Gender quotas in the French bureaucratic elite: the soft power of restricted coercion

Drawing on a collective study of gender and executive careers in the French Civil Service, this article provides an account of the genesis and implementation of the gender quota introduced by the 2012 “Sauvadet law” in senior executive positions of the State bureaucracy. First, it shows the crucial, yet unobtrusive, role played by feminist actors in the process, in spite of the absence of any strong open politicization of the issue, from agenda setting to implementation. Second, the article points to the dual character of the reform. On the one hand, the gradual character and limited perimeter of the quota (to senior executive “initial appointments”) restricted its immediate quantitative impact. On the other hand, the constraining, technical dimension of the tool endowed it with a potential to foster incremental social change.

On March 12, 2012, 1 with almost no media coverage and without prompting any parliamentary and public controversy, the "Sauvadet law" (named after François Sauvadet, minister for Civil Service at the time) introduced a series of mandatory gender quotas in decision-making positions of the French public administration, notably in appointments to senior executive positions. 2 This last provision (named here "the Sauvadet quota") imposed to all public administrations a gradual gender quota (up to an expected 40% in 2018) in the first appointments of senior executive positions in the three branches of the French Civil Service. In the Nationallevel Civil Service (Fonction publique d'Etat), 3 which is at the center of our study, the gender quota provision applies to an average of 600 appointments a year, out of 3600 senior executive positions. At the time the provision was adopted (2012), women made up half of the National-level Civil Service workforce and accounted for less than 26% of senior executive positions.
The absence of visible controversies and struggles over the Sauvadet quota is not entirely surprising. The provision expanded to a new social field, a gender policy tool that had gained increased legitimacy since the early 2000s (Bender et al. 2015;Bereni and Revillard 2015;Bui-Xuan 2012;Lépinard 2016). A year earlier, the "Copé-Zimmermann" statute imposed incremental gender quotas in private sector corporate boards. Advocates of the Sauvadet quota presented the provision as a means to align the public sector on the private sector, against the backdrop of an ongoing bureaucratic reform under the auspices of new public management (Bezes et al. 2012).
Yet, on the other hand, the Sauvadet quota appeared as "a small revolution" (une petite revolution), as Sauvadet himself presented it at the Assemblée nationale at the time. 4 Indeed, the degree of constraint imposed by the quota represented a "turning point" in gender equality policies in the public sector (Bui-Xuan 2012; Edel 2013). It introduced a breach into the ideology of meritocracy and equal treatment particularly associated with the French Civil Service, whose recruitment and promotion principles are deemed strictly neutral and bounded by statutory rules. 5 Moreover, whereas the quota put in place by the 2011 Copé-Zimmerman law only targeted corporate boards, leaving the more strategic executive (Comex) and directory boards (Codir) untouched, the Sauvadet quota targeted the highest positions of power (3600 senior executive positions) in the National-level Civil Service. Because of this unprecedented constraining dimension, it faced strong-yet covered-resistances among most administrative elites and government members at the time, behind the public display of consensus. These resistances could only be overcome by the personal arbitration of President Sarkozy, according to Sauvadet's recent public account of the reform.
How was such a constraining policy tool adopted? To what extent have feminist actors and framings influenced its adoption and shaped its implementation? Did it favor gender transformation, in terms of both the access of women to the highest positions, and gender representations (Engeli and Mazur 2018)? Is it a coercive policy tool producing substantive effects, or a symbolic measure with only limited influence on state organizations' daily functioning (Edelman 2016;Mazur 1995)?
This article provides an account of the genesis and implementation of the Sauvadet quota, drawing on a collective research on gender and executive careers in the French public bureaucracy (Marry et al. 2017;Revillard et al. 2018). We conducted a series of 100 in-depth interviews (one-third with men and two-third with women) with executives and senior executives working in four administrative departments pertaining to two ministries (Ministry of Economic and Financial Affairs, Ministry of Social Affairs), between 2011 and 2013. While our research focus was mainly the mechanisms behind the production of the "glass ceiling," we also granted attention to the reception of the Sauvadet quota by high civil servants and to its influence on their career perspectives (Bereni and Revillard 2015). These data were complemented by a set of materials collected since the introduction of the provision (2012-2019), documenting its genesis and its implementation: general and professional press articles, administrative reports, parliamentary debates, documentation issued by female high civil servants organizations.
This research reveals the crucial, yet unobtrusive role played by feminist actors in the process, in spite of the absence of any strong open politicization of the issue, from agenda setting to implementation. In the terms of the Gender Equality Policy in Practice (GEPP) project (see Engeli and Mazur, this issue), this case reveals high substantive and moderate descriptive women's empowerment: (upper-class) women's interests have been taken into account, although women openly acting as women's rights advocates have become less visible. We draw on an analytical framework that envisions feminist mobilization in a broad perspective, including nongovernmental as well as governmental actors (Banaszak 2010;Bereni and Revillard 2018;Katzenstein 1998). Both the legitimization of the quota and its rigorous implementation can largely be explained by structural transformations of the "women's cause field," defined as the relational structure of groups mostly devoted (although restrictions persisted for many years), and the legal discriminations based on sex to enter high ranked Civil Service corps and functions had gradually been removed throughout the following decades.
Footnote 5 (continued) to the advancement of women in a variety of social settings, cutting across the line between civil society and political institutions (Bereni 2019). The transformations at stake followed the adoption of the parity law in 2000 and were threefold: (1) a diffusion of the "parity grammar" (Bereni and Revillard 2007) in dominant institutions; (2) a strengthening of the institutional pole of the women's cause field (women's policy agencies), and (3) a burgeoning of women's mobilizations among the business, professional, and bureaucratic elite. We argue that these structural transformations combined with the opening of political and discursive opportunities produced the conditions of a possible consensus of legislators around this constraining measure and the overcoming of resistances stemming from administrative elites. Two opportunities were particularly key: the rise of a managerial rhetoric praising "gender diversity" (mixité) in the private sector, and the implementation of new public management reforms and "state modernization" in the public sector.
The paper is divided into six sections, tracing the reform from its genesis to its outcomes. The first two sections deal with the pre-adoption context, examining the slow emergence of equal employment policy in the public administration and the rise of the quota as a policy tool throughout the 2000s. The third section focuses on the adoption of the reform in 2011-2012. As demonstrated in Sect. 4, this genesis helps explain the tool's dual character, both restricted and relatively constraining. We then turn to the implementation and evaluation of the quota in practice: even though the actors in charge of implementation and evaluation were mainstream bureaucratic actors, they pressed in favor of a reinforcement of the measure, in a broader context of increased feminist mobilization in France. Finally, the last section assesses the outcomes of the reform, showing that in spite of its limited immediate effect, the tool is likely to result in significant gender transformation within the bureaucratic elite, over the years, due to its incremental effects.

The slow construction of equal employment policy in the public administration Prior to 2000
The issue of women's access to senior executive positions in the Civil Service was not put on the political agenda before the early 2000s. During the 1970s and 1980s, gender equality policies in the public sector focused on equal employment principles, following a series of reforms in the private sector initiated by women's policy structures. The first equal employment public policies mainly aimed at ensuring formal equal rights for men and women, be it in the private sector or in the public administration. A combination of external pressure from the UN and the European commission, and internal lobbying on the part of women's policy structures, favored the gradual legal assertion of a principle of non-discrimination (Mazur 1995;Edel 2013;Revillard 2016). As of the end of the 1960s, the Comité du Travail Féminin (a women's rights advisory board) raised the issue of gender discrimination in access to administrative exams (concours de la Fonction publique) and issued reports on gender inequalities in the public administration in 1974 and 1978 (Revillard 2009). A 1975 law eliminated the possibility of gender distinctions in exams, except in limited cases listed by a Conseil d'Etat decree.
Governmental action in favor of equal employment gained more visibility with Yvette Roudy's appointment as minister for women's rights in 1981. At the beginning of the 1980s, women already accounted for half of public employees, but they were almost absent from the highest positions: in 1983, there were not a single female prefect, and there were only 3% female ambassadors and 5% female heads of national-level ministerial departments. 6 In 1983, general Civil Service regulations were modified to prohibit all distinction based on sex, following a law adopted on the same day as the more famous "Loi Roudy" on equal employment in the private sector. Beyond formal non-discrimination, attempts to promote more actively the advancement of women in public administration were very limited and took the form of rather vague incentives, lagging behind the more structured equal employment policy put in place in the private sector by Roudy's equal employment law. For example, a 1983 executive memorandum 7 prompted ministerial departments to act to "tackle disparities" between women and men through various measures in terms of recruitment and promotion. Up until the 2000s, equal employment policy, at the heart of French state feminism (Revillard 2016), mainly targeted the private sector. Because of a dominant framing in terms of (formal) equal employment, the issue of women's access to senior executive positions in the public administration remained marginal, leading to no significant policy outcomes (Mazur 2002). Only when it started being framed in terms of access to power, in line with the diffusion of the parity grammar throughout the 1990s, did a coercive measure start being envisioned.

From the gender parity reform to the routinization of gender quotas in decision-making sites (2000-2010)
French political and intellectual elites opposed for a long time any form of affirmative action, conceived as an infringement of the Republican ideals of meritocracy and of "indifference to differences" between citizens (Scott 2005). Yet the campaign for gender parity shattered this consensus at the end of the 1990s, progressively making gender quotas compatible with republican universalism (Bereni 2007;Lépinard 2007). The controversy around gender quotas, which was initially strong even among feminist advocates, suddenly faded after the passing of the gender parity law in 2000 (see Mazur et al. in this issue). Increasingly justified by "pragmatic" reasons, quotas became the default solution to the "underrepresentation" of women in power positions (Bereni and Revillard 2015;Lépinard 2016), while the term "quota" itself, still contentious, was replaced by more consensual phrasings, such as "balanced representation" (représentation équilibrée), "parity" (parité), or "gender diversity" (mixité). In 2008, a new provision was added into the first article of the Constitution, stating that "the law favors the equal access of women and men to elected offices and positions, as well as professional and social responsibilities." The new paragraph expanded beyond the political field a principle that had been initially introduced in 1999 into article 3 of the Constitution as part of the parity reform and thus paved the way for the use of gender quotas in non-political decision-making sites. Those were first introduced in the private sector: the Copé-Zimmermann law (named after the two deputies who introduced the bill), passed in 2011, mandated a 40% quota of persons of each sex in corporate boards of the largest firms (Bender et al. 2015), before being extended to the first appointments of the senior executive positions in the public sector by the Sauvadet law (2012). The gender parity campaign contributed in three respects to the routinization of gender quotas in decisionmaking sites throughout the 2000s, through framing, institutional, and mobilization effects.

A "balanced representation" at the top: the diffusion of the parity grammar
The gender parity reform participated in the diffusion of a parity grammar (Bereni and Revillard 2007), i.e., a new standard of gender equality based on the numerical co-presence of men and women in historically male-dominated sites. Against this discursive backdrop, gender equality policies that gained momentum in the bureaucracy during the 2000s particularly turned attention to women's access to the highest positions (Edel 2013;Jacquemart et al. 2016).
In 1999, high civil servant Anne-Marie Colmou delivered a report to the Civil Service minister documenting the "glass ceiling" (at the turn of the 2000s, women accounted for only 12% of executive level positions in public administration 8 ) and making the case for a "balanced representation of men and women" (Colmou 1999). Importing the "parity grammar" from the political to the administrative field, the report paved the way, with subsequent reports (e.g., Versini 2004), for legitimizing the idea of "representative bureaucracy" (Peters et al. 2015), historically unfamiliar in the French Civil Service (Bui-Xuan 2015; Calvès 2005; Meier and Hawes 2009). The report issued a series of recommendations, including gender monitoring and ministerial plans of actions with quantified targets. A 2000 memorandum issued by the prime minister 9 prompted ministerial departments to adopt "multi-annual plans to improve women's access to executive positions," setting goals in terms of "feminization rates." The same year, a steering committee for the equal access of women and men to senior management positions in the Civil Service was set up. The committee issued three reports developing a quantitative assessment of the glass ceiling as well as recommendations (Le Pors and Milewski 2002, 2003. A 2001 statute (the "Génisson" law) 10 provided that the government must deliver every 2 years to parliament a report on men's and women's employment in the different branches of the Civil Service and at different hierarchical levels-this provision, however, was suppressed in 2007.
Therefore, throughout the 2000s, efforts to promote women in the higher ranks of the public bureaucracy rested mainly on soft, incentive tools, such as gender monitoring. In the report issued at the end of its mandate, in 2005, the Le Pors committee stressed the inefficiency of the legal provisions, the lack of political will, and the weakness of the ministerial plans that were adopted (Le Pors and Milewski 2005). Throughout the decade, the presence of women among senior executives of the National-level Civil Service (Fonction publique d'Etat) increased, from 12% in 1999 to 24% in 2010, although women still accounted for less than 10% of prefects and 15% of ambassadors in 2010. 11

Reinforced women's policy agencies mobilizing for gender quotas
Another significant effect of the campaign for gender parity was to broaden the range of women's policy agencies within the government: The Observatoire de la parité was created in 1995, and Parliamentary Delegations for women's rights were set up at the Assemblée nationale and the Sénat in 1999 (Savinel 2019). These women's policy agencies, although merely advisory, played a major role in the process of legitimation of the gender quota beyond the political field, through their "critical expertise" (Revillard 2009) and through their discreet, unobtrusive activism in the hallways of republican institutions. They played the role of "support structures" for the "entrenchment and diffusion of gender quotas" (Lépinard 2016).
Marie-Jo Zimmermann, a Gaullist (UMP) deputy who had been a key advocate of the parity reform against most of her own political party and who had been appointed at the same time the head of both the Delegation for Women's rights at the National Assembly and the Parity Observatory, orchestrated this feminist campaign within the parliament and the government. She seized president Sarkozy's project of constitutional reform, launched in the name of «state modernization» in the first year of his tenure, as an opportunity to extend the constitutional basis for gender quotas beyond the electoral sphere. Indeed, a provision of the 2005 equal pay law setting quotas in corporate board had been ruled out by the Conseil constitutionnel in 2006 based on the argument that the 1999 constitutional reform only authorized the enactment of gender quotas in political offices and functions. A constitutional reform was thus needed in order to extend this tool to other domains. To this effect, Zimmermann mobilized her nongovernmental, governmental, and parliamentary networks across party lines (notably through the parliamentary Delegations for women's rights). As a result of this successful mobilization, the paragraph opening the possibility for gender quotas was extended and moved from article 3 to article 1 of the Constitution in 2008, without any major parliamentary opposition.
Made possible by this constitutional reform, the 2011 Copé-Zimmermann law imposing gender quotas in corporate boards also resulted from the parliamentary and bureaucratic activism of a coalition of state feminist actors. Marie-Jo Zimmerman notably drew on a 2009 report on equal employment delivered by Brigitte Grésy, former head of the women's rights bureau (Service des droits des femmes et de l'égalité), arguing in favor of a law imposing an obligation of quasi-parity (40%) in a "reasonable timeframe" to feminize the executive boards of major firms.
As the Copé-Zimmermann law was enacted in January 2011, right-wing deputy and member of the Parliamentary Delegation for women's rights Françoise Guégot delivered to President Sarkozy a report entitled "Equal employment between men and women in public administration." One of her recommendations was to adopt quotas in order to produce a "balanced representation" of men and women at the highest levels of the state. She made the case for "an obligation to succeed" in the attempt to increase the share of women, given the "strong initiative" developed in parallel in the private sector through the Copé-Zimmermann law. Alignment with the private sector thus became a strong rationale for the promotion of quotas at the highest levels of the public administration.

Women elite networks promoting the women's elite cause
Mobilizations of MPs and bureaucrats acting within women's policy offices unfolded hand in hand with (equally unobtrusive) mobilizations from members of female upper civil servant organizations. Throughout the 2000s, in a wider context of feminist revival, elite professional women's associations mushroomed. The most visible "women's networks" put together female alumni from French elite universities (Grandes écoles) and corporate executives, lobbying to improve women's access to strategic business positions. Fueled by the parity grammar, they also drew on the "diversity" rhetoric that gained momentum in the French business world by the mid-2000s (Bereni 2009). On the one hand, these networks promoted incentive tools pertaining to "market feminism" (coaching, mentoring, benchmarking, good practices, rankings, etc.) (Blanchard et al. 2013;Kantola and Squires 2012;Pochic 2017). But on the other hand, they also more discreetly mobilized in favor of "constraining measures," including gender quotas. In parallel, similar networks also burgeoned in the public sector. By 1998, the professional women's organization "Administration moderne" (Admod) brought together women upper civil servants, mostly énarques, working in various administrations, to promote an increased representation of women among the bureaucratic elite. These mobilizations at the top of the public administration unfolded against the backdrop of the diffusion of new public management (NPM) in the bureaucracy, presenting the adoption of management practices coming from the private sector as a way to improve government functioning (Bezes et al. 2012). Endorsing this dominant discourse, they argued that the feminization of the Civil Service was a key driver for modernizing the bureaucracy . At the end of the 2000s, while the Copé-Zimmermann bill was to be adopted, the leadership of Admod began lobbying in favor of the adoption of a quota. Two of them were called to give testimony by Guégot in preparation for her report, notably Nathalie Tournyol Duclos, who was to be appointed head of the women's rights bureau a few month later. Admod members also helped minister Sauvadet rally members of parliament and higher civil servants to the idea of quotas behind closed doors.

The adoption of the Sauvadet law: an opportunity under constraints (2011-2012)
The bill François Sauvadet introduced in the Senate on September 7, 2011, was mainly devoted to favoring access to tenure in the Civil Service. Only one-largely symbolic-provision dealt with gender equality: the obligation for ministerial departments to deliver an annual report on gender equality policy in the public service. Even though the government had officially taken a stand in favor of gender quotas in "professional responsibilities" as of 2011, the introduction of a gender quota provision into the Sauvadet law was the outcome of a new intra-institutional mobilization of women's rights advocates. It was also made possible by avoiding the term "quota," which was replaced by the more consensual idea of "balanced representation" (Bui-Xuan 2012).
In 2010, when President Sarkozy commissioned deputy Françoise Guégot to prepare a report on women's access to executive positions in the Civil Service, what was at stake for him was to promote measures that would publicize his presidencies action on gender equality and bureaucratic modernization when he would run for reelection in 2012. A measure such as the quota had the advantage to signal a commitment to gender equality, without entailing any major budgetary cost; this was particularly crucial in the aftermath of a financial crisis. One month before the report was officially delivered in March 2011, the State Secretary for Civil Service, Georges Tron, announced that he endorsed the idea of quotas and hinted at the possibility of financial sanctions in case of non-compliance. 12 On March 8, after a bargaining meeting with public employees unions, he presented the Assemblée nationale with a series of measures, including the gender quota, which would be included into a future bill on "the fight against insecurity" in the Civil Service. Public employee unions strongly called for such a bill in the context of accelerating New Public Management reforms, and their representatives were formally involved in the preparation of the bill. However, the gender quota provision disappeared from the bill that François Sauvadet, Tron's successor (the latter resigned because of charges of sexual assault), 13 introduced in parliament in September 2011.
One reason for the omission was that public employees unions, including gender equality advocates within them, were ambivalent about introducing a gender quota for senior executive positions: even though this policy tool broke with the incentive, soft approach that had prevailed so far, most union representatives saw it as a minimal provision, focused on a narrow administrative elite, and which would turn the attention away from the broader structural reforms needed to favor gender equality in the bureaucracy (equal pay, work-family reconciliation, career assessment, etc.). 14 More importantly, the absence of a gender quota provision in the first draft of 12 "Fonction publique: Georges Tron veut instaurer des quotas de femmes aux postes de direction", Les échos, 26 janvier 2011. 13 Tron was accused of sexual assault by two of his female collaborators in May 2011. 14 As example, see http://ufsec gt.fr/IMG/pdf/Note_avis_negoc iatio n_egali te_v3-1.pdf the bill, while Sauvadet declared himself in favor of the measure, 15 was due to resistances from administrative elites in the upper echelons of the bureaucracy, and even more crucially, to the ambivalence of Prime Minister François Fillon. 16 Still, at the beginning of 2012, women's rights advocates in parliament renewed their efforts to introduce a gender quota provision within the Sauvadet bill. Appointed as Rapporteur for the bill, socialist Catherine Tasca, who had been one of the most prominent advocates of the parity law in parliament at the end of the 1990s, criticized in her report the weakness of the measures regarding gender equality. 17 In the senatorial debate over the bill, communist Senator Brigitte Gonthier-Maurin, president of the Delegation for women's rights, pressed for the introduction of quotas, quoting the Guégot report. 18 While the Senate did not introduce any quota provision when it first adopted the bill on January 26, 2012, Sauvadet announced a few days later that the government would introduce a gender quota amendment during the parliamentary debate. The quota provision would set a target of 40% by 2018, and Sauvadet alluded to possible sanctions. This announcement was made possible by president Sarkozy's arbitration against Fillon's reluctance toward quotas. 19 Women's rights advocates from the right-wing majority, notably deputies Marie-Jo Zimmermann and Françoise Guégot, backed by minister Sauvadet, introduced new gender equality provisions into the bill during its examination in parliamentary commission at the Assemblée nationale. François Sauvadet finally introduced a gender quota in senior executive appointments during the public debates at the Assemblée. As he argued, "even though quotas are no panacea, to this day they remain the only way to really make things change; in the government we are convinced of this." 20 He also stressed the impact of women's rights advocates (mentioning Françoise Guégot and Marie-Jo Zimmermann) in putting the reform on the agenda and changing his own opinion on the issue. 21 Meanwhile, the union bargaining process (which included demands in favor of broader gender equality measures) was put to a halt with the 2012 presidential election looming.
Cheered by women's rights advocates across the political spectrum, backed by promoters of neo-managerial reforms keen on "modernizing" the public 15 "Egalité hommes-femmes: François Sauvadet n'écarte pas l'idée de dispositifs contraignants", La gazette des communes, 22 septembre 2011. 16 "Fonctionnaires: le plan de l'Etat pour promouvoir les femmes", Les échos, 23 novembre 2011. 17 Sénat, Rapport de Catherine Tasca sur le projet de loi relatif à l'accès à l'emploi titulaire et à l'amélioration des conditions d'emploi des agents contractuels dans la fonction publique, à la lutte contre les discriminations et portant diverses dispositions relatives à la fonction publique, 17 janvier 2012, pp. 32-33. 18 Sénat, Journal officiel de la République française, Compte rendu intégral de la séance du 25 janvier 2012, pp. 542-543. 19 "Parité: la révolution culturelle des quotas gagne la haute fonction publique", Les échos, 3 février 2012 ; François Sauvadet, "Il y a encore du chemin à parcourir pour féminiser la haute fonction publique", Le Monde, 8 mars 2019. 20 Assemblée nationale, Journal officiel de la République française, Compte rendu intégral de la séance du 7 février 2012, p. 773. 21 Assemblée nationale, Journal officiel de la République française, Compte rendu intégral de la séance du 7 février 2012, p. 800. administration to align it with private sector standards, the quota was only publicly opposed by conservative deputy Hervé Mariton 22 and was adopted by a large majority. While the consensus over the Sauvadet quota can be accounted for by the diffusion of the parity grammar within institutional circles, it could also be explained by the technical restrictions of the legislation. First, the statute is provided for a gradual implementation of the quota, going from 20% in 2013 to 40% in 2018. This contributed to convince skeptics about the "feasibility" of the provision and dismissed the argument that "there are not enough female candidates" to these positions. This incremental framework was also in line with the mandated quota introduced a year earlier in private sector corporate boards; this policy precedent helped mainstream the idea of gender quotas as a relevant policy tool. Second, the choice to target only the flux of "first appointments" (unlike the Copé-Zimmermann law on corporate boards), rather than the whole stock of senior executive positions, reassured already appointed men regarding the absence of threat on their personal situation, as clearly stated at the time by François Sauvadet: "We need to make sure that men who have already been appointed will be renewed within the limit defined by Civil Service regulations, or that they will be appointed to a similar position. For example, a prefect moving to another department must escape the quota system." 23 This consensus, however, was no more than a public façade; Sauvadet recently reported how higher civil servants and members of parliament discreetly but vehemently opposed the quota: "nobody can imagine what I had to face, as minister for Civil Service: the resistance was as numerous as it was fierce, coming from all administrative bodies, with relays on the benches of the National Assembly and the Senate." 24 In this respect, what the public consensus reflected was also the increased illegitimacy of an open criticism of the gender quota, in the context of its diffusion as a tool of gender equality policy.

A coercive and monitored policy tool
Following the parliamentary debate, the Sauvadet legislation of March 12 introduced a series of mandatory gender quotas in several decision-making positions in the Civil Service: in executive boards of public institutions, in entrance examination committees, in joint administrative committees, in Civil Service advisory boards, and, last but not least, in appointments to senior executive positions. The statute mandated a minimal representation of "each sex category" of 20% in 2013, 30% in 2015, and 40% in 2018 in appointments to senior executive positions. The law provided that the quota was to be reached on a yearly basis and at the level of 23 Ibid., p. 843. 24 François Sauvadet, "Il y a encore du chemin à parcourir pour féminiser la haute fonction publique", Le Monde, 8 mars 2019. 22 Assemblée nationale, Journal officiel de la République française, Compte rendu intégral de la séance du 8 février 2012, pp. 840-841. each ministerial department 25 for initial appointments, that is, "excluding cases of renewal of the person in the same position or appointment to a similar type of position." In the National-level Civil Service, the gender quota applied to first appointments among about 3600 senior executive positions (including government appointments). 26 Finally, the law provided for financial penalties to administrations that would not comply. This policy tool thus appeared as a negative incentive: sanctions would be applied in case of non-compliance with the goal (Ingram andSchneider 1990 cited in Engeli andMazur 2018, 114-15).
However, even negative incentives do not necessarily have a binding character, depending on if and how they are implemented. A closer look at implementation is therefore needed in order to better characterize this quota. The first striking element in the implementation of the Sauvadet quota is that the implementing decree was adopted quickly, unlike what is often the case for gender equality legislation. The Sauvadet implementation decree was adopted less than 2 months after the law was passed. It provided a precise list of the positions included in the perimeter of the law, and set the amount of the fine to 30.000€ for each "missing unit" (that is, for each woman-or man-missing in order to fill the quota) in 2013, 60.000€ in 2015, and 90.000€ in 2018. The decree therefore confirmed the coercive potential of the law by giving more substance to the idea of sanctions in case of non-compliance.
Moreover, the decree specified a mechanism of accountability and monitoring. By April 30 of each year, administrative agencies were required to deliver the list of initial appointments from the previous year to their assessors and to the ministry for Civil Service, as well as the sex ratio of these appointments and, if applicable, the amount of the fine due. A "Senior executive mission" hosted by the General Secretariat of the Government (SGG in French) was to be set up to collect the information for each ministerial department and take steps to "favor the attainment of this goal." The Civil Service and administration general management (Direction générale de l'administration et de la fonction publique, DGAFP) was to circulate each year a public report documenting the sex ratio for new appointments and, if applicable, the amount of the fine due.
Arguably, the existence of precise amounts of fines and the definition of a mechanism of monitoring and accountability could favor a strong implementation, leading to a policy situation of full coercion. The profile of the organizations in charge of implementation (SGG and DGAFP), however, could introduce doubts in this respect. Indeed, after being promoted by feminist advocates, quotas were integrated as part of administrative routines. Their implementation and the supervision thereof were assigned to Civil Service administrative services, far from state feminist institutions. Neither the women's rights bureau, nor the parliamentary Delegations women's rights, nor networks of women's higher civil servants, were formally included in the policy implementation process. Initially conceived as a tool to fight against gender inequalities, quotas thus became an administrative tool "just like any other," implemented by "laypeople in the field of equality" (Perrier 2015), whose mission is defined in terms of legal compliance rather than in terms of feminist advocacy.

A feminist implementation by mainstream bureaucratic actors
The fact that quota monitoring was not devoted to Women's policy structures did not impede a strong implementation. Each actor followed the letter of the law. A procedure was defined to facilitate the declaration of nominations, with a spreadsheet made available on the Civil Service ministry website (see image below). Men and women were counted, and at the end of 2016, for the first time since the introduction of the quota, two ministers-the Department of Justice and the department of Defense-were imposed financial penalties for non-complying with the 30% quota.
How can we account for such a diligent implementation on the part of state actors not endowed with a feminist mission? The format of the policy instrument at stake certainly helped: as argued earlier, the quota is a very simplistic and rather straightforward translation of the ideal of equality. The margin of interpretation is limited. Hence, the characteristics of the implementing actors seem less likely to influence the outcome. While an absence of implementation could have been a possible outcome, the instrument was implemented as planned. Moreover, the DGAFP's involvement in the implementation process was not limited to monitoring the gender breakdown of new senior executive appointments. In its yearly reports on the Sauvadet quota implementation, the DGAFP developed more qualitative comments, pressing in favor of a more coercive implementation process. In its 2016 report, it pointed to two main limits of the legal provision. 27 On the one hand, ministries were expected to voluntarily pay their fines, but no enforcement procedure was provided for in case of a failure to pay, so that "the coercive potential of the sanctions may turn out limited." On the other hand, the DGAFP pointed out that the fines contribute to the state's general budget, when they could be used to fund gender equality policy in the Civil Service. Such comments would typically have been expected on the part of women's policy agencies, and here they were voiced by the DGAFP. In other words, not only did mainstream administrative actors implement the law as planned, but they also pressed in favor of its reinforcement, fueling a policy feedback effect (Pierson 1993). In GEPP terms (see Engeli and Mazur, this issue), implementation was marked by high substantive empowerment without any clear evidence of descriptive empowerment.
Does this mean that feminist advocacy had no influence on the process? Such a conclusion should be nuanced in several ways. First, this success in the mainstreaming of gender equality concerns in policy implementation can be analyzed as the result of the long-term diffusion of feminist ideas in non-feminist circles (Albenga et al 2015): beyond feminist advocacy structures, people working in mainstream bureaucratic settings such as the DGAFP in the 2010s had been socialized to some extent to feminist ideas, through higher education for the youngest (via the diffusion of gender studies programs at the university level) and through the mass media or through peer influences. Moreover, the adoption of the Sauvadet law coincided with an upsurge in feminist mobilizing in civil society, which increased this diffusion of feminist ideas in mainstream settings (notably through increased discussions of feminist ideas in the media).
Against the backdrop of the 2012 presidential campaign, feminist groups pressed in favor of the creation of a strong Women's policy agency, a new "Ministry of women's rights" (Ministère des droits des femmes), in reference to the Roudy's episode in the 1980s. François Hollande had committed to this prior to his election. Once elected, he appointed Najat Vallaud-Belkacem as minister for Women's rights, at a high rank and with all the more visibility since Vallaud-Belkacem was also the government spokesperson, and soon became very present and vocal in the media. Her position as government spokesperson helped reinforce the legitimacy of women's rights as a policy issue, which she also made sure to embed more firmly in the different ministerial departments, with the appointment of senior executives in charge of gender equality (hauts fonctionnaires à l'égalité) in each ministry. While marginal, these positions helped diffuse feminist ideas in the different ministerial departments. In parallel, both Women's policy agencies and the more informal feminist networks described in the previous section contributed to monitoring the implementation of the Sauvadet quota by favoring the maintenance of a perceived risk of shaming in case of non-compliance.
This context favored the reinforcement of gender equality policy in the public sector, and even the reinforcement of the quota itself. The government signed with public employees unions a "protocol agreement on equal employment in the Civil Service" in 2013, and a new "Charter for the promotion of equality and the fight against discriminations in the Civil Service" was adopted that same year. The 2014 law on "real equality between women and men" reinforced the provisions of the Sauvadet law: it increased to 50% the minimum share of people of each sex in the executive boards of public institutions (extending to the public boards the principle of the Copé-Zimmermann law), and set for 2017 (instead of 2018) the obligation of at least 40% of appointees of each sex for new appointments to the highest positions.
Finally, the 2018 protocol agreement on equal employment in the Civil Service created a separate gender employment equality fund, meant to receive the fines paid by non-compliant administrations. 28 The new fund was to fuel other gender equality initiatives in the public administration, as advised by the DGAFP report mentioned above.

A restrictive yet coercive tool with incremental effects
The Sauvadet law introduced a coercive policy tool, with financial sanctions in case of non-compliance, but restricted it to a limited number of positions: out of about 3600 positions in the State Civil Service workforce within the perimeter of the law, only a few hundreds were filled with first appointments each year. The quota applied to 641 initial appointments in 2013 and 587 in 2017.
Between 2013 and 2016, first appointments to senior executive positions within the perimeter of the Sauvadet quota were systematically filled with 30-33% of women. With the exception of the departments for Justice and Defense, all ministerial departments complied with the legal obligation during the four first years of its implementation. Yet, at the end of 2017, the first year of implementation of the 40% quota, the rate was only 36%, with 6 out of 11 ministerial departments being in non-compliance. The Ministry of Economic and Financial Affairs (Bercy) alone, with the lowest rate (26%), had to pay a contribution of 1.710.000€ (out of a total of 2.340.000€). 29

A mitigated quantitative effect: a smoothed curve
It is not so easy to assess the extent to which the Sauvadet quota has led to an increased share of women among senior executive positions of the State Civil Service. Since the perimeter of application does not match any preexisting administrative and statistical category, it is impossible to have a precise assessment of the gender breakdown in these positions prior to the implementation of the law. It is therefore difficult to assess the quantitative impact of the measure. Available figures since the early 2000s point to the hypothesis of a limited impact: rather than a turning point increasing the pace of feminization of senior executive positions, the Sauvadet quota led to securing and smoothing the curve of feminization. According to Civil Service reports between 2015 and 2017, the share of women among senior executive positions within the perimeter of the Sauvadet quota went from 26 to 28%. This demonstrates the limitations of a policy that only targets the flux (first appointments) rather than the stock (all the positions).
Moreover, if one looks at statistical categories that had been, until the Sauvadet law, used as indicators of the presence of women in the administrative elite (concerning around 1.900 senior executive positions included in the perimeter of the law), reports show a very limited impact of the Sauvadet law on the rate of women. As shown in the graph below, and mentioned earlier, the share of women in senior executive positions started increasing as of the middle of the 2000s, and particularly as of 2007, when Sarkozy came to office and hastened the pace of new public management reforms. In the years preceding the Sauvadet law, women promoting gender equality at the highest level of the bureaucracy had already managed to frame feminization as a key marker of the administration's "modernization" policy. The share of women in all senior executive positions increased by 1.8 points (from 21.9 to 23.7%) between 2007 and 2012-prior to the Sauvadet law-and by 2.1 points (from 25.1 to 27.2%) between 2013 and 2016 (the last year with available figures). This comparison of pre-and post-Sauvadet law levels of feminization suggests a limited immediate impact of the quota: its main effect was to secure and smooth an underway pattern of feminization of senior executive positions, rather than to accelerate its pace. The gradual dimension of the Sauvadet quota (rising from 20% in 2013 to 40% in 2017, and applying to first appointments only) partly accounts for this mitigated quantitative impact 5 years after the law was passed. While it does not entail an immediate, sharp increase in the gender ratio among senior executive positions, it allows for an incremental increase. In order to increase the potential impact of this policy tool, networks of higher executive women in the Civil Service started lobbying in favor of a new law by 2017, demanding more coercive measures, notably increasing the perimeter of the positions considered and increasing the fines. 30 The context, however, was not favorable. The implementation of the quota at the level of 40% for the first time in 2017 reactivated a strong, albeit invisible, hostility toward the measure: "we are still in a hostile environment," argued Admod president. 31 These oppositions remained very discreet: seldom publicly stated, they took the form of unobtrusive resistance on the part of men in the higher ranks of the public administration, as informally reported by several women's network's members. This resistance had concrete effects, translating into a regulatory backlash. First, following its poor results in 2017 (26%), the minister of Economic and Financial Affairs, in the hope of improving compliance without having to appoint more women, signed a decree reducing by one-third the perimeter of positions concerned by the Sauvadet quota on December, 27, 2018. Second, more significantly, the 2019 law on the public service 32 introduced several technical modifications that significantly weakened the law's impact. It notably legalized the established practice of the "rounding down" in setting the number of new women's appointments necessary to achieve 40%. Given the low numbers at stake, this rule leads to allow the appointment of less than 40% of women. For example, in the case of nine new appointments to fill, at least 3, 6 persons of the underrepresented sex would have to be appointed, theoretically; yet, in practice, the "rounding down" rule authorizes to appoint only three women (33%), rather than four (44%).

Beyond numbers: gender transformation through symbolic impact
Beyond the limited quantitative impact of the Sauvadet quota, one should also stress its intrinsic limitations as a gender equality policy tool. Targeting only a small number of elite women, the quota mainly represents an opportunity for women who cumulate a series of resources (notably in terms of social class and education). It does not act upon the organizational mechanisms that produce the glass ceiling (Rouban 2013;Jacquemart et al. 2016;Marry et al. 2017). These limitations reflect the elitist profile of the feminist actors who have been advocating in favor of gender quotas in decision-making sites since the 2000s.
Nevertheless, the Sauvadet quota did not leave the gender regime unchanged (Connell 2006;Acker 2009) of the French Civil Service. Indeed, gender transformation does not boil down to quantitative measures: the symbolic dimension of the tool must be taken into account, and the implementation of the quota also paved the way for significant gender transformation in this respect. First, the Sauvadet law and concomitant "soft" gender equality reforms (labels, charters, etc.) turned the question of the glass ceiling in the Civil Service into a public problem (Gusfield 1981). For example, the fines imposed to the departments of Justice and Defense in 2017 received extensive media coverage, stressing the shocking persistence of gender inequality. Second, the introduction of gender equality policies within the bureaucracy 30 See for example the joint press release by 9 organizations on June, 1st, 2018. 31 Entretien avec Nathalie Pilhes, AEF Info, 14 février 2018. 32 Loi no. 2019-828 du 6 août 2019 de transformation de la fonction publique. favored the denunciation of gender inequality. For sure, women's capacity to be vocal on these issues depends on their professional position (often correlated with their upper-class background and/or education in the most prestigious training tracks such as the ENA). However, the emergence of gender equality claims among women bureaucrats reflected the new legitimacy of a gendered perspective on careers within the public administration, challenging a long tradition of a gender-blind universalism (Marry et al. 2017). Third, the Sauvadet quota and other gender equality policies also impacted men. While in the first years of its implementation, the quota had a limited impact on men's odds to access higher bureaucratic positions, the new visibility of gender equality goals in the Civil Service had a symbolic effect on men: as the most conservative attitudes toward the gender equality agenda were delegitimized, men holding executive positions were to redefine their professional identities in relation to gender equality norms .

Conclusion
At first, by introducing a minimum quota of 40% of persons of each sex in new appointments to the highest administrative positions, the Sauvadet law appeared like a turning point, away from the "symbolic policy" (Mazur 1995) that had been promoted by the French state so far. Analyzing the details of this policy tool and its implementation, however, leads to a more complex assessment.
A closer look at the immediate effects of the measure points to the idea of a limited gender transformation: first, because the increase in the share of women in senior executive positions had in fact started prior to the enactment of the law, and second, because of the limitation of the quota's perimeter to "first appointments," rather than the whole stock of senior executive positions. This very technical feature, however, has endowed this tool with a potential to foster incremental social change. In other words, its effect was limited at first, but it is likely to increase over the years, as appointments made since 2012 represent a larger share of the stock of senior executives. The gender quota's relative strength also derives from the continuous involvement of state feminist actors who, in the years following the adoption of the law, kept lobbying to maintain its feminist framing. In this respect, a few years after its introduction, this reform cannot be categorized as a "policy failure," unlike other cases of gender quota reforms (Verge and Lombardo 2019).
The persistence of this positive gender transformation, as shown by the first years of implementation, will depend upon a close monitoring of the process by women's networks and feminists within the state apparatus, even though Women's policy agencies are not formally part of the implementation mechanism. Indeed, as illustrated by the 2018 decree adopted by the Ministry of Economic and Financial Affairs (Bercy) to limit its own obligations toward the law and by the technical modifications introduced in 2019, legal backpedaling always is a threat. On the other hand, the government's endorsement of a coercive provision is likely to have a symbolic effect. It fuels a representation of inequalities as illegitimate, while legitimizing their denunciation and the state's intervention to correct them. This strong gender equality statement, however, also runs the risk of diffusing an idea of equality as "already in place." The evolution of both the legislation and its implementation in the future years will confirm whether the quota has a transformative potential.
From a more theoretical perspective, this case study reveals the limit of focusing on the genesis and effects of a single reform to assess the transformation of gender relations. Indeed, as we have seen in this case, the same factors [mainly institutional feminist mobilization in the aftermath of the parity laws, and the broader transformations of the women's cause field (Bereni 2019)] that help explain the adoption of the quota also influenced the feminization of the higher levels of the public administration prior to its adoption, thus conversely limiting its direct measured impact. Moreover, the adoption of the Sauvadet law cannot be explained short of the broader context of the diffusion of the quota as a gender policy tool in other sectors, through other legal reforms (Lépinard 2016). Finally, the fact that the Sauvadet quota was simultaneously reinforced (for example, by the creation of the equal employment fund in 2018) and weakened (for example, by the Bercy decree of December, 2018 and the 2019 law on public service) in the aftermath of its adoption points to the ongoing political struggles around this policy tool.