Introduction

The introduction of quotas is meant to help women overcome certain obstacles to political office that thwart an egalitarian representation. Several countries have adopted positive action in order to grant a gender-balanced composition of political office (see Krook, 2009). Quotas address a particular structure within the political process: the vertical distribution of power within institutions (Krook and Squires, 2006, p. 8). The efficacy of quotas in promoting the election of women depends on the interaction of the reforms taking place in specific sets of institutions, be it normative institutions (definitions of equality at the constitutional, legal, party or public opinion level), systemic institutions (electoral rules, district size and so on), or practical institutions that inform candidate selection (Krook, 2010). It also depends on the willingness of individuals within the system to act on the opportunities inherent in these structures (Davidson-Schmich, 2006, p. 228; Holli et al, 2006).

Parties are the main ‘gatekeepers’ to political office in most countries and their decisions are structured within the context of formal party rules and informal norms and practices (Norris and Lovenduski, 1995; Kenny, 2011). Likewise, political parties are not gender-neutral structures and institutionalise definitions of femininity and masculinity, arrange gender hierarchies and define gender-appropriate jobs (Lovenduski, 2005). Indeed, research into the implementation of quotas shows that their effectiveness is constrained by several discriminatory manoeuvres that parties use to circumvent the law (Troupel, 2002; Bird, 2003; Murray, 2008; Verge, 2010). Therefore, as Mateo Díaz and Millns (2004, p. 301) argue, ‘even where parity measures are introduced at the highest legal and constitutional levels they have little chance of producing effective outcomes where they operate within a climate of gender biased political structures and do not adjust to national electoral systems’.

This article aims at unveiling the broad array of strategic discrimination practices that political parties apply, which diminish the equalising potential of positive action. In order to examine parties’ gendered practices, the article builds on a comparative analysis of France and Spain. Both cases make for an intriguing comparison. Although a quota law informed by the principle of ‘parity democracy’ has been recently introduced in both cases, these countries have followed different routes in their respective quota reform processes. Whereas France took the ‘fast track’, seeking to enact overnight change through the introduction of legislative quotas, Spain followed an ‘incremental track’. In this case, the country experienced a steady increase in the levels of female representation over a considerably long time lag thanks to the adoption of targets and party quotas and a statutory quota was only introduced very recently (for a review of routes to women's representation see Freidenvall et al (2006) and Tripp and Kang (2008)). Differences between the two countries also stem from the dissimilar leading role of social and political actors within the movement for parity, its constitutional fit, as well as the extent to which party quotas adopted before the parity laws succeeded in their goal.

The article is structured as follows. The first section compares the dynamics of the mouvement pour la démocratie paritaire in France and Spain. It briefly presents the advocacy coalitions that supported quota reforms and the respective legislative paths, including the differences in the definition of equality that informed the reforms and the requisites imposed to political parties. The following section describes the evolution of women's presence prior and after the implementation of statutory quotas. The fourth section examines the interaction between institutional constraints and gendered party practices. The final section summarises the main findings and reflects on potential corrections to quota laws with a view to prevent partisan strategic (ab)uses.

Parity Movements and Quota Laws

Whereas in France the movement for parity basically came from a bottom-up organisation of women activists, feminist scholars and female politicians from the entire political spectrum (Bereni, 2007, p. 197), in Spain the initiative was promoted top-down by left-wing political parties with the support of the feminist movement, although the latter played a secondary role (Valiente, 2005). Contrary to France, when the quota law was passed, there was already a certain degree of comfort with the notion of gender quotas – the left using the language of parity when calling for quotas since the mid-1990s. Therefore, the leap from party to legal quotas did not require the terminological switch that parity advocates had to make in France (see Verge, forthcoming).

In France, despite the fact that the Parti Socialiste (PS) had already adopted voluntary quotas over the course of the 1970s and 1980s they were never implemented to their fullest extent. Lack of enforcement mechanisms and exclusive application to elections held under proportional representation (PR) produced very low gains in women's political presence. Smaller left parties, such as the Greens and the French Communist Party (PCF), presented a stronger commitment and included a high proportion of women among their candidates and elected representatives. Conversely, quotas were not accepted by right-wing parties until the mid-1990s (Opello, 2006).

By the beginning of the 1990s, the claim for parity was framed in terms of citizenship and it engaged female elected officeholders as well as female social activists. Calling for parity presented several rationales. First, equality among the two constituent parts of citizenship, men and women, adapted to the republican motto – albeit slightly changed to ‘liberté, égalité, parité’ (see Gaspard et al, 1992). Second, parity advocates defined gender difference as a universal one, accepted by the French constitution (Bereni, 2007, p. 199). Third, many proponents of parity explicitly rejected quotas due to the annulment by the Constitutional Court in 1982 of the 25 per cent quota for municipal elections on the grounds that the equality principle established in the French constitution did not permit to divide the electoral corps by categories (Bird, 2003). In order to overcome opposition to positive action, activists promoted simple equality. Finally, the parity rhetoric assumed a ‘pragmatic bypass’ emphasising higher representation of women would change the way of doing politics ‘as usual’ (see Lépinard, 2006, pp. 37–40).

By 1995, parity activists undertook signature campaigns and instigated joint actions with women's associations. The increasing strength of the movement forced parties to take a stance on the issue and to invest in more women candidates. In 1994, the PS presented, for the first time, zip lists (where men and women alternate from top to bottom of the list) for European elections and in 1997 it assumed a 30 per cent quota for national elections (although the percentage of women elected just reached 17 per cent), which fostered a contagion effect on the party system (Jenson and Valiente, 2003, pp. 75–82).

Conversely, in Spain, the role played by party feminists within the main national left-wing party, the Spanish Socialist Workers’ Party (PSOE), has been decisive in boosting the movement for parity democracy through the adoption of voluntary quotas in the late 1980s (Verge, 2006; Threlfall, 2007). Most of the Spanish women's movement has involved party feminists, and therefore access to elected office very soon came as one of its top priorities (Jenson and Valiente, 2003, p. 85). The adoption of quotas in Spain followed an incremental process. First, left-wing parties introduced a 25 per cent quota for women in party and elected offices. This quota was enlarged over the 1990s and it reached parity (40 per cent/60 per cent for either sex) in 1997. Second, as it has been found in other countries, a ‘contagion’ effect took place. By 2004, all parties but the conservative Popular Party (PP) had adopted either gender quotas, targets or recommendations for women's representation, which eased the way to the adoption of a statutory quota in 2007 (Verge, 2011).

In both cases the parity movement was heavily influenced by international events, such as the Athens Summit Women in Power (1992), the United Nations Women's Conference (1995) and the IV Plan of Action on Equality of Opportunities of the European Union, which urged states to implement measures to make equality effective. The parity movements were also helped by their ties to the European Women's Lobby. Therefore, international tipping played a key role in instilling quota reforms (see Krook, 2006).

Let us now turn to the specific legislative path followed by statutory quotas in each country. In France, political parties were neither able to reform themselves or to effectively apply parity (Appleton and Mazur, 1993). A law appeared to be the only way to put an end to women's under-representation. As a matter of fact, several laws were necessary to increase the numbers of elected women. The political change that brought the left to power in June 1997 precipitated the reform, as one of the central issues trumpeted by PS leader Lionel Jospin was the renewal of political institutions – that is, parity and term limits (Sineau, 2002a, p. 2). The first law, the constitutional amendment of July 1999, set the framework within which the parity reform would be based. The revised article 3 of the Constitution states that ‘the law favors equal access for women and men to electoral mandates and executive functions’. Then it stipulates (article 4) that political parties ‘contribute to the implementation [of this principle] under the conditions set by the law’. The word ‘parity’ does not explicitly appear. This switch from ‘ensuring’ equal access to ‘favoring’ equal access was not trivial as it turned the parity law into a policy of equal opportunity and not a policy of positive action, and neutralised the argument used by the Constitutional Council in 1982 when the quota was annulled (Halimi, 2003; Scott, 2005; Bereni, 2007). Furthermore, Prime Minister Lionel Jospin conceded that the electoral system would not be changed. This obeyed to the need to reach a consensus with right-wing parties, which held the majority of seats in the Senate. Whereas the left considered the law should leave no loophole for political parties, the right preferred instead to provide incentives. This concession seriously affected the effectiveness of the parity law (Troupel, 2006).

In June 2000, a new law tailored the application of parity to the various types of elections. Concerning regional and local elections, electoral authorities do not accept party lists that fail to reach parity (50 per cent of each sex) in every consecutive stretch of six candidatesFootnote 1, as Table 1 indicates. At the local level, the law only affects municipalities with over 3500 inhabitants. These towns represent about 7 per cent of French municipalities, although they include two-thirds of French citizens (Bird, 2003, p. 14). With respect to European elections and senators elected under the proportional system zip lists were introduced. The cost of defection is also the withdrawal of the party list. Regarding the lower house, parties failing to comply with the required provisions see a portion of their public subsidy withheld but remain eligible to contest the election. The economic penalty is calculated on the difference between male and female candidates listed by parties. The public subsidy linked to the number of votes obtained in the first round is reduced by a percentage equivalent to half the difference between the number of candidates of each sex with respect to the total number of candidates. The weakest provision for the National Assembly illustrates the extent to which the main parties were resistant to change the status quo (Bird, 2003), specifically by being unwilling to deselect male incumbents given that they could more easily absorb the losses in state funding than minor parties (Murray et al, forthcoming).Footnote 2 Cantonal elections remained unaffected by the 2000 law (Troupel, 2005).

Table 1 Parity reform process in France (2000–2007)
Full size table

Finally, in January 2007 some corrections were made, although the main shortcomings affecting the implementation of parity for deputies and senators elected by majority vote were not addressed (Troupel, 2007a, 2007b). Zip lists were introduced for municipal elections and parity was extended to regional and municipal cabinets (50 per cent of each sex). Regarding cantonal elections, parties have to include a substitute for the leading candidate who is of the opposite sex, otherwise the candidacy is annulled. Last, the financial penalty applied for legislative elections held from 2012 onwards was increased.

As regards Spain, starting in 1996, left-wing parties presented several bills aimed at increasing women's presence in decision-making positions but on all occasions conservative MPs, who held the majority of seats in the lower house, blocked their discussion. Regional laws incorporating parity, passed in 2002, were not enforced in the 2003 regional elections as the conservative national government lodged an appeal to the Constitutional Court that effectively suspended their implementation (Verge, 2008, p. 128). In 2004, the PSOE went back into central government and equality policies were put at the forefront of the political agenda. On 15 March 2007, with the support of all parties but the PP, the Organic Law for the Effective Equality of Women and Men was passed. It reforms the General Electoral Regime Law forcing parties to incorporate a minimum of 40 per cent and a maximum of 60 per cent of either sex into candidates’ lists for all elections. As Table 2 shows, this proportion must be respected in each stretch of five candidates. For Senate elections, parties present three-candidate lists, and therefore at least one woman has to be included. Electoral authorities do not accept party lists that fail to meet the quota provisions. This applies to all elections. The PP lodged an appeal to the Constitutional Court, which was eventually rejected as the 1978 Spanish Constitution explicitly mentions political participation among the aspects public authorities not only must remove the obstacles preventing or hindering its full enjoyment but also promote and facilitate the conditions to make equality of all individuals effective (article 9.2). It also supported the constitutionality of all regional quota laws, including those that had introduced zipping (see Verge, forthcoming).

Table 2 Parity reform process in Spain (2007)
Full size table

The Implementation of Quota Laws: Outcomes and Institutional Constraints

In this section we present the evolution of women's numerical representation at the four layers of government in France and Spain (local, regional, national and European) since the late 1970s. On the one hand, we describe the institutional constraints, in particular the different electoral systems in use, which favour or hinder women's representation and the effectiveness of quotas. On the other hand, we compare women's representation before and after the quota laws came into effect.

Institutional factors

In France, several electoral systems coexist (see Table 3). European elections are the only ones that use PR. Local and regional elections (both held every 6 years) combine PR with majority vote and the winner's list automatically obtains a quarter (regional elections) or half (local elections) of the seats. The remainder of seats is allotted using PR. Owing to this majority bonus, few lists get enough number of seats to attain parity. Given that lists are often led by male candidates, women's chances to get elected are reduced. Both PR and majority vote are in use for the Senate (Sénat) elections: whereas in the smaller departments senators are elected by a two-round simple majority system, in the most populated departments they are elected by PR. In addition, the upper house is partly renewed by thirds every three years for a 6-year term.Footnote 3 The lower house (Assemblée Nationale) is elected for a 5-year term in single-seat constituencies (two-round runoff). This is by far the most difficult election for new candidates as incumbency is a key asset (Sineau, 2002b; Murray, 2008).

Table 3 Electoral systems
Full size table

In Spain, elections at all levels are held with closed and blocked party lists and under PR (D’Hondt system), with the exception of the upper house (see Table 3). Yet, proportionality is diluted at the lower house (Congreso de los Diputados) due to low district magnitude, with an average of 6.7 seats. At the regional level, proportionality is higher precisely because of larger district magnitude (Llera Ramo, 1998). Municipal elections take place on the same day throughout the country. Each town forms a single district. The number of council members is proportional to the population of the municipality. For European elections, the whole country constitutes a single district. Regarding the upper house (Senado), 208 senators are directly elected by voters while 56 of them are appointed by the parliaments of the 17 Spanish regions (Comunidades Autónomas). Regarding elected senators, in general, parties nominate three candidates and voters cast three votes, although four senators are finally elected. Typically, the three most-voted candidates come from the same party, and the fourth senator is the most-voted candidate in the runner-up (Montabes and Ortega, 2002).

Existing literature on electoral systems and minorities’ representation has highlighted three features: first, district magnitude and minorities’ representation are strongly connected. In other words, intra-party competition over available offices is critical to female representation (Alozie and Manganaro, 1993; Lovenduski and Norris, 1993). The fewer offices there are in dispute, the more difficult it is that the (political) majority – men – shares power with the minority – women. Second, proportional systems applied in multinomial districts tend to advantage minority groups whereas majority systems produce a low representation for these groups (Rule and Zimmerman, 1994; Welch and Studlar, 1996; Norris, 1997; Farrell, 2001, p. 166). Third, quotas are most successful when applied to closed party lists and when the minimum percentage quota is coupled with placement mandates (Htun and Jones, 2002).

The salience of elections may also affect women's chances to be recruited (and subsequently elected). In Bird's terms (2003, p. 15), ‘the greater the power at stake, the less likely it is to be ceded voluntarily’. Taking these considerations into account, we can now turn to examine which institutions have elected more women over time.

Women's numerical presence prior and after quota reforms

In France, the proportion of female regional councillors increased from 12 per cent in 1992 to 27.5 per cent in 1998, and it reached 47.6 per cent in 2004 when the quota was applied, as Table 4 shows. Female European deputies held 22.2 per cent of seats in 1979 up from just 7 per cent in previous elections, when the PS introduced the first quota. In 1999, the French delegation to the European Parliament already counted with 40.2 per cent of women. The equality law served to reinforce the ascending trend, and in 2004 women constituted 43.6 per cent.

Table 4 Women's representation in France, 1979–2008
Full size table

Regarding municipal elections, the parity law triggered a massive effect. In municipalities with over 3500 inhabitants, the percentage of female councillors almost doubled in one term (from 26 per cent in 1995 to 47.3 per cent in 2001). A positive effect was also observed in towns with under 3500 residents, which are not subject to the law. Since 2001, over 30 per cent of women have been elected – 32.2 per cent in 2008 (Troupel, 2011). However, whereas 56 per cent of elected male councillors were not affiliated to any political party this was the case of 72 per cent of female councillors, thus the former have greater access to political networks and supportive ties (Sineau, 2001, p. 273; Lépinard, 2006, p. 51).

The national level has remained much more masculinised. Up to the 1990s, women made less than 6 per cent of national deputies. Women's presence increased to 10.9 per cent in 1997 thanks to the application of a voluntary quota by the PS and it only mounted to 12.3 per cent when the quota law was applied in 2002. Political parties preferred to pay the financial penalty rather than substantially increasing the numbers of female candidates and claimed not to find a sufficient number of qualified candidates (Murray, 2004, p. 349). The upper house has followed a similar evolution: female senators represented less than 6 per cent of the Senate until 2001. In that year the effect of the parity law on the seats renewed was evident, increasing women's presence to 10.6 per cent. The two next partial elections (2004 and 2008) also boosted women's representation although female senators still constitute less than a third of upper house members.

In Spain, despite their voluntary nature, party quotas have made a significant difference at all layers of government, their impact being more limited for the upper house elections. As it can be seen in Table 5, female representation increased in the national lower house from 6.3 per cent in 1982 to 12.9 per cent in 1989 after the 25 per cent quota was adopted by left-wing parties. Women's presence in the lower house reached 28.3 per cent in 2000, 3 years after the quota was enlarged to the 40 per cent/60 per cent proportions by the left, and it mounted to 36 per cent in 2004, when all parties but the right-wing PP had adopted either gender quotas or targets. The quota law maintained this percentage almost intact, 36.3 per cent. Given that the political context was already quite egalitarian, the contribution of the statutory quota to the increase of female deputies was rather secondary, in tune with Meier and Verlet's findings (2008).

Table 5 Women's representation in Spain, 1979–2008
Full size table

In the upper house it was not until 1999 that a significant increase occurred. The Senate has performed well below the levels of women's representation found in the lower house due to the constraining effect of the plurality system. Yet, the quota law instilled a significant increase in 2008 when female senators mounted from 25.1 per cent to 30.5 per cent.

Regional parliaments have followed a similar evolution, which is also explained by the role played by party quotas. Women's representation increased over 500 per cent in two decades, from 5.6 per cent in 1983 to 35.6 per cent in 2003. Progress was especially significant in 1999 when PSOE's central leadership exerted a strong pressure over regional party branches with a view to comply with gender equality in representation (Threlfall, 2007). In 2007, when the quota law was applied, women obtained 42.8 per cent of regional seats. Only 6 out of 17 regions failed to reach at least 40 per cent of female deputies, but all of them experienced a considerable increase and two of them even elected more women than men. In 2011, parity was also attained in all but four regional parliaments with an average of 41.1 per cent women elected. Owing to larger district size, regional parliaments have achieved a gender-balanced composition, whereas the national lower house has not.

With regard to the local level, we observe a slow but steady increase from the late 1970s to the most recent elections. The sharpest increase was also found in 1999, with 21 per cent of female councillors elected, after the parity quota was assumed by the left. In 2003, women already obtained 25.6 per cent of local offices. In 2007, representation of women in cities with over 5000 inhabitants, where the gender quota was mandatory and which account for 85 per cent of the national electoral census, reached 39.4 per cent (compared to 32 per cent in 2003). But 84 per cent of Spanish municipalities have less than 5000 inhabitants, so the average of women's representation for all municipalities was reduced to 31 per cent. Towns with under 5000 inhabitants only elected 27 per cent of women (FEMP, 2007, p. 80).

Finally, with respect to European elections, very soon after Spain's membership (1986) women obtained a significant proportion of seats, reaching 32.8 per cent of seats in 1989. Successive elections did not produce a relevant variation in women's levels. In 2009, the Spanish delegation incorporated 36 per cent of female deputies, despite the fact that party lists included 49 per cent of women candidates.

In both countries, the European Parliament is the arena in which more women have traditionally been elected. Large district size (eight districts in France and a single district in Spain) and PR are obvious facilitating conditions. Furthermore, being a European deputy does not help to build a political career. First, it is an off-centred institution. Voters do not usually know European deputies and distance from the most relevant national power centres might erode politicians’ networks. Second, the European mandate is more restrictive than other institutional mandates with regard to the accumulation of offices, with men generally accumulating more mandates (for example, senator and mayor).

Partisan Strategic Discrimination Practices

In this section we look at the interaction between systemic and demand-side factors, those that depend on party selectorates (see Lovenduski and Norris, 1993); that is, we pay attention to the combination of institutional constraints and partisan strategic discrimination practices that reduce the effectiveness of quotas. In doing so, the comparative framework allows us to investigate what variables are more important in explaining strategic discrimination, namely the electoral system, the gender quota legislation, party ideology or incumbency.

Analyses of the implementation of quota laws should not be restricted to formal processes but should also include informal practices. Several gendered practices applied by political parties have been identified. Some of them are shared across borders and others are particular to the French or Spanish institutional settings and parties. We have occasionally focused on the main left-wing and right-wing parties in both countries in order to provide a more detailed picture.

Allocation of top positions and safe seats

The most common practice used by political parties is the gendered allocation of safe and winning positions. This allows parties to comply with quota provisions but simultaneously keep the bulk of the most precious positions to men. In France, this practice was widely used in the 2002 and 2007 legislative elections, as Table 6 reports. Parties disproportionately located women in unwinnable districts. This imbalance was more acute in parties with larger proportion of incumbents such as the PS, the UDF and the UMP, which preferred to pay a financial penalty rather than selecting unknown female candidates – even if no strong evidence has proved the impact of incumbency on election results (Murray, 2008).Footnote 4 Public funding allocated to parties is linked to the votes a party obtains at the national aggregate and to the seats it wins. The penalty for failing to meet the parity requirements affects exclusively the first portion. Therefore, the law basically constrained small parties that could not afford to side-step the law (Murray, 2007, p. 575). In the 2007 legislative elections, the PS made an effort to field in more female candidates, although just half of them were actually elected. For senatorial and departmental elections, the reservation of winnable seats to male candidates is even more extended.

Table 6 Lower house, France, 2002–2007 (percentage of women)
Full size table

For the 2008 Spanish legislative elections we present disaggregated data for the two main parties. With respect to the lower house, Table 7 shows that only 31.7 per cent of safe positions were allocated to conservative female candidates, and therefore the levels of women in PP's parliamentary bench experienced a marginal increase from 28.4 per cent to 30.5 per cent even if women represented 45.1 per cent of the candidates. Regarding the PSOE, although the party placed more women in winning positions than in the previous election, the percentage of females elected descended from 46.3 per cent to 42.3 per cent due to intense competition with the PP in some districts. Women's presence as heads of party lists at the constituency (provincial) level was very low in both parties (less than a fourth). Concerning the upper house, in 2008 only about a third of PP's and PSOE's ballots allocated the first position to women. In the second and third positions the proportions by sex were more balanced. Nonetheless, it should also be noted that all-men lists vanished in the last election whereas they represented a non-negligible percentage in 2004 (that is, 30.5 per cent of PP's lists) (Verge, forthcoming). Regarding safe seats, in the period 1996–2008, 20 per cent of expected winning lists were fielded with women whereas women's share of expected losing lists was 53 per cent (Esteve-Volart and Bagues, 2010, p. 3).

Table 7 Legislative elections, Spain, 2008 (percentage of women)
Full size table

At the local level, winning seats are mainly reserved for men too: in 2007, of the first five positions of parties’ lists, in the first stretch in which parity had to be applied, women's share was 33.8 per cent (national average irrespective of town size). Women were mainly located in the fourth and fifth positions. Only the main left-wing party, the PSOE, almost reached parity in the first three positions, with 39.3 per cent of women – the PP filled 34.7 per cent of women in these positions (see Verge, 2011). In the 2011 local elections, women occupied the first position in about a fifth of party lists: 22 per cent of PSOE's and 19 per cent of PP's.

Women in executive office

In France, before the 2007 law almost a third of cabinet positions in town councils and regional assemblies were already accorded to female councillors. The 2007 law mandates that executive offices have to be evenly split between women and men. After the 2008 elections, women constitute 47.6 per cent of regional councillors and 35 per cent of local councillors with government responsibilities (average for all municipalities irrespective of size). Nonetheless, just 3.8 per cent of the presidencies of regional councils are occupied by women and only 13.8 per cent of mayors are women. Although the law does not apply to the national government, it is evident that it has influenced its composition too. Whereas the percentages of women in previous governments ranged between 20 per cent and 30 per cent, in 2007 the right appointed a cabinet with an equal number of male and female ministers. Yet, since November 2009, the percentage of women ministers has fallen below 30 per cent.

The Spanish Equality Law does not impose quotas to executive offices. However, since March 2004 the national government has presented a gender-balanced composition: 50 per cent of women under the first PSOE government (2004/2008) and 52.9 per cent under PSOE's government re-elected in 2008 – it decreased to 47 per cent in October 2010 when the cabinet was reshuffled. Previously, no cabinet had reached more than 31 per cent of female ministers. At the regional level, after the May 2007 regional elections, the average of female presence in cabinets attained 39.5 per cent. Indeed, the majority of the 17 regional governments had a gender-balanced composition. We find a statistically significant variation based on party ideology: on average, PSOE-led cabinets included 45.6 per cent of women whereas PP's contained 34.5 per cent of female regional ministers. After the 2011 regional elections the percentage of female regional ministers has decreased to 35.1 per cent.Footnote 5 However, the only region governed by a woman (Madrid) was led by the PP. In the previous two decades, there was only one precedent of female leadership at this level, that of the PSOE cabinet in the region of Murcia (1993–1995). The May 2011 regional elections saw three women elected as regional presidents, two of them belonging to the PP and the third one to a conservative regionalist party. At the local level, gender balance is still out of reach in the executive. In 2007, women accounted for just 14.3 per cent of mayors, having experienced a very limited increase (12.6 per cent in 2003). Municipal cabinets were also predominantly filled with men, 78.3 per cent (Verge, 2011, p. 65).

Turnover

In both countries, women's turnover is significantly larger than men's. This is a systematic practice that has persisted after the statutory quotas were passed. In France, as Table 8 shows, it particularly affects local and regional female candidates, who are often substituted by young women. Women suffered a higher turnover too at the European Parliament until 2004, when more men were renewed than women (see Table 9). Regarding the national level, women are much more likely than men to sit for the first time at the National Assembly and Senate. For example, in 2002, whereas 69.4 per cent of men elected that year had previous parliamentary experience only 47.1 per cent of women did so (Murray, 2004, p. 354).

Table 8 Incumbents in three French regions, 1992–2004 (percentages)
Full size table
Table 9 Incumbents in the European Parliament, French delegation, 1989–2004 (percentages)
Full size table

In the Spanish lower house, women's turnover was also higher than men's in the last election (2008): 56.1 per cent compared to 45.6 per cent. At the local level, in 2007, turnover affected women disproportionately, 61 per cent, whereas men's turnover was 40 per cent. This generalised pattern seems to imply that women are seen as interchangeable candidates by political parties that can assume to increase female numbers without giving women the chance to consolidate their careers. As a consequence, female networks remain weak or are more difficult to articulate (Verge, 2011, pp. 67–68).

Same-party competing lists

As district magnitude is very low for French senatorial elections, it is extremely easy to hinder the effectiveness of parity. With an average of three seats in competition, only the first and the second place of the list are safe. The implementation of zip lists meant that male candidates located on the second or third positions risked not being elected. To neutralise this possibility, many of them decided to launch their own candidacy and competed against the official party list (Murray, 2007, p. 572). Thus, small departments saw two PS's or two UMP's ballots in competition. The votes received by these parties were split in two lists, resulting in the election of the heads of each list, namely male candidates. This practice was already in use before the parity law came into effect but it has become more extended since 2001 (Troupel, 2006, p. 528).

Ordering of candidates in the list

Although preferential voting systems, among them the panachage system used for the Spanish Senate elections, allow voters to choose their preferred candidates, the personalised dimension produces a psychological rather than an effective impact, given that the predominant criteria remains party closeness. In the Spanish case, as voters tend to pick up candidates in the same order their preferred party drafted the ticket, being at the top of the list translates into obtaining most votes. We have already seen that in 2008 only about a third of PP's and PSOE's top positions belonged to women. If within each party ticket candidates are listed in an alphabetical order, why do we find such a low female presence at the top positions? Researchers have found that when drafting ballots parties select candidates with surnames that place them below the incumbent senator (Montabes and Ortega, 2002). In addition, surname order suffers from an acute gender bias. A recent study shows that female candidates’ position on the ballot depends on parties’ electoral expectations. Female candidates are placed at the top of the ballot in 7 per cent of races where the party expects to win just one seat whereas women head 47 per cent of lists where the party expects to win all three seats. Whereas male candidates’ surnames present no significant differences from the distribution of surnames in the Spanish population, women's surnames tend to be drawn from the end of the alphabet in ballots where the party expects to win a single seat and from the beginning of the alphabet when three positions are expected to be won (Esteve-Volart and Bagues, 2010, p. 14).Footnote 6 Thus, parties secure that their incumbent candidates – most of them men – get elected, especially when safe positions are scarce, that is, when only one seat might be obtained.

Concluding Remarks

The article has offered an extensive comparison of two countries that have recently undertaken statutory quota reform processes. France and Spain have followed different routes to adopt their respective quota laws. Although the introduction of party quotas in France inspired Spanish left-wing political parties to incorporate them, only in the latter case were they effectively implemented.

The enforcement of quota laws has boosted female representation in all institutions. Right-wing parties in both countries were more reluctant to quotas than the left, which is translated into a higher tendency to circumvent the law thereby reducing its effectiveness, as data for legislative elections shows. Institutional constraints, particularly the different electoral systems in use, are to a great extent responsible for the dissimilar success rates of the statutory quotas. In tune with past research on the field, we have observed that PR electoral systems produce more egalitarian results than plurality systems. In the latter, parties tend to protect their (male) incumbent candidates, despite the fact that empirical evidence does not support voter bias against female candidates. The effectiveness of quotas also increases if placement mandates are established and if the electoral cost of non-compliance is high – as it occurs when non-compliant party lists are withdrawn.

Nonetheless, demand-side factors significantly determine the level of female candidates and elected women, irrespective of electoral rules. In addition to shared practices such as the gendered allocation of safe and winning positions and turnover rates, in both countries parties apply specific strategic discrimination practices, both affecting the respective upper houses. These are the presentation of same-party competing lists (France) and a specific ordering of candidates on the list according to candidates’ surnames (Spain).

Overall, no single variable captures all the effect for strategic discrimination though there is a strong relation between the electoral system in use for a particular institution and the extent to which candidate selection processes suffer from party gendered practices. This suggests, as previous research has noted, that parity measures need to be carefully drafted in order to adjust to the electoral system in use for different elections, and thus to prevent parties from discriminating female candidates.

The extended use of strategic discrimination forces researchers to identify the reforms that would make harder for parties to implement these practices. With regard to elections held under PR and closed party lists, zipping is highly recommended because it prevents parties from filling the first positions mainly with male candidates. The withdrawal of the lists by electoral authorities is preferred over financial penalties because parties’ defection cost is then tougher. In any case, the financial penalty should always exceed the financial loss perceived by not fielding enough female candidates.

Elections that use plurality vote are harder to evaluate. Even if it could be established by law that at least half the seats a party has in competition should be fielded with female candidates, nothing prevents parties from presenting women in unwinnable or unsettled districts. Thus, even high numbers of female candidates do not necessarily translate into more women being elected. One direction of reform would be to establish a minimum number of women elected, which would forcibly lead parties to present female candidates in winnable or safe seats.

In conclusion, parity reforms might be insufficient to resolve the disadvantages women face when competing for political office unless parties get rid of their reluctance to make access truly egalitarian.