widespread consensus that women are not liabilities but strategic assets as well as the voluntary promises by parties to promote female candidates, feminist activists can exert pressure on party gatekeepers to recruit female candidates. Moreover, parties’ quotas have given their rivals ammunition to use against them should they not implement their own affirmative action policies. Interviewees across parties and states agreed that attention is now paid to the gender balance of lists by voters, NGOs, parties’ own women’s auxiliary organizations, journalists, and members of rival factions. Interviews indicate that gatekeepers across the political spectrum believe voters care about women’s presence on the ballot; when a Green state legislator was asked whether people would notice if her party did not fulfill its parity quota she maintained that, yes, reporters know the Greens are supposed to be a party with a high quota and an attention paid to women’s issues. She claimed that there is an “expectation” on the part of the media and feminist voters that the list will be at least half women; a Green candidate from another state agreed the same was true in her area. Page 171 →Across the aisle, interviewees from the CDU, the party with the weak quorum, believed a genderbalanced slate was less important to their core voters; however, they maintained that it was still vital for the party to adhere to its own quorum. One state legislator argued that “If you try to compete without women—for example we have a list with 20 percent women and the others have 40 percent or 45 percent women on the list—that is going to catch people’s eyes in a negative way, and it’s also going to have a negative impact” on election results. One interviewee reported that his Christian Democratic Party organization didn’t want to be seen as “an old boy’s club” and thus always tried to fulfill its quorum by putting together a party list that reflected the underlying population. As a result, he argued, party leaders tried to “target” women active in the community as potential candidates, claiming they “enjoy advantages” in candidate selection. Interviewees also cited NGOs as a source of pressure to recruit female candidates. When asked whether voters would notice or care whether her party implemented its quota, a state legislator responded in the affirmative, mentioning a feminist organization in her Land that monitored all party lists and public appointments. She outlined past instances when this group was able to mobilize feminists across the state to make an “outcry” against quota violations. In one German Land, the state-level Women’s Policy Agency (WPA) writes to each party before elections asking them what steps they are taking to ensure that they will have enough female candidates to achieve the promised level of women’s representation in the state parliament. Each party’s answer is published in a newsletter that is sent to the members of over 50 women’s organizations of all stripes—not just to feminist groups but church, farming, and professional groups representing more conservative voters as well. The head of the WPA reported that she was happy to publicize this information as ammunition for feminist activists within parties with no or weak quotas. She claimed that these activists in turn were able to use evidence of women’s successes in parties with higher quotas to pressure their own organization into also

giving women some good spots on the party list. Interviews suggest that these partisan women’s auxiliary organizations play a key role in enforcing quotas. A long-time member of the Social Democratic ASF observed, “people accept the quota and know if I try some tricks [to avoid the quota] there are going to be actively engaged women who are going to put the brakes on me, ” recalling: There once was a case when a Bundestag nomination was being awarded and there were only men on the short list. Then we [the Page 172 →women’s organization] said we want the quota to be implemented, and we also at the same time suggested a woman, and she was also ultimately selected. Then there were some long facesВ .В .В . but that’s the way it is with us. We have the party statutes on our side. When women have not been appropriately represented on draft party lists, state-level party activists have not hesitated to report gatekeepers’ quota violations to the national party leadership for correction (“Wolfgang Rose” 2007) or to the media (“Liberale Frauen” 2012; “Der Aufstand” 2012). Were such efforts to fail, the Social Democratic ASF could bring the matter to a court of law as it is illegal for German political parties not to follow their own statutes. Not only do feminists pressure party leaders to comply with their own quotas, other parties routinely chastise their rivals, especially quota-less parties, for having too few female candidates. Journalists often echo these critiques in the media. In January 2013, one of Germany’s newspapers of record, the SГјddeutsche Zeitung, ran the headline “The FDP Has a Problem with Women,” publicizing a critique of the party made by the SPD’s general secretary (“Die FDP” 2013). The Handelsblatt (a conservative business paper) trumpeted “The Women’s Problem in the CSU Is Getting Worse” (“Frauenproblem” 2012). The surprise 2011 victory of the high-tech Pirate Party in Berlin’s state elections—leading the party to gain 15 seats, 14 of which were held by men—led to a firestorm of criticism from other parties including the Greens (“Warum” 2012) and the SPD (“Ich suche” 2011), as well as many media commentators (e.g., Gathmann and Naumann 2011). Even Germany’s widely derided far-right parties are disparaged in the media, not only for their controversial policy stances but also for their lack of female candidates (“Frauen sind” 2007). Quota Diffusion These watchdogs have created a diffusion effect that has spread across the political spectrum to parties not employing affirmative action. A city council member representing the quota-less Christian Social Union in a very small, rural, socially conservative town maintained, “In the last electoral cycle there were no women at all in [our town’s council] but this was not the case in other nearby villages. This really hurt the reputation of [the town] and it was also bad for the image of the party. People thought we were some kind of exclusive club that shuts out women.” As a result, she maintained, she was asked by the party to run for the council in the next Page 173 →election and they placed her first on the party’s list; the election after that the party made efforts to ensure that a quarter of their candidates in top spots were women. A state-level manager of the quota-less FDP responded to a question about his party’s track record of electing women by glancing over his shoulder, quietly confessing, “You can’t say this in public but [our party is] bad in terms of women. This is a problem” for attracting voters. Indeed, FDP women in another state reported that, due to the dearth of female party members, the party made efforts to include women on its electoral lists. A German-language study of female politicians from the quota-less CSU also found gender being a common reason women in that party were asked to run for elective office. A party member commented, “We don’t have a quota, but informally there really is one. There’s this saying, вЂgroup picture with a woman.’ .В .В . That isn’t a quota but it’s a so-called must-have.” Another agreed, “I mean, we don’t have a quota but it’s really de facto like a quota” (KГјrschner 2009, 185; author’s translation). Survey data confirms these interview responses are not aberrations. When asked if their local party

group made an attempt to place women in promising places on their party lists, 80 percent of CIS respondents from parties not using a quota agreed. Female Gatekeepers Another mechanism through which quotas drive gatekeepers to ask women to run for elective office is by creating a class of female gatekeepers who can recruit other women and promote their candidacies. A Green state legislator argued that the quota has been a big help to women in the party, not only because it gave them access to innerparty posts that are the qualifications for elective offices, but also, she stressed, the quota has put women into positions of power themselves. Such posts give women access to information—which is “half the battle” in being prepared to contest an open seat—and opportunities to “get to know each other, develop trust, and work together on strategies” to act on behalf of other women. The head of a WPA reported a similar trajectory, saying that today when a female aspirant needed to be promoted, her organization could “pick up the phone and act” via their network of women in various political parties. A veteran state legislator also noted that women were increasingly able to defend their interests as more and more women became active in politics. Several other female gatekeepers discussed how simply being present as lists were developed allowed them to intervene on behalf of female aspirants. One state legislator from the CDU recalled a party meeting where Page 174 →leaders claimed they could not fulfill the quorum because they had already looked and been unable to locate any available female candidates. Previously at the same meeting, however, a recently retired woman had let the assembly know she had more time to devote to party work. The state representative “spontaneously” suggested this retiree for the ballot position and she was awarded the nomination. To the extent that female gatekeepers exist in parties without quotas, they can perform similar functions. For example, one CSU leader interviewed by Isabelle KГјrschner asserted, “I swore to myself, when I get to have a say in things, then I must promote [other] women” (2009, 182; author’s translation). However, given the low percentages of women in German parties and the lack of mechanisms to promote them within quota-less organizations, female gatekeepers are in the minority in parties such as the CSU and FDP. In sum, there are several quota-driven mechanisms that lead gatekeepers—both in parties with quotas and in those without—to ask women to run for elective office, including quotas’ improving perceptions of female candidates’ viability, creating “watchdogs” who ensure quota compliance, and elevating more women to gatekeeper status. The latter mechanism is particularly strong in parties with high and binding quotas as they possess the highest percentage of female gatekeepers and are required to locate more female candidates than other parties. However, simply asking existing female aspirants may not yield enough female candidates to fulfill a party’s quota.

Gender Quotas and Party Gatekeepers in Practice: Changing Recruitment Structures As German parties with quotas have sought to recruit more women, they have come up against obstacles, finding fewer female than male aspirants through traditional channels. When quotas were first adopted in the early 1990s, the extant evidence suggests, parties with quotas may have resorted to simply turning to any available woman to run, bypassing the traditional Ochsentour (Kolinsky 1993, 143–44). Research conducted at that time found that women were recruited for elective office based solely on their sex. For example, one study of women active in local politics in the state of Niedersachsen found that the most common reason these women said their parties gave for asking them to run was that the party wanted to increase the representation of women in the local council; this was more common Page 175 →than being asked because they held a leadership position within the party, had ties to an important interest group, or because they had expertise in a particular policy area (HorstkГ¶tter 1990, 119).10 From the perspective of gatekeepers seeking the most electorally viable candidates possessing high degrees of party loyalty, however, this outcome was clearly suboptimal. My more recent interviews and the CIS establish that gatekeepers in parties with quotas have now altered their practices in order to identify and train qualified women

to run for elective office. These efforts are particularly systematic where a large gap exists between the percentage of women in a given party and the percentage of female candidates required by the quota. Redefining Selection Criteria: The Ochsentour One way in which quotas have changed recruitment structures is to redefine what constitutes the Ochsentour necessary to be nominated. Rather than eliminate the Ochsentour altogether, quotas have begun to relax—but not eliminate—their requirement of long-term service to the party. One study of Bundestag members finds that women and men representing the parties with quotas discussed here have statistically significant differences of means in their tenures as party members, with women on average having been members for fewer years than men when they first entered the national parliament. This was not the case for the quota-less Free Democrats where on average male and female MPs had equal tenures in the party upon Bundestag entry (Kintz 2011, 119).11 However, in Germany’s two largest parties with quotas (the CDU and SPD), the average man was a party member for 17 years before being elected to federal office; for women the figure was 14 years—hardly an elimination of the Ochsentour (Kintz 2011, 125).12 Qualitative evidence finds a clear-cut relationship between quotas and changing recruitment structures. Interviewees from parties with quotas indicate that if female party members take a break from party work for childrearing, their interrupting an Ochsentour is no longer punished. Women who become active in their parties again after their children are older are no longer discriminated against for having temporarily withdrawn from party life. Especially the parties with high and binding quotas report trying to capitalize on the women who are currently active in their ranks. A college student elected vice president of her local Social Democratic Party group found nothing unusual about her rapid ascent through the ranks but noted that older women in the organization were particularly excited Page 176 →about her selection, coming up to her and telling her what a great thing it was. Another young Social Democrat explained that she joined her party in early 2005 and then was asked to run for city council in the fall of 2006; “it went off relatively quickly,” she summarized; in late 2012 she was selected to run for an open Bundestag constituency. Interviewees from quota-less parties did not report such rapid political career trajectories. My results parallel the findings of a German-language study of female politicians in Berlin. One woman from a party employing a quota, who initially had no political ambition, recalled, “The party organization was looking for suggestions of people to nominate and then they said, вЂYou’ve not been with us very long but what you’ve accomplished within the party—you could also do that on the city council. Don’t you want to [run]?’” (Geissel 2000, 125; author’s translation). One of her colleagues observed, “It’s clear there are too few women in party politics. I think that naturally played a considerable role in my case. It went pretty fast and then they asked me too” to run for city council (Geissel 2000, 126; author’s translation). Redefining Selection Criteria: What Constitutes Visibility? German quotas also have refined recruitment structures other than the Ochsentour. In addition to party loyalty and service to the party, one of the main selection criteria for German candidates is visibility in the community, which Germans refer to as a prospective candidate’s Bekanntheitsgrad (literally, “the degree to which someone is known”), somewhat akin to the American concept of name recognition. Quotas have altered recruitment structures in this respect as well. Traditionally, membership and leadership roles in certain professions and social groups were the litmus test of an aspirant’s visibility; popular sources of candidates included the volunteer fire department, and high-visibility professions such as law enforcement, jobs important to local economy, or sports clubs. Often these organizations are either exclusively male (e.g., the men’s soccer club) or predominantly male (e.g., the volunteer fire department), making it less likely that there would be high-profile female party members in their ranks. Where quotas have forced parties to select female candidates, party leaders now report casting a broader net to locate well-known women in the community. One of the first places they have thought to look for women is through the school parent-teacher or youth sports organizations (Holuscha 1999, 164). A Christian Democratic

woman from the south of Germany who ultimately went on to become a member of the state cabinet claimed her party recruited her because she was a mother of four. She Page 177 →served as the room mother in each of their classes and took part in almost all local efforts to improve the lives of children, meaning that she knew many, many other parents of school-aged children. One Social Democratic leader in a small northern town reminisced, “I’ve sought out new candidates and I’ve often found them through my children. I looked for women active in the preschool or school” and encouraged them to become active in the party. Parties seeking to implement their quotas have also sought female candidates from the ranks of citizen initiatives or other volunteer groups such as movements for safer streets or environmentally friendly zoning, as well as organizations of professional women such as lawyers or city planners (Holuscha 1999, 164, 180–81, 198).13 A Green candidate believed she had been nominated because she had high visibility in feminist circles through her work in a women’s policy agency. Another interviewee noted that women who worked as nurses or in services related to her town’s tourist-based economy had been tapped as candidates. Women active in church groups were also recruited. Parties without quotas, or those willing and able to sidestep their quorum, have less of an incentive to seek out nontraditional applicants. A case study of Berlin women in elective office found that parties employing quotas had recruited women from a broader range of backgrounds than quota-less organizations (Geissel 2000, 54). While the latter tended to look for women in similar places as they found male candidates—women from politically active families, with public-service related educations, and in public sector workplaces—the former also recruited women whose households were not particularly politicized, whose educations had little to do with politics, and who worked outside the civil service. Training and Mentoring Women Once gatekeepers in parties with quotas locate women who they believe will make good candidates, they may encounter difficulties convincing them to run for office—especially for high-level posts. Because German women in parties with quotas have been found, on average, both to be less confident in their own qualifications for elective office than similar men and to be less willing than men to run for low-level elective offices, one way in which party gatekeepers can try to recruit women is to offer special training programs to turn eligibles into aspirants, by making women more confident of their own abilities. Prior research on Germany shows that national party organizations have pledged to do just this; in the late Page 178 →1990s parties with quotas including the Greens, the Social Democrats, and today’s Left Party began formal mentoring programs pairing aspiring women with female officeholders and offering training classes in areas such as public speaking (McKay 2004). My research indicates that these programs created at the national level have indeed had tangible grassroots outcomes. Quantitative evidence from the CIS reveals a statistically significant difference in the percentage of women reporting access to training programs in parties with and without quotas (see figure 4.4).14 While 68 percent of women in parties employing quotas claimed to have access to such career development initiatives, the figure fell to 59 percent for women in quota-less organizations. Clearly both these percentages are quite high and a testament to the efforts German parties put forth in developing candidates for elective office; however, women in parties without quotas do find themselves in more of a sink-or-swim atmosphere than women in other parties. Moreover, higher percentages of women in the parties with binding (near) parity quotas reported access to official party training programs than did women in the CDU with its circumventable one-third quorum. The qualitative evidence concurs. Of the women interviewed for this project, even when asked directly not a single member of a quota-less party indicated that her party offered any type of training program for women. Christian Democrats’ answers to my questions reflected their party’s middling commitment to increasing female representation: approximately half of the CDU members interviewed mentioned efforts to increase the numbers of female candidates, but these were informal attempts at candidate development done by interested gatekeepers rather than systematic party programs; one Christian Democrat even argued that women did not need such encouragement. In contrast, every single interviewee from a party with a parity or near-parity quota could discuss her organization’s formal efforts to develop female candidates. Here I depict responses from

interviewees across the political spectrum. Interviewees from the Greens and Social Democrats detailed their organizations’ formal attempts to develop women candidates. A Social Democratic state representative who headed her party’s women’s auxiliary organization argued, It is difficult to activate and engage [female party members] in their middle years. Because the party is a hindranceВ .В .В . we must try to develop attractive opportunities for women members.В .В .В . Our party caucus at the state level began a mentoring programВ .В .В . for female party members who will be promoted by other women who already Page 179 →have one or more offices with the party. We have had very much success with our mentoring program.В .В .В . It was received so favorably that even our male colleagues in the caucus leadership, the ones who used to say “well that’s an expensive program,” are now totally convinced and have given us the green light to continue. Fig. 4.4. Percent of Female Respondents Reporting that their Party Offers Training Programs for Aspirants. (Source: Candidate Interest Survey [N = 213]; * p ≤ 0.10.) A Green state legislator also described a similar program: “We’ve had mentors for years.В .В .В . We also try again and again to get young women excited about taking on party posts.” These programs seem to be effective. A female member of the SPD described her entrГ©e into public life: “I have always been interested in politics. The problem was вЂhow do you break in’ to a party? This was very difficult.” Then she read a newspaper article about a Social Democratic mentoring program for women, which was “exactly what I was looking for.” She applied and was accepted, attended some workshops and training programs and, most important from her perspective, was paired with a local (female) mayor as a mentor. Subsequently she was elected to local-level office and then to the Bundestag. In interviews, representatives of the CDU’s Frauen Union consistently mentioned this women’s auxiliary group’s efforts to mentor other women, rather than official CDU party programs.15 A state legislator and former state-level Frauen Union head responded to a question about whether the party was able to identify enough female aspirants, stating, “that’s where we break down. We break down because of women’s selfconfidence, when it comes to them believing in themselves. And that’s an assignment the Page 180 →Women’s Union has to take on, right? We have to do a little bit more motivational work.” A woman with leadership positions in her local, county, and state Christian Democratic Party groups argued, “At first I was opposed [to the quorum]. I said to myself, вЂwhy do we need a crutch? We can already walk, ’” but, after the quorum was implemented, her experiences convinced her that we need [the quorum]. Sometimes women lack self-confidence and sometimes it’s not easy [to convince them to assume a new role]. I don’t know why. With the quorum you have to try to build up women’s confidence by saying “you can do it just as well as any man.” If the quorum forces the party to ask women to run for elective office, [female eligibles] will say “man, if you need me, if you think I can do it” and agree to be nominated. Women will run if you support them. Moreover, she argued, the quorum encourages women by making them think they have a “better chance” of getting this support. The head of a county-level Frauen Union organization reported that her group’s main job was trying to encourage women to run for office and groom them into viable candidates for elections several years hence: “we have to consider which candidates we want to promote and then strategic groundwork is very important to build up” a credible candidacy. In the election following the interview, one of the women being cultivated was indeed elected to the state legislature. She agreed that the “political seminars” and mentoring she had received from the party were “very important” to her decision to throw her hat into the ring. She described

the role played by her mentors and concluded that “without these people I would have never had the courage” to run for office. Interviews with grassroots members of quota-less parties found no evidence of such development programs for female eligibles.16 Instead, many respondents simply lamented the lack of qualified women in their party. One study of women in the quota-less CSU also found this party’s lack of support for female aspirants a common theme among interviewees. One elected official lamented, “It was always learning by doing. There were no mentors to give one advice. I was always left as a lone warrior.” Another recalled, “Yes, I told Minister X [that I was going to run for office]. Then he wished me well, but nothing more. You don’t get any more support than Page 181 →that from others, even from your own county-level party group. They say, вЂtry it’ and if you win you’re a big hero; if not, then that’s fine with them too.” (KГјrschner 2009, 160; author’s translation) Thus both quantitative and qualitative evidence establish that the parties with the highest and most binding quotas for women have made the strongest systematic efforts to train female candidates for office, followed by the party with the lower, weaker quorum. Parties without quotas have made fewer efforts in this regard. Moreover, the evidence presented here likely underestimates the degree to which training programs for women have proliferated in Germany since the adoption of quotas. Above I discuss only political parties’ efforts to develop their own candidates. Public funds also have been spent in many states and localities on initiatives to induce women, regardless of party, to run for office. For example, the state-level Civic Education Office (Landeszentrale fГјr politische Bildung) in the Christian Democratic–dominated state of Baden-WГјrttemberg ran seminars before local elections in the 1990s to teach aspiring women how local governments function, how to become a candidate, and how to conduct an election campaign. City-level women’s policy agencies in that state developed seminars for women on topics such as how to get involved in local politics and how to read a budget (Holuscha 1999, chap. 5). In sum, German parties with quotas—especially with those with binding, (near) parity quotas—have changed recruitment structures in order to obtain the highly visible, reliably partisan women they need to fill their ballot slots, resulting in more gender egalitarian behavior on the part of gatekeepers. German parties without quotas did not consistently make such efforts.

The Results of Asking Women to Run and Changing Recruitment Structures As a result of asking women to run for elective office, relaxing the Ochsentour, casting a wider net for candidates, and implementing mentoring and training programs, parties with quotas indeed have little trouble locating suitable female aspirants when it comes time to nominate candidates. Interviewees from parties without quotas, in contrast, more frequently maintained that there were no (or few) qualified women in their organization who could become candidates. Rather than reporting strategies to develop female candidates, they appeared resigned to having few or mentionedPage 182 → attempts to meet very minimal numerical targets for female candidates, such as ensuring one woman was on the list (see also KГјrschner 2009, chap. 4). The state leader of a Free Democratic women’s organization claimed that her organization tried to promote female candidates—if they could locate someone qualified. She expressed opposition to picking women just because of their sex, fearing it could lead to the selection of a candidate ill-suited for a position who was not skilled enough to handle it. Responses given by interviewees representing parties employing affirmative action measures were quite different. Many female interviewees attributed their own decision to run for elective office to encouragement by party gatekeepers. Moreover, virtually all gatekeepers interviewed reported being able to locate qualified female candidates for high-level elected offices, although the situation at the local level was sometimes different. Here I discuss responses across the various quota types, beginning with the Greens, the originator of the parity quota. A Green gatekeeper from southern Germany reported that “we haven’t had a problem so far” locating

female candidates. A Green state representative in a northern Land boasted that his party organization never had difficulty finding women to run for the state legislature and attributed this outcome directly to quotas, arguing that requiring women to hold inner-party offices at rates equal to men was “an encouragement for women to run for office.” He scornfully noted this was not the case for the [quota-less] Free Democrats whose state parliament delegation at the time contained no women. A Green woman from a small town in northern Germany concurred that “[n]aturally we can find enough women to run. It’s logical—the quota has proven itself and we see fights [among party members] for the women’s slots on the ballot.” In fact, she reported, female aspirants often also compete for the even-numbered positions on the party’s list.17 A local leader from the SPD, which employs a 40 percent quota, agreed that “we have enough women” to fill ballot slots. Another Social Democratic gatekeeper reported that “there are also sometimes situations when men are happy that there’s a gender quota becauseВ .В .В . there are more women available as potential candidates than men.” A Social Democratic state legislator reported: [Since adopting the quota] we have cities where there are more than 50 percent women on the list, where men have to watch out that they get their share of list places.В .В .В . In the past ten years there have been clear changes.В .В .В . The women’s quota has unambiguously led to this advancement. If I think about many precinct organizations, Page 183 →they used to just say there were no female candidates available but now they must search for eligible women and give a credible excuse if they can’t locate any. Responses from the party utilizing the more flexible quorum were more mixed. On the one hand, one state gatekeeper argued that “we have very many good women” within our ranks, and another CDU member responsible for putting together the party list for elections to her town council agreed that when the quorum was first adopted it took several attempts to put together a list containing the requisite number of women but now the quorum had become routine and there were not problems locating aspirants. On the other hand, in small towns where parties had few members this proved more difficult. This was true not just for Christian Democrats but also for Social Democrats. One SPD leader confessed, “but again and again there are small communities” where there are “a lot of problems” finding enough female candidates for local office. Overall, however, quantitative evidence indicates that gatekeeper recruitment efforts have paid off (see table 4.3). The Candidate Interest Survey asked eligibles who had run for elective office, or had considered doing so, what factors had prompted their interest in throwing their hats into the ring. The three most frequent answers included a motivation to change something in the community, the desire to represent citizens’ interests, and the goal of pursuing a certain political agenda. Marked sex differences emerged when eligibles were asked if they became candidates as a result of being asked by their party or encouraged by a political mentor. In parties with quotas, women were twice as likely as their male colleagues to agree that they had become candidates at the urging of their party or a political mentor; these sex differences did not manifest themselves in a statisticallyPage 184 → significant manner in the quota-less parties investigated here (see also KГјrschner 2009, 87). While both men and women reported wanting to change something at similar rates, in parties with quotas men were more likely to want to push forward a political agenda while women were more likely to express a desire to represent their fellow citizens. Table 4.3. What Prompted You to Run (or Consider Running) for Elective Office? Parties with Quotas Parties without Quotas Reason for Running Women Men Women Men To change something 85% 85% 90% 83% To represent citizen interests 73%* 68% 69% 72% To accomplish political goals 58% 66%** 56% 59%

Because my party offered me a ballot slot 43%***

25%

37%вЂ

28%

A mentor encouraged me to do so

5%

8%

6%

13%***

Source: Candidate Interest Survey. Women’s and Men’s Difference of Means Significant: †p ≤ 0.20, * p ≤ 0.10, ** p ≤ 0.05, *** p ≤ 0.01 Disaggregating these responses by quota type sheds light on how the different affirmative action policies manifest themselves among gatekeepers (see figures 4.5 and 4.6). In parties with parity or near-parity quotas, nearly twice as many women than men claimed to have run as a result of being asked by party gatekeepers; this difference of means is statistically significant at conventional levels. A gender gap also emerges in the CDU, which uses a weakly enforced 33 percent quorum, but the difference of means is not as large and is significant at a lower confidence level. The gender gap in respondents attributing their decision to run for elective office to a mentor’s encouragement also exists in both parties with (near) parity quotas and in the party with the weaker quorum. In contrast to expectations, however, this gap is larger in the latter party. As suggested by interviews, this may be a result of the Frauen Union’s more ad hoc, personalized mentoring arrangements in contrast to the SPD’s and the Greens’ more formally institutionalized party programs, which lead party leaders to ask women to run in an official capacity, rather than leaving the task to informal mentors. The end result, however, is that after asking their female members to run for office, relaxing Ochsentour requirements, expanding the social networks from which candidates are drawn, and developing mentoring and training programs for women, parties employing quotas—especially those that are binding and well above the percentage of women in the organization—can locate female candidates to fulfill their quotas, at least for highlevel offices; this is less often the case in parties without affirmative action measures.

Implications: Have Quotas’ Goals Been Fully Met? This research is one of the first systematic candidate pool studies conducted in a setting in which gender quotas are used. The results establish that, as their adopters intended, quotas shape the practices of party gatekeepers. Parties employing candidate gender quotas ask their female eligibles more often than their male members to run for elective office. This finding holds regardless of the level of the party quota, but some Page 185 →measures indicate that higher, more enforceable quotas result in more frequent requests for women to run for office. No significant sex differences in being asked to run were observed in parties without quotas, which in itself is noteworthy, as it sets the results off from those found in countries where gender quotas are not used and where men are asked more often than women to run for office. Fig. 4.5. Eligibles Reporting they had Run for Office (or Considered it) because they were Asked by their Party. (Source: Candidate Interest Survey [N = 330]; * p ≤ 0.10, *** p ≤ 0.01.) Fig. 4.6. Eligibles Reporting They had Run for Office (or Considered It) because They were Encouraged by a Mentor. (Source: Candidate Interest Survey [N = 330]; ** p ≤ 0.05, *** p ≤ 0.01.) Quotas prompt gatekeepers to ask women to run for office and exert a Page 186 →diffusion effect on quota-less parties by making women appear to be viable candidates and creating quota watchdogs who “name and shame” parties that fail to contest elections with well-placed female candidates. Quotas also result in the creation of female gatekeepers who can promote other women’s candidacies. In addition, the data presented here demonstrate that parties implementing quotas have begun to alter their recruitment practices in order to develop a cadre of loyal, locally visible women who will be able to garner votes for their party. Such parties, especially those with high and binding quotas set above the percentages of women in their organizations, have ceased penalizing interrupted party careers due to caregiving, widened the social circles from which candidates are recruited, and implemented mentoring and training programs for women. This chapter also offers evidence to contradict quota opponents who argue that parties will be forced to select unqualified women in order to generate sufficient female candidates to fill ballot slots allotted to women; many

women have opposed gender quotas precisely because they fear being labeled tokens rather than serious politicians (Franceschet, Krook, and Piscopo 2012). I provide clear evidence that, in the decades since quota adoption, parties with quotas have taken steps to ensure themselves a supply of qualified female candidates. While the average man asked to run for elective office had been a party member three years longer than the average woman, it would be difficult to categorize a woman who had served her party for 16 years as unqualified to hold elective office. In addition, I also present evidence that parties do take women’s professional qualifications into consideration just as they do men’s, for example, recruiting the self-employed and women visible in their communities. In this case, quotas have not led to the recruitment of unqualified women selected solely on the basis of sex. However, being asked by her party to run for elective office is of course no guarantee that a woman will actually accept a ballot nomination or, if she does, that she ultimately will be elected. Indeed, it is possible that gatekeepers’ recruitment efforts may simply be designed to target women to serve as “sacrificial lambs” in unwinnable ballot slots or for insignificant local posts. For example, a study of women in the quotaless CSU found that they were selected as candidates to run in hopeless elections. One mayoral candidate reported, “it was a situation where the CSU said, we probably won’t win even with a man so we’ll try something new. We’ll put forth a woman”; another candidate for a district-level post believed, “I was selected because the chance of the CSU winning this office was equal to Page 187 →zero. Otherwise they’d never have given [the nomination] to a woman.” A third recalled, “In the CSU there was a quiet discussion about whether we should even put up a candidate against this incumbent. Many believed we’d already lost the election. Then the head of the party asked me to run” (KГјrschner 2009, 144–45; author’s translations). It is to this possibility that the next chapter turns.

Page 188 →

Five Candidates and Elected Officials Yes, naturally there are enough women to run. It’s logical. The quota has proven itself. When it really counts [i.e., at the top of the list] there are fights for the women’s list positions. After a while the zipper comes to a stop, however. We try when it’s possible [to include women further down the list] but there are not as many women as men in the party. —Green city council member In 2007 I was elected via the eighteenth position on the list. My goal was to obtain the twenty-first list place and I would have been happy if I had been nominated for the twenty-fourth. —Female CDU state legislator It was really more like an alibi function that one had. One or two [female candidates] or so were toleratedВ .В .В . but even that couldn’t be taken for granted. —Female CSU politician describing how she came to be elected (quoted in KГјrschner 2009, 143; author’s translation) The past three chapters have discussed gender quotas’ impact on political recruitment in Germany. Chapter 2 established that, while the percentage of women in German political parties has edged up slightly since quota adoption, female members remain a minority in all German parties—even those with parity or close-to-parity gender quotas. Quotas have nonetheless helped to propel female party members into the positions of innerpartyPage 189 → leadership that are the prerequisites for a successful candidacy in Germany, creating many female eligibles in the Greens, Social Democrats, and the CDU. Moreover, chapter 3 found that women occupying these leadership positions consider running for local-level elective offices at similar rates to their male counterparts, although they are less likely than their male peers to claim they would be willing to actually become a candidate for local-level office. Gender does not exert an independent effect on the willingness of an eligible in a party with a quota to agree to campaign for high-level elective offices; however, female eligibles in these parties are on average less confident about their own qualifications to run for office and more likely to be responsible for domestic work than similarly placed men—both factors that reduce political ambition. Chapter 4 noted that low percentages of politically ambitious female party members, combined with higher quotas for female candidates, produce a situation where women are often asked by their parties to run for elective office. Moreover, gatekeepers in parties with parity and near-parity quotas have begun to allow for interruptions in the Ochsentour, seek female candidates from a variety of backgrounds, and have started informal and formal training and mentoring programs to encourage female eligibles to become aspirants. This chapter delves into quotas’ impact on the final phases of the political recruitment process: becoming a candidate and, eventually, an elected official. Here I investigate whether female eligibles in parties with both (near) parity quotas and the 33 percent quorum are as likely to seek a nomination for elective office as male eligibles in their parties, whether they are as likely as men to actually be selected as candidates, and whether they are as likely as men to be elected in the end. I ask the same of male and female eligibles in quota-less parties. The answers to these questions are important to assessing whether quotas’ goals have been met, but these topics are rarely addressed in the existing literature on quotas and women’s representation. Most research on the subject begins at the stage when women have been successfully chosen as candidates (e.g., Jones 2004; Dahlerup 2006; Krook 2009; Tripp and Kang 2008). While this conventional approach does allow us to compare whether

male and female candidatesВ are equally likely to win elective office, it cannot tell us anything about whether or not equally qualified male and female party leaders actually run for office at the same rates or whether gatekeepers select male and female aspirants to become candidates at equal rates. This information is critical to assessing whether or not quotas’ original goals have been met. Page 190 →In contrast, the pathway case research design developed here allows me to identify members of the eligibility pool and investigate gendered patterns of political recruitment in the “secret garden” of politics—the phases before aspirants become candidates. Based on my Candidate Interest Survey and personal interviews with German eligibles, I find evidence that quotas exert their strongest impact at this stage of the recruitment process, exercising a significant impact on men’s and women’s chances of seeking and receiving a local-level ballot nomination and ultimately being elected in their communities. Parity or near-parity quotas improve women’s chances vis-Г -vis men’s, whereas the 33 percent, difficult-to-enforce quorum worsens women’s chances of election vis-Г -vis their male colleagues. These patterns are similar—but less pronounced—for the state, national, and European parliaments. In parties without quotas, where sex is not a formal consideration in candidate selection, few significant inner-party gender differences emerge. I develop these conclusions as follows. First, I discuss the importance of studying the eligibility pool rather than only individuals actually appearing on the ballot. Second, I theorize the relationships between quotas and women’s and men’s propensities to seek a ballot nomination, to become a candidate, and to be elected in Germany. Third, based on the Candidate Interest Survey, I empirically establish the positive effect that parity and near-parity quotas have on female eligibles’ fates and demonstrate the quorum’s negative impact on women. Finally, I conclude by expounding upon these findings’ implications.

Why Study the Eligibility Pool? To date we know little about the relationship between quotas and male and female eligibles’ relative propensity to seek elective office, gain ballot nominations, or ultimately win election. Eligibility pool studies conducted where gender quotas are not used find qualified women less likely to run for elective office than similarly placed men (Norris and Lovenduski 1993; Lawless and Fox 2005; Evans 2008; Lawless and Fox 2010; Lawless 2012). This literature suggests that one mechanism through which quotas may have led to an increase in women’s representation is by increasing the number of women who step forward to seek a ballot nomination. The extant literature on quotas suggests that this is the case. This work generally proceeds from the candidate stage, examining those who are nominated by their party to run for elective office. It finds that that when Page 191 →quotas are enforced, enough qualified women come forward to fill ballot slots allotted to female candidates (Brown 2001; Lovenduski 2005; Casas-Arce and Saiz 2011; but for a more nuanced view, see Murray 2010b). Indeed, as demonstrated in chapter 1, German parties have been able to comply with their quotas at the highest levels of government. The number of female Bundestag candidates has steadily increased since the adoption of quotas in the mid-1980s (Manow and Flemming 2012, 772).1 Even in very traditional cultures such as India and Bangladesh, where sociocultural barriers to political participation are quite high, enough female aspirants can be found to fill quotas (Dahlerup 2006)—suggesting that when quotas are in place, some women will indeed be willing to become politicians. However, given the small number of political posts compared to the very much larger underlying population of a particular country, finding enough female aspirants to occupy available seats should not be particularly challenging, even when women are far less likely to seek elective office than their male counterparts. For example in Germany there are only 598 seats in the Bundestag but almost 400,000 female party members (Niedermayer 2011) and over forty million female citizens. Locating aspirants sufficient to achieve gender parity in the Bundestag would still be possible even if extremely low percentages of female eligibles harbored political aspirations. Thus in order to assess whether gender quotas lead qualified women to run for elective office at similar rates to their male counterparts, a scholar must begin by examining the eligibility pool rather than the candidates on the ballot.

It is also not possible to determine from the existing literature whether male and female aspirants have equal chances of being selected by gatekeepers to appear on the ballot in auspicious positions. Scholars focusing on gender quotas virtually always study those successful in gaining a nomination (Matland 2006; Jones 1998, 2004; Htun and Jones 2002; Meier 2004; Manow and Flemming 2012; Hennl and Kaiser 2008; Fortin-Rittberger and Eder 2013) or who ultimately get elected (Franceschet, Krook, and Piscopo 2012), rather than those individuals who present themselves to the selectorate. Understanding aspirants’ odds vis-Г -vis gatekeepers is normatively important, however, as quotas’ goal of “justice” or gender-equal participation in democracy depends on whether both male and female citizens aspiring to political office have equal odds of appearing on the ballot, regardless of their sex. In Germany understanding how aspirants fare is especially important because German voters do not discriminate against female candidates (Holtkamp, Wiechmann, and Schnittke 2009) and the most important step in gaining elective office in the Federal Republic is Page 192 →not being elected by voters, but selected by gatekeepers to run (Manow and Flemming 2012). If female party members are less likely to be nominated than their equally qualified male peers, they are also unlikely to ever be elected by voters. Because women’s underrepresentation was so severe in the prequota era, gender quotas could have increased the percentages of women in elective office while female party members still retain poorer odds than their male peers of gaining a ballot nomination. Only by comparing the fates of eligible male and female party members can we determine whether gatekeeper discrimination exists. Beate Hoecker’s work on Germany in the prequota era indicates that far higher percentages of female party members harbored political ambitions than the percentage of women in the Landtag, but her study could not distinguish whether this result occurred because women were unlikely to attempt to gain a nomination or because they were unlikely to be selected by gatekeepers, or both (1986). In contexts where quotas are not used, comparative research indicates that male eligibles are more likely than female eligibles to receive ballot nominations (Evans 2008; Sanbonmatsu 2006; Ashe 2015). To date, however, no large-scale, English language studies examine whether this is the case where successful quotas are in place. Finally, the existing literature cannot tell us whether male and female eligibles are equally likely to ultimately be elected. Quotas could conceivably increase the numbers of women elected to public office while female party members still remain less likely than men in their parties to win office. Work on political recruitment in countries without quotas reveals that female members of the candidate pool are less likely to ever be elected than equivalent male members (Lawless and Fox 2005; Lawless and Fox 2010; Lawless 2012). Where quotas are in place, female and male candidates’ chances of winning have been studied (e.g., Holtkamp, Wiechmann, and Schnittke 2009) and women’s odds of election are good, but these studies do not tell us whether male and female eligibles are equally likely to win office. As a result, we have only limited knowledge of quotas’ effects. Here, in contrast, using the eligibility pool as a baseline, I focus on the final three steps of political recruitment: an aspirant seeking the nomination for elective office, gatekeepers granting this nomination, and a candidate being elected to public office. This research design allows me to shed light on the effectiveness of candidate gender quotas in meeting their goals of promoting gender-equal participation in democracies, creating role models, and improving the substantive representation of women. Investigating who does and does not seek a ballot nomination helps Page 193 →establish whether or not quotas have led men and women to participate equally throughout the political process—not just in the final stage—achieving what Jane Mansbridge refers to as “justice” (1999). Quotas were developed in order to increase women’s participation in democracy (Dahlerup 2006), and if they render male and female aspirants equally likely to step forward as candidates, they will have succeeded in their mission. If women in parties with quotas remain reluctant to campaign, however, quotas will have only partially succeeded—even if they do increase the numbers of women in elected bodies. Under such circumstances quotas could also be expected to have little symbolic, or role model, effects (Mansbridge 1999; Schwindt-Bayer and Mishler 2005; Dahlerup and Freidenvall 2010). Moreover, if women retain a low propensity to throw their hat into the ring, women’s substantive representation may suffer. Comparative research has shown that the more women who participate in the political process, the greater the range of women’s interests raised in public debates (Celis 2006 and 2009;

Weldon 2002). If female eligibles are less likely to seek elective positions than male eligibles, voters will suffer from a narrower range of potential representatives and may in turn see fewer of their interests articulated. Finally, the more women who come forward to contest a list place, the more likely the winning nominee/officeholder will be considered by others in the party as a legitimate politician in her own right, rather than a token selected only to fill a quota. Even if quotas do increase the number of women in elective office, they may still fall short of their potential if female aspirants remain less likely than their male counterparts to be selected for a party list. If this were the case, quotas would have failed to accomplish gender-equal participation and women’s substantive representation could be limited. Moreover, if quotas were to encourage women to come forward as candidates only to be disproportionately turned down by party gatekeepers, a negative feedback loop could be created, actually discouraging female candidates in the long run. In other words, quotas could have negative symbolic effects, deterring other women from attempting, and girls from considering, a run for office rather than inspiring them to do so. The data below are thus a pioneering analysis of whether quotas improve female aspirants’ odds of appearing on the ticket in winning slots vis-Г -vis their male counterparts. Quotas can only be considered a partial success at achieving gender-equal participation in a democracy if women are less likely than comparably placed men to be nominated and elected, even if the numbers of women in elective office increase from a prequota era. Focusing on the overall candidate pool allows me to probe these possibilitiesPage 194 → in a way that studying the percentages of women on the ballot or among winning candidates does not.

Gender Quotas, Seeking and Receiving a Ballot Nomination, and the Odds of Being Elected In the earlier phases of political recruitment discussed in this book, the recruitment environment played a large role in shaping women’s experiences. Domestic responsibilities reduced women’s propensity to join political parties, to assume leadership roles within them, and to develop political aspirations. Masculinized informal recruitment structures also dampened women’s interest in joining parties and running for elective office. However, the above effects represent average outcomes. Not all women were deterred from joining parties, taking on leadership functions, and developing political aspirations. Female aspirants—those women comfortable in parties and hoping to win elective office—have overcome gendered hurdles in both the recruiting environment and in informal recruitment structures. Once a woman (or man) has decided in principle to run for elective office, formal recruitment structures have been found to exert considerable impact both on the decision of when to run and on the likelihood that an aspirant is successful in his or her quest for elective office (Lawless 2012). One of the main reasons politically ambitious individuals decide to actually step forward—as opposed to merely considering a run for office—is that they detect a favorable political opportunity structure; in other words, eligibles become aspirants when they think they can win (Schlesinger 1966; Maisel and Stone 1997; Maestas et al. 2006). While positive political opportunity structures are important in encouraging all potential candidates, they have been found to be particularly important in inspiring women to run for elective office (Fulton et al. 2006). Quotas directly shape women’s (and men’s) political opportunities by determining which eligibles may seek and receive which specific nominations; the higher and more binding the quota, the better the opportunities for women are (and the more limited they are for men). It is thus at this stage of political recruitment that quotas are expected to exert the most clear-cut effect on political recruitment. Because specific quota rules create distinct political opportunity structures, it is necessary to distinguish between the different affirmative action policies in place in Germany. Below I discuss in turn the hypothesized relationships between (near) parity gender quotas and the CDU’s 33 percent quorum on women’s and men’s propensity to seek elective office, to become Page 195 →candidates, and to be elected; I also discuss expectations for parties without quotas. I expect that candidate gender quotas create within-party gender differences in the propensity to seek a ballot nomination, receive one, and eventually be elected: in parties with

parity or near-parity quotas these gender differences are hypothesized to favor female eligibles, but in the CDU gender differences are predicted to favor men. Because parties without quotas do not have rules associating certain ballot slots with a particular sex, male and female aspirants’ experiences are thought to be similar. (Near) Parity Quotas In the Green and Social Democratic parties, female members are expected to enjoy advantages over their male colleagues in terms of being nominated for and winning elective office—especially at the local level. As a result, female aspirants are expected to be more likely to seek ballot spots than men. This expectation is driven first and foremost by the sharp imbalance between the low supply of female party members and the high quota-driven demand for female candidates that parity or near-parity candidate selection rules create. The Greens’ 50 percent quota is a minimum percentage for women (rather than an upper bound); while oddnumbered ballot slots are reserved for women, the latter may also appear on even-numbered places. Indeed, interviewees report that in nominating meetings women do routinely compete for the even-numbered positions. In contrast, the Greens’ national membership is only 37 percent female, creating a minimum 13 percent gap between the percentage of female members available and the percentage of female candidates required. This creates a very favorable party-member-to-winning-ballot-slot ratio. As depicted in table 2.1, at the time the CIS was conducted, in the five representative LГ¤nder surveyed, the Greens won one elected “women’s” seat at the local level for every eight female party members (compared with fourteen for men). By virtue of their inner-party leadership posts, the women surveyed here possess the qualities sought by gatekeepers and are thus eligible for these ballot spots. Moreover, as discussed in chapter 2, female members are not spread evenly throughout the country, but tend to be concentrated in densely populated urban areas, making the ratio of local seats to eligibles even more favorable in rural areas. While there is fierce competition for Landtag, Bundestag, and European Parliament ballot nominations, Green women still enjoyed better odds of being elected than either men in their party or women in other parties (see table 2.1). There were 202 Green women for every one Page 196 →winning Landtag, Bundestag, or European Parliament mandate in contrast to 344 male Greens per winning “men’s” spot. Moreover, Green party statutes make it difficult to violate this quota; a majority of the women (not men) present at the nominating meeting must approve any deviation from the quota (Davidson-Schmich and KГјrschner 2011, 29). Because it is difficult for Green gatekeepers to avoid selecting female candidates, they have taken steps to ensure themselves a supply of qualified women aspirants by broadening the social circles from which candidates are recruited, recognizing interrupted party careers due to caregiving, and developing training and mentoring programs for potential female candidates (see chapter 4). Green gatekeepers are also more likely to ask women to run for elective office than they are to ask men. These efforts are important because recruitment efforts by party gatekeepers have been found in other contexts to outweigh women’s lack of confidence in their abilities in inspiring women to run for elective office (Lawless 2012, 180); this should aid in overcoming the barriers to women’s political ambitions observed in chapter 3. In sum the Greens’ quota has created excellent political opportunity structures for women and should inspire them to come forward as candidates, especially at the local level, where the odds are particularly in their favor. The SPD’s near-parity quota functions in a similar fashion. The Social Democrats’ affirmative action policy requires that between 40 percent and 60 percent of their candidates are female and their statutes permit no violations of this regulation. Activists within the party have taken quota violations to inner-party adjudication bodies for correction (“Wolfgang Rose” 2007), threatening legal action should the party not abide by its own regulations. This translates into a 9 percent to 29 percent gap between the percentage of female party members available nationally and the required percentage of female list candidates.2 At the local level, in the five states surveyed by the CIS, there was at least one winning “women’s” seat per 28 female party members at the time the survey was conducted, compared to one man’s seat per 39 men3 (see table 2.1 for details). Again, this ratio was even more favorable in rural areas where the percentage of female party members was lower than in cities. Social Democratic women also faced less competition (one winning “female” ballot spot per 500 party members) than SPD men (one winning “male” slot per 750 members) for high-

level posts, although both sexes held long odds of winning a nomination for these powerful bodies. The individuals surveyed by the CIS are particularly attractive candidates for these positions by virtue of their innerparty leadership posts. Like the Greens, the SPD has also taken steps to identifyPage 197 → promising female candidates and developed training and mentoring programs for them; gatekeepers are more likely to ask Social Democratic women to run for office than they are to ask Social Democratic men. Here too, then, women face very promising political opportunity structures encouraging them to come forward as candidates. Thus Green and SPD women who aspire to local-level elective office, especially in thinly populated areas, enjoy high chances of being selected for winning ballot slots. These very favorable political opportunity structures, combined with their parties’ extensive efforts to identify and train potential female candidates, are expected to encourage female candidates to come forward. While there are fewer female than male aspirants in these parties, those women who do harbor political ambitions are highly likely to come forward when observing these propitious circumstances—especially at the local level. In contrast, the odds faced by Green and SPD men at the local level, while good, are not as auspicious as those for women. There are fewer ballot slots designated for Green men (less than or equal to 50 percent of the total) than for women, and more men (63 percent of the total party membership, more in rural areas) to contest these positions. Moreover, Green men are on average more likely to be willing to accept their party’s local-level nomination than Green women. Similarly, men make up 69 percent of the SPD’s members but are guaranteed only 40 percent of the list positions; they are also more willing than women to claim that they would accept a local-level ballot nomination. Thus Green and SPD men’s odds of receiving a local-level nomination are poorer than women’s and these opportunity structures should dampen men’s propensity to throw their hat into the ring for a local race; while there are more male than female aspirants, they are expected to be less likely than women to express their ambitions by seeking a ballot nomination. In sum, against the backdrop of the Green’s parity and SPD’s near-parity quota, women are expected to be more likely than men to seek a local-level nomination and become a candidate. Moreover, to the degree that not enough female aspirants can be found to fill all “women’s” list positions, men are expected to appear more often than women at the bottom of the electoral list, and hence more likely than women to lose a local race. Green and SPD women’s advantages over their male colleagues are hypothesized to be similar—but far less pronounced—for higher levels of elective office. Women’s absence from low-level list positions is less likely for high-level elective offices because larger electoral districts and higher ratios of party members to available seats make finding female aspirants at this level easier.Page 198 → Moreover, women are more willing to accept ballot nominations for professional high-level offices than they are amateur local posts. The 33 Percent Quorum Gender is expected to play a significant, but quite different, role in the party with the lower, weaker quorum: the CDU. Here the party is also officially bound to place a higher percentage of women on the ballot than present in the candidate pool; while women make up only 26 percent of party members, the quorum allots them 33 percent of list positions. On paper, female CDU members face better odds of receiving a winning ballot position than CDU men (see table 2.1). However, in practice Christian Democratic women’s chances of being nominated, and their chances vis-Г -vis CDU men of actually being elected, are less favorable than SPD and Green women’s odds. This occurs for two reasons. First, the CDU’s statutes require a lower percentage of women on the ballot than do the Greens’ and SPD’s: only one in every three slots must to go to a woman and interviews suggest that the earliest ballot position seen as a “woman’s” is the third slot, rather than the first or second as in the parties with (near) parity quotas. Moreover, although the quorum does not specify any upper bound to the percentage of women appearing on the ballot, interviews indicate that CDU members believe the quorum was not designed to give more than 33 percent of list positions to female candidates. This creates a narrower gap (7 percent) between the percentage of female CDU members and the percentage of “women’s” ballot slots than exists in either the SPD or the Greens.

Second, the Christian Democrats’ bylaws allow considerable wiggle room to avoid selecting even one-third female candidates. The CDU’s form of affirmative action, adopted in 1996, was deliberately termed a “quorum” rather than a “quota” after a party congress voted down a stricter quota measure the year before (Wiliarty 2010, 152–59). The weaker quorum, pushed through by the party leadership against considerable rank-and-file resistance, calls for one woman among every three ballot slots. If this condition is not met in the first list drawn up during a nominating meeting, the list is invalid, additional names of female candidates must be sought, and a second list voted on; however, if this second list is approved by a majority of members present (female and male), it is valid regardless of whether or not the quorum is fulfilled (DavidsonSchmich and KГјrschner 2011, 29). Given that women make up a minority within the CDU, gender-imbalanced lists can easily be passed even against united female opposition;Page 199 → this situation stands in stark contrast to the Green’s and SPD’s more stringent regulations. Indeed, the quorum’s effect is so minimal that 11 percent of Christian Democrats responding to the Candidate Interest Survey erroneously believed their party had no affirmative action policy at all; in comparison, less than 1 percent of Greens and only 2 percent of Social Democrats surveyed were ignorant of their party’s quota. When asked how their party selected candidates, interviewees from the CDU at times failed to mention to quorum or wrongly claimed it did not apply to certain nominations; this was not the case in interviews with members of the Greens and the SPD. Other CDU interviewees noted the weak nature of their party’s affirmative action policy. As demonstrated in chapter 3, female eligibles in the CDU are less likely than their male counterparts to be willing to accept a local-level ballot nomination. Moreover, there are fewer women in the CDU’s rank and file than required by the quorum—especially in thinly populated rural areas. Because the quorum can easily be circumvented, however, rather than taking additional steps to locate or encourage promising women as occurs when more stringent quotas are in place, Christian Democratic gatekeepers can simply fall back on the more ambitious men in their party, arguing that there are no suitable female candidates (see also Holtkamp, Wiechmann, and Schnittke 2009; Eder and Fortin-Rittberger 2013, 12). Chapter 4 found the CDU less likely to engage in the formal training and mentoring programs for women that the Greens and the SPD have adopted. In sum, while the Christian Democrats’ quorum does direct the CDU to select female candidates from a maledominated pool, the percentage of woman candidates is not as high as in the other parties with quotas, the ballot slots reserved are not as promising, and the pressure to select women is not as strong as within the parties with higher, more binding quotas for women. Efforts to groom women for elective office have been less intense. Because the quorum obligates the CDU to place women in the third, rather than first or second, intervals on the ballot, if they do manage to secure their party’s nomination, female candidates may routinely receive less auspicious list places than their male counterparts and for every woman near the top of the list, two men will be nominated. As a result of these political opportunity structures, I expect that male CDU eligibles will be more likely than their female counterparts to pursue their party’s nomination for local office, to be selected as a candidate, and to win at this level. Just as with the Greens and the SPD, the situation is expected to be similar, but less acute, for high levels of elective office. Both Christian Democratic men and women face more competition for Page 200 →Germany’s top elected bodies, and federal and European electoral districts are more likely to include urban areas where more female party members can be found. Moreover, at this level sex is not an independent predictor of willingness to accept a ballot nomination. It is also more difficult for CDU gatekeepers to violate the quorum for such high-profile offices as there is far more publicity at nominating conventions and quota watchdogs are more focused on top slots than lower ones (Eder and Fortin-Rittberger 2013, 15). Because the chance of locating female aspirants is greater at this level, men’s political opportunity structures are less good and, while they are still expected to be more likely than women to step forward, be nominated, and win, I hypothesize that their advantages are not as great for top positions as they are at the local level. No Quotas In the CSU and FDP, because sex is not a criteria for selecting candidates, significant differences in male and

female eligibles’ experiences are not expected. While women in the candidate pool have been found to face poorer odds than their male counterparts in countries where quotas are not used by any parties, the situation in Germany, where some parties do employ affirmative action, should be different. In parties without quotas, female party leaders (i.e., female eligibles) are few and far between since there are few female members and the CSU and FDP do not make systematic attempts to promote women within their organizations’ ranks. Nonetheless, as established in the previous chapter, other parties’ use of gender quotas pressures quota-less parties to appear female-friendly; as a result, the latter have an incentive to nominate at least token female candidates from among their few women officers. As detailed in chapter 4, gatekeepers in parties not employing affirmative action are equally likely to ask their male and female members to run for elective office. Thus, quotas are expected to have spillover effects, reducing discrimination against women seeking a ballot nomination in other parties as well (for comparative evidence on quota diffusion see Matland and Studlar 1996; Kittilson 2006; Thames and Williams 2013). Moreover, as detailed in chapter 3, the women in parties without quotas who do manage to acquire inner-party office have prevailed in a very male-dominated environment; they are politically ambitious individuals who have fought for the qualifications that will enable them to compete for elective office. They are as likely as men to say they are willing to accept Page 201 →a local-level or high-level ballot nomination (see chapter 4). Women in parties without quotas were also on average more confident about their qualifications to run for elective office than either their male counterparts or women from parties without quotas. As a result there is no reason to expect that female eligibles in parities without quotas will be any less likely than their male colleagues to step forward and seek a ballot nomination for either local-level or high-level positions. It is important to note, however, that there are very few female eligibles in these organizations. As outlined in chapter 1, local races in Germany—with the exception of some mayoral contests—are conducted using a form of proportional representation. Under these rules, interviewees report, parties generally present electoral lists longer than the number of seats they expect to win, sometimes considerably so. This occurs for several reasons including trying to appear confident of a strong electoral performance, attracting voters from the community who would turn out to vote for their friends and acquaintances, offering voters a wide choice of candidates, and rewarding active party members with public recognition. Similarly, many state legislatures, half of the Bundestag, and the European Parliament are all elected through variations of proportional representation. As a result, most candidates seeking a nomination for both local-level and high-level political offices can be accommodated somewhere on the party list. Thus, even if they are not required by a quota to be included among the top list candidates, most ambitious women in parties without quotas can at least be placed somewhere on the ballot. In fact, including women in inauspicious slots on the ticket is a low-cost method for parties such as the FDP and CSU to appear women-friendly to voters. Token female candidates can be placed on publicity materials and attend campaign events, demonstrating the party’s (ostensible) commitment to women’s descriptive representation (Schulte 2013; KГјrschner 2009, 142–43). Hence I do not expect to observe significant gender differences in the experiences of male and female party leaders in parties without quotas in terms of receiving the nomination to run for political office, either at the local level or for higher offices. Because quota-less parties are not obligated to put women in specific ballot positions, however, I expect that women in the FDP and CSU should enjoy no systematic advantages vis-Г -vis their male colleagues (as women in parties with parity quotas do) nor any regularized disadvantages (as women in the CDU do). Instead, women in these organizations are expected to be as (un)likely as their male colleagues to be elected to local- and high-level Page 202 →public offices. In sum, while German quotas are expected to create different, gendered experiences in seeking ballot nominations, being selected as candidates, and being elected, this is not expected to be the case for quota-less parties.

Empirical Evidence: Quotas, Candidates, and Elected Officials I now investigate these expectations empirically using data from the Candidate Interest Survey. The CIS asked

respondents whether, for a particular elective office, they had ever (1) sought their party’s nomination, (2) received the nomination and had become a candidate for that level of government, and (3) been elected. Here I aggregate responses for all substate offices including city or town council, mayor, county council, county mayor, or regional council and refer to these posts as local offices.4 A respondent who had taken one of these steps for any of these positions was considered to have sought, received a nomination for, or been elected to local-level office, respectively. Similarly, if a person surveyed had pursued one of these actions for her state legislature, the Bundestag, or for the European Parliament, she was considered to have pursued high-level office. Here I discuss steps 1–3 in turn, comparing the responses of women and men in parties with the above range of affirmative action rules. Before presenting the results, however, it is important to note that this survey is biased against finding any influence of quotas on women’s and men’s experiences. The quotas adopted in Germany apply only to the proportional representation portions of the ballot; any seats contested using plurality are not subject to quotas and parties with quotas often fail to nominate women for directly elected posts (Eder and Fortin-Rittberger 2013; Deutscher StГ¤dtetag 2012). The CIS did not ask respondents to distinguish whether they had sought, received, or won nominations via PR or plurality, however, meaning that at least some of the Green, Social Democratic, or Christian Democratic women reporting their experiences here will not have been (dis)advantaged by a quota, muting any effect the latter may have. In addition, some of the older respondents in parties employing quotas may have given answers referring to experiences they had prior to quota adoption. Despite this difficult research design, I find support for the above hypotheses. As expected, women and men in parties utilizing candidate gender quotas had significantly different experiences running for elective office at the substate level, while men and women in quota-less parties did not exhibit such marked divergence. Gender differences in terms of campaigningPage 203 → for high-level political office followed the hypothesized patterns, but did not always reach conventional levels of statistical significance. Fig. 5.1. Eligibles Ever Seeking a Nomination for Local-Level Elective Office. (Source: Candidate Interest Survey [N = 445]; * p ≤ 0.10.) Figure 5.1 conveys the intersection of partisanship and gender in terms of seeking a nomination to run for local level elective office. As expected, women in parties with (near) parity quotas for female candidates were significantly more likely than their male counterparts to have attempted to gain a ballot spot for a substate election; 76.9 percent of SPD and Green women had done so in comparison to 69.2 percent of men in these parties. As depicted in figure 5.2, larger percentages of women (19.0 percent) than men (15.8 percent) in these organizations had sought a state, federal, or European-level nomination as well. This difference of means did not achieve a conventional level of statistical significance. However, given the fact that women in these parties are on average less likely to aspire to high-level office than their male counterparts in the first place, the lack of meaningful gender difference in actually pursuing a nomination is a testament to the encouraging influence of a high quota for women. The weak quorum adopted by the Christian Democratic Union has also created significant gender differences in running for local-level elective office. As expected, in this party the men were significantly more likely than their female counterparts to have sought a spot on a substate ticket; 87 percent of the male CDU leaders surveyed had done so compared to only Page 204 →73.5 percent of the female CDU respondents. The pattern for high levels offices was similar, but less pronounced. While 24.6 percent of CDU men had attempted to appear on the ballot for the state, federal, or European legislatures, the percentage fell to 20.6 percent for women; this difference was not statistically significant, though. At the local level, where gender gaps in political ambition exist, the party’s systematic efforts to encourage female candidates are limited, and the quorum easily ignored, women are less likely than their male counterparts to step forward for elective bodies. At higher levels of office, where the candidate pool is larger and the quorum more difficult to violate, Christian Democratic men’s and women’s experiences are more similar. Fig. 5.2. Eligibles Ever Seeking a Nomination for High-Level Elective Office. (Source: Candidate Interest Survey [N = 445]; differences of means not significant.)

Finally, as predicted above, no statistically significant differences in the propensity for men and women from parties without quotas were observed in terms of running for either high- or low-level elective office. Seventyeight percent of both male and female eligibles in the CSU and FDP had run for elective office. The percentage of women running for high-level posts (36.0 percent) was higher than that for men (32.7 percent), but this difference was not statistically significant. Notably, higher percentages of female eligibles in parties without quotas sought ballot nominations than in parties with either type of affirmative action measure. The women who rise to positions of inner-party leadership in the CSU and FDP are clearly Page 205 →politically ambitious; however, there are few such women present. In contrast, where quotas are employed the female eligibility pool is larger and more diverse in terms of political ambition. Fig. 5.3. Eligibles Ever Becoming Candidates for Local-Level Elective Office. (Source: Candidate Interest Survey [N = 444]; * p ≤ 0.10, ** p ≤ 0.05.) Figure 5.3 and figure 5.4 trace what transpired during the next phase of the candidate selection process. Here again, significant gender differences emerge at the local level in parties that have adopted candidate gender quotas; this variance is not seen in quota-less parties. Quota-driven gender differences were less stark compared to higher-level nominations, however. As hypothesized, women in parties with parity or near-parity gender quotas were significantly more likely than their male counterparts to have been selected as a candidate for local-level office; while 75.4 percent of the women surveyed had once appeared on a substate ballot, only 68.3 percent of their male colleagues had gained an opportunity to run for a local-level post. The direction of differences was the same in terms of high-level elective office—17.6 percent of female eligibles had campaigned for high-level office compared to 13.3 percent of male eligibles—but the difference of means did not reach conventional levels of significance. As expected, in the party with the weakly enforceable one-third quorum, men and women had significantly different experiences being selected as candidates for low-level offices, but the direction of the difference was the opposite than in parties with (near) parity quotas. Eighty-seven percent of the CDU men surveyed, but only 70.5 percent of the female CDU respondents,Page 206 → had been granted a place on a local ballot. Although roughly equal percentages of men and women had sought high-level elected office, a higher percentage of men (21.8 percent) than women (14.7 percent) had appeared on the ticket for such posts, not surprising given that only onethird of the ballot slots at this level are considered “women’s”; this gender variance did not reach conventional levels of statistical significance, however. Fig. 5.4. Eligibles Ever Becoming Candidates for High-Level Elective Office. (Source: Candidate Interest Survey [N = 445]; * p ≤ 0.10.) As predicted, given the lack of explicit gendering of ballot slots in parties without quotas, no significant differences in men’s and women’s experiences gaining a local-level candidacy were observed in the CSU and FDP. While equal percentages of male and female eligibles had sought a nomination, a higher percent of men (76 percent) than women (70 percent) had appeared on the ballot for local-level elective office; however, this difference did not achieve statistical significance. Contrary to expectations, women from quota-less parties were significantly more likely than their male counterparts to have actually appeared on the ballot for high-level elective office; 30 percent of female CSU and FDP respondents had appeared on the ballot compared to only 21.8 percent of men. This finding may result from quota contagion—pressure on parties without quotas to place their few female eligibles somewhere on the ballot. Figures 5.5 and 5.6 depict the ultimate results of the gendered political opportunity structures created by candidate quotas. As hypothesized, parity or near-parity gender quotas combined with few women in the candidate pool create a situation advantageous to female office seekers. While Page 207 →60.8 percent of the Green and Social Democratic women surveyed had won elective office at the local level, only 47.5 percent of the men had. Contrary to expectations, women were also significantly more likely than men to win higher level office. The ratio was almost 4:1: while 11.4 percent of the Green and Social Democratic women surveyed at been elected to a state, federal, or European legislature, only 3.3 percent of the men in these parties had won high office. Fig. 5.5. Eligibles Ever Elected to Local-Level Office. (Source: Candidate Interest Survey [N = 444]; ** p ≤

Conversely, 77.8 percent of the men from the party with the weak quorum had won a substate office compared with only 58.8 percent of the women in the CDU sample. The direction of differences was the same for state, federal, and European-level parliaments—20 percent of CDU men and 11.8 percent of Christian Democratic women had obtained a high-level parliamentary seat—although the difference of means was not significant at conventional levels. As expected, in quota-less parties, no significant gender differences in the propensity of men and women to win elective office emerged, either at the local or high level. Forty percent of FDP and CSU men had been elected at the local level and the figure for women was quite similar: 43.1 percent. In the CIS sample, more women had actually won high-level office than men: 2 percent had received a Landtag, Bundestag, or European Parliament seat compared to none of the men. However, this difference of means was not significant at conventional levels. Because CSU and Page 208 →FDP women were significantly more likely than men to have run for these offices, it suggests that they received poorer ballot placement than their male peers.5 Fig. 5.6. Eligibles Ever Elected to High-Level Office. (Source: Candidate Interest Survey [N = 445]; †p ≤ 0.20, *** p ≤ 0.01.) These results indicate that gender quotas create significant within-party gender differences in terms of seeking and receiving local-level ballot nominations and winning substate office. Parity and near-parity quotas encourage women to run for local elective office, ensure their quest for a ballot nomination will be successful, and make them more likely than their male colleagues to be elected to community office. The weak quorum studied here, in contrast, systematically disadvantages local-level female aspirants who are less likely than their male counterparts to seek or receive a place on the local electoral list and to eventually be elected to local-level office. Although the strong competition for high-level ballot nominations reduces the intensity of these patterns in terms of seeking and gaining ballot nominations for top political posts, the direction of difference remains the same. At the highest political levels, women in parties with (near) parity quotas continue to be more likely to win elective office than their male counterparts. The weaker quorum employed by the Christian Democrats offers women no such advantage. While women in parties without quotas are more likely than their male counterparts to appear on the ticket for top positions, they are not more likely to win these (or other) elections, nor are there gender differences in pursuing and gaining local-level ballot nominations among members of these parties. This lack of gender differencePage 209 → is salient as well, however, and suggests a process of quota diffusion; elsewhere where quotas are not used, women are significantly less likely than their male counterparts to be selected as candidates and then elected (Evans 2008; Lawless 2012; Lawless and Fox 2005, 2010; Ashe 2015).

Explaining Gender Variance in Candidacies Local-Level Office These bivariate comparisons also hold at the local level controlling for some of the major variables that can influence a candidate’s electoral success (see tables 5.1–5.3). Here the (categorical) dependent variables are ever having run, been selected as a candidate for, or been elected to any local-level political office. In order to most clearly depict the fates of male and female aspirants under various affirmative action rules, I ran separate models for parties with (near) parity quotas, the 33 percent quorum, and no quotas at all. In addition to sex, I controlled for a respondent’s age as a proxy for her candidate quality; because long-term service to the party is often rewarded with a promising ballot spot in Germany, older candidates are likely to have better odds of being nominated and winning than younger ones who more recently joined the party. Further, older aspirants will have had more occasions to contest elections than younger ones. I also controlled for candidates’ perceptions of their party’s strength in their electoral district on the assumption that candidates running in a district where their party is highly popular enjoy a better chance of election than those contesting a seat where their party is weak.6 The model also included a measure of the population of a respondent’s community along a seven-point scale (less than 20,000 inhabitants through over 500,000). There is considerable variation in the relationship between the number of local council seats available and the local

population. In smaller communities, aspirants enjoy better ratios of available seats to citizens, rendering it more likely that an ambitious party member living in a small town will be able to obtain a ballot nomination and win a seat than one residing in a large city. I also controlled for a respondent’s state of residence because some German LГ¤nder have better ratios of available offices to citizens than others and because there are some state-tostate variations in electoral systems. Finally, I controlled for party size. In Germany’s dual-ballot system, small parties such as the Greens and Page 210 →the Free Democrats rarely, if ever, win direct mandates via plurality, instead gaining all their Bundestag seats via the proportional representation portion of the ballot.7 The models presented here explain only a low portion of the variance in aspirants’ fates; however, they also confront a very difficult task because they attempt to predict whether an aspirant had ever, over the course of a lifetime, sought and obtained one of several local offices in five different states with an array of political opportunity structures. Any variable uncovered to have a significant effect below exerts a very strong effect that would likely be even more pronounced in explaining the outcomes of a more precisely defined electoral contest. Even under these circumstances, (near) parity quotas are found to exert a significant positive impact on female party leaders’ chances of electoral success at the local-level (see tables 5.1 and 5.2). As expected, women in the SPD and the Greens were more likely than their male counterparts to seek a ballot nomination, appear on the ballot, and be elected. Holding all other independent variables at their means, female eligibles in parties employing near-parity quotas are 8.5 percent more likely than their male counterparts to throw their hat into the ring for a city, county, or district-level contest, Page 211 →7.3 percent more likely than men in the SPD and the Greens to appear on the ballot, and 14.1 percent more likely to actually be elected to local-level office. These findings are particularly striking given that the CIS did not distinguish between directly elected seats (for which quotas are not employed) and positions elected via party lists (for which quotas do apply). The extent of the SPD and Green quotas’ effect is similar to a respondent moving to a smaller community where there are fewer inhabitants contesting a similar number of seats. Women’s chances of being elected vis-Г -vis men’s improve more than that of a person moving from a swing electoral district to one dominated by her party. Table 5.1. Predictors of an Eligible in a Party with a (Near) Parity Quota Pursuing Local-Level Elective Office: Logistic Regression Results Model 1: Model 2: Model 3: Sought Nomination Became Candidate Elected Coefficient Coefficient Coefficient Variable (Standard Error) (Standard Error) (Standard Error) Sex (1 = Female) 0.50 (0.32)* 0.41 (0.32)†0.61 (0.28)** District Population в€’0.35 (0.11)*** в€’0.36 (0.11)*** в€’0.44 (0.11)*** Age 0.02 (0.01) †0.02 (0.01)* 0.03 (0.01)*** Party Strength в€’0.11 (0.32) в€’0.20 (0.32) 0.52 (0.28)* Small Party 0.37 (0.35) State: Baden- WГјrttemberg в€’0.06 (0.65) State: Bremen 0.18 (0.73) State: Hamburg в€’0.30 (0.72) State: Nordrhein- Westfalen в€’0.81 (0.66)* Constant 1.77 (0.95)* N 243 0.141 Pseudo R2 0.000 Prob > chi2

0.29 (0.34) 0.31 (0.63) 0.54 (0.70) 0.17 (0.70) в€’0.43 (0.53) 1.44 (0.91)* 243 0.137 0.000

Source: Candidate Interest Survey. †p < 0.20 * p ≤ 0.1 ** p ≤ 0.05 *** p ≤ 0.01

0.33 (0.30) 0.14 (0.44) 1.96 (0.71)*** 1.99 (0.72)*** 1.26 (0.48)*** в€’1.5 (0.81)* 243 0.103 0.000

As hypothesized, the chances of an aspirant having sought or received a candidacy, or ever having been elected, increased with age. The control for party size was not significant nor did most state-level controls significantly effect the chances of an aspirant seeking and receiving a ballot nomination; Green and SPD members were significantly more likely to have ever been elected in Bremen, Hamburg, and Nordrhein-Westfalen, however. CDU women fared quite differently, as anticipated (see tables 5.2 and 5.3). Here female eligibles were significantly less likely than their male counterparts to have ever sought a local-level ballot nomination, become a candidate, or been elected in their communities. The relative impact of sex was even greater here than where more stringent affirmative action rules were Page 212 →used. CDU women were 13.6 percent less likely than Christian Democratic men to have sought a local-level nomination, 16.4 percent less likely to have actually been a candidate, and 25.4 percent less likely to have ever won election at the local level. Just as in parties with (near) parity quotas, the magnitude of the effect of an aspirant’s sex was roughly akin to moving from a swing district to a precinct where a party was popular. However, in this case it was male aspirants who enjoyed the electoral boost, not women. Given that the quorum is interpreted as calling for two higher-placed men for every one woman on a CDU party list, these results are not surprising. Table 5.2. The Relative Impact of Sex on Pursuing Local-Level Elective Office (Near) Parity Quota 33% Quorum Sought Became Candidate Elected Sought Nomination Became Candidate Elected Nomination +8.5% +7.3% +14.1% в€’13.6% в€’16.4% в€’25.4%

Sex (Change Male в€’ Female) Population в€’5.9% в€’6.5% of Community (1 unit change, 7point scale) Age +0.3% +0.4% (1 standard (SD = 11.5 deviation years) change) Party not significant not significant Strength (0 to 1 change)

в€’10.9%not significant

в€’8.8%

not significant

+0.8%

+0.6% (SD = 12.8 years)

+0.7%

+0.2%

+12.8%

not significant

not significant

+24.2%

Source: Candidate Interest Survey. Note: Results based on Models 1–3 in tables 5.1 and 5.3, using Stata command “prchange” varying sex and holding all other variables at their means. Older CDU eligibles were more likely to have ever run for office than younger ones. Christian Democrats in large cities were less likely to have been become a candidate than those in smaller communities with relatively fewer aspirants and more council seats. The model does not contain a control for party size or a fourth Land because the CDU, a large party, was the only party included here and it is not present in the state of Bayern. In the small, Leftdominated city-states of Bremen and Hamburg, CDU candidates were less likely to come forward or be nominated than in other states, and local-level CDU candidates were less likely to have been elected in Baden-Württemberg than in other Länder.

Finally, tables 5.2 and 5.4 present the results for the quota-less CSU and FDP. In these parties, where sex was not a criteria when drawing up electoral lists, male and female eligibles were equally likely to seek a ballot Page 213 →nomination, receive one, and to ultimately be elected. However, female aspirants in these parties are few in number. Instead, age was a significant predictor of winning local-level office, as expected in a system of political recruitment where gatekeepers reward those having completed an Ochsentour. State-level political opportunity structures also proved significant. Table 5.3. Predictors of an Eligible in the Party with a 33% Quorum Pursuing Local-Level Elective Office: Logistic Regression Results Model 1: Model 2: Model 3: Sought Nomination Became Candidate Elected Coefficient Coefficient Coefficient Variable (Standard Error) (Standard Error) (Standard Error) Sex (1 = Female) в€’1.14 (0.64)* в€’1.36 (0.65)** в€’1.27 (0.56)** District Population в€’0.64 (0.51) в€’0.83 (0.53)* 0.10 (0.21) Age 0.06 (0.03)** 0.06 (0.03)*** 0.06 (0.02)*** Party Strength 0.33 (0.62) 0.29 (0.62) 1.00 (0.51)** State: Baden-WГјrttemberg 0.61 (1.14) 1.3 (1.0) в€’1.17 (0.83)* State: Bremen в€’4.14 (2.85)* в€’4.6 (3.0)* в€’0.42 (1.10) State: Hamburg в€’4.00 (2.84)* в€’4.6 (3.0)* в€’0.76 (1.1) Constant в€’1.96 (2.12) в€’3.6 (2.12)* в€’2.68 (1.76)* N 88 87 87 0.174 0.208 0.170 Pseudo R2 0.042 0.013 0.012 Prob > chi2 Source: Candidate Interest Survey. †p < 0.20, * p ≤ 0.1, ** p ≤ 0.05, *** p ≤ 0.01. In sum, even when controlling for a range of variables known to shape candidate emergence, male and female eligibles have significantly different experiences becoming local-level candidates when gender quotas are employed. As hoped by quota proponents, when parity or near-parity rules for selecting local-level candidates are employed, women were very likely—more likely than their male counterparts—to participate in the democratic process. Rather than being less likely to appear on the ballot, as in countries where quotas are not present, women were more likely than men to be selected as candidates; moreover, placement mandates ensure that women appear in winning ballot slots and they are more likely than their male counterparts to ultimately be elected at the local level. These results are driven by the facts that women are less likely to join a party than men and that female eligibles are less likely to aspire to local-level elective office than their male counterparts. As a result, while enough women can be found to occupy top ballot slots, women are underrepresented lower on Page 214 →electoral lists, meaning that they are less likely to lose elections than their male counterparts. Table 5.4. Predictors of an Eligible in a Party with No Quota Pursuing Local-Level Elective Office: Logistic Regression Results Model 1: Model 2: Model 3: Sought Nomination Became Candidate Elected Coefficient Coefficient Coefficient Variable (Standard Error) (Standard Error) (Standard Error) Sex (1 = Female) в€’0.20 (0.60) в€’0.55 (0.55) 0.56 (0.52)

0.09 (0.20)

в€’0.05 (0.17)

Age 0.01 (0.02) Party Strength 0.18 (0.55) Small Party 0.16 (1.13) State: Baden-WГјrttemberg 1.16 (1.32)

0.02 (0.02) 0.35 (0.52) 0.94 (01.03) 0.28 (1.02)

0.05 (0.02)** 0.21 (0.54) в€’0.76 (0.84) в€’0.54 (0.75)

State: Bremen 0.39 (1.53) State: Hamburg в€’2.01 (1.21)* State: Nordrhein-Westfalen в€’0.93 (0.93)

в€’0.27 (1.56) 3.12 (1.46)** в€’3.06 (1.25)*** в€’2.4 (1.4)* в€’ 0.93 (0.93) 0.81 (0.72)

Constant N

1.35 (1.60) 104

в€’0.16 (1.47) 104

в€’2.3 (1.6)* 104

0.200 0.013

0.170 0.019

0.233 0.000

District Population

Pseudo R2 Prob > chi2

0.07 (0.21)

Source: Candidate Interest Survey. †p < 0.20, * p ≤ 0.1, ** p ≤ 0.05, *** p ≤ 0.01. The CDU’s 33 percent quorum, in contrast, hinders rather than helps women’s chances vis-Г -vis their male counterparts at the local level. Given that the quorum is widely interpreted as—if enforced at all—granting men two higher-ranked local-level ballot slots for every one a woman receives, women were less likely than men to come forward as candidates, to be selected for the ballot, and to have won local-level elective office. Finally, where no affirmative action rules were in effect, no gender differences were observed in eligibles’ propensity to seek or receive a local-level ballot nomination or to be elected in their communities. High-Level Office The situation in terms of seeking high-level elective office was somewhat different. Here, once the abovementioned controls are accounted for, women and men in the Greens and the SPD were equally likely to come forward as candidates and be selected to appear on the ballot for positions at the state, federal, and European levels. Similarly, once the above-mentioned controls are accounted for, male and female CDU party leaders were also equally likely to come forward as candidates and be selected to appear on the ballot for positions at the state, federal, and European levels. And, as at the local level, in parties where candidate quotas were not deployed, sex played no significant role in predicting whether an eligible would seek a nomination, be nominated, or win elective office at any level of government. These conclusions were reached by running models (not shown here) identical to the first two in tables 5.1, 5.3, and 5.4 for high-level elective offices. In this analysis, the coefficient for sex was not significant in any of the models. Only 36 CIS respondents had ever actually been elected to high-level office, illustrating the sharp competition aspirants desiring a state legislative, Bundestag, or European Parliament seat face. These low numbers render meaningful multivariate analysis of eligibles’ determinants of winning high-level office impossible here; a larger sample size would be needed for such an analysis. However, a German-language study of all Bundestag candidates between the founding of the Federal Republic and 2009 obtained results similar to those presented here for local-level office (Manow and Flemming 2012; see also Bieber 2013).8 Because the research here indicates that male and female eligibles, regardless of quota type, have equal chances of becoming candidates for the national parliament, Manow Page 215 →and Flemming’s conclusions about male and female candidates’ odds of election likely represent male and female eligibles’ chances as well. Since 1998 the percentage of women elected to the Bundestag has consistently exceeded the percentage of all candidates who were female, indicating that women are more likely than men to receive auspicious ballot placement. In addition, female Bundestag candidates are more likely than men to win office the first time they attempt to do so. Manow and Flemming demonstrate that these results are driven by the behavior of parties with parity and near-

parity quotas (see also Fortin-Rittberger and Eder 2013). Prior to quota adoption, SPD women had, on average, poorer Bundestag list placement than their male colleagues; when the party phased in a 33 percent quota for women, men’s and women’s list placement equalized, and since the adoption of the minimum 40 percent quota for women, SPD women have on average had higher list placement than their male colleagues. The authors also find Green women more likely than their male counterparts to receive a promising double nomination for the Bundestag (a direct nomination in a constituency backed up with a secure list place); the situation in the quota-less FDP is the reverse and the authors note that the FDP has the fewest female Bundestag candidates of all parties. The authors do not distinguish between the CDU and CSU but the joint results indicate that the introduction of the quorum has done little to change the fate of CDU/CSU women; they have consistently received worse list places than their male colleagues, both before and after the introduction of the quorum. The authors do depict the propensity of CDU and CSU candidates to receive coveted double nominations; here the negative effects of the quorum on women can indeed be seen. While roughly equal percentages of CSU male and female candidates received double nominations, 20 percent fewer CDU women received such secure ballot placement than CDU men. In short, the CIS’s findings at the local level are paralleled by those of other scholars working at the national level.

Implications: Have Quotas’ Goals Been Fully Met? This chapter indicates that quotas exert a direct impact in the final stages of the political recruitment process, shaping whether female aspirants seek ballot nominations, become candidates, and whether they are subsequently elected. Voluntary party quotas shape the chances of participating in the democratic process for the women who have overcome gendered hurdles to political participation in the recruitment environment and in informal Page 216 →recruiting structures. Women in parties employing enforceable parity or near-parity gender quotas enjoy advantages vis-Г -vis their male colleagues at each of these final stages of seeking a local-level elective office. CDU women, in contrast, are systematically disadvantaged by the easily circumventable 33 percent quorum at the local level. For higher levels of elective offices, these patterns repeat, but they are not as pronounced. In contrast to settings in which no parties use quotas, female eligibles in quota-less German parties have experiences similar to their male colleagues. Instead, they face odds similar to their male counterparts when seeking and receiving ballot nominations for elective offices at all levels. Moreover, men and women from these parties are equally likely to win office. Below I discuss each of these patterns and their normative implications for women’s descriptive, symbolic, and substantive representation. While the adoption of quotas in Germany failed to greatly increase the numbers of women who joined political parties, these affirmative action measures did increase the percentages of women acquiring leadership positions within political parties. As a result, even politically unambitious women have become eligible for elective office and are recruited by their parties to fill the many available “women’s” places on electoral lists. Against this backdrop, in parties employing (near) parity quotas, women are more likely than men to seek a ballot nomination, to receive one, and to be elected at the local level. On average, significantly higher percentages of female than male eligibles in the Greens and the SPD won election to their Landtag, the Bundestag, or the European Parliament. Thus, (near) parity quotas are inspiring female party leaders to step forward at least as, if not more, often than similarly placed men, even though they may not be as ambitious as men in the first place. Parity or near-parity quotas increase women’s political participation as intended and “justice” has been achieved in the later phases of political recruitment. Their effects do not stop here, however. Because female aspirants are just as, if not more, likely than their male counterparts to step forward for office, be nominated, and win, female citizens can be assured of a broad range of female representatives to represent their concerns. Although, as previous chapters have demonstrated, female aspirants are drawn from a far narrower subset of the population than male aspirants to begin with, the range is at least not narrowed further at this stage of the recruitment process. Quotas produce a number of female aspirants to contest high-level elective offices. Moreover, the Candidate Interest Survey found that the women and men who are elected from parties with (near) parity quotas express equal levels of progressive ambition (Schlesinger 1966, 10), or Page 217 →the desire to rise to a higher elective office than the one they currently occupy (see figure 5.7).9 Hence women from the Greens and SPD parties are unlikely to be derided as tokens tapped to provide an

“alibi” for a party’s commitment to women10 and, once elected, are just as likely as their male peers to desire to become career politicians. Finally, parity or near-parity candidate gender quotas not only increase women’s participation, but to the degree that quotas have significant symbolic effects, other female aspirants will be encouraged by the prior successes of their female colleagues who have contested elections. Fig. 5.7. Progressive Political Ambition. (Source: Candidate Interest Survey [N = 258]; *p ≤ 0.10, **p ≤ 0.05.) Such quotas have an additional important effect: they help level male and female aspirants’ odds of election in quota-less parties as well. While women have been found to be less likely than their male counterparts to receive a ballot nomination and ultimately be elected when quotas are not used, the Green, SPD, and CDU affirmative action rules have also succeeded in eliminating most gender differences in the FDP and CSU. When compared to male eligibles in their organizations, female eligibles in quota-less parties were just as likely to seek either local- or high-level ballot nominations, to become a candidate for city, county, or district elections, and to be elected at the local level. Women who were elected via parties without quotas were significantly more likely than their male counterparts to express progressive political ambition as well (see figure 5.7). A higher percentage of FDP and CSU women than men were nominated for high-level races. Thus other parties’ use of quotas has aided CSU and FDP female eligibles’ participation in the democratic process. However, it is important to underscore that there are few female eligiblesPage 218 → where quotas are not employed, keeping women less likely than their male counterparts to be present at this stage of political recruitment. Conservative women’s substantive representation may therefore be limited as well. Moreover, although FDP and CSU women were more likely than their male peers to appear on the ballot for high-level offices, women from these parties were not significantly more likely than men from their organizations to win such races, suggesting poorer ballot placement. To the extent symbolic effects occur, these results may indicate to FDP and CSU women that, at the local level, should they be able to obtain a leadership position, they face at least similar odds to their male counterparts. However, for high-level offices, it would appear that the symbolic message to female aspirants in quota-less parties is that, while the party is happy to have them on the ballot, it is unlikely to put them in a winning slot. This negative role model effect may further limit women’s inroads into quota-less parties. Finally, this chapter offers some lessons about the impact of a difficult-to-enforce 33 percent quorum. While the quorum was designed to aid women’s political participation in democracy, and while it has led to increases in the percentages of Christian Democratic women represented in elected bodies across Germany, the analysis presented here indicates that this type of affirmative action measure has had unintended consequences. Women in the Christian Democratic Union, who are allotted one, lower-ranked ballot slot for every two men receive, are less likely than men to run for office, be nominated, or win at the local level. The same pattern was found, albeit without conventional levels of statistical significance, in elections to Germany’s most powerful political bodies. As a result, the quorum fails to achieve what Mansbridge refers to as “justice” or women’s equal role in democracy. Rather than increasing, or at least holding steady, the substantive choice of representatives open to female citizens, the quorum further constricts the already narrow range of potential representatives available in Germany.11 Moreover, to the degree that quotas exert a symbolic impact, they will have a negative impact on CDU women who can see their female counterparts’ lack of electoral success. The results here are driven by the fact that few women enter parties with quotas, especially in thinly populated areas, and female eligibles in these parties are less prone than men to develop political aspirations. As a result, the existing female aspirants in the Greens and SPD have excellent odds and the CDU gatekeepers are more prone to ignore their quorum. The concluding chapter now turns to policy recommendations to rectify this situation.

Page 219 →

Conclusion A Glass Half Full Quotas’ impact differs across the various stages of the political recruitment process depending upon whether the influence of formal recruitment structures—the candidate selection rules that quotas directly change—outweigh the influence of informal recruitment structures and the recruitment environment—gendered societal norms and patterns of interaction. These informal norms matter the most early on in political recruitment, hindering women from joining parties and limiting their subsequent propensity to become aspirants. Quotas’ biggest impact comes later in the recruitment process, when gatekeepers select inner-party officers and candidates for elective office. This concluding chapter briefly summarizes my findings and draws out policy recommendations, discusses my methodological contributions, and finally turns to avenues for future research.

A Glass Half Full Voluntary gender quotas in Germany, especially the difficult-to-violate (near) parity quotas adopted by the Greens and the SPD, have indeed succeeded in improving women’s descriptive representation as well as women’s participation in many phases of the political process. This rise in political participation in turn offers female citizens greater possibilities of being substantively represented today than they had in the prequota era. Finally, some German women and girls do enjoy positive role models of successful Page 220 →female politicians. Below I summarize quotas’ accomplishments in terms of descriptive, substantive, and symbolic representation, respectively. Descriptive Representation Political party quotas in Germany have markedly improved women’s descriptive representation since the Greens first adopted affirmative action in the mid-1980s. The percentage of elected positions held by women has risen at every level of government and, at all levels except the European Parliament, parties employing quotas exhibit higher percentages of female MPs than quota-less organizations do; Germany’s EP delegation contains equally high numbers of female MEPs from parties with and without affirmative action policies. Quotas have exerted diffusion effects, increasing the numbers of women that parties without quotas send to state legislatures and the Bundestag as well. Women’s participation has improved not only in this final stage of the political recruitment process but also in the phases immediately preceding election. In political parties featuring (near) parity quotas, female eligibles are more likely than their male counterparts to have sought and received a ballot nomination for local-level elected offices and are more likely than their male counterparts to have actually been elected to communal bodies. Women and men in such parties are equally likely to seek and receive ballot nominations for parliaments at the state, federal, and European levels. Some indicators suggest female aspirants in parties with (near) parity quotas are even more likely than their male peers to win seats in these bodies. Here too, quotas have exerted a diffusion effect and women and men in parties not employing quotas are equally likely to throw their hats into the ring for elected positions at all levels of government, equally likely to appear on the ballot for these posts, and equally likely to be elected. Female aspirants have benefited from a quota-driven change in party gatekeepers’ actions. Anecdotal evidence suggests that, in the prequota era, women were rarely asked by their parties to run for elective offices, especially for top positions. The data presented here paint a very different picture. In parties employing quotas, gatekeepers are now more likely to ask female eligibles than to ask male eligibles to run for elective office. Moreover, especially in parties with binding, parity or near-parity quotas, gatekeepers have begun to reevaluate the selection criteria used to select candidates. While party loyalty remains a priority for German gatekeepers, rather than requiring a long, uninterrupted party career as evidence of fealty, party leaders have begun to accept

caregiving-related breaks in a Page 221 →woman’s track record of service to her organization. Gatekeepers in such parties have also broadened their search for candidates beyond traditional sources (e.g., business groups or male sports teams) into social niches more often occupied by women (e.g., parent-teacher associations or charity organizations). Finally, all parties with affirmative action measures, but especially parties with binding (near) parity quotas, have begun systematic efforts to train and mentor the women within their ranks in order to develop the pool of qualified female aspirants needed to fulfill their quotas. Where quotas are not employed, gatekeepers have not engaged in any of the above candidate-development activities in a pronounced manner. However, in sharp contrast to both the prequota era in the Federal Republic and to the United States, German gatekeepers in quota-less parties are now equally likely to ask their male and female eligibles to run for elective office. As parties employing affirmative action successfully contested elections with highly placed female candidates, gatekeepers in quota-less organizations learned that women could indeed be viable candidates. Moreover, German quotas have created watchdogs in the press, NGOS, women’s policy agencies, and within quota-less parties themselves, who monitor not only whether parties with quotas comply with their own regulations but also track other parties’ propensity to contest elections with multiple women in top ballot slots. Party organizations failing to do so are now publicly “named and shamed,” creating incentives for even quota-less parties to approach female eligibles about running for office. Where affirmative action is employed, party gatekeepers have numerous female eligibles whom they can approach about running for office because German quotas apply not only to ballot positions but also to inner-party offices. Quotas exert an “elevator effect,” raising the limited number of women among rank-and-file party members to (almost) half of those eligible to run for elective office by virtue of their leadership positions within the organization. Thus one of quotas’ main achievements has been increasing women’s participation in positions that render them eligible to become elected officials. The closer the quota is to parity, the more female eligibles are present in this early phase of the political recruitment process. In parties not employing quotas, very few women rise to positions of inner-party leadership, perpetuating the traditional, extremely male-dominated eligibility pool for elective office. The quota-induced rise in female eligibles is also important because it places women in situations conducive to developing political ambition. The results presented here clearly indicate that while fewer female eligibles than male eligibles have always wanted to be politicians, serving Page 222 →in an inner-party leadership post renders women as prone as their male counterparts to consider running for local-level elective office. Women also become as likely as similarly placed men to be willing to accept a candidacy for the state, federal, or European parliaments. The experiences and skills gained while holding inner-party office awakens party members’ interest in running for office and helps them develop the skills and confidence needed to do so. Again, the higher the quota for inner-party office, the more women who become aspirants. Where quotas are not used, female party members are often overlooked in leadership selection and hence never ascend to ambition-inducing positions of responsibility. In short, voluntary gender quotas in Germany have gone far in changing gatekeeper behavior and elevating women’s participation in many phases of the political recruitment process, improving the opportunities for women to take part in democracy. Substantive Representation The increased opportunities to become inner-party leaders that quotas provide women ensure a larger number of qualified candidates from which female representatives can be selected than when quotas are not used. These female eligibles, most of whom had not previously held political ambitions, are often asked to run for elective office by party leaders, are trained and mentored by their parties, and face favorable odds of winning; in turn, many become aspirants considering running for elected office. Gatekeepers in parties with (near) parity quotas also search for candidates in a wider range of nontraditional places than other parties, again diversifying the group from which candidates are selected, especially for top levels of elective office. As a result, quotas expand the aspirant pool. The more potential female candidates, the broader and more diverse a range of women’s viewpoints are

included in the democratic process. In short, quotas contribute to women’s substantive representation by generating more female eligibles from among whom representatives of women’s interests can be picked. Because parties with (near) parity quotas devote resources to developing suitable female candidates and because quotas create high numbers of female eligibles, the likelihood that “quota women” are considered less qualified or tokens, or both, declines. Binding quotas encourage the parties employing them to fulfill their democratic function of identifying and grooming a steady stream of able individuals—both male and female—to assume positions of political power and substantively represent their fellow citizens. Page 223 →Symbolic Representation Finally, voluntary party quotas in Germany appear to have had some, albeit limited, symbolic effect. They have indeed demonstrated that women can be viable candidates for elective office. The parties adopting quotas, first the Greens, then the Social Democrats, and finally the Christian Democratic Union, experienced electoral success after adopting affirmative action policies, pressuring nearby parties in the political spectrum to improve their images by including visible women on their party lists. Even parties without quotas now seek to promote women’s candidacies to avoid the wrath of quota watchdogs and the scorn of citizens. No one interviewed for this project expressed a belief that women were incapable of, or unsuited to, holding high political office. Virtually all interviewees concurred that female candidates were not an electoral liability with the vast majority of the electorate. The positive effects of candidate quotas on women’s political careers described above certainly will not have a negative role model effect or discourage any potential aspirants in Germany. The success of quotas for women has spawned discussion of quotas for racial and ethnic minorities as well, and Berlin’s branch of the SPD has now adopted a quota for migrants (Geissel 2013) and in 2015 Germany passed a national quota for women on corporate boards.

A Glass Remaining Half Empty Clearly, then, voluntary party quotas in Germany have met many of their original goals and can be considered a partial success, especially when comparing the experiences of women in parties with quotas to those of women in the prequota era or to those of women in parties not utilizing affirmative action. While progress has been made, however, not all of quotas’ hoped-for effects have materialized. When women’s experiences are compared not to women’s in an earlier era, or to situations where quotas are not employed today, but instead to men’s in contemporary political recruitment, the glass can also be seen as remaining half empty in terms of women’s descriptive, symbolic, and substantive representation. Descriptive Representation While German parties consistently fulfill their quotas for high-level elective bodies such as the European Parliament, the Bundestag, and Germany’sPage 224 → state legislatures, and while quota diffusion also has increased the numbers of MPs from parties not employing affirmative action measures, the situation in many of Germany’s smallest localities—the places with the highest total number of elected officials—differs considerably. The smaller the community, the less likely parties with quotas are to comply with their own selfimposed rules regarding female candidates. While the percentages of women holding community-level office have increased since the adoption of quotas, the increase has been far smaller than for top-level posts and diffusion effects to quota-less parties have not occurred. The percentage of women that quota-less parties send to elective offices in Germany’s smallest towns and villages has hardly budged since the mid-1980s. Quotas have thus been less successful in achieving justice, or women’s equal participation in democracy, at lower levels of government, and men’s overrepresentation is most pronounced at this level. In addition, although women’s representation has increased considerably at high levels of elective office, it has stagnated in these bodies at the level of the lowest affirmative action policy (approximately 33 percent) and the percentages of women in directly elected positions remains below even this level (Fortin-Rittberger and Eder 2013). One cause of the local-level gender imbalance can be found early on in the political recruitment process: despite over a generation’s use of quotas in Germany, female German citizens remain far less likely than their male

counterparts to join a political party in the first place—even parties with (near) parity quotas. While quota proponents often anticipate that quotas’ “role model” effects will inspire women and girls to become politically interested and take part in political life, Germany offers little empirical evidence to support these claims. Women make up a minority of party members in all German parties, including among new entrants and party youth wings. While this is true across Germany, women’s underrepresentation among party members is particularly pronounced in thinly populated areas of the country. Where quotas for inner-party leadership positions are not employed, or easily circumvented, this low percentage of female members translates into low percentages of female inner-party leaders, perpetuating men’s overrepresentation in the eligibility pool. The research presented here identifies two primary causes of quotas’ failure to attract women to German political parties. First, quotas cannot change the recruitment environment, or the gendered division of labor within society. Female citizens have far more household and care responsibilities than their male peers and, as a result, less time to devote to voluntary groups such as political parties. Responsibility for childcare and Page 225 →domestic tasks such as doing laundry, cleaning, and cooking reduces both male and female eligibles’ willingness to consider or accept a nomination to run for elected office, especially at the local level where the voluntary nature of elected positions would create a triple burden beyond uncompensated household labor and paid employment. Women more often assume these responsibilities than men, however, resulting in overall lower levels of female political ambition. Second, quotas have done little to change the informal character of party life, and political party organizations retain their masculinized ethos, especially in small communities. Male-dominated party groups tend to meet at times and in places not conducive to women’s participation; moreover, the substantive style and content of discourse at such gatherings often alienates even politically interested women with time to participate in public life. Not only are women less likely to join parties—even those with quotas—than men, but the women who emerge as party leaders, or eligibles, when quotas are in place are less likely than their male counterparts to be willing to accept a nomination to run for amateur local-level elected office. While gender does not make a difference in willingness to accept a ballot nomination for professional, high-level positions, an eligible’s assessment of his or her own qualifications exerts a significant impact on political ambition at this level. Those who do not believe they are qualified are unlikely to aspire to the Landtag, Bundestag, or European Parliament; in parties employing quotas for leadership posts and for candidates, women are less likely, on average, than their male peers to believe they are highly qualified to run for such positions, again dampening their political ambition. Here again, then, quotas’ symbolic effects are limited. The gendered recruitment environment in Germany accords masculinized life experiences higher status and prestige than feminized ones, making it easier for men to view themselves as competent. Moreover, quotas have not created a change in informal recruitment structures such as the type of self-promotion required of a successful candidate; here, too, political rituals are based on masculinized behavior at odds with what is considered appropriately feminine. The low percentages of female party members, especially in thinly settled areas, combined with lower levels of female political ambition for amateur posts, render it difficult for parties to fulfill their quotas in small communities, even decades after quota adoption. Finally, quotas have created significantly different political opportunity structures for men and women in the final phases of political recruitment. Men in parties employing (near) parity quotas face steeper odds of receiving a promising ballot nomination and being elected than women in their Page 226 →organizations and, as a result, they are on average less likely to step forward to run for elective office than their female peers. In contrast, the CDU’s weak 33 percent quorum gives women rather than men less of a chance of receiving a promising local ballot nomination and being elected at this level; as a result female CDU eligibles are less likely than their male counterparts to throw their hat into the ring for community office. The situation is similar, but less pronounced, for higher level parliaments. Thus while the quorum has increased the percentage of female Christian Democratic officials in Germany, it has not leveled the playing field for female aspirants. Lastly, while quotas have exhibited some diffusion effects, they have not improved the odds of women in quota-less parties ultimately being elected. While on average a higher percentage of female than male CSU and FDP eligibles appear on the ballot for highlevel posts, women from these parties are not more likely to win, indicating that these parties award male aspirants

better ballot slots than they do females. In sum, despite the improvements in women’s political participation in Germany following the Greens’ introduction of gender quotas, women and men still do not participate equally throughout the democratic process. Symbolic Representation Quotas’ limited success in achieving descriptive representation highlights flaws in some of the arguments made regarding symbolic representation. While quotas may create female role models—something not guaranteed given the sexist media coverage of female candidates (e.g., Murray 2010a)—and create a societal belief that women are capable of ruling, quotas cannot grant women the resources that they require, such as time to act on the political ambitions they may possess. The gendered division of domestic and care work in Germany, along with the persistence of masculinized political party routines, limit quotas’ symbolic effects. Moreover, for Christian Democratic women, who generally receive one lower-ranked ballot position (if any) for every two higher-up slots awarded men, the quorum may indeed teach female aspirants negative lessons rather than positive ones. Finally, to the extent that quotas do create role models, the women successful in politics (just as their male counterparts) are not typical of the underlying population. They tend to have higher levels of self-confidence in their abilities, be unencumbered by domestic responsibilities, and hail from politicized social networks. Eligibles are well-educated, ethnically German women from upper-income households, pursuing either careers Page 227 →in the public sector or self-employment. If they have children, the children are out of their first decade and they are likely to be raising them together with a spouse or partner. Moreover, while both the women and men who come forward to run for elected office in Germany are not typical of the underlying population, the women who do so are even less typical because they are unburdened by domestic responsibilities common to many women and highly self-confident in a context where many women doubt their abilities. Immigrant women, women with lower levels of educational attainment and household incomes, single mothers, and women pursing private sector careers are much less likely than other women to enjoy female role models in Germany. Substantive Representation Thus while voluntary party quotas in Germany have certainly been successful in increasing both women’s descriptive representation and the numbers of women’s voices heard in deliberative bodies over the past several decades, female representatives are today drawn from a smaller subset of citizens than male representatives are, as there are lower percentages of women in all parties, even those with (near) parity quotas. As a result, while they definitely are granted more choice in representatives than they were in the 1980s, Germans still do not enjoy the same numbers of potential female MPs as they do possible male MPs. While (near) parity quotas for inner-party office create gender balance in the eligibility pool, female eligibles are selected from a smaller number of underlying rank-and-file party members than are male party leaders. In addition, the CDU’s circumventable 33 percent quorum does not even go this far in qualifying women for elected office. Thus Christian Democratic women’s alternatives in potential representatives are even narrower than those of female Green, SPD, and Left Party adherents. As a result of the FDP and CSU’s lack of (local-level) quotas, libertarian women and conservative women from Bayern are offered the fewest options of all citizens. Similarly, because of the skewed nature of who joins political parties in Germany, while voluntary party quotas have increased women’s descriptive representation, the women who are elevated to positions of eligibility—just as the men—are likely to have had a narrower range of life experiences than the citizens they represent. They are not likely to have firsthand knowledge of the problems faced by immigrant women, single mothers, or women in the private sector challenged by globalization—especially those with low levels of education and skills. Page 228 →The amateur nature of local-level elected office in Germany further constrains the already-narrow range of potential women representatives. Because serving in a local-level council position comes in addition to pursuing paid employment, many women with domestic responsibilities are deterred from accepting ballot

nominations for the first steps on the German political career ladder. Thus men, already a majority among party members, become further overrepresented among aspirants for local-level offices. Because the CDU’s quorum can be bypassed, rather than search for ambitious women to fill their affirmative action measures, gatekeepers can simply fall back on more ambitious male eligibles. As a result, Christian Democratic women’s already limited choice of female candidates narrows further, as female eligibles are less likely than their male counterparts to seek a local-level ballot nomination. Thus, even several decades after the adoption of quotas in Germany and the corresponding rise in women’s descriptive representation, women remain underrepresented compared to men in earlier stages of the political recruitment process. This finding demonstrates the limits of quotas’ symbolic effects and has implications for the representation of women’s interests in German democracy. I now turn to policy suggestions to remedy these problems, both in the Federal Republic and elsewhere.

Policy Recommendations The results presented here offer some clear policy ideas for those seeking to improve upon voluntary party quotas. Before raising these points, however, it is important to stress that this book confirms that gender quotas for innerparty offices and ballot slots have achieved many successes. Those wishing to underscore women’s capacity to rule and desiring to improve women’s descriptive representation, participation in the policy process, and their chances of substantive representation should not abandon a commitment to quotas. Compared to the situation German women faced in the prequota era, and the conditions experienced by contemporary women in parties without quotas, women in today’s SPD and Green parties, and at times in the CDU, enjoy some advantages in the political recruitment process. The evidence here supports the utility of voluntary party quotas and should not be interpreted as a case against them. In fact, one of the main suggestions supported by this study is that political parties do adopt binding parity or nearparity quotas; in Germany, those interested in gender equality should call on the Free Democrats to Page 229 →begin employing affirmative action and on the CSU to expand its inner-party quota to elected positions at all levels.1 Quotas are most effective at changing gatekeeper behavior when they cannot be circumvented. “Zipper” style quotas have the advantage that they do not create “battles of the sexes” in which men and women competing for the same ballot spot may debate the merits of a candidate’s sex as a qualification for elective office. My research also cautions against the use of below-parity quotas as they create worse odds for female than male aspirants and, at least without specific placement mandates, gender the struggle for promising ballot positions. Raising the CDU’s one-third quorum to the 40 to 50 percent range, or adding a placement mandate specifying that women occupy the second, fifth, eighth, and so forth ballot slots, rather than the current assumption of multiples of three, would improve female Christian Democrats’ odds of becoming a candidate and winning elective office. Amending the CDU’s easily circumventable quorum to make it more difficult to violate would also help. This could be accomplished by requiring a majority of women present in nominating meetings approve any deviation of the quorum—as is done in the Greens—or by eliminating any escape clause at all—as the SPD’s statutes do. The evidence presented here indicates that high and binding quotas have prompted party gatekeepers to undertake more extensive efforts than gatekeepers in other parties to identify female candidates. These steps include redefining selection criteria to accommodate female life experiences while still tapping into the qualities needed to succeed in Germany’s political system as well as initiating systematic mentoring and training programs to develop female eligibles’ self-confidence and skills. In contrast, quota-less party simply fall back on the excuse that there are few available women to run for office. Were quota-less parties to adopt binding quotas, gatekeepers would gain greater incentives to expand the scope and seriousness of candidate identification and development programs. Even with difficult-to-violate parity quotas, however, additional steps are necessary to ensure gender-equal

participation in democracy in Germany and elsewhere. First, in order to give women more time to participate in civil society, the gendered division of domestic labor must become more equal, with men assuming a greater share of household tasks. Instead, other public policies practiced in Scandinavia, such as gender-neutral early childhood education (Tagliabue 2012), extensive paid leave allowing both male and female workers to take time off for caregiving of sick family members (Esping-Andersen 2009), or improved school offerings in home economics for both boys and girls, would appear more conducive to increasing women’sPage 230 → political participation in the earliest phase of political recruitment—joining a political party. However, it will take more than free time to attract women to parties. Over the past several decades all German parties have undertaken multiple national-level campaigns to increase their female membership. At the time of this writing, the Greens were running a “Fifty Fifty” Project endeavoring to achieve gender parity in their rankand-file party membership,2 the Berlin Social Democrats’ “More Women in the SPD” initiative placed postcards and buttons featuring attention-grabbing slogans such as “bitch” and “diva” in women’s restrooms across the city in an attempt to increase their female membership,3 and the CDU’s YouTube Channel featured multiple videos of a staged, 1,500-person rally entitled “Women for Merkel” in which woman after woman came forward to sing the CDU’s praises.4 These efforts are likely to yield the same results as the parties’ previous attempts: very limited success. Clever slogans, eye-catching paraphernalia, and entertaining videos may inspire women to look into attending a meeting of the local party organization, but if such get-togethers are held in masculinized spaces filled mainly with men, or if they drag on and on as male members repeat each other’s comments, it is unlikely that interested women will continue attending the party functions in their own community. A reconfiguring of party life would go far to increase women’s willingness to participate and such restructuring appears more feasible when women lead, or colead, local party organizations. Attending a meeting chaired or cochaired by another woman would ensure that female newcomers do not find themselves the only woman in the room; encouraging existing female members to be sure to attend as many meetings as possible, and to bring a female friend when they do, could bolster women’s presence, especially in areas where party groups are quite small. Changes to make meetings more female-friendly could include organizing party activities in locations other than pubs or back rooms, such as community or youth centers, cafes, public places of business, or (weather permitting) parks. Organizing breakfast meetings, perhaps immediately following the start of the school day, could also create a less-masculinized ambiance than a late-evening session. Shifting party work or decision-making away from in-person meetings held on evenings and weekends and into cyberspace instead would allow women to participate from home or at times more conducive to their schedules. For example, e-votes could be taken to decide issues, conversations could be held via conference call or Skype, or work could be performed by smaller groups of party members meeting during Page 231 →the day when childcare is available. These changes would be especially helpful to rural women who otherwise find party life overly time consuming as it often requires travel from one end of a large precinct to another. Critical to the success of such endeavors, however, is also the presence of women in virtual party life. Germany’s Pirate Party, famed for its use of social media, has had difficulty attracting female members in part because of its highly masculinized, conflictual, and at times openly sexist online discourse (Meiritz 2012). Women are often deterred from participation in face-to-face party groups due to the long-winded, self-promoting nature of discourse in party meetings. Calling members’ attention to this habit and consciously trying to limit verbosity or pomposity can be done in a humorous manner, and streamlined meetings would give all members more time to pursue other interests. Doris Bucholz, the head of the Free Democrats’ Liberal Women auxiliary organization, recounted her efforts to this end: One time in a [county-level] Board Meeting someone said in making a point, “Let’s pretend we are as dumb as a woman.В .В .В .” After that we put a “punishment piggy bank” (Strafschwein) on the table. You had to put in a Euro if you took a cell phone call during the meeting and two Euros if you made a sexist comment. It helped. (Schulte 2013) Such piggy banks could be used to deter repetitive comments as well. In regions where some local party groups

are dominated by obstinate men it may be preferable to encourage politically interested women away from their local party group toward engagement with a county-level organization instead. Because women in Germany and elsewhere (Dow 2009; Westle 2009) remain on average less interested than men in “politics”—but not in certain issue areas—and because women report dissatisfaction with the lack of substance at political party meetings, running membership campaigns to attract “party members” may not be the most fruitful way in which to attract women. Instead, parties should consider trying to first engage women with their substantive working groups (called Arbeitsgemeinschaften or Arbeitskreisen) on various political topics. These bodies’ primary function is to develop policy stances on issues of concern to many women such as education, health, poverty, gender inequality, and foreign policy. Publicizing meetings of these bodies and opportunities to get involved in tackling problems of the day may both better attract and hold politically interested Page 232 →women’s attention than endeavoring to pull them into local party groups. Party membership may follow, rather than precede, involvement with a working group. Given that parties have such groups devoted to issues of interest to women underrepresented in parties—such as low-income women, businesswomen, or female immigrants—this strategy appears particularly useful in helping broaden the diversity of female party members. In addition to altering the ethos of party life and making efforts to recruit female members through working groups, changes can also be made to render those women who do join parties more likely to develop political aspirations, especially for local-level offices. One possibility suggested by this study would be to improve the remuneration local officeholders receive—lessening women’s need to pursue both unpaid domestic labor and paid employment in addition to political office. Another possibility would be to modify the type of speech required to obtain ballot nominations. Many women find the process of standing in front of their peers and making a case why their candidacy would be preferable to a fellow party member’s an uncomfortable process; moreover, psychological research indicates that women are indeed perceived less favorably than men when espousing such self-promoting rhetoric. A minor alteration in nomination procedures—giving aspirants the option of delivering a sample stump speech promoting the party’s platform or their particular policy concerns, rather than offering up a defense of their own candidacy—would offer female aspirants a less off-putting option and one that would likely be better received by other party members. Either type of rhetoric would allow selectors to assess the public speaking ability of potential candidates, the underlying purpose of this candidate selection mechanism. In sum, difficult to violate (near) parity quotas combined with the recommendations made here—changing the recruitment environment, altering the tenor of party life, modifying the methods used to recruit women to parties, compensating local officeholders, and varying procedures of candidate nomination—should help continue to improve women’s participation in all phases of the democratic political recruitment process.

Methodological Contributions Not only can practical policy recommendations be drawn from the results presented here, so too can some methodological lessons for social scientists interested in gender and political recruitment. This study makes a novel contribution to the literature on quotas by extending the systematic study Page 233 →of affirmative action’s impact earlier in the process that is used to whittle the entire citizenry down to a few elected officials. While quotas have been shown to raise the numbers of women in elective office, the “black box” through which citizens must pass before becoming candidates has remained more or less closed. Without studying whether quotas enable more women to obtain the qualifications needed to be an eligible, and whether they motivate female eligibles to develop political aspirations and gatekeepers to change their actions, it is not possible to determine whether quotas’ original goals have been met. The numbers of women in elective office may rise after quota adoption, but females may remain less likely than male citizens to become eligible to run for office, to desire to do so, and to be selected by gatekeepers for candidacies. Only by opening the black box of political recruitment can scholars determine whether or not is the case. The research methodology employed here allows for exactly such an exploration of the “secret garden” of candidate selection (Gallagher and Marsh 1988). By extending the candidate pool survey technique developed by

students of Anglo-American politics (e.g., Norris and Lovenduski 1995; Lawless and Fox 2005) to a country in which affirmative action policies are implemented, quotas’ effect on the early stages of political recruitment can be assessed. Surveying eligibles—those with the basic qualifications required to be nominated for elective office—enabled me to assess men’s and women’s experiences at each stage of political recruitment, both in parties employing affirmative action and in those not doing so. By comparing the sex breakdown of a given party’s rank-and-file membership to the percentage of its female inner-party leaders, I could document quotas’ elevator effect in creating eligibles. By surveying and interviewing male and female eligibles about their propensities to have always wanted to be a politicians, to have thought about running for certain public posts, and their willingness to accept a nomination to become a candidate for public office, I was able to document gender differences in political ambition. Interviewing these eligibles allowed me to discover the gendered and nongendered mechanisms shaping their ambition. My survey of eligibles also helped me to document potential male and female candidates’ direct experiences with party gatekeepers—for example, whether they had ever been asked by their party to run for elective office—as well as their perceptions of gatekeepers’ preferences—for example whether they believed men and women in their party were equally likely to be selected as candidates. Here again, personal interviews with Page 234 →eligibles and gatekeepers allowed me to discover the changing strategies gatekeepers have employed to identify and groom promising female eligibles since quota adoption. Finally, by surveying eligibles I was also able to examine the fortunes of similarly qualified male and female inner-party leaders by comparing whether they had actually endeavored to become a candidate, had been selected to appear on the ballot by gatekeepers, and had ultimately been elected. I could also document the motivations for such activities, including whether party encouragement to run for office played a role. The conventional approach to studying quotas, examining who appears on the ballot or gets elected, cannot shed light on the above aspects of political recruitment. My research methods not only allow me to understand the process of candidate recruitment in a setting where affirmative action is used, they also enable me to help contextualize others’ findings regarding gender and political ambition in the United States (e.g., Lawless 2012; Lawless and Fox 2012). In the United States’ quota-less, entrepreneurial system of candidate selection, female eligibles have been found to be less politically ambitious than their male counterparts, less often asked to run for office by party gatekeepers, and less likely to become candidates or democratically elected officeholders than similarly qualified men. The findings presented here indicate that these latter gender gaps are not universal, but rather highly conditioned by the institutional context in which eligibles find themselves. The pathway case research design utilized in this study further allowed me to rule out alternative explanations as causes of the political recruitment patterns I observe. By holding factors such as the level of economic development, the nature of the welfare state, the electoral system, political culture, and candidate selection procedures constant, I was able to go far in isolating quotas’ effects on the political recruitment process in Germany. While my work has a very high level of internal validity, future studies in other contexts are now needed to assess its external validity and better address a potential endogeneity problem. I conclude by outlining some such future research ideas.

Avenues for Future Research The methodological approach utilized here can be employed in various national contexts with other types of quotas to determine how different affirmative action measures—such as reserved seats, electoral law quotas, a Page 235 →binding 33 percent quorum, or an easy-to-violate 50 percent quota—shape men’s and women’s experiences throughout the political recruitment process. Comparing these and other quota types deployed in various electoral systems will allow for greater assessment of this study’s external validity. Examining countries in which legislative quotas have been adopted would solve the possible endogeneity problem

here. Electoral law quotas are imposed upon political parties via state mandates rather than freely chosen to promote gender equality. As a result, observed changes in a party’s political recruitment practices following the enactment of a quota law could not be caused by party ideology and would seem to be attributable to quotas.5 Of special interest in future studies would be issues such as the threshold at which quotas acquire “teeth” and can force gatekeepers to alter their behavior. What types of sanctions or incentives are most effective in this regard? How large of a gap between the percentage of women among party members and the percentage of female candidate required by the quota must exist before gatekeepers are spurred to take actions to develop viable female candidates? Moreover, the changes in gatekeeper actions observed here—such as relaxing the Ochsentour or identifying new avenues through which local visibility can be obtained—are intimately related to Germany’s electoral system and how it determines the qualities required of a successful candidate. Other electoral systems are expected to generate different alterations in gatekeeper behavior in line with the electoral incentives that they create; for example, in some contexts where personal loyalty is a highly valued candidate trait, quotas have led gatekeepers to select their wives and daughters rather than the trusted male friends and relations they had previously tapped (Kenny and Piscopo 2015). An additional question raised here is what types of quota “watchdogs” emerge in other contexts to push quota-less parties to promote female candidates. In Germany these were largely found in political parties, in the media, and in women’s policy agencies and women’s organizations. Watchdogs might assume other forms in different contexts, however; in a democratizing country, for example, international actors might play this role. The research presented here also indicates that quotas’ impact has been limited in the Federal Republic because voluntary party regulations are unable to change the gendered division of labor in society. As a result, the unpaid nature of local office in Germany appears to be a deterrent to women’s political ambition. Studies of eligibles in contexts where local-level political positions are paid offices can shed light on whether professionalizing community office can increase women’s participation in earlier stages of the democratic process. In addition, future research is needed to Page 236 →determine which public policies or initiatives are best suited to altering gender roles and increasing men’s participation in the domestic sphere. Many countries, including recently Germany, have begun to adopt paid paternity leave to establish routines early in parenthood in which men play a greater role than in the past in childrearing.6 Whether such policies have an impact on the domestic division of labor, and, in turn, on women’s participation in civil society as a whole and on women’s participation in parties in particular, merits future scholarly consideration. Even if women do have time to devote to voluntary activity, however, this book argues, they are deterred by German political party life in its current form, especially in thinly populated regions of the country. Additional research is required in order to best determine how this problem is to be overcome; examining local party organizations, especially those in rural areas, that have indeed managed to attract a critical mass of female members will be useful in this regard. Understanding the rituals that draw women into precinct-level party groups—such as the meeting times and locations or the substantive focus and type of discourse used in gatherings—can aid in the development of best practice guidelines for local party leaders interested in inviting more female members to their area’s party group. Future research in other national contexts can also help further elaborate on how affirmative action regulations shape men’s political ambition. The evidence presented here shows, on the one hand, that when quotas improve women’s odds of local-level election compared to their male peers, male party members become less likely than female members to seek a ballot nomination in their community. Men are also less often asked by party gatekeepers to run for elective office than women. On the other hand, the candidate training programs parties develop to encourage women to run for elective office have positive externalities for male party members; men in parties with quotas report significantly more access to such professional development programs than their male counterparts in quota-less organizations. Thus, gender quotas, even those targeted specifically at women, can have a spillover effect and shape men’s willingness and ability to participate in democracy as well. Future research can explore in more detail how men’s careers are impacted by various affirmative action policies. For the German case, more data is needed in order to confirm that (near) parity quotas reduce male eligibles’ propensity to run for top-level posts and that 33 percent quotas diminish women’s likelihood of stepping

forward for high level office. Large-scale studies of the German eligibility pool for Landtag and European Parliament elections would be particularly welcome. Future eligibility pool surveys like the one conducted here could select Page 237 →a survey sample with an eye toward intersectional analysis. Conducting surveys of the (very few) racial and ethnic minority members of German parties—or parties in other democracies—would allow for an assessment of the role other descriptive characteristics play in political recruitment. As quotas extend beyond gender to other groups, it will also be possible to study the intersection of race/ethnicity, gender, and affirmative action in political recruitment. Such work could also be conducted comparing women across generations and regions to determine how gender role socialization impacts political recruitment. For example, some studies indicate that eastern German women are more politically ambitious than their western sisters as a result of communist-era socialization (Glatte and deVries 2014) whereas others find fewer and fewer eastern Germans entering politics at all (Kintz 2014, 19–20). Additional candidate interest surveys comparing eastern and western Germans to each other could shed light on this debate. Finally, the results presented here suggest that quotas should have an impact on women’s substantive representation. Because parties with (near) parity quotas have the most women in their eligibility pool, prior work on women’s substantive representation (e.g., Celis 2006, 2009) suggests that MPs from these parties will express the broadest range of issues of interest to women, followed by parties with less stringent quotas. Because quota-less parties feature the fewest female elected officials, their MPs could be expected to bring up the narrowest range of issues of interest to women. Conversely, because women still make up a minority in both parties and elected bodies in Germany, it is also likely the case that a broader range of men’s than women’s interests are articulated in parliamentary discourse. By systematically comparing the range of women’s and men’s interests raised by MPs from parties with various (or no) quota policies,7 we can empirically investigate the relationship between the two affirmative action types and women’s substantive representation. In her research, Christina Xydias (2008) establishes that “quota women” in the Bundestag are more likely to advocate for women’s interests in committee debates than women from parties with no quotas; moreover, in a later study she also finds that men from parties employing affirmative action are more likely than male MPs from quota-less parties to engage in substantive representation for women (Xydias 2014). Future investigations in this vein could investigate the breadth of issues that are raised by parties before and after quota adoption, or as affirmative action rules change. Such scholarship should also pay particular interest to documenting which women’s voices are not heard. The research presented Page 238 →here indicates these are likely to be the voices of immigrant, poor, or less-educated women, single mothers, and those employed in the private sector. German quotas intersect with a political recruitment process to create a situation where a certain type of women’s descriptive representation has likely increased, but other groups of women are likely to remain underrepresented politically. Future studies of quotas and women’s substantive representation can investigate the intersectional relationships between political recruitment procedures and various groups of women’s substantive representation. In conclusion, voluntary party quotas have done much to improve women’s participation in democracy, created female role models, and increased the prospects of substantive representation for many women. However, the glass is still half full and more than quotas is required to create gender equality throughout the democratic political recruitment process.

Page 239 →

Appendix A Translation of Survey and Descriptive Statistics Survey-Related Correspondence Surveys were sent in a staggered fashion to correspond with the end of the summer school vacation in each state. Each survey was preceded by a brief German-language letter on University of Miami letterhead with the following text (author’s translation appears here): Re: Candidate Interest Survey Esteemed Party Member: I am a political science professor at the University of Miami in Florida, USA and am currently researching why some party members decide to run for elective office at the local, county,1 regional,2 state, federal, or European level while others do not. Supported by the Alexander von Humboldt Foundation and the University of Bremen I am conducting a national survey on this subject. Because of your leading role in your political party, I would very much like to hear about your experiences with, and thoughts on, this matter—regardless of whether you have run for elective office or not. For this reason, I kindly request that you fill out the following survey and return it in the pre-paid envelope to my P.O. Box at the University of Bremen. I assure you that your answers will be evaluated anonymously and your name will not appear any place in the results. Page 240 →If you are interested in the results of my research, have thoughts on this matter, or are interested in a personal meeting with me, please write your name and address on a separate piece of paper and return it along with the survey. I will send you a report about the results after I have evaluated the survey responses. In case you have any questions for me you can reach me via e-mail or telephone or at the address on the envelope. Thanks in advance for your cooperation! Best, Louise K. Davidson-Schmich A reminder postcard was sent a week later and stated: I sent you a survey about candidate interest a few days ago. Many thanks if you have already filled this out and returned it. Your opinions are very important to this research project. If you have not yet filled out the questionnaire or have not yet returned it, I kindly request that you do this soon. Your answers will remain anonymous and will only be used for academic purposes. In case you lost your survey or did not receive one, I would be happy to send you a new one. Please write or call. Best, Louise K. Davidson-Schmich Assistant Professor of Political Science University of Miami

Translation of Survey Questions and Coding of Variables Below is an English-lanvguage translation of relevant survey questions in the order in which they appeared; survey questions not used in this book are omitted. Survey about Candidate Interest in State Name Each respondent’s state of residence was determined by looking at the title of the questionnaire she or he returned (all respondents received a questionnaire with Page 241 →their state’s name in the survey title). This information was used to create the indicator variables for each state: Baden-WГјrttemberg (N = 101) Bayern (N = 91) Bremen (N = 87) Hamburg (N = 84) Nordrhein-Westfalen (N = 104) N = 467 Part I: Questions about your Political Environment 1.In which party are you a member? BГјndnis 90/Green CDU CSU FDP SPD This question was recoded as follows to create the Quota variable: 0.CSU or FDP respondents ( N = 110) 1.BГјndnis 90/Green, SPD, and CDU respondents ( N = 353) Mean: 0.44 Standard Error: 0.43 N = 463 It was also recoded to create the (Near) Parity Quota variable: 0.CDU, CSU, or FDP respondents ( N = 202) 1.BГјndnis 90/Green and SPD respondents ( N = 261) Mean: 0.56 Standard Deviation: 0.50 N = 463 It was also recoded to create the Quorum variable: 0.BГјndnis 90/Green, CSU, FDP, and SPD respondents ( N = 372) 1.CDU respondents ( N = 91) Page 242 →Mean: 0.20 Standard Deviation: 0.40 N = 463 It was also used to create the Small Party variable. 0.CDU, CSU, and SPD members ( N = 260) 1.BГјndnis 90/Green and FDP members ( N = 203)

Mean: 0.44 Standard Deviation 0.50 N = 463 2.How many inhabitants does your town or city have? 1.Less than 20,000 ( N = 121) 2.20,000–50,000 ( N = 57) 3.100,000 ( N = 35) 4.200,000 ( N = 24) 5.300,000 ( N = 18) 6.400,000 ( N = 17) 7.500,000 and above ( N = 190) Mean Category 4.23 Standard Deviation = 2.60 N = 462 The following question was used to create the party strength variable: 3.Which of the following apply to your town or neighborhood? Your party’s ideas and its program 0.are rejected by the majority of the inhabitants ( N = 64) 1.find both acceptance and rejection among inhabitants ( N = 302) 2.are approved of by the majority of the inhabitants ( N = 92) Mean: 1.06 Standard Deviation: 0.58 N = 458 4.If you were interested, would there be continuing education opportunities for you to prepare for a candidacy (for example, public speaking courses, mentoring programs, etc.)? 0.No, not offered or I don’t know ( N = 161) 1.Yes, offered by my party ( N = 298) Page 243 →Mean: 0.65 Standard Deviation: 0.48 N = 459 5.If your party does not have a quota/quorum, does your party group nevertheless attempt to place women in promising list places? 0.No ( N = 27) 1.Yes ( N = 136) Mean = 0.84 Standard Deviation = 0.38 N = 163 6.When it comes to selecting candidates, who would have the better chance of getting the nomination for a safe constituency seat and a double-nomination on the party list? 0.A qualified man ( N = 94) 1.Both would have an equal chance ( N = 287) 2.A qualified woman ( N = 78) Mean 0.97 Standard Deviation: 0.61 N = 459 7.What percentage of your local party organization is male? _____% Minimum: 15 Maximum: 100 Mean: 67 Standard Deviation: 11 N = 443 8.In your electoral district who has a better chance of winning a directly elected seat?

0.A qualified man ( N = 100) 1.Both would have an equal chance ( N = 308) 2.A qualified woman ( N = 49) Mean 0.89 Standard Deviation: 0.56 N = 457 Page 244 →9.For how many years have you been an active member in your party? _____ Min: 1 year Max: 52 years Mean: 16.8 years Standard Deviation: 11.1 years N = 451 Part II: Questions about your Political Career 10.Which political experiences have you had? For each elected body please give one answer: City/Town Council 0.I never considered running for this body 1.I considered running for this body but did not 2.I attempted to get my party’s nomination for this body but did not 3.I was nominated by my party to run for this office but did not win 4.I have been elected to this body County Council (in Baden-WГјrttemberg, Bayern, and Nordrhein-Westfalen only) 0.I never considered running for this body 1.I considered running for this body but did not 2.I attempted to get my party’s nomination for this body but did not 3.I was nominated by my party to run for this office but did not win 4.I have been elected to this body District Council (in Bayern only) 0.I never considered running for this body 1.I considered running for this body but did not 2.I attempted to get my party’s nomination for this body but did not Page 245 →3.I was nominated by my party to run for this office but did not win 4.I have been elected to this body Local Mayor 0.I never considered running for this office 1.I considered running for this office but did not 2.I attempted to get my party’s nomination for this office but did not 3.I was nominated by my party to run for this office but did not win 4.I have been elected to this office County Mayor (in Baden-WГјrttemberg, Bayern, and Nordrhein-Westfalen only) 0.I never considered running for this office 1.I considered running for this office but did not 2.I attempted to get my party’s nomination for this office but did not

3.I was nominated by my party to run for this office but did not win 4.I have been elected to this office Responses to the above five questions were combined to create the following variables: Considered Running for Local Office: Giving a response of ≥1 to any one of the above five questions was coded as 1; if responses were 0 for all five offices, the respondent was coded as 0 0.Never considered running for all of the above five offices ( N = 54) 1.Considered running for at least one of the above five offices ( N = 396) Mean: 0.88 Standard Deviation: 0.33 N = 449 Page 246 → Sought Nomination for Local Office: Giving a response of ≥2 to any of the above questions about the five local-level offices was coded as 1; if all responses were 0–1 the respondent was coded as 0 0.Did not attempt to receive the party’s nomination for all of the above five offices ( N = 106) 1.Attempted to receive the party nomination for at least one of the above five offices ( N = 343) Mean: 0.76 Standard Deviation: 0.42 N = 448 Became a Candidate for Local Office: Giving a response of ≥3 to any of the above questions about the five local-level offices was coded as 1; if all responses were 0–2 the respondent was coded as 0 0.Did not receive the party’s nomination for all of the above five offices ( N = 115) 1.Received the party nomination for at least one of the above five offices ( N = 333) Mean: 0.74 Standard Deviation: 0.44 N = 448 Won Local Office: Giving a response of 4 to any of the above questions about the five local-level offices was coded as 1; if all responses were 0–3, the respondent was coded as 0 0.Did not win the election to any of the above five offices ( N = 203) 1.Won election to at least one of the above five offices ( N = 245) Mean: 0.55 Standard Deviation: 0.50 N = 448 State Legislature 0.I never considered running for this body 1.I considered running for this body but did not Page 247 →2.I attempted to get my party’s nomination for this body but did not 3.I was nominated by my party to run for this office but did not win 4.I have been elected to this body Bundestag 0.I never considered running for this body

1.I considered running for this body but did not 2.I attempted to get my party’s nomination for this body but did not 3.I was nominated by my party to run for this office but did not win 4.I have been elected to this body European Parliament 0.I never considered running for this body 1.I considered running for this body but did not 2.I attempted to get my party’s nomination for this body but did not 3.I was nominated by my party to run for this office but did not win 4.I have been elected to this body Responses to the above three questions were combined to create the following variables: Considered Running for High Office: Giving a response of ≥1 to any of the above three questions was coded as 1; if all responses were 0 the respondent was coded as 0 0.Never considered running for any of the above three offices ( N = 252) 1.At least considered running for any one of the above three offices ( N = 197) Mean: 0.44 Standard Deviation: 0.50 N = 449 Page 248 → Sought Nomination for High Office: Giving a response of ≥2 to any of the above three questions was coded as 1; if all responses were 0–1 the respondent was coded as 0 0.Did not attempt to receive the party’s nomination for any of the above three offices ( N = 346) 1.Attempted to receive the party’s nomination for at least one of the above three offices ( N = 103) Mean: 0.23 Standard Deviation: 0.42 N = 449 Became a Candidate for High Office: Giving a response of ≥3 to any of the above three questions was coded as 1; if all responses were 0–2 the respondent was coded as 0 0.Did not receive the party’s nomination for any of the above three offices ( N = 363) 1.Received the party’s nomination for at least one of the above three offices ( N = 86) Mean: 0.19 Standard Deviation: 0.39 N = 449 Elected to High Office: Giving a response of 4 to any of the above three questions was coded as 1; if all responses were 0–3, the respondent was coded as 0 0.Did not win any of the above three offices ( N = 413) 1.Won election to at least one of the above three offices ( N = 36) Mean: 0.08 Standard Deviation: 0.27 N = 449 11.If you would decide to run for office, how would other members of your party react? 0.They would reject my candidacy ( N = 10) 1.They would not care one way or another ( N = 22) 2.They would react positively ( N = 392)

Mean: 1.9 Standard Deviation: 0.37 N = 424 Page 249 →12.If you’ve already run for elective office, or if you can imagine doing this in the future, why did/would you do so? Please check as many as apply I have always wanted to be a politician 0.Not checked ( N = 411) 1.Checked ( N = 26) Mean: 0.06 Standard Deviation: 0.24 N = 437 I wanted to change something in the community 0.Not checked ( N = 64) 1.Checked ( N = 373) Mean: 0.85 Standard Deviation: 0.35 N = 437 I wanted to represent the citizens’ interests 0.Not checked ( N = 129) 1.Checked ( N = 308) Mean: 0.70 Standard Deviation: 0.25 N = 437 I was encouraged by a mentor 0.Not checked ( N = 402) 1.Checked ( N = 35) Mean: 0.08 Standard Deviation: 0.27 N = 437 My party offered me the nomination 0.Not checked ( N = 293) 1.Checked ( N = 144) Mean: 0.33 Page 250 →Standard Deviation: 0.47 N = 437 I wanted to accomplish political goals 0.Not checked ( N = 172) 1.Checked ( N = 265) Mean: 0.61 Standard Deviation: 0.49 N = 437 13.Has your party ever asked you to run for elective office?

0.No ( N = 83) 1.Yes ( N = 361) Mean: 0.81 Standard Deviation: 0.39 N = 444 14.Imagine that your party offered you the chance to become a candidate for the following bodies. What would you do? City/Town Council 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so County Council (Baden-WГјrttemberg, Bayern, and Nordrhein-Westfalen only) 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so Regional Council (Bayern only) 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so Page 251 →Local Mayor 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so County Mayor (Baden-WГјrttemberg, Bayern, and Nordrhein-Westfalen only) 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so The responses to the above five questions were combined to create the following variable: Willingness to Accept a Hypothetical Nomination for Local Level Office Respondents answering 1 to any of the above five questions were coded 1, all others 0 0.Not willing to accept a hypothetical nomination for local-level office or not sure ( N = 125) 1.Willing to accept a hypothetical nomination for at least one local-level office ( N = 320) Mean: 0.72 Standard Deviation: 0.45 N = 445 State Legislature 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so

Bundestag 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so Page 252 →European Parliament 1.I would definitely accept the nomination/I have already done so 2.I am not sure what I would do 3.I would turn down this nomination/I have already done so The responses to the above three questions were combined to create the following variables Willingness to Accept a Hypothetical Nomination for High-Level Office Respondents answering 1 to any of the above three questions were coded 1, all others 0 0.Not willing to accept a hypothetical nomination for high-level office or not sure ( N = 282) 1.Willing to accept a hypothetical nomination for at least one high-level office ( N = 162) Mean: 0.36 Standard Deviation: 0.48 N = 444 15.Please assess how qualified you are to run for elective office at any level. 0.I am not at all qualified ( N = 6) 1.I am partially qualified ( N = 115) 2.I am qualified ( N = 267) 3.I am very qualified ( N = 60) Mean: 1.85 Standard Deviation: 0.65 N = 448 16.Which of the following reasons would deter you from running for an elected position? Please check all that apply: An incumbent from my party holds the seat I would be interested in (Note: this was also used as a measure of party strength) 0.Not checked ( N = 339) 1.Checked ( N = 91) Mean: 0.21 Page 253 →Standard Deviation: 0.41 N = 430 Running would be too expensive 0.Not checked ( N = 373) 1.Checked ( N = 57) Mean: 0.13 Standard Deviation: 0.34 N = 430 Running would hurt my professional career 0.Not checked ( N = 332) 1.Checked ( N = 108)

Mean: 0.25 Standard Deviation: 0.43 N = 430 Running would take time away from hobbies or free time 0.Not checked ( N = 331) 1.Checked ( N = 99) Mean: 0.23 Standard Deviation: 0.42 N = 430 Running would take time away from social life with friends 0.Not checked ( N = 321) 1.Checked ( N = 109) Mean: 0.25 Standard Deviation: 0.43 N = 430 Running would take time away from family life 0.Not checked ( N = 411) 1.Checked ( N = 26) Mean: 0.44 Page 254 →Standard Deviation: 0.50 N = 430 I see no deterrents to running 0.Not checked ( N = 347) 1.Checked ( N = 84) Mean: 0.19 Standard Deviation: 0.40 N = 430 17.In your opinion, how important are the following traits for a candidate? Substantive policy knowledge 0.Unimportant ( N = 4) 1.Not very important ( N = 78) 2.Important ( N = 236) 3.Very important ( N = 125) Mean: 2.09 Standard Deviation: 0.70 N = 443 High public visibility 0.Unimportant ( N = 1) 1.Not very important ( N = 48) 2.Important ( N = 197) 3.Very important ( N = 200) Mean: 2.33

Standard Deviation: 0.67 N = 446 Public speaking skills 0.Unimportant ( N = 0) 1.Not very important ( N = 17) 2.Important ( N = 241) Page 255 →3.Very important ( N = 189) Mean: 2.38 Standard Deviation: 0.56 N = 447 Involvement in community voluntary associations 0.Unimportant ( N = 7) 1.Not very important ( N = 48) 2.Important ( N = 238) 3.Very important ( N = 153) Mean: 2.20 Standard Deviation: 0.69 N = 446 18.How important are the following goals for your life? Time with Family 0.Unimportant ( N = 3) 1.Rather unimportant ( N = 13) 2.Important ( N =194) 3.Very important ( N = 223) Mean: 2.48 Standard Deviation: 0.59 N = 443 Part II: Questions about You 19.When you were growing up did your family often talk about politics? 0.No ( N = 160) 1.Yes ( N = 293) Mean: 0.65 Standard Deviation: 0.48 N = 453 Page 256 →20.Has anyone from your family ever run for office? Please mark all relevant choices: 0.No one ( N = 135) 1.My father or grandfather ( N = 129) 2.My mother or grandmother ( N = 29) 3.My spouse or partner ( N = 0) 9. Another relative (N = 70) N = 307 This was recoded as an indicator variable: 0. No one (answer 0 above or missing value to question 20 but responses given to questions 19 and 21)

1. At least one family member (answering 1–9 above) Mean: 0.35 Standard Deviation: 0.48 N = 452 21.Are you married or in a committed relationship? 0.No ( N = 60) 1.Yes ( N = 338) Mean: 0.87 Standard Deviation: 0.34 N = 448 22.How would your spouse or partner react if you decided to run for elective office? 0.My partner would oppose the decision ( N = 11) 1.My partner would give me only minimal support ( N = 46) 2.My partner would support me ( N = 156) 3.My partner would wholeheartedly support me ( N = 170) Mean: 2.27 Standard Deviation: 0.78 N = 383 23.Is your spouse or partner a member of a political party? 0.No ( N = 199) 1.Yes, a member of the same party as I am ( N = 191) Page 257 →2.Yes, a member of a different party than I am ( N = 9) Mean: 0.52 Standard Deviation: 0.54 N = 399 24.Do you have any minor children living with you? (If not, skip to question 28) 0.No ( N = 314) 1.Yes ( N = 147) Mean: 0.30 Standard Deviation: 0.46 N = 451 25.How many minor children do you have? Min: 1 Max: 5 Mean: 1.70 Standard Deviation: 0.77 N = 143 26.How old are these children? Min: 3 months Max: 18 years Mean: 9.95 years Standard Deviation: 5.49 years N = 130 27.Who has primary childcare responsibility? 0.Your spouse ( N = 23) 1.You and your spouse share equally ( N = 80) 2.You ( N = 32) 9.Primarily someone else (e.g., other family member, nanny) ( N = 5) Mean: 1.35 Standard Deviation: 1.60 N = 140 This was recoded as an indicator variable:

Page 258 →0. Answers 0, 1, and 9 above or missing value here but a response given to Question 24 (N = 416) 1. Answer 2 above (N = 32) Mean: 0.07 Standard Deviation: 0.26 N = 448 28.Who does the following tasks in your household? Cleaning 0.Primarily your spouse ( N = 80) 1.You and your spouse share equally ( N = 114) 2.Primarily yourself ( N = 133) 9.Primarily someone else (e.g., other family member, housekeeper) ( N = 113) Mean: 3.18 Standard Deviation: 3.50 N = 440 This was recoded as an indicator variable: 0. Answers 0, 1, and 9 above (N = 307) 1. Answer 2 above (N = 133) Mean: 0.30 Standard Deviation: 0.46 N = 440 Laundry 0.Primarily your spouse ( N = 152) 1.You and your spouse share equally ( N = 72) 2.Primarily yourself ( N = 191) 9.Primarily someone else (e.g., other family member, housekeeper) ( N = 33) Mean: 1.68 Standard Deviation: 2.24 N = 448 This was recoded as an indicator variable: Page 259 →0. Answers 0, 1, and 9 above (N = 257) 1. Answer 2 above (N = 191) Mean: 0.43 Standard Deviation: 0.50 N = 448 Cooking 0.Primarily your spouse ( N = 132) 1.You and your spouse share equally ( N = 105) 2.Primarily yourself ( N = 186) 9.Primarily someone else (e.g., other family member, housekeeper) ( N = 20) Mean: 1.48 Standard Deviation: 1.84 N = 443

This was recoded as an indicator variable: 0. Answers 0, 1, and 9 above (N = 257) 1. Answer 2 above (N = 186) Mean: 0.42 Standard Deviation: 0.49 N = 443 Shopping 0.Primarily your spouse ( N = 183) 1.You and your spouse share equally ( N = 80) 2.Primarily yourself ( N = 175) 9.Primarily someone else (e.g., other family member, housekeeper) ( N = 9) Mean: 1.48 Standard Deviation: 1.84 N = 443 This was recoded as an indicator variable: 0. Answers 0, 1, and 9 above (N = 272) 1. Answer 2 above (N = 175) Page 260 →Mean: 0.39 Standard Deviation: 0.49 N = 447 Responses to the above four questions were combined to create a measure of the number of household tasks for which a respondent was primarily responsible. For every task a respondent answered “2” to, one point was awarded. For every task a respondent shared primary responsibility for (answer 1), one half a point was awarded. All other responses (choices 0 and 9) were coded as 0s. Min: 0 Max: 4 Mean: 2.05 Standard Deviation: 1.29 N = 450 29.What is the highest level of education you have obtained? 0.None ( N = 0) 1. Hauptschule (N = 2) 2. Mittlere Reife (N = 48) 3. Abitur (N = 42) 4. Lehre (N = 59) 5. MeisterprГјfung (N = 15) 6. Fachhochschule (N = 78) 7. Examen (N = 178) 8. Promotion (N = 31) 9.Other: please specify ( N = 9) Mean: 5.55 Standard Deviation: 1.92 N = 458 This variable was used to create the classifications used in the text:

Those with a university degree were those who answered 6–7 to the above question Those who held a PhD were those who answered 8 to the above question Page 261 → Those with high education answered 6–8 above Mean: 0.64 Standard Deviation: 0.48 N = 458 Those with low education answered 0–2 above Mean: 0.11 Standard Deviation: 0.31 N = 458 30.Please describe your employment position: 0.Unemployed ( N = 7) 1.Tenured civil servant ( BeamteR) (N = 81) 2.Public sector ( AngestellteR) (N = 65) 3.Worker in private sector ( ArbeiterIn) (N = 3) 4.Salaried employee in private sector ( AngestellteR) (N = 106) 5.Self-employed ( N = 99) 6.Retired ( N = 47) 7.Trainee/student ( N = 7) 8.Homemaker ( N = 23) 9.Other. Please describe: ( N = 9) Mean: 3.93 Standard Deviation: 2.15 N = 454 This variable was used to create the classifications used in the text: Those working in the public service were those who answered 1 and 2 to the above question Those working in the private sector were those who answered 3 and 4 to the above question Those who were self-employed were those who answered 5 to the above question. This indicator variable was also used in multivariate analysis Page 262 →Mean: 0.22 Standard Deviation: 0.41 N = 454 31.What is your personal monthly income? 1.Under €1,000 ( N = 74) 2.€1,000–€1,500 ( N = 68) 3.€1,500–€2,000 ( N = 64) 4.€2,000–€2,600 ( N = 69) 5.€2,600–€3,200 ( N = 57) 6.€3,200–€4,500 ( N = 55) 7.€4,500 or above ( N = 36) Mean: 3.65 Standard Deviation: 1.92 N = 423 What is the monthly income of your household?

1.Under €1,000 ( N = 10) 2.€1,000–€1,500 ( N = 14) 3.€1,500–€2,000 ( N = 31) 4.€2,000–€2,600 ( N = 44) 5.€2,600–€3,200 ( N = 65) 6.€3,200–€4,500 ( N = 114) 7.€4,500 or above ( N = 122) Mean: 5.4 Standard Deviation: 1.6 N = 400 This variable was also used to create two additional measures of income:

High Income 0. All respondents answering