Introduction

Chile presents an interesting paradox. In 2021, it made worldwide headlines for electing the first Constitutional Assembly with an equal number of men and women delegates. However, it is among the latest Latin American countries to incorporate affirmative action for women in the national legislature. Chilean elites were very reluctant to address the issue of women’s subrepresentation. From 1991 to 2015, when Argentina debuted its legislative gender quotas at the national level, seventeen Latin American countries enacted either gender quota systems or parity laws (^{Archenti and Tula 2017}). In 2015, amidst a reform that overhauled the electoral system in place since Augusto Pinochet’s dictatorship, the center-left government and women’s rights activists were able to include the prohibition for political parties of running over 60 % of candidacies of the same gender to have their lists registered. The reform was pushed by President Michelle Bachelet during her second administration (2014-2018) and resisted by the center-right coalition in Congress. This resistance is hardly surprising, given that gender issues have frequently been part of a leftist agenda (^{Caul 1999}), but also due to the scarce presence of women in both legislative chambers.¹

In 2013, Michelle Bachelet ran against Evelyn Matthei, the presidential candidate of the then-incumbent right-wing government coalition. Ms. Matthei explicitly underlined her opposition to gender quotas during the campaign: “I do not believe in a quota law that requires women to be included in the ballots regardless of their abilities. It is more important that there is the freedom to choose among those who think like us and give incentives to overcome existing barriers rather than having candidates imposed because of their gender.”²

Without incentives to foster women’s political representation, the Chilean Congress was among the chambers with fewer women parliamentarians in Latin America. Until the 2017 elections, the inaugural contests for gender quota, women represented less than 16 % of both the Senate and the Lower House.³ In congressional committees where the electoral reform was debated, difference between men vs. women parliamentarians was abysmal. In the Committee on Constitutional Affairs, there was only one congresswoman out of thirteen members, and the Committee on Financial Affairs had an all-male composition. The balance of power among genders was similar in the Senate; all-male senators sit on the Committee on Constitutional Affairs and the Committee on Financial Affairs. Once the bill reached the floor, the government coalition fully supported the presidential proposal.⁴

Women from right-wing parties are faced with contradictory expectations for being women, on the one hand, and for espousing a conservative ideology, on the other. Comparative evidence has shown, however, that attitudes about gender quotas are subject to a gender gap that transcends the ideological divide. In the United Kingdom, a study finds that party lines explain attitudes on free-market issues among women and men politicians. Nevertheless, this does not hold for positions on values related to women’s interests, such as affirmative action. In fact, “significant differences between women and men politicians, and differences that are consistently found within each of the major parties. The results confirm that on the scales concerning support for equal opportunities and affirmative action there are strong and significant gender differences among Conservative, Labour and Liberal Democrat politicians” (^{Norris and Lovenduski 2003, 95}). Even more, the authors uncover that Conservative women are slightly more in favor of affirmative action than male Labor politicians.

In Latin America, where populist politics, movements, and leaders tend to erase the explanatory power of the left-right divide (^{Mudde and Rovira Kaltwasser 2017}; ^{Ostiguy 2017}), Chile stands as one of the few countries with a more ideological party system in traditional terms, at least at the national level (^{Siavelis 2014}). Thus, in a context of high party discipline and political organizations divided along ideological lines, conservative party elites are expected to display a homogenous position on most political issues. Following the above-cited findings of ^{Norris and Lovenduski (2003)}, I investigate whether conservative Chilean party elites shared a similar stance on legislative gender quotas. Using original data from a representative survey of center-right party elites at the national, regional, and local levels, I estimate whether right-wing men and women leaders share attitudes vis-à-vis legislative gender quotas. My findings for Chile resemble those in the United Kingdom.

The shift in perspective among right-wing women regarding gender quotas sheds light on the paradox between the challenging path to implementing quotas from 2015-2017 and the swift adoption of gender parity for the 2021 constitutional delegates. In five years, Chile passed from a weak quota system to a full-fledged Constitutional Assembly in which half of its members were women (^{Suárez-Cao 2021}). Following the social uprising of October 2019,⁵ the demands of many protesters revealed the limitations of Dictator Augusto Pinochet’s constitution as an “institutional straitjacket” that hindered many popular demands for reforms (^{Heiss 2020}; ^{Piscopo and Siavelis 2021, 45}). Because of these constraints, the idea that only a new constitution could achieve actual social development gained strength.

Even though the proposal to change the constitution has been present in Chilean politics for the past decade (^{Escudero 2021}), it took off after the 2019 social uprising. Civil society organizations, social movements, academics, and experts mobilized to change the electoral rules for constitutional delegates to include independent candidate lists, reserved seats for indigenous peoples, and gender parity. This latter was achieved thanks to two conditions: (i) the insufficient quota system in place for legislative elections and (ii) a coalition of female deputies from left to center-right that pushed for it (^{Suárez-Cao 2023}).

These center-right women deputies challenged their government, which opposed gender parity, and many of their male party colleagues. My argument is that the change in attitude can be attributed to a gender gap within conservative party elites in Chile. Through my research, I found that women leaders on the right in Chile have consistently and significantly supported legislative quotas since 2015, regardless of their individual traits and perceptions on party issues. This level of support marks a significant departure from the past few decades, during which influential women from conservative parties were opposed to affirmative action.

The argument develops as follows. The next section engages with the literature on descriptive representation and gender quotas. It also deals with comparative findings about conservative elites and women’s interests. The third section concentrates on a brief history of women and politics in Chile. The fourth section displays the results of a series of logistic regressions with a finite-sample correction to tap into the determinants of conservative party elites’ support of legislative gender quotas. The results highlight a gender gap among conservative elites to support affirmative action in electoral politics. The final section concludes and delineates the avenues for future research.

1. Women in Politics: Presence and Interests

Representation is a defining element in contemporary democracies (^{Urbinati 2012}). Amidst the crisis of representation (^{Manin 1997}), the politics of ideas is not enough to meaningfully link citizenry with the political system.⁶ However, amidst contemporary discredit and citizen distrust of political parties, historically underrepresented groups started to demand the right to be present in democratic institutions (^{Phillips 1995}). These claims for descriptive representation, in which “representatives are in their persons and lives in some sense typical of the larger class of persons whom they represent” (^{Mansbridge 1999, 629}), are a way to make citizens visible when the politics of ideas do not convey their interests.

Descriptive representation, then, deals with the importance of the politics of presence (^{Phillips 1995}). As argued by Phillips, “liberal democracy makes its neat equations between democracy and representation, democracy and universal suffrage, but asks us to consider as irrelevant the composition of our elected assemblies” (2003, 354). This situation needs to be corrected by including women and other underrepresented groups. Not only is descriptive representation a goal in any democratic society,⁷ but also because the literature has shown that it impacts the legitimacy of the political system and the representation of substantive interests.⁸

The political system’s legitimacy increases when significant numbers of women participate in politics and when they are visible and competitive (^{Atkinson 2003}). More women in legislatures are associated with higher levels of trust in the political system declared by citizens regardless of gender (^{Schwindt-Bayer and Mishler 2005}). In sum, democratic legitimacy benefits from institutions where women are not grossly underrepresented (^{Hill 2011}; ^{Norris 2002}).

However, when political party agendas do not include substantive interests, such as most women’s issues, descriptive representation helps bring them to the floor. In fact, “particularly on issues that are uncrystallized or that many legislators have not fully thought through, the personal quality of being oneself a member of an affected group gives a legislator a certain moral force in making an argument or asking for a favorable vote on an issue important to the group” (^{Mansbridge 1999, 648}).⁹

Institutional designs to improve descriptive representation in politics frequently deal with quotas, reserved seats, and parity laws in the specific case of gender.¹⁰ Since 1995, over one hundred and thirty countries around the world have adopted some gender quota system (^{Krook 2009; 2017}), whose effectivity varies according to other institutional features of the electoral rules and the strength of quota laws (^{Caminotti and Freidenberg 2016}; ^{Dahlerup 2006}; ^{Jones et al. 2012}). Latin American congresses first implemented mandatory legislative quotas and later parity laws (^{Archenti and Tula 2017}). Features of design and informal institutions and practices affected the effectiveness of these quota systems, which have shown ample variance.¹¹ However, quotas have indeed promoted the descriptive representation of women in political institutions (Jones 2009).¹²

Post-material issues, such as environmental and women’s rights, are usually considered part of a leftist agenda. Unsurprisingly, quota systems are typically associated with leftist political organizations (^{Beckwith and Cowell-Myers 2007}). ^{Lovenduski and Norris (1993)} argue that right-wing parties extend market logic to the political market and are, therefore, against intervening to balance the historical underrepresentation of women. ^{Caul (2001}) shows that, for political organizations in advanced industrial societies, a leftist ideology significantly affects the likelihood that a party will voluntarily implement quotas.

Conservative women are then between a rock and a hard place. On the one hand, conventional wisdom holds that the fulfillment of feminist interests measures women’s substantive representation. Yet “given that ‘women’ are not a homogeneous but a heterogeneous group … good representation is enhanced by ‘the making present of’ complementary, competing and conflicting views on what women, and their interests and needs, are” (^{Celis and Childs 2012, 214}). On the other hand, studies on advanced industrial societies have shown that right-wing women and men hold different attitudes regarding gender equality issues and affirmative action, and in some cases, even traditional left-right matters, being conservative women more liberal than their male counterparts (^{Campbell and Childs 2015}; Lovenduski and ^{Norris 2003}; ^{Broughton and Palmieri 1999}).¹³

The gender gap within elites of the same organization is relevant because if right-wing (or left-wing) men and women politicians are indistinguishable in their positions on critical issues, the link between descriptive and substantive representation is severely weakened. Previous research has shown that gender does matter, and congresswomen and congressmen do not always see eye to eye, even when they share party membership. Following Campbell and Childs, “we might question how a disproportionately male Conservative Party can substantively represent conservative women when their attitudes (perceived interests) appear to diverge significantly” (2015, 634). Thus, a discussion about gender legislative quotas is a subject in which we can contrast right-wing men’s and women’s stances. It is an issue that fulfills the criteria laid out by ^{Wängnerud (2000)} to define a shared interest of women because it (i) recognizes women as a social category, (ii) acknowledges the unequal balance of power between genders, and (iii) implies a policy to increase the autonomy of women.¹⁴

2. Women and Politics in Chile: Lagging in Legislative Representation

Chile is an interesting case regarding women and politics. Michelle Bachelet (from the Socialist Party) ruled the country for two presidential mandates (2006-2010 and 2014-2018) and was able to push for a women’s rights agenda,¹⁵ to a certain extent, in one of the most conservative countries in the region (^{Reyes-Housholder 2018}). However, the legislative representation of women was subpar and only reached 16 % until 2017, when the country finally implemented legislative gender quotas. Therefore, Chile is a case where a woman was the head of government on two occasions, but without a critical mass of women legislators.

The role of Chilean women in politics is not a recent phenomenon. Conservative Chilean women have been politically active well before women’s right to vote in 1949. In 1865, for instance, they created a newspaper, El Eco de las Señoras de Santiago (The Echo of Santiago’s Ladies), to defend religious rights. In the early twentieth century, they were vital in pushing for the extension of suffrage to women. Parties on the left were suspicious that women would vote for conservative options, mainly supporting Catholic values and candidates, and therefore reluctant to endorse progressive feminist movements. Meanwhile, conservative political parties upheld the fight for Catholic women (^{Errázuriz Tagle 2005}).

A maternalistic perspective is helpful to understanding the inclusion of women in politics in Chilean history.¹⁶ According to ^{Goldsmith-Weil (2020, 87)}: “In spite of some movements toward defamilialization and co-responsibility, there are little signs of demotherization of carework, and therefore, Chile can still be defined as a strong case of maternalism.” Maternalism rings particularly true with women on the right. Conservative women played a substantive role in bringing down the democratic government of Salvador Allende in 1973, rallying in a movement called Poder Femenino (Female Power), which supported strikers against the government and organized the distribution of anti-government propaganda (Power 2008). Yet they did it without questioning their role as mothers and wives. In this sense, right-wing women managed to express demands of women from other social sectors as they appealed to elements common to most women, such as the defense of the family, the problem of food shortage, and the reinforcement of gender roles (^{Toro Céspedes 2015}).

After democratization in 1989, women could not capture the center stage of politics until the election of Michelle Bachelet in 2006. Data from the UNDP shows that from 1989 to 2006, the political representation of women in Congress went up from 5.3 % to 15 %, a proportion that would remain unchanged for a decade.¹⁷ Legislative quotas seem to be the remedy for this situation. However, the inclusion of quotas in the new electoral law was met with resistance. In the first decade of the twenty-first century, center-left and left representatives introduced several bills about gender quotas, but they never reached the floor.¹⁸

An ethnography of all conservative women politicians in 2003 concludes that they share an “absolute rejection of gender quotas, even though they acknowledge their practical utility” (^{Letelier Kramer 2006, 162)}. Conservative women politicians, in particular, were on record showing doubts about the desirability of a quota law. Lily Pérez, at the moment the Secretary of the National Renovation Party (Renovación Nacional, RN), argued that “we believe that this generates a detrimental situation for women (...) to establish a quota law that sets a floor and a ceiling for women to enter elected office frankly, it seems to us that it is on the opposite line to what is to involve women in the public sphere, which ultimately leads to greater discrimination” (^{Rudnik 2006}). Evelyn Matthei, from the other traditional party of the right, Independent Democratic Union (Unión Democrática Independiente, UDI), also expressed her reservations about a quota system because political parties “will run women in those districts or regions where they have no chance of winning” (^{Vargas 2012}).

The second term of Michelle Bachelet (2014-18) opened a policy window to reform the two-member electoral system inherited from Pinochet’s dictatorship and include affirmative action measures to bolster women’s representation (^{Arce 2018}).¹⁹ The two-member electoral system heavily discouraged women’s representation because it required established candidates to win the majority of votes, and women were usually seen as a gamble (^{Franceschet 2005}). In 2015, the Congress passed a new electoral law establishing a multi-member proportional electoral system. It also included the requirement to nominate at most 60 % of candidates of the same gender, enacting a 40 % electoral quota for women in practice.

The Deputies Chamber passed the provision for gender quota with a vote of ninety against twenty-two (and five abstentions). All abstentions and almost all negative votes, except for one, were from men, all of them from right-wing parties.²⁰ Out of six, five conservative women deputies supported the quota system in the Lower Chamber. In the Senate, however, the gender divide faded: the bill passed with a vote of twenty-eight against six (and four abstentions). The ten non-positive votes were from right-wing senators. Out of three conservative women senators, two voted explicitly against the bill.²¹

Therefore, conservative women in Chile have always been politically active, with a robust maternalistic perspective, and not always supporting democracy. After democratization, right-wing parties sponsored more women to local elected offices than their leftist counterparts (^{Hinojosa 2009}), but they did so without questioning traditional gender roles. Indeed, the right has presented the role of the homemaker as highly compatible with local politics since the municipal government entails “the administration of a larger house” (Hinojosa 2009, 394). These women have shared the ubiquitous concern about the severe underrepresentation of women in democratic institutions. Yet, their most prominent figures did not univocally endorse affirmative action when discussing the 2015 electoral reform.

In 2019-2020, women deputies from RN were pivotal to a broad coalition of representatives for enacting a solid gender parity provision for the election of constitutional delegates (^{Suárez-Cao 2023}). Bravely facing the opposition of their own government, they allied with their women colleagues from the left and convinced their male counterparts to vote for gender parity.²² In sum, there has been a notable shift among right-wing women in Chile over the past few years. Without their support, gender parity in the Constitutional Assembly would not have been possible.²³ While they previously opposed affirmative action, they have since changed their views, now even advocating for gender parity. In doing so, they have set themselves apart from the male elites within their own political parties.

3. Conservative Party Elites in Chile: Attitudes on Gender Quotas

Conservative Chilean women were adamantly on record against legislative gender quotas; nevertheless, right-wing congresswomen voted in favor of the reform in the Chamber (against their own party line and conservative women senators who also opposed). This ambivalence may be a product of the electoral connection; conservative congresswomen could support quotas because their constituents wanted them as well, even though their ideology “promotes individual freedom and refuses any state intervention in political participation” (^{Letelier Kramer 2006, 162-163}). But, taken as a whole, were right-wing party elites for or against quotas?

This question is assessed thanks to a survey of Chile Vamos party elites (UDI, RN, and the recently created EVOPOLI), applied right after the law was passed (between May 2015 and April 2016). This survey permits evaluating the determinants of conservative party elites’ support for quotas-a matter understudied outside advanced industrial democracies (^{Alenda et al. 2020}, ^{Campbell and Childs 2015}; Childs et al. 2009; ^{Webb and Childs 2012}).

Several hypotheses are drawn to gauge support for quotas among right-wing party authorities in Chile. The null hypothesis states that there is no gender gap among party elites. This hypothesis is supported by ^{Barnes and Cassese’s (2017)} reasoning. Party elites are engaged partisans, and individuals with strong partisan inclinations tend to share more agreement than those with weaker partisan affiliations.

The central hypothesis is related to a gender gap. It seeks to provide evidence to assess the importance of descriptive representation, given the different positions of men and women from the same political coalition. Therefore, hypothesis #1 states that support for legislative gender quotas can be explained by gender, with women more in favor than men. Hypothesis #2 postulates a positive relationship between left-leaning attitudes regarding the role of the State in the economy and support for gender quotas. ^{Campbell and Childs (2015)} found a gender gap among UK conservative elites in economic issues; if this is the case, gender alone may not account for variation in support for quotas. It may be the case that women hold less conservative attitudes than men. Finally, hypothesis #3 states that there may be a party effect within the right-wing coalition. If this is so, the expectation is that members of the new party EVOPOLI are more in favor of quotas than members of the traditional parties RN and UDI. If differences of opinion arise within the coalition, these would be assessed based on party affiliation rather than gender, as the organizations themselves are those who align their leadership with similar ideological positions.²⁴

4. Data and Methods

Based on complete lists of party authorities provided by the organizations, UDI and RN elites were selected via a stratified random sample. In contrast, a census was used for EVOPOLI authorities and congresspersons. The response rate for party leaders was 47.2 % for RN, 65.3 % for UDI, and 91.2 % for EVOPOLI; for congresspersons, it was 90 % for RN and 72.2 % for UDI (^{Le Foulon et al. 2020}). Six hundred twenty party authorities and elected officials (303 in UDI, 213 in RN, and 104 in EVOPOLI) responded to the survey. The key demographics of the sample show center-right elites similarly divided by gender (76 % of men in UDI, 77 % in RN, and 71 % in EVOPOLI) and education in all parties (the median is college educated). RN and UDI elites also share a similar age average (50 years of age for UDI and 52 years for RN) and a high percentage of religious people (87 % and 93 %, respectively). EVOPOLI, on the other hand, is notoriously younger (a mean of 36 years) and less religious (64 %).

The dependent variable is a dummy variable based on the question: “Do you agree or disagree with a reform that makes it mandatory for parties to run a fixed percentage of female candidates to Congress?” Party membership adopts the value reported (UDI, RN, EVOPOLI). For the measurement of left-leaning attitudes, I use a variable called political sensitivity²⁵ that assesses the respondent’s perspective on the desirable role of the State in the economy (^{Alenda et al. 2020}). Thus, this variable takes the value of libertarian,²⁶ subsidiary,²⁷ and solidary,²⁸ showing an increasing degree of agreement with the State’s involvement in economic affairs. This variable concentrates on the subsidiary sensitivity (55 %), then solidary (31 %), and lastly, libertarian (14 %). Interestingly, the distribution does not correlate with political parties (UDI is 12 % libertarian, 59 % subsidiary, and 29 % solidary; RN is 14 % libertarian, 57 % subsidiary, and 36 % solidary; EVOPOLI is 17 % libertarian, 47 % subsidiary, and 31 % solidary).²⁹

I include four variables measuring attitudes and perceptions of party leaders about their organization (selectorate, campaign, party discipline, and party doctrine) to evaluate whether women benefit from party-centralized nominations (as observed by ^{Franceschet 2005}; ^{Hinojosa 2009}). The selectorate variable indicates whether candidates are selected via a primary election among party members, by the vote of party counsels, by a group of high-ranked party authorities, or by the highest-ranked party leader.

For the campaign variable, I use a question that asks whether a candidate’s election campaign is managed by the party, by the candidate, or coordinated by both. Party discipline is measured by a question that interrogates whether congresspersons should follow the party line or their position-if they are discordant. Party doctrine relies on the responses to whether the party’s electoral program should promote the party manifesto or adapt to the current political circumstances to appeal to more voters. Finally, control variables are the following: elected official (a dummy for whether the respondent is a congressperson, mayor, regional or local councilperson), age (included as a continuous variable both as linear and as a squared term to capture non-linear relations), education (a dummy distinguishing those who report having a college degree), and religion (a dummy distinguishing those who state a religious affiliation from those who do not).³⁰

Given that the universe of Chile Vamos elites is known and finite, I estimate a series of multivariate logit regressions with finite-sample correction to maximize discriminating power and goodness-of-fit (see Table A1 in the Appendix). Since, in non-linear models, the coefficients do not have a direct interpretation, and the marginal effects of independent variables change according to the value that all explanatory variables take, the discussion will focus on the estimated marginal effects of variables anchored on the median and mode values.³¹ It is worth noting that marginal effects demonstrate the change in predicted probabilities concerning the critical variable of interest estimated at the median or mode of the other variables. Therefore, they allow for evaluating whether a change from one category to another has a statistically significant impact on the predicted probability. The presented results are from a full model. Results are robust and stable through different specifications.³²

Figure 1 plots the marginal effects of the full model (see Table A3 in the Appendix for predicted probabilities).³³ Among conservative Chilean party elites, the gender gap in the support for quotas is substantively and statistically significant. Men have a 28 % lower probability of favoring affirmative action measures compared to women. Furthermore, elites with a solidary sensitivity are 22 % more likely to support quotas than those with a subsidiary perspective. Interestingly, the difference in support for quotas between those with solidary and libertarian perspectives is not statistically significant.

The results indicate a party effect, given the statistically significant differences between all parties. Leaders from the UDI have the lowest predicted probability of supporting quotas, at 34 % (see Table A3 in the Appendix). In comparison, an EVOPOLI party leader is 29 % more likely to support quotas than a UDI party leader, and an RN party leader is 16 % more likely than a UDI party leader. Additionally, the belief that candidate nominations are determined within parties and that party discipline is highly valuable has a positive marginal effect on quota support compared to those who prioritize individuals over organizations. On the other hand, leaders who report that parties do not manage political campaigns are more likely to disapprove of gender quotas.

Regarding the controls, age is also a significant factor; at the same time, having a college degree and reporting not having a religion have negative and statistically significant marginal effects on the predicted probabilities of supporting quotas among conservative Chilean elites. Please refer to the Appendix, Table A2 for detailed descriptive statistics.

Since finite-sample correction does not support the assumption that cases are all independent of each other, traditional goodness-of-fit indicators are not appropriate. Therefore, alternative measures for logistic regression using complex survey data are employed, such as the area under the ROC curve and the Hosmer-Lemeshow test (^{Archer and Lemeshow 2006}). The area under the ROC curve shows that the models have good discriminating power. However, the Hosmer-Lemeshow tests show that they do not have a good fit.

Source: own elaboration

Figure 1.  Marginal effects of determinants for quota support with respect to base category, full sample 

Given the power of the gender variable in the logistic regression, a reason for this poor calibration could be that determinants for the support of gender quotas are different between men and women. To assess this, the models are estimated again for the two separate samples of men and women (see Table A4 in the Appendix), which confirms that explanatory variables do not work homogeneously in right-wing men and women. The goodness-of-fit of the models substantially increases, showing a better calibration and power of discrimination than the models using the whole sample.

Figure 2 compares the marginal effects in the full model for separated samples of women and men (see Table A3 in the Appendix for predicted probabilities). For women, several explanatory variables lose statistical significance. The marginal effect between EVOPOLI and UDI party membership, as well as RN and UDI, are still positive and statistically significant (with a predicted probability of 93 % for a woman EVOPOLI party leader to supporting quotas vs. 59 % for UDI at with a 95 % confidence interval), and a solidary sensitivity (78 % vs. 46 % libertarian). The marginal effect of non-affiliation to a religion is large and negative.

Attitudes and perceptions about parties lose their explanatory power among women.³⁵ And the other individual characteristics are not significant either. In sum, Chilean women leaders in right-wing parties agree with gender legislative quotas, especially those who espouse a solidary sensitivity, belong to EVOPOLI and RN, and have a religion.

The picture is more nuanced for men since additional factors are associated with supporting legislative quotas. As we observe among women, both party and sensitivity matter. A male leader with a solidary sensitivity has a 21 % higher probability than one associated with a subsidiary sensitivity but exhibits no statistical difference with one with libertarian positions. If he belongs to EVOPOLI, his predicted probability of supporting quotas is 60 % (and 51 % for RN, difference that is only significant with a 90 % confidence interval). A party leader holding a popularly elected position has a 15 % higher probability (positive and statistically significant marginal effect) of favoring affirmative action for women in politics.

Interestingly, for a party leader with a college degree, the predicted probability plummets to 24 %. Perception variables also play a role for right-wing men. In contrast to women, a leader who believes that candidates are responsible for their own campaigns is 47 % less likely (or 46 % for mixed) to agree with quotas than a party leader who believes that parties are in charge of their candidates’ campaigns. Additionally, male party leaders who value party discipline are 10 % more likely to support affirmative action or quotas than those who do not.

Source: own elaboration

Figure 2.  Marginal effects of determinants for quota support with respect to base category, separated samples 

In sum, support for quotas among right-wing party leaders in Chile shows a substantial gender gap. Belonging to the new party of the coalition and having a left-leaning view on economic affairs also increase the probability of endorsing affirmative action mechanisms.³⁶ For men, however, support is also mediated by other factors that are not significant for women, such as holding an elected position and valuing party discipline, among others. Given the uneven number of observations between men and women, it is not possible to rule out that the statistical significance is a product of the larger size of the men sample. However, these findings align with previous research indicating that women within Chilean right-wing party elites tend to hold more progressive views than their male counterparts, particularly on issues related to human rights, including but not limited to same-sex marriage (^{Suárez-Cao et al. 2019}).

Conclusions

Despite having had a woman serve twice as head of government and an equal representation of men and women in its Constitutional Assembly, Chile continues to lag in women’s representation in Congress. Conservative women have long been active in politics, yet they have staunchly opposed affirmative action measures that have proven effective in increasing women’s presence in democratic institutions across Latin America. However, this situation was rectified in 2015 with a new law that included gender quotas, following multiple attempts to introduce electoral reform legislation.

Surprisingly, conservative congresswomen in the Lower Chamber voted in favor of the bill, against their party lines. In the Senate, however, women from right-wing parties voted against the proposal. This article attempts to explain this contradiction by examining the position on gender quotas. Within Chilean right-wing parties, gender does matter to understand the support for legislative quotas. Conservative women are more consistently in favor of affirmative action, mainly if they belong to EVOPOLI and have a solidary sensitivity; neither their education nor age matters for this approval. Whereas for conservative Chilean male leaders, numerous factors increase the probability of a party authority approving quotas.

The findings for right-wing women in Chile echo the results from advanced democracies. Similar to those studies on right-wing women in industrial democracies, conservative elites have a significant gender gap on issues related to affirmative action and gender equality (^{Campbell and Childs 2015}; ^{Norris and Lovenduski 2003}). Interestingly, this gap seems to have developed over the past few years in Chile since conservative women politicians have always worked to reinforce traditional gender roles that relegate women to private spaces. The final shift for conservative women occurred when prominent congresswomen from the right fought for gender parity rules for the election of constitutional delegates showing the power of a broad coalition with women from the left on women’s shared interests.

Further research should contribute with qualitative data that helps explain this stark change in right-wing women politicians in Chile regarding affirmative action. Similarly, it would be interesting to investigate attitude congruence with the conservative women electorate. Likewise, comparing women from the left would shed light on cross-cutting positions, such as the coalition that made gender parity possible in 2020.

Descriptive representation matters because it fulfills the definition of a shared interest of women, regardless of ideological differences. Quotas increase women’s autonomy and acknowledge that women constitute a historically disadvantaged group. Suppose right-wing men and women elites show stark differences regarding supporting a woman’s shared interest. In that case, the politics of ideas is not enough to guarantee the substantive representation of women. The findings of this article endorse the importance of descriptive representation of women because right-wing men and women politicians’ attitudes significantly diverge on issues that are close to women’s interests.