性別比例原則、保障平等的制度、與拉丁美洲的婦女參政 - 政大學術集成

全文

(1)International Master’s Program in International Studies National Chengchi University 國立政治大學國際研究英語碩士學位學程. Gender Quota Laws, Egalitarian Institutions, and Political 治政 Women’s 大. 立. ‧ 國. 學. Representation in Latin America ‧. 性別比例原則、保障平等的制度、. n. al. er. io. sit. y. Nat. 與拉丁美洲的婦女參政 Ch. engchi. i n U. v. Emelyn Peralta (埃米琳) Advisor: Professor Yen-Pin Su (蘇彥斌), Ph.D.. July, 2019. DOI:10.6814/NCCU201900318.

(2) Abstract Existing work regarding women’s political representation in Latin America have focused on the effect of electoral institutions. Nevertheless, there is little research that takes into account the impact of political institutions that facilitates egalitarianism. This thesis aims to fill the empirical gap by testing a theory based on the interactive effect of gender quota laws and egalitarian institutions. I maintain that the combined effects of gender quota laws and egalitarian political institutions is key in explaining the degree of. 政治大 lower chamber elections in 18 Latin American countries from 1988 to 2018, the analysis 立 empirically conducted proves that in a country that adopts a gender quota law, the level female representation in the parliament in Latin American countries. Built on data on. ‧ 國. 學. of women’s political representation tends to be much higher when the political institutions of this country are egalitarian. The result is solid and robust through different. ‧. model specifications and with the two predicting techniques used. In addition, this thesis adopts a case study about women’s political representation in the lower chamber of the. y. Nat. sit. Dominican Republic after the adoption of a quota law in 1997, examining the historical. er. io. background of feminism in the Dominican Republic, how the gender quota law was implemented, and the effectiveness of this mechanism on female political representation. al. n. in the lower chamber.. Ch. engchi. i n U. v. Keywords: gender quotas, egalitarianism, women’s representation, lower chamber elections, Latin America.. 1. DOI:10.6814/NCCU201900318.

(3) 摘要. 目前探討女性政治代表性的文獻，大多探討選舉制度的效果。然而，很少研究考量政治制度對於平等的保障如何影響女性的政治代表性。為了彌補既有文獻的不足，本論文建構一個結合「性別比例原則」與「保障平等的政治制度」的理論，以解釋女性在國會政治代表性的跨國差異。本文以 18 個拉美國家、跨越 1988 年至 2018 年的資料為基礎，利用量化方法對於本文的理論假設進行檢. 政治大. 驗。本文的實證研究分析顯示發現，當一國採取性別比例原則，且該國政治制. 立. 度較能保障平等的情況下，該國女性在下議院的政治代表性愈高。除了量化分. ‧ 國. 學. 析之外，本文亦針對多明尼加進行案例研究，探討該國女權意識的歷史發展，性別比例原則的制定過程，以及性別比例原則對於女性國會代表性的影響。. ‧ er. io. sit. y. Nat. 關鍵字: 性別比例原則、平等、女性政治代表、國會選舉、拉丁美洲.. n. al. Ch. engchi. i n U. v. 2. DOI:10.6814/NCCU201900318.

(4) Acknowledgements When finishing a work as arduous as it is the development of a master’s thesis, it is inevitable not to feel happy and proud of the work done. However, it is important to remember that all this work would not have been possible without the exceptional contribution of individuals and institutions. That is why I want to express my words of thanks in the following lines. First and foremost, I am grateful to God who inspired me and gave me strength and. 政治大 who supported me unconditionally 立 throughout these two years of arduous work, always. faith to believe that I could achieve it, even in times of weakness. To friends and family,. ‧ 國. 學. offering me their love, understanding, patience, and encouragement. Thank you for making this stage of my life a journey of experiences that I will never forget.. ‧. To my advisor Dr. Yen-Pin Su, for his support and trust in me, in addition to his. y. Nat. guidance, follow-up, and supervision during the course of this investigation. For the. sit. members of my thesis committee, Dr. Wen-Chin Wu and Dr. Chiung-Chu Lin, for the. al. er. io. time invested and their insightful criticism, which helped to improve the quality of this. v i n (IMPIS), who during these years C transmitted that we boast about U h e n gto cushthei knowledge n. thesis. To all the professors at the International Master's Program in International Studies today. Thank you for igniting and fanning the flame of research and the desire to learn in us. I also want to express my gratitude to the institutions that opened their doors to me to obtain information, especially to the Dominican organization Participación Ciudadana (PC) for the great support received by them that was very useful for the elaboration of the case study in this thesis. To all of you, my greatest acknowledgment and gratitude.. 3. DOI:10.6814/NCCU201900318.

(5) Table of Contents Abstract ..................................................................................................................................... 1. 摘要 ........................................................................................................................................... 2 Acknowledgements.................................................................................................................... 3 List of Acronyms ....................................................................................................................... 6 List of Figures............................................................................................................................ 8 List of Tables ............................................................................................................................. 9 Chapter One: Introduction ..................................................................................................... 10 1.1 Puzzle.............................................................................................................................. 10. 政治大 Chapter Two: Theoretical Perspectives.................................................................................. 15 立 2.1 The Socioeconomic and Cultural Barriers ........................................................................ 15 1.2 Why Study Women’s Political Representation? ............................................................... 13. ‧ 國. 學. 2.2 Electoral Systems and Political Institutions ...................................................................... 18 2.3 Gender Quota Systems .................................................................................................... 20. ‧. 2.3.1 What are Quotas and How Do They Work ................................................................ 20 2.3.2 Gender Quota Laws in Latin America ....................................................................... 23. sit. y. Nat. 2.3.3 The Combined Effects of Gender Quota and Egalitarian Institutions..........................27 Chapter Three: Research Design............................................................................................30. al. er. io. 3.1 Large-N Tests .................................................................................................................. 30. n. v i n Ch 3.1.2 Independent Variables...............................................................................................31 engchi U 3.1.3 Control Variables ...................................................................................................... 32 3.1.1 Dependent Variable .................................................................................................. 30. 3.1.4 Methods and Estimation Techniques ......................................................................... 35 3.2 Case of Study: The Dominican Republic.......................................................................... 39 Chapter Four: Statistical and Qualitative Analyses............................................................... 43 4.1 Empirical Results ............................................................................................................ 43 4.1.1 Robustness Check ..................................................................................................... 48 4.2 The Path to Equality: Dominican Women in the Parliament. ............................................ 50 4.2.1 Political Historical Background ................................................................................. 50 4.2.2 Women’s Suffrage in the Dominican Republic .......................................................... 54 4.2.3 Gender Quota in the Dominican Legislation .............................................................. 58. 4. DOI:10.6814/NCCU201900318.

(6) 4.2.4 The Gender Quota Reform ........................................................................................ 62 4.2.5 Current Situation of Women’s Representation in Politics........................................... 63 4.2.6 Conclusion................................................................................................................ 73 Chapter Five: Conclusion ....................................................................................................... 76 5.1 Research Summary .......................................................................................................... 76 5.2 Research Findings ........................................................................................................... 77 5.3 Research Limitations and Future Studies.......................................................................... 78 References ............................................................................................................................... 80 Appendices .............................................................................................................................. 98 Appendix A: Interview Questionnaire.................................................................................... 98. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. 5. DOI:10.6814/NCCU201900318.

(7) List of Acronyms AFD. Acción Feminista Dominicana (Dominican Feminist Action). CEDAW. Convention on the Elimination of All Forms of Discrimination Against Women. CEG-INTEC. Center for Gender Studies of the Technological Institute of Santo Domingo. CIM. Inter-American Commission of Women. CIPAF CSP. 治政大 Center for Women’s Action) Center立 for Systemic Peace. DEMOS. Encuesta Nacional de Cultura Politica y Democracia (National. Centro de Investigación para la Acción Femenina (The Research. ‧ 國. 學. Survey of Political Culture and Democracy). FMD. European Institute for Gender Equality. ‧. EIGE. Federación de Mujeres Dominicanas (Federation of Dominican. y. sit. HDI. Nat. Women). Human Development Index. io. Instituto Nacional Electoral (National Electoral Institute). International IDEA. International Institute for Democracy and Electoral Assistance. INTEC. Instituto Tecnológico de Santo Domingo (Technological Institute. n. al. er. INE. Ch. engchi. i n U. v. of Santo Domingo) ILO. International Labour Organization. IPU. Inter-Parliamentary Union. JCE. Junta Central Electoral de la República Dominicana JCE (Central Electoral Board of the Dominican Republic). JD. Juventud Democrática (Democratic Youth). LAPOP. Latin American Public Opinion Project. MODA. Movimiento Democrático Alternativo (Alternative Democratic Movement). NDI. National Democratic Institute for International Affairs 6. DOI:10.6814/NCCU201900318.

(8) PC. Participación Ciudadana (Citizen Participation). PLD. Partido de la Liberación Dominicana (Dominican Liberation Party). PQDC. Partido Quisqueyano Demócrata Cristiano (Quisqueyano Christian Democratic Party). PRD. Partido Revolucionario Dominicano (Dominican Revolutionary Party). PRM. Partido Revolucionario Moderno (Modern Revolutionary Party). PRSC. Partido Reformista Social Cristiano (Social Christian Reformist Party). PSP. Partido Socialista Popular (Popular Socialist Party). TSE. Tribunal Superior Electoral (Superior Electoral Tribunal). UASD. 政治大 Universidad Autónoma de Santo Domingo (Autonomous 立. University of Santo Domingo). ‧ 國. 學. UCSD. Universidad Católica Santo Domingo (Catholic University Santo Domingo). y. Nat. United Nations Economic Commission for Latin America and the. sit. UNECLAC. United Nations Development Programme. io. Caribbean. n. al. er. UNDP. United Nations. ‧. UN. Ch. engchi. i n U. v. 7. DOI:10.6814/NCCU201900318.

(9) List of Figures. Figure 1: Women in Parliament by Regions, 2018.………………………………….…12 Figure 2: How Important is Religion for Latin American Citizens, 2014.……….…….18 Figure 3: Percentage of Women Legislators in Lower Chambers in Latin America, 19882018…………………………………………………………………………………..…26. 政治大. Figure 4: Predicted Percentage of Women Legislators in Latin America………………46. 立. Figure 5: Difference in Predicted Percentage of Women’s legislators…………………47. ‧ 國. 學. Figure 6: Main Historical Events in Dominican Feminism…………………………….51 Figure 7: Women in the Dominican Legislature Before and After the Implementation of. ‧. the Gender Quota Law………………………………………………………………….62. y. Nat. Figure 8: Rejection of Male Supremacy in Politics in the Dominican Republic, 2008-. er. io. sit. 2014………………………………………………………….………………………….66 Figure 9: Women in the Dominican Parliament, 1990-2016…………………………...68. al. n. v i n C hof the Dominican Institutions, Figure 10: Degree of Egalitarianism 1990-2016........….70 engchi U. Figure 11: Female Legislators Elected by Political Party, 1998-2016……………….....74. 8. DOI:10.6814/NCCU201900318.

(10) List of Tables. Table 1: Latin American Gender Quota Laws by Country………………………………25 Table 2: List of Variables…………………………………………………………….….38 Table 3: Profile of the Interviewees……………………………………………………...43 Table 4. Effects of Egalitarianism of Political Institutions and Gender Quota Laws on. 政治大. Women’s Representation in Lower Chambers in Latin America, 1984-2018………...…45. 立. Table 5. Additional OLS Models.……………………………………………………......48. ‧ 國. 學. Table 6: OLS Models for Robustness Check.………………………………………...….50. ‧. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. 9. DOI:10.6814/NCCU201900318.

(11) Chapter One: Introduction 1.1 Puzzle Historically, politics has been an area of public life exclusive for men. This has meant that political positions of leadership have been generally occupied by men and that politics must be comprehended in male codes and guidelines, thus marginalizing women in the decision-making processes on issues of interest to the population. In the twentieth century, this situation of exclusion and injustice started to change. In fact, this has been. 政治大 struggles of the suffragettes 立 in favor of the recognition of women as citizens. During the recognized as the century of women as it was witnessed from the beginning of the. ‧ 國. 學. last three decades, the incorporation of women into the public space through their. participation in social movements of all kinds was evident and, finally, the efforts of women to gain access to areas of leadership and political responsibility have been seen. ‧. (International IDEA1 2003).. y. Nat. sit. The gender gap in political representation remains a tenacious global challenge. As. al. er. io. we can observe in Figure 1, worldwide, women represent 24 percent of all members of. v i n C h across different is great variation in women’s representation e n g c h i U regions. For instance, in the n. parliament in single or lower houses (Inter-Parliamentary Union 2018). Regionally, there Nordic countries, the representation of women is 42.3 percent, while in other regions,. such as Sub-Saharan Africa and Asia, it is less than 20 percent. In the Americas region, the representation of women is less than 30 percent. Why is the level of female political representation higher in some countries but not in others? The lack of women legislators in political leadership positions is a big problem that, in the absence of measures to solve it, is probable that will continue for several years. To address these challenges, many countries have established governance structures, support mechanisms, and strategic plans to improve women’s political representation and gender equality. One of this. 1. International Institute for Democracy and Electoral Assistance.. 10. DOI:10.6814/NCCU201900318.

(12) strategies is the gender quota policy. In 1991, Argentina was the first country of the region to implement a gender quota system when a multiparty group of women involved in political action joined in order to persuade their male colleagues to help and defend the implementation of a system in this sense. The women were inspired by the system used in the Partido Socialista Obrero Español (Spanish Socialist Workers Party), as well as by international norms and conventions on gender equality. This pioneering step by Argentines triggered debates on the subject throughout the region. However, the decisive momentum came with the Fourth Conference on Women, celebrated in Beijing in 1995. The Beijing Platform for Action supported the right of women to take part in decisionmaking. In particular, governments were urged to guarantee “women’s equal access to. 政治大 81), as well as to enhance the capacity of women to participate in these areas. 立. and full participation in power structures and decision-making” (United Nations 1995, p.. ‧ 國. 學. Figure 1: Women in Parliament by Regions, 2018.. ‧. al. n. 35.0. sit. io. 40.0. er. 45.0. y. Nat. The Proportion of Women in Parliament (Single or Lower House) by Regions. 30.0 25.0. Ch. engchi. i n U. v. 20.0 15.0 10.0 5.0 0.0. %. World. Nordic countries. Americas. Europe (excluding Nordic countries). Sub-Saharan Africa. Asia. Arab States. Pacific. Source: The Inter-parliamentary Union (IPU). Available on the internet at http://archive.ipu.org/wmn-e/world.htm.. 11. DOI:10.6814/NCCU201900318.

(13) In Latin America, sixteen countries adopted a gender quota policy after the Fourth Conference on Women for the purpose of increase female representation in politics: Mexico and Paraguay in 1996; Bolivia, Brazil, Costa Rica, Dominican Republic, Ecuador, Panama, and Peru in 1997; Colombia and Honduras in 2000; Uruguay in 2009; Nicaragua in 2012; El Salvador in 2013; Chile and Venezuela in 2015. This mechanism should allow effective equality to be achieved since it implies that women must add a defined number or percentage in the lists of candidates for local and national congresses. However, gender quota policies do not always achieve their objectives on their own. For instance, Honduras and Ecuador introduced a gender quota of 50% in 2012 and 2008,. 政治大 after introducing the gender quota policy is 23% in Honduras and 36% in Ecuador. 立. respectively. However, the average proportion of seats occupied by female legislators Colombia and Peru introduced a gender quota of 30% in 2000. However, the average. ‧ 國. 學. proportion of seats occupied by female legislators after introducing the gender quota policy is 14% in Colombia and 24% in Peru. Clearly, even though for countries that greatly.. ‧. adopt the same gender quota policy, women’s political representation might differ. sit. y. Nat. al. er. io. This thesis aims to address the puzzle of why gender quota is more effective in some. n. countries but not in others. I argue that quotas are only the first step to adapt the ground. Ch. i n U. v. for the development of policies and programs to deal with the demands of women, in. engchi. order to have more egalitarian societies. In other words, some factors might affect the effectiveness of the gender quota policies. More specifically, I argue that the combined effects of gender quota laws and egalitarian institutions matters for explaining women’s representation in the lower chambers. To test my theoretical argument, I conduct quantitative analysis based on Latin American countries from 1988 to 2018. I also conduct a case study on the Dominican Republic for illustrating the mechanism of my theoretical argument. Overall, this research aims to facilitate a better comprehension of the causes of women’s underrepresentation in politics in Latin America.. 12. DOI:10.6814/NCCU201900318.

(14) 1.2 Why Study Women’s Political Representation? Why does the participation of women in politics matters? The achievement of the full participation of women in the decision-making process and gender equality are some of the most important indicators of democracy and its quality. The equal participation of women and men in politics is a relevant requisite for good governance and an efficient democracy (Childs and Dahlerup 2018; EIGE 2017; Inglehart and Welzel 2010; Welzel 2013). The participation of women in all aspects of politics creates more equal societies and allows for a stronger democracy (IPU 2008). Therefore, the degree to which a woman has the opportunity or not to get involved in politics and in decision-making. 政治大. processes is largely related to the quality of democracy.. 立. In addition, there is an association among the female presence in government and a. ‧ 國. 學. lower degree of corruption. Previous research provides strong evidence to support the statement that countries with a larger amount of elected female legislators in the lower. ‧. chamber generally have lower levels of corruption (Dollar, Fisman, and Gatti 2001; Esarey and Schwindt-Bayer 2019; Sundström and Wängnerud 2016; Treisman 2007).. Nat. sit. y. Some scholars argue that this is because corruption shows the presence of shady deals. io. er. that profit the people who already have privileges in the system, which in most countries tend to be men, as women are often excluded from opportunities to exercise power. n. al. i n U. v. (Barnes 2016; Bjarnegård 2013; Branisa and Ziegler 2011; Goetz 2007; Grimes and. Ch. engchi. Wängnerud 2012; Johnson et al. 2013; Stockemer 2011; Tripp 2001). On the other hand, other scholars argue that this is because women are treated differently by voters or are more risk-averse than their male counterparts and, therefore, are less frequently involved in corruption (Alatas et al. 2009; Esarey and Chirillo 2013; Esarey and Schwindt-Bayer 2018). Another reason is that female and male legislators put first distinct policy sectors, and these divergences can mold social spending criterions (Bolzendahl and Brooks 2007; Funk and Garthmann 2006; Svaleryd 2009). Clayton and Zetterberg (2018) found that as more women are elected, government spending on public health increases, which is offset 13. DOI:10.6814/NCCU201900318.

(15) by the relative decline in military expenditures and other expense categories. Therefore, female political representation influences both the range of policy issues that get considered and the types of solutions that are proposed. Existing research demonstrates that whether a legislator is male or female has a distinct impact on their policy-making and social spending priorities, making it critical that women are present in politics to represent the concerns of women and other marginalized voters and help improve the responsiveness of policy-making and governance (Campbell et al.2010; Volden, Wiseman, and Wittmer 2010). There is strong evidence that as more women candidates are elected into legislation, there is an increase as well in policy design that prioritizes families and minorities (Dahlerup 1988; Dingler, Kroeber, and Fortin-Rittberger 2019;. 政治大 2012; Schwindt-Bayer and Mishler 2005; Wängnerud and Sundell 2012; Yoon 2013). 立 Iverson and Rosenbluth 2006; Mansbridge 1999; McAllister and Studlar 1992; NDI There is also evidence that the more women there are in the parliament, the more. ‧ 國. 學. confidence women have in themselves and the more men look at women in positions of power in a more positive way (Barnes and Burchard 2013; Gilardi 2015; Verge and parliaments, political parties, democracy and the lives of citizens.. Nat. n. al. er. io. sit. y. ‧. Pastor 2018). Therefore, female political representation has positive impacts on. Ch. engchi. i n U. v. 14. DOI:10.6814/NCCU201900318.

(16) Chapter Two: Theoretical Perspectives 2.1 The Socioeconomic and Cultural Barriers What explains the level of women’s representation in the political arena? As many political scientists would define politics as an activity primarily concerned with the allocation and distribution of power (Berenskoetter and Williams 2007; Buchanan and Badham 2008; Keohane and Nye 1977; Wight et al. 2002), Blumberg (1984) argues that the most important source of power that affects women’s position in a system of social stratification is economic. Further, he asserts that the greater relative economic power of. 政治大. women, the greater the probability is that it will result in influence in the political sphere. Reynolds (1999) showed that socioeconomic development levels are notably associated. 立. with the percentage of women in parliaments worldwide. In other words, women find. ‧ 國. 學. harder to enter in the parliaments when they are in a situation of socioeconomic disadvantage due to the burden of low access to education, lack of financial resources,. ‧. poor health care, and unemployment or underemployment. In the same way, Stockemer (2017) identified that high-income inequalities diminish growth in female political. y. Nat. representation. More precisely, he found that the political representation of women in the. io. sit. parliament in countries with high-income inequalities has grown at a rate 20% lower than. al. er. that of countries with low-income inequality in the last 50 years. This shows that the fight. n. v i n C a disadvantage inUthe economy, but also from a high-income inequalities put womenhate ngchi. against income inequalities is not only important from a development perspective, since gender perspective since these inequalities can also affect women in the political arena. Moreover, Goodnow, Madrid, and Moser (2019) found that, compared to the male candidates, it is usually harder for female candidates to collect funds for their campaigns. Some scholars found that this issue is related to the level of economic development the country has (Matland 1998; Schmidt 2008; Schwindt-Bayer 2005; Stockemer and Byrne 2012; Tripp and Kang 2008). They argue that women tend to occupy a larger proportion of seats in parliament in more developed countries. They found that development is positively correlated with female representation in parliament because women enjoy. 15. DOI:10.6814/NCCU201900318.

(17) greater economic opportunities in more developed societies and because citizens of more economically developed countries are more likely to support gender equality. Several social and cultural attitudes have also been argued to influence the level of women’s political representation (Krook and O’Brien 2012; Paxton and Kunovich 2003; Teczar et al. 2018; Tripp and Kang 2008). Paxton and Kunovich (2003) argue that “ideas about women’s role and position in society can enhance or constrain women’s ability to seek political power” (p.90). They found that cultural attitudes were the most important factor explaining variance in the election of women in the legislature. Therefore, cultural norms can hinder women’s representation in politics. Research on the Nordic countries. 政治大 region is due to a political culture that gives more importance to economic and social 立 also shows that the comparatively greater number of women in the legislature in this. equality (Miller 2013). In addition, Inglehart and Norris (2003) show that culture matters:. ‧ 國. 學. countries with egalitarian attitudes, such as Sweden and Norway, have more women in parliament. This egalitarian culture is a reflection of broader patterns of socio-economic. ‧. development and cultural modernization. In contrast, in much of Latin America, traditional social norms prevail in which it is understood that the role of women is linked. y. Nat. sit. to the parental role, while that of men is associated with paid labor. As a result of this, a. al. er. io. large part of the population believes that politics is not an appropriate environment for. n. women as this would meddle with their parental obligations. According to the survey. Ch. i n U. v. conducted by Latinobarómetro (2018) in eighteen Latin American countries, 28 percent. engchi. of respondents do not agree that half of the members of parliament should be women. Thames (2017) found strong evidence that, in the long term, socio-cultural factors are relevant in increasing the percentage of women elected to the legislature. Mcculloch (2012) emphasizes that another important attitudinal factor that can affect the representation of women in politics is the degree of religiosity (the importance of religion for the citizens of a country). Religious rules and standards have traditionally been very involved in the formulation of public policies in many countries of the world and the Latin American region is no exception. As we can observe in Figure 2, according to a survey conducted in the year 2014 by the AmericasBarometer developed by the Latin 16. DOI:10.6814/NCCU201900318.

(18) American Public Opinion Project (LAPOP), a 58% of Latin American Citizens think that religion is very important while only a 6% consider that it is not important at all. Thus, we can see that religion plays a key role in the life of Latin American society.. Figure 2: How Important is Religion for Latin American Citizens, 2014. Importance of Religion for Latin American Citizens. Not important at all. 立. Not very important. ‧ 國. y. io. al. 20. 30. 40. 50. sit. 10. 60. 70. er. 0. Nat. %. ‧. Very important. 學. Rather important. 政治大. n. Source: The AmericasBarometer by the Latin American Public Opinion Project (LAPOP). Available on the internet at http://lapop.ccp.ucr.ac.cr/LapopDummies.html. Ch. engchi. i n U. v. Religion gives its devotees a diversity of civic abilities and faster methods of communication. In addition, it fosters a psychological commitment to domestic affairs. For this reason, it is argued that religiosity promotes political participation (Djupe et al. 2007; Verba, Sidney, and Brady 1995; Wald et al. 1988). Nevertheless, this positive influence of religion on the political participation is not general throughout the citizens. According to Cassese and Holman (2016), despite women tend to go to church with more consistency and regularity than men and their degree of religiosity is higher, the level of female political representation is not uniformly higher than that of men. Further, existing literature argues that, countries where the population is especially religious, citizens tend 17. DOI:10.6814/NCCU201900318.

(19) to be more traditional, conservative and less likely to advocate for female representation in politics (Burns, Schlozman, and Verba 2001; Inglehart and Norris 2003; Kenworthy and Malami 1999; Mcculloch 2012; Morgan and Buice 2013; Plutzer 1991; Tripp and Kang 2008). In brief, these arguments show that religious beliefs and cultural and socioeconomic development are relevant factors to explain why in some countries there is more support for the representation of women in politics than others. More specifically, they argue that in countries with greater socio-economic development and cultural modernization, citizens are more likely to support gender equality. Therefore, these factors are crucial to. 政治大. explain female political representation.. 學. ‧ 國. 立. 2.2 Electoral Systems and Political Institutions. ‧. In the process of explaining the degree of women’s representation in politics, many academics also place great emphasis on the role of institutional design (e.g. Goodnow,. y. Nat. sit. Madrid, and Moser 2019; Paxton, Hughes, and Painter II 2010; Schwindt-Bayer 2005;. er. io. Tripp and Kang 2008). The political institutions are the organizations, the rules, and the procedures by which politics is made (Lovenduski 2005). These institutions influence the. n. al. Ch. i n U. v. way in which political representation works in reality. Existing literature has found that. engchi. institutions can be decisive in explaining gender divergences in political participation because of disproportionate electoral rules (Kittilson and Schwindt-Bayer 2010). Heath, Schwindt-Bayer and Taylor-Robinson (2005) argue that “the political and institutional environment of legislative chambers and characteristics of individual legislators and political parties will mediate the ability of male legislators to marginalize women” (p. 421). They found that to attain full inclusion into the parliament, women cannot resign themselves with only winning seats, but must reform the political institutions that traditionally allow male politicians to control limited political resources. Therefore, political institutions can handle who manages and controls the political appointments, as well as considerations to balance appointments between members of different genders. 18. DOI:10.6814/NCCU201900318.

(20) In addition, electoral laws will play a meaningful function in the political appointment with respect to women. Many scholars point out the importance of proportional representation (PR) electoral systems because of their beneficial ramifications for female candidates, especially in the long-term (e.g. Krook and O’Brien 2012; Larserud and Taphorn 2007; Matland 2005; Thames 2017; Womack 2018). The main reason why PR systems are much more female-friendly is that such systems have higher district magnitudes, and this generally produces higher party magnitudes. The district magnitude is basically the number of seats per district. On the other hand, the party magnitude are the seats that a party wins within a district (Matland and Montgomery 2003). Every district race is considered a separate election. Usually, the. 政治大 winning. Most of the time, this tend to be men of the prevailing ethnic group. Therefore, 立 political parties will propose a candidate they believe has the highest possibility of. district and party magnitudes are relevant due to their leading role in party strategy when. ‧ 國. 學. choosing candidates (Matland 1995; Southwell 2013; Verge and Wiesehomeier 2018).. ‧. At its base, the main difference between the PR and the majoritarian systems is the size of the district magnitude. The majoritarian systems are composed intrinsically by. y. Nat. sit. uninominal districts. On the other hand, in PR systems, several candidates are elected by. al. er. io. district (Došek et al. 2017). In PR systems, where seats are designated according to the. n. portion of total votes obtained, political parties are likely to hold an incentive to offset. Ch. i n U. v. their party-lists, incorporating candidates who belong to a diversity of social groups (for. engchi. example, women and other minorities) and to nominate applicants that embody distinct factions and sectors of the electorate. In majoritarian systems, however, the incentives are different: parties are more prone to nominate those who have realistic chances of winning more votes than any of their contenders and, in most cases, leaders tend to lean more towards male candidates. Therefore, the larger district magnitudes encourage the election of candidates who are at inferior positions within the party-lists, which are generally women. For this reason, PR systems are considered hypothetically more beneficial for women's election in the parliament.. 19. DOI:10.6814/NCCU201900318.

(21) However, within the PR systems, there are many variants, and some may be better than others from the point of view of women. Most scholars agree that closed-list systems, where a party-list is voted on and headmen manage the order in which applicants are placed on the party-list, is generally more favorable in choosing female legislators than in the open-list systems, where candidates are voted individually (Goodnow, Madrid, and Moser 2019; Htun and Jones 2002; Krook 2006; Larserud and Taphorn 2007; Norris 2004; Thames and Williams 2010; Valdini 2013). However, some academics argue that open-list systems increase the likelihood that candidates will be elected, by allowing the voter to choose them regardless of the position in which they are placed by the leaders within the list (Rule and Shugart 1995). Mari and Kang (2008). 政治大 most important decisive factor of female political representation, the introduction of 立. found that although in the past, party list PR electoral systems were considered one of the gender quotas from the mid-1990s implies that although the type of electoral system is. ‧ 國. 學. very relevant in some regions of the world, the gender quota law is even more important to raise women’s political representation worldwide.. ‧. In short, these arguments posit that institutions can affect the extent to which the. y. Nat. sit. party leaders may feel forced to incorporate female politicians in less conventional. er. al. n. by the country.. io. positions, and this will also depend highly on the electoral system and type of list adopted. Ch. engchi. i n U. v. 2.3 Gender Quota Systems 2.3.1 What are Quotas and How Do They Work Given the slow growth in the number of women in politics, there is an increase in requests for more efficient methods to achieve a gender balance in political institutions. Gender quota laws are one of these mechanisms. These quotas are regulations to set that a certain proportion of women must be selected for positions in government institutions. They are implemented to ensure that women are not marginalized from political life or have a merely decorative presence. Most quota systems seek to increase women’s 20. DOI:10.6814/NCCU201900318.

(22) political representation since the problem they need to address is usually the underrepresentation of women. In general, gender quotas systems are created to give women more power. However, according to Dahlerup (1998), to implement quotas for women against severe resistance it is necessary that women have already gained some power. There are different methods or types of gender quotas systems to guarantee that women are represented in the legislatures. According to Lovenduski (2005) the main types of gender quotas are legal quotas, which specify that every political party must appoint a determined minimum amount of female candidates; reserved seats, which. 政治大 these are internal requirements of the political parties to propose at least a minimum 立. formally reserve a share of seats for women; and gender quotas stipulated in party rules, percentage of women as candidates for the legislation. Of these, the most common are the. ‧ 國. 學. party quotas (Norris 2004). Each type of quota is supported by some requirements that might be more or less efficient in different countries. Even though the objective of the. ‧. three types of quotas is to increment the presence of women in politics, the question of whether they offer promotion or guarantees of equality will depend on the details of their. y. Nat. sit. requirements and, in particular, on the sanctions used in case of non-compliance. n. al. er. io. (Dahlerup and Freidenvall 2013).. Ch. i n U. v. The analysis of gender quotas shows interesting variations: they are very. engchi. controversial in some countries and regions, while proposals for quotas have been approved with little discussion in others. Some countries consider gender quotas as a form of infringement of the principle of fairness, while other countries see them as compensation for socioeconomic and cultural obstacles that impede fair competition between man and women. The objectors of gender quota policies observe that gender quota laws violate the concept of merit and that they diminish the quality of politicians because they restrict the candidates’ composition. On the other hand, the arguments supporting gender quota laws are centered on the principle of equity. Quota systems allow a more equitable representation of women in management positions, which improves their descriptive representation and this, in turn, translates into a more equitable 21. DOI:10.6814/NCCU201900318.

(23) representation of the interests of women in the decision-making processes, thus the objective is to improve the substantive representation of women (Jalušič and Antić 2001). The implementation of effective gender quota laws shows a change in the approach of “equal opportunities” to “equality of results”. Therefore, gender quota laws are a mean to achieve equality of results. This argument is built on the experience that equality is an objective that cannot be achieved through formal equitable treatment. If the barriers exist, it is argued, it is necessary to introduce compensation measures to achieve said equality of results (Ballington and Binda 2005). Genevieve Zingg (2015) claims that gender equality in politics has incremented due to a combination of wider economic, political,. 政治大 Before 1995, very few countries have gender quotas laws, but today more than 120 立. and social changes and the notable support of gender quotas laws in the legislature. countries have implemented a type of gender quota system. Despite the fact that the. ‧ 國. 學. percentage of women in the legislations around the world has increased continuously in the last decades, women still occupy a disproportionately smaller number of seats than. ‧. men (IPU 2008). However, quota laws are a relevant measure in fast-tracking female political representation (IPU 2012).. sit. y. Nat. al. er. io. Current empirical evidence implies that when gender quota laws are well-designed. n. and political parties cannot cheat them, they assist to increment the number of female. Ch. i n U. v. candidates elected (e.g. Allen and Cutts 2018; Baltrunaite et al. 2016, Celis, Krook, and. engchi. Meier 2011; Dahlerup and Freidenvall 2013; Esteve-Volart and Bagues 2012; Jones 2009; Matland 2006; Paxton, Hughes, and Painter II 2010; Schwindt-Bayer 2009; Stockemer 2018; Tripp and Kang 2008). Nevertheless, there is not much evidence on the effectiveness of gender quotas to increment the female presence in political positions of leadership or on their influence on the legislators’ quality. The greatest prominent evidence is from Sweden, where the willful adoption of gender quota system by the Social Democrat Party headed to an increment of women’s participation in political positions of leadership and in politicians’ the qualifications (Besley et al. 2017; O’Brien and Rickne 2016).. 22. DOI:10.6814/NCCU201900318.

(24) 2.3.2 Gender Quota Laws in Latin America In the last twenty years, much has been written about gender quotas and the participation of women in politics. However, much of this literature has focused mainly on Europe and the United States (Childs and Dahlerup 2018; Darcy, Welch and Clark 1994; Davis 1997; Matland 1993; Mrsevic 2004; Rule 1990; Russell and DeLancey 2002; Stockemer and Sundström 2019). Little research has extended the focus beyond these advanced democracies. As in the rest of the world, historically in Latin America, women have been extremely underrepresented in popularly elected positions, since the presence. 政治大 presents a risk to the democracy, an increasing number of people have supported 立 mechanism and proposals to increment the female participation in the spheres of power.. of women and men has never been equaled. In recognition that such underrepresentation. ‧ 國. 學. During the nineties, ten Latin American countries implemented laws for the establishment of gender quotas that guaranteed women minimum levels of political. ‧. representation as candidates in national elections. In the 1990s and 2000s, the gender quota required in almost all Latin American countries was between 20 and 30 percent.. Nat. sit. y. Nevertheless, at the end of the 2000s, the concept of “parity” began to take root in some. er. io. of the Latin American countries (IPU 2014). As Table 1 shows, the first State to increase its quota to 50 percent was Ecuador in 2008. In addition, since the 2000s, we can observe. n. al. Ch. i n U. v. that other Latin American countries also began to adopt gender quota laws.. engchi. Since its implementation, the gender quota system has proven to be an effective legal measure to guarantee a growing political presence of women. However, the results of this mechanism are not the same for all Latin American countries, despite the fact that most of them implemented it in a similar period of time. In Figure 3, we can observe that women have higher level of political representation in lower chambers in some countries. For example, in Bolivia, Mexico, and Costa Rica, the proportion of seats held by women is between 45.6 to 53.1 percent. However, the level of women’s political representation is relatively low in Brazil, Colombia, and Paraguay. The proportion of women in the lower chambers of these countries have not even reached 20 percent. 23. DOI:10.6814/NCCU201900318.

(25) Table 1: Latin American Gender Quota Laws by Country. Year of the Reform. Minimum Quota for Women by Law. 1991 2017 1997 2010 1997 2015 2000. 30% 50% 30% 50% 30% 40% 30%. 1997 2009 1997 2000 1997 2008 2013 2000 2012 1996 2014 2012 1997 2012 1996 1997 2000 2009 2015. 40% 50% 25% 33% 20% 50% 30% 30% 50% 30% 50% 50% 30% 50% 20% 25% 30% 33% 50%. Argentina Bolivia. 立. El Salvador Honduras. Uruguay Venezuela. al. n. Paraguay Peru. io. Panama. Nat. Nicaragua. Ch. engchi. ‧. Mexico. 學. ‧ 國. Ecuador. y. Dominican Republic. 政治大. sit. Brazil Chile Colombia Costa Rica. er. Country. i n U. v. Source: The United Nations Economic Commission for Latin America and the Caribbean (ECLAC). Available on internet at: https://oig.cepal.org/en/laws/3/count. 24. DOI:10.6814/NCCU201900318.

(26) Figure 3: Percentage of Women Legislators in Lower Chambers in Latin America, 1988-2018. Argentina. Bolivia. Brazil. 60. 60. 60. 40. 40. 40. 20. 20. 20. 0. 0. 20. 40. 40. 20. 20. 0. 0. sit. io. al. n. Dominican Republic. y. Nat. 0. 60. ‧. 40. 60. Costa Rica. Ch. Ecuador. engchi U. er. 60. Colombia. 0. 學. Chile. ‧ 國. 立. 政治大. v ni. El Salvador. 60. 60. 60. 40. 40. 20. 20. 20. 0. 0. 0. 40. 25. DOI:10.6814/NCCU201900318.

(27) Guatemala. Honduras. Mexico. 60. 60. 60. 40. 40. 40. 20. 20. 20. 0. 0. 0. 立. 60 40. 0. Paraguay. 60 40 20 0. 0. 40. 20. 20. 0. 0. Ch. engchi U. sit. y. n. 40. Uruguay. er. io. al. 60. 60. 40 20. Nat. Peru. 60. ‧. 20. Panama. 學. ‧ 國. Nicaragua. 政治大. v ni. Venezuela. 60 40 20 0. Source: Gender Statistics by the World Bank and the Archived Data from The Inter-Parliamentary Union (IPU).. 26. DOI:10.6814/NCCU201900318.

(28) 2.3.3 The Combined Effects of Gender Quota and Egalitarian Institutions As Table 1 and Figure 3 demonstrate, after implementing the gender quota laws, there are improvements in women’s opportunities to participate in the parliamentary sphere in Latin America. However, women’s political representation is different across Latin American countries, despite the fact that they all implemented a gender quota law. Women have obtained political power in some of the Latin American countries. However, this is not the case of all countries. Why, if most of these countries implemented a gender quota law through the 1990s, they present different outcomes? The first possible answer suggests that countries that require a higher percentage of women in. 政治大 However, as we can observe in Figure 3, the level of women’s political representation 立 might be rather different even for countries with the same gender quota policy. For their gender quota law should have a higher level of women’s political representation.. ‧ 國. 學. instance, Honduras and Ecuador introduced a gender quota of 50% in 2012 and 2008, respectively. However, the average proportion of seats held by women after introducing. ‧. the gender quota policy is 23% in Honduras and 36% in Ecuador. Colombia and Peru introduced a gender quota of 30% in 2000. However, the average proportion of seats held. Nat. sit. y. by women after introducing the gender quota policy is 14% in Colombia and 24% in. er. io. Peru. Therefore, the above discussion suggests that gender quotas alone cannot ensure higher participation of women in the national parliament. A further question that needs to. n. al. Ch. i n U. v. be addressed is: Why are gender quota laws more effective in some Latin American. engchi. Countries than in others in facilitating women's access to seats in lower chamber? In order to answer this puzzle, I argue that the causal mechanism that explains different levels of the women’s political demands additional attention and study on how institutional settings and attitudes interact. Existing work on the political representation of women in Latin America have focused on the combined use of gender quotas with placement mandates and closed-party lists (Htun and Jones 2002; Jones 2009; Krook 2005; Larserud and Taphorn 2007; Norris 2004). Nevertheless, there is few research focusing on the relationship between egalitarian institutions and the use of gender quotas in Latin America. I argue that there is a nexus between the gender quota laws, the 27. DOI:10.6814/NCCU201900318.

(29) presence of the egalitarian principle in political institutions, and higher participation of women in politics. As Christiano (2008) claims, a part of the functions of political institutions is to distribute resources to citizens so that they can define their interests and create opportunities for themselves. In a democracy, the ideal and fair procedure is that each citizen must have the same resources to influence the results of the collective decisionmaking process. This roughly indicates that everyone must have an equal vote and the same resources to participate in the communities decision-making process (that is, basically, the principle of political equality). In addition, they must have equal access to. 政治大 in defining the opportunities that citizens can receive in society. Thus, political equality 立. public resources such as education, health, water, etc., because these resources are crucial implies that each and every citizen must have a voice in the choice of the collective. ‧ 國. 學. characteristics of the population in the decision-making process that affects all the citizens. The lower poverty rates and the more egalitarian allocation of goods and. ‧. services (for example, education, medical care, and so on) guarantee that all the citizens can participate in politics and government. Therefore, egalitarian political institutions. y. Nat. sit. (which are in charge of the distribution of resources that can give citizens access to. n. al. er. io. political power) play a key role in the achievement of parity in the political sphere.. Ch. i n U. v. A greater participation of women in politics is highly linked to the fact that political. engchi. institutions can or cannot provide equal protection of the rights and freedoms of citizens, an equitable distribution of resources (education, health care, housing, etc.) and equal access to power regardless of gender, socioeconomic class or social group. To the extent that the gaps in these factors are widened, many women in Latin America will need and require the essential skills and will prevail marginalized from the sectors with leadership positions. As long as public institutions fail to provide egalitarian conditions and norms, it will be hard for novel policies such as gender quota systems to have a greater result in the political participation of women. Furthermore, if public and political institutions do not provide egalitarian conditions (distribution of resources, protection of the rights and freedom, and access to power), fewer women will be candidates, and consequently, fewer 28. DOI:10.6814/NCCU201900318.

(30) women will be parliamentarians. Therefore, in order to reach gender equality in the political sphere, it is indispensable to guarantee that the engagement to equality is illustrated at the normative and national policy level. In summary, I maintain that it is crucial that countries in which women are underrepresented in the political sphere implement well-designed gender quota laws, but the effect of this mechanism is conditional on the degree of gender equality that their political institutions possess. With the adoption of gender quota systems and more egalitarian institutions and norms, countries can observe much higher female participation in politics.. 政治大 In trying to answer this thesis’s research question, why are gender quota laws more 立. effective in some Latin American countries than others in facilitating women’s political. ‧ 國. 學. representation? I argue that the conditional relationship between gender quota laws and egalitarian political institutions is key in explaining the degree of women’s representation. ‧. in the parliament in Latin American countries. Based on the discussion, I generate a testable hypothesis: Latin American countries with gender quota laws and a higher degree. y. Nat. n. al. er. io. national parliament.. sit. of egalitarian institutions are more likely to have greater participation of women in the. Ch. i n U. v. The causal relationship of gender quota laws and egalitarian political institutions is. engchi. quite overlooked in the Latin American context. For this reason, there is the need for research analyzing this relationship. This study aims to fill the gap in the literature and provide policy implications for gender politics.. 29. DOI:10.6814/NCCU201900318.

(31) Chapter Three: Research Design 3.1 Large-N Tests For this research, the unit of analysis is country-elections. I gathered data on lower chamber elections in 18 Latin American countries from 1988 to 20182. Due to the unavailability of data regarding the proportion of seats held by women before 1988, I begin to observe a country from 1988 or in the first legislative election after the beginning of democracy.. 3.1.1 Dependent Variable. 立. 政治大. ‧ 國. 學. The dependent variable for this study is women’s representation in the lower chamber, which I operationalize through the proportion of seats held by women in unicameral national parliaments or lower chambers (%). The data for this variable was. ‧. collected from two sources: The World Bank’s Gender Statistics Data Base and the. y. sit. io. al. er. Union (IPU).. Nat. Statistical Archive of Women in National Parliaments from The Inter-parliamentary. v i n elaborated by Elisa María Carrio C published International Institute for Democracy U h e nongthe i h c n. For Argentina’s 1989, 1991, 1993, 1995 Elections, the data is from a case study. and Electoral Assistance in 2002. For Bolivia’s 1993 Election, the data is from a case study about Bolivia's Legislative Power developed by the Naciones Unidas para el Desarrollo (PNUD, United Nations Development Programme). In the case of Chile’s 1989, 1993, 1997, 2001 Elections, the data was collected from a candidacy report. conducted by Candidaturas Chile (www.candidaturaschile.cl); and for Guatemala’s 1994 Election, the data is from a case study elaborated by ATENEA, a project lead by UN Women, UNDP, and International IDEA, which seeks to measure the real exercise of women’s political rights in Latin America and the Caribbean. Lastly, for Mexico’s 1991 2. Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay, and Venezuela.. 30. DOI:10.6814/NCCU201900318.

(32) Election, the data was gathered by the Instituto Nacional Electoral (INE, National Electoral Institute).. 3.1.2 Independent Variables One of my main independent variables is whether or not a gender quota law has been implemented to establish that a certain proportion of women must be selected for positions in the national parliament. This will be a dummy variable coded as “0” when there are no regulations of quota for women and “1” when there are regulations of quota. 政治大. for women. For this variable, the data used is from the Gender Quota Database provided by the International Institute for Democracy and Electoral Assistance (International. 立. IDEA) and from the Gender Equality Observatory for Latin America and the Caribbean. ‧ 國. 學. developed by the United Nations Economic Commission for Latin America and the Caribbean (UNECLAC). In addition, I include a variable with a different measure for. ‧. gender quota law. This variable shows the actual percentage of seats required by the. io. sit. Nat. Commission for Latin America and the Caribbean (ECLAC)3.. y. gender quota laws. The data for this variable is from the United Nations Economic. al. er. My second main independent variable is the level of the egalitarianism of political. n. v i n C h (V-Dem) Institute. developed by the Varieties of Democracy e n g c h i U The main question of this institutions. To operationalize this variable, I use the Egalitarian Component Index index is: to what extent is the egalitarian principle achieved? According to this. component (Coppedge et al., 2018), The egalitarian principle of democracy maintains that inequalities hinder the exercise of rights and freedoms, and decrease the ability of citizens to participate in public life decision making processes. Egalitarian democracy is reached when the following factors are equally distributed by public institutions regardless of gender and socioeconomic class: 1) protection of the rights and liberties of the citizens; 2) resources distribution; and 3) access to positions of power.. 3. Available on internet at: https://oig.cepal.org/en/laws/3/country 31. DOI:10.6814/NCCU201900318.

(33) The V-Dem Institute developed the Egalitarian Component Index4, which is labeled Egalitarianism of Political Institutions in this thesis, by averaging their others following indices: . Equal Protection Index, which is formed by taking the estimates of the indicators of equality of social classes when respecting civil liberties, the degree of equality of social groups, and the percentage of the population with the weakest civil liberties.. . Equal Distribution of Resources, that includes the point estimates of indicators of the distribution of public or particularistic goods, the implementation of universalistic welfare versus means-tested policies, the degree of educational equality, and how equal is. 政治大 And, because neither the protection 立 of rights and liberties or the equitable resources health care access.. distribution is enough to guarantee an equitable representation of women, there was also. 學. ‧ 國. included the Equal Access Index, which basically analyses the following questions: do all citizens have equal opportunities to access power? Or is power distributed by. Nat. y. ‧. socioeconomic position, social group or gender?. sit. In order to test the hypothesis of this research, I incorporate an interaction term of the. io. two variables mentioned before: Gender Quota Laws*Egalitarian Political Institutions.. al. er. . v i n my theoretical argument, I expectC a positive association U h e n g c h i between the interaction term and the percentage of seats occupied by women in the lower chambers. n. Both variables interact in increasing women’s political representation and so, as stated in. 3.1.3 Control Variables. 4. According to the V-Dem V9 Dataset Codebook, there is no information showing that the V-Dem research team considers women’s political representation in constructing the Egalitarian Component Index. In my dataset, the Pearson’s correlation coefficient between the Index and women’s representation in lower chambers is 33.3, suggesting that the correlation of these two variables is not high. Moreover, there is also no information showing that the V-Dem research team considers the adoption of gender quota laws in constructing the Index. The Pearson’s correlation coefficient between the Index and Gender Quota Laws is 0.298, suggesting that the two variables are not highly correlated.. 32. DOI:10.6814/NCCU201900318.

(34) Previous literature (Htun 2003; True and Mintrom 2001) argues that advanced education prepares people to be ministers or parliamentarians. However, data concerning women’s attainment of tertiary education are not available for most Latin American countries. Therefore, I will measure female education as the proportion of women in secondary education age enrolled in secondary school. The literature also posits that more female candidates will get appointments to government offices as the percentage of women professionals increments (Hughes and Paxton 2008; Jaquette 1997; Rosenbluth et al. 2006). Because data regarding women holding professional jobs are not available, the percentage of women in the workforce is commonly substituted. Thus, my second control variable is workforce participation which I measure with the percentage of the labor force. 政治大 Bank’s Gender Statistics Data Base. 立. comprised of women. I will gather the data for these two variables from the World. ‧ 國. 學. In addition, I utilize Human Development Index (HDI), elaborated by the United Nations Development Programme (UNDP), to measure the overall education level of the. ‧. countries. As Escobar-Lemmon and Taylor-Robinson (2005) suggest, the percentage of the population of a country with higher education could better measure the more open-. y. Nat. sit. minded and egalitarian perspective that society is expected to have with the increment of. al. er. io. education levels, but because tertiary education statistics are missing for some years, I. n. will employ the HDI to estimate the general level of education of the population.. Ch. engchi. i n U. v. Moreover, previous literature shows that women are more likely to receive political appointments from a leftist president (Escobar-Lemmon and Taylor-Robinson 2005) hence I will control for whether the president is from a left party. I will operationalize this through a dummy variable coded “1” for a leftist government and “0” otherwise. The data will be gathered from The Database of Political Institutions 2018 developed by the Inter-American Development Bank. I also control the district magnitude of the countries which is basically the number of seats per district (Matland and Montgomery 2003). According to Matland (1995), district magnitudes are relevant due to their leading role in party strategy when choosing 33. DOI:10.6814/NCCU201900318.

(35) candidates. Thus, I include a control variable with the average district magnitude size. The data for this variable is collected from the dataset of district magnitude developed by Laura Wills Otero5. Most scholars agree that closed-list systems, where a party-list is voted on and party heandmen manage the order in which the applicants are placed on the party-list, are more likely to be favorable in choosing women other types of voting systems (Htun and Jones 2002; Krook 2006; Larserud and Taphorn 2007; Norris 2004). I will include a dummy variable coded as “1” when the country has a closed-list system, and “0” otherwise. The data is from the Gender Equality Observatory for Latin America and the. 政治大. Caribbean developed by the United Nations Economic Commission for Latin America. 立. and the Caribbean (UNECLAC).. ‧ 國. 學. Existing literature argues that, countries where citizens are particularly religious, tend to be conservative and less likely to support women’s representation in politics (Burns,. ‧. Schlozman, and Verba 2001; Inglehart and Norris 2003; Morgan and Buice 2013; Mcculloch 2012; Plutzer 1991). Thus, I will include a variable to control the religiosity of. y. Nat. sit. a country. The data for this variable comes from survey data available from 2004 to 2014. al. er. io. from the AmericasBarometer developed by the Latin American Public Opinion Project. n. (LAPOP). Specifically, LAPOP includes one question asking how often do you attend the. Ch. i n U. v. religious services which can be used to measure the degree of religiosity of the citizens.. engchi. Some scholars highlight that the level of democracy shapes women’s access to political positions (Friedman 1998; Paxton, Hughes, and Painter II 2010; Pierson 2000; Waylen 1994; Viterna and Fallon 2008). They argue that that initial levels and growth in democracy produce gains in women’s political representation because the level of democracy in a country establishes the general context in which women are placed in political positions or placed in political positions. Thus, I control the level of democracy lagged by one year, which I measure with the Polity Score in the Polity IV dataset. 5. Available online at https://uniandes.academia.edu/LauraWillsOtero.. 34. DOI:10.6814/NCCU201900318.

(36) developed by the Center for Systemic Peace (CSP). The Polity Score examines the concurrent qualities of democratic and autocratic authority in government institutions. Lastly, existing literature argues that the overall level of economic development in the country has positive effects on women’s political opportunities (Duflo 2012; Matland 1998; Stockemer 2015). In short, their argument is that development leads to a weakening of traditional values, greater urbanization, greater participation of the labor force and education for women and changes in attitude in the perceptions of appropriate roles for women. To control the level of economic development of the countries, I control for GDP per capita, which basically is the gross domestic product divided by the midyear. 政治大 the year prior to the election year. The data is from the World Bank’s World 立. population. This variable is operationalized as the logged value of the GDP per capita for Development Indicators.. ‧. ‧ 國. 學. 3.1.4 Methods and Estimation Techniques. sit. y. Nat. In this research, I employ “mixed methods” which combines quantitative and qualitative approaches into a new methodology. As a method, it focuses on collecting,. io. n. al. er. analyzing and combining quantitative and qualitative data into a single study (Johnson. i n U. v. and Onwuegbuzie 2004). The reason why I employed this methodology is that the. Ch. engchi. utilization of qualitative and quantitative approaches gives a better comprehension of the research problem than any approach by itself. In addition, mixed methods research provides an opportunity to compensate for the inherent weaknesses and unavoidable biases of the methods, as well as to take advantage of the inherent strengths of the methods (Greene 2007; Harwell 2011; Johnson and Turner 2003). For the quantitative approach, I use two estimation techniques. Firstly, I employ Ordinary Least Squares regression (OLS), often called lineal regression, in order to estimate women’s representation in the lower chambers in Latin America. I have chosen this technique because I use time-series–cross-section (TSCS) data (i.e., my dataset includes multiple observations of the same country over time). These observations may 35. DOI:10.6814/NCCU201900318.

(37) not be absolutely independent and, therefore, the analysis has risks of serial correlation and heteroskedasticity (Beck 2001). Hence, to take care of these threats, I employ robust standard errors grouped by country.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. 36. DOI:10.6814/NCCU201900318.

(38) Table 2: List of Variables. Measurement. 立. 政治大. Percentage of seats held by women in the unicameral national parliaments or lower chambers of the national legislatures.. ‧. A dummy variable coded as “0” when there are no regulations of quota for women and “1” otherwise. n. al. Ch. engchi. sit. io. Independent Variables. y. Nat Gender Quota Laws. For this variable, the data is from two sources: The World Bank’s Gender Statistics Data Base and the Statistical Archive of Women in National Parliaments from The InterParliamentary Union (IPU). For Argentina's 1989, 1991, 1993, 1995 Elections, the data is from a case study elaborated by Elisa María Carrio published on the International Institute for Democracy and Electoral Assistance in 2002. For Bolivia's 1993 Election, the data is from a case study about Bolivia's Legislative Power developed by Naciones Unidas para el Desarrollo (PNUD, United Nations Development Programme). In the case of Chile's 1989, 1993, 1997, 2001 Elections, the data was collected from a candidacy report conducted by Candidaturas Chile (www.candidaturaschile.cl); and for Guatemala's 1994 Election, the data is from a case study elaborated by ATENEA, a project lead by UN Women, UNDP, and International IDEA, which seeks to measure the real exercise of women's political rights in Latin America and the Caribbean. For Mexico's 1991 Election, the data is from the Instituto Nacional Electoral (INE, National Electoral Institute). The data is from the Gender Quota Database provided by the International Institute for Democracy and Electoral Assistance (International IDEA) and from the Gender Equality Observatory for Latin America and the Caribbean developed by the United Nations Economic Commission for Latin America and the Caribbean (UNECLAC).. er. Women’s Representation in Lower Chambers. Sources. 學. Dependent Variable. Labels. ‧ 國. Variables. i n U. v. Actual Gender Quota (%). The percentage of seats required by the Gender Quota Laws.. The United Nations Economic Commission for Latin America and the Caribbean (ECLAC). Available on internet at: https://oig.cepal.org/en/laws/3/country. Degree of Egalitarianism of Political Institutions. An interval from low to high (0-1).. To operationalize this variable, I use the Egalitarian Component Index developed by the Varieties of Democracy (V-Dem) Institute.. 37. DOI:10.6814/NCCU201900318.

(39) The women’s percentage in secondary education age that is matriculated in secondary school. The percentage of the labor force comprised of women. An interval from low to high (0-1).. President’s Ideology. A dummy variable coded “1” for a leftist government and “0” otherwise.. Log of Average District Magnitude. Logged average district magnitude size.. Closed-Party List System. A dummy variable coded as “1” when the country has a closed-list system, and “0” otherwise.. 政治大. Degree of Religiosity. io. n. al. Polity Score. Logged GDP per capita (current US$). Ch. engchi. The data for this variable is from the dataset developed by Laura Wills Otero. Available online at https://uniandes.academia.edu/LauraWillsOtero. The Gender Equality Observatory for Latin America and the Caribbean developed by the United Nations Economic Commission for Latin America and the Caribbean (UNECLAC). The data for this variable comes from survey data available from 2004 to 2014 from the AmericasBarometer developed by the Latin American Public Opinion Project (LAPOP).. y. Nat. An interval variable that is the average of the available data based on the ordinal answer (0-4) for the AmericasBarometer survey question how often do you attend the religious services. The “Polity Score” evaluates the type of regime in a country with a range from -10 (hereditary monarchy) to +10 (consolidated democracy). This variable is lagged by one year. GDP per capita is gross domestic product divided by midyear population. This variable is logged and lagged by one year.. The Database of Political Institutions developed by the InterAmerican Development Bank.. ‧. ‧ 國. 立. This variable is operationalized by the Human Development Index (HDI), developed by the United Nations Development Programme.. 學. Control Variables. Overall Education Level of the Countries. The World Bank’s Gender Statistics Data Base.. sit. Female Labor Force Participation. The World Bank’s Gender Statistics Data Base.. er. Women’s Secondary Education. i n U. v. The Polity IV dataset developed by the Center for Systemic Peace (CSP).. The World Bank's World Development Indicators.. 38. DOI:10.6814/NCCU201900318.

(40) Second, I use pooled Time-Series Cross-Sectional (TSCS) analyses along with panelcorrected standard errors. This technique is especially useful because my data is available only for a short time (two decades or three) for the 18 Latin American countries. Also, TSCS analyses gives me the possibility of capturing variation of time and space simultaneously (Fortin-Rittberger 2014). In this research, I employ two different series of models. First, I use one set of models that considers the variables’ independent effect. Regarding the second series of models, I incorporate all the variables, including the interaction term of gender quota laws and the egalitarianism of political institutions. Because there are two time-invariant variables in. 政治大 estimate country fixed-effect models. 立. my models, which are closed-list system and religiosity, it will be inappropriate to. ‧ 國. 學. 3.2 Case of Study: The Dominican Republic. ‧. Following the statistical analysis, I proceed with a qualitative case study about the. y. Nat. sit. political representation of women in the lower chamber of the Dominican Republic after. er. io. the adoption of a quota law in 1997. I examine the historical background of feminism in the Dominican Republic, how the gender quota law was implemented, and the. n. al. Ch. i n U. v. effectiveness of this mechanism in its goal of increasing female political representation.. engchi. In addition, I would like to explore the challenges that the Dominican Republic faces in increasing the political representation of women in the lower chamber. The Dominican Republic was one of the pioneering countries in Latin America to adopt a system of gender quotas in 1997. However, it has lagged behind in terms of gender parity. At present, women represent 26.8% of the lower house. The countries that implemented the gender quota more than a decade after the Dominican Republic, such as Costa Rica (2009) and Nicaragua (2013), have achieved greater participation of women in the parliament: 45.6% and 41.3% respectively. This shows that it is necessary to explore what factors have affected the efforts of the Dominican Republic to increase the 39. DOI:10.6814/NCCU201900318.