The candidacies of former first ladies for national office are becoming commonplace in Latin America. Between 1999 and 2016, 20 of them ran for office 26 times, being elected on 19 occasions. These candidates became presidents (twice), vice presidents (3 times), and legislators (14 times). Some unsuccessful candidacies were controversial: three former first ladies (Panamanian Marta Linares and Guatemalans Sandra Torres and Raquel Blandón) challenged the constitutions of their countries when running for the presidency, and a fourth (Honduran Xiomara Castro) did not accept losing the presidential election. However, despite the increasing number of these candidacies, their electoral success, and the controversies that have surrounded some of them, the scholarly literature has failed to explain when former first ladies run for national office. Instead, analysts and academics have often dismissed these candidates as delegates of an outgoing or former president, implying that former first ladies lack political independence and represent an indirect form of continuismo (Serrafero Reference Serrafero2015; Zovatto Reference Zovatto2014; Grondona Reference Grondona2007; Leuco Reference Leuco2015; Pérez Salazar Reference Salazar and Carlos2013). However, this conventional view obviates the point that some first ladies are politicians on their own merit, obscuring our understanding of their candidacies and their implications on women’s representation.
We argue that first ladies who have previously been elected to office are more likely to become candidates because they are politicians who have already shown their political ambition. Furthermore, we expect that these politicians are more likely to run in the first election after leaving the executive to capitalize on the public support amassed as first ladies—irrespective of the approval of the president they accompanied in office.
The absence of studies analyzing the candidacies of former first ladies is surprising because nontraditional presidential and legislative candidates have been studied in the region. Research has examined the electoral rise of newcomers (Carreras Reference Carreras2017; Corrales Reference Corrales2008), ex-presidents (Corrales Reference Corrales2008; Corrales et al. Reference Corrales, Hernández and Penfold2014), and women (Htun Reference Htun2000; Htun and Piscopo Reference Htun and Piscopo2014).
The candidacies of first ladies have also been overlooked despite their growing electoral participation and the politicization of their roles in the US (e.g., Burns Reference Burns2004; Watson Reference Watson1997) and Latin America (Sefchovich Reference Sefchovich2003; Balcácer Reference Balcácer2010). This politicization has also reached subnational politics. In federal countries such as the US, Argentina, Brazil, and Mexico, the wives of governors are also called first ladies and often support the campaigns of governors, develop networks with civil society organizations, and participate in public events. Sometimes they even run for office (Martin Reference Martin2018).
Furthermore, former first ladies are running for office across world regions. Shortly after the Uruguayan María Julia Pou became the first to run for Congress in the region in 1999, Hillary Clinton became the first American former first lady to run for the Senate in 2001, a presidential primary in 2008, and the presidency in 2016. In Africa, Janet Museveni of Uganda was the first to run for a seat in Parliament in 2006, Nana Konadu Agyeman-Rawlings was the first woman to run for Ghana’s top office in 2016, and Dlamini-Zuma ran for the South African presidency in 2017.
Although there are no legal restrictions on appointing any woman as first lady, the conventional wisdom has centered on the associates of male presidents. Therefore, we focus on the wives, sisters, and daughters whom male presidents have appointed to the role.
Former first ladies are a unique type of candidate. First, they enjoy significant public recognition and media coverage despite not holding a political position (Borrelli 2011; Burns Reference Burns2004; Winfield, Reference Winfield and Norris1997). This visibility allows them to influence the public agenda, promote their policy positions, and expose their political skills.
First ladies are frequently regarded as role models for women, mothers, activists, and even the fashion industry (Winfield Reference Winfield and Norris1997). To be sure, they can receive harsh criticism: Cristina Fernández de Kirchner of Argentina faced criticism that focused on her femininity rather than her involvement in public policies and her marriage to Néstor Kirchner rather than her political career (Piscopo Reference Piscopo and Murray2010). However, the press, interest groups, and politicians rarely antagonize first ladies because they are seldom engaged in controversies, do not have formal decisionmaking powers, and are not politically accountable.
Even when first ladies engage in public policies, they tend to do it in noncontroversial, gendered roles centered on women’s issues, childcare, and family values (Borrelli Reference Borrelli2011; Van Wyk Reference Van Wyk2017, 170). For example, Chilean Leonor Oyarzún in 1991 created the Integra Foundation, which provides childcare to low-income families. In Brazil, Ruth Cardoso promoted Comunidade Solidária (Solidarity Community), a program that fought extreme poverty. In the Dominican Republic, Margarita Cedeño implemented four major public policies to promote health and cultural training and to fight extreme poverty (more examples in Guerrero Valencia and Arana Araya Reference Guerrero Valencia and Araya2019).
Second, these women enjoy privileges due to their access to the apex of the executive branch, allowing them to increase their political capital. Their personal connections to governing elites allow them to build specialized political knowledge and an influential network. First ladies also enjoy material resources: they have institutionalized offices in 11 countries. These offices provide them with staff, budget, organizational subunits, and tasks.
Third, the public persona of first ladies is inevitably connected to the most powerful politician in the country. This has precisely fed the prejudice that the only distinctive characteristic of this group of candidates is that they represent a former president in the public sphere, independent of their own merits. The name connection they have with heads of state can influence their electoral prospects, for better or worse.
We contend that the growing electoral participation and success of former first ladies, the politicization of their role, and their uniqueness as candidates makes understanding when they run for office a pressing query.
The electoral emergence of former first ladies has deep implications. If, as conventional wisdom suggests, these candidates are political delegates, then their careers may simply reproduce the patriarchal view that a woman’s success is tied to her male companions. Furthermore, if former first ladies serve as delegates, they contribute to personalizing politics in a region where several countries already suffer institutional weaknesses (McAllister Reference McAllister, Dalton and Klingemann2007). Moreover, they would make powerful political families even more influential, dampening political competition (Baturo and Gray Reference Baturo and Gray2018; Jalalzai Reference Jalalzai2013; Folke et al. Reference Folke, Rickne and Smith2017; Arana Araya Reference Arana Araya2016).
However, if some former first ladies who run for office are politicians in their own right, they may foster the engagement of women in politics and contribute to more gender-balanced public policies. Research has shown that female politicians encourage other women to run for office and increase the social acceptability of women in positions of authority (Thames and Williams Reference Thames and Williams2013; Escobar-Lemmon and Taylor-Robinson Reference Escobar-Lemmon and Taylor-Robinson2005). Furthermore, congresswomen tend to prioritize legislation related to women, children, and families (Schwindt-Bayer Reference Schwindt-Bayer2006; Jones Reference Jones1997; Taylor-Robinson and Heath Reference Taylor-Robinson and Michelle Heath2003).
First Ladies as Candidates
Thus far, scholars, pundits, and the media have treated most former first ladies who run for office as little more than the “covert reelection” of their husbands or fathers. Serrafero (Reference Serrafero2015, 99) describes the candidacies of former first ladies as a resource used by ambitious presidents who cannot be reelected. Zovatto (Reference Zovatto2014) argues that Latin America is increasingly moving toward “conjugal reelectionism.”
When we examine individual cases, we invariably find that former first ladies who run for office are depicted as delegates of former presidents, even though some of them are seasoned politicians. For instance, Cristina Fernández served four terms as a legislator (deputy and senator) before becoming first lady, but analysts often described her presidential candidacy as an attempt at covert reelection of her husband (Grondona Reference Grondona2007; Leuco Reference Leuco2015). Fernández’s election was also mocked as the “Kirchner dynasty” (Gallo Reference Gallo2008), “marital succession” (Serrafero Reference Serrafero2015), and “diarchy” (Grondona Reference Grondona2007). Other politically experienced first ladies, such as the Argentine Hilda González and the Paraguayan Emilia Alfaro, received a comparable treatment when they ran for office.
Similar descriptions abounded for the candidatures of politically inexperienced first ladies. A scholar described the candidacy of Marta Linares as a strategy of former Panamanian president Ricardo Martinelli to “essentially reelect himself through his wife” (Schipani Reference Schipani2014). Xiomara Castro was labeled as Manuel Zelaya’s “puppet” (Pérez Salazar Reference Salazar and Carlos2013), while Margarita Cedeño was described as the mere “shadow” of her husband, Dominican president Leonel Fernández (Arroyo Reference Arroyo2012).
In sum, these candidacies have been associated with a type of continuismo. This concept, introduced by Fitzgibbon (Reference Fitzgibbon1940), alludes to the practice of presidents to extend their power beyond their terms by the removal, extension, or avoidance of term limits. Former presidents would use their first ladies to continue governing (or remain influential in Congress), a type of continuismo via proxy that, according to Baturo (Reference Baturo2019, 81), occurred nine times in the world from 1945 to 2017.
Based on the conventional view, it could also be argued that politicized former first ladies may extend political dynasties, which dampens democratic representation by concentrating power in a few families (Dal Bó et al. Reference Dal Bó, Dal Bó and Snyder2009, 115). This subject has been studied in Asia. For example, Lee (Reference Lee2017, 387) suggests that the former first lady of South Korea, Park Geun-Hye, capitalized on her kinship to her father (president from 1963 to 1979) to become president in 2013.
The simplistic explanations of the electoral emergence of first ladies do not reflect the increasing involvement of this group of women in political affairs. Scholars have shown the relevance of American first ladies in domestic politics by focusing on their roles (Burns Reference Burns2004), impact on presidential campaigns (Burrell et al. Reference Burrell, Elder and Frederick2011), types of representation (Borrelli Reference Borrelli2011), absence of accountability (Eksterowicz and Roberts Reference Eksterowicz and Roberts2004), histories (Watson Reference Watson1997), press framing (Burns Reference Burns2004; Winfield Reference Winfield and Norris1997), and general political influence (Borrelli Reference Borrelli2002; Watson Reference Watson1997). The political involvement of Latin American first ladies can be found in biographies (e.g., Gordinho Reference Gordinho2009; Ruiz Reference Ruiz2012; Wornat Reference Wornat2005), historical reviews of specific countries (e.g., Balcácer Reference Balcácer2010; Sefchovich Reference Sefchovich2003), and case studies (Piscopo Reference Piscopo and Murray2010). Also, the literature about women in politics has increasingly included illustrative cases of first ladies (Reyes-Housholder Reference Reyes-Housholder2018; Krook and O’Brien Reference Krook and O’Brien2012).
To understand the electoral behavior of former first ladies, we examine associated literatures with the caveat that not all the logics, assumptions, and findings can be transferable among bodies of scholarship. Since a defining characteristic of first ladies is their gender and since some become candidates only after leaving the executive branch, we review literature on women in politics and on the emergence of new candidates. We draw from research centered on the United States and Latin America because they are presidential systems in which first ladies have become more politically engaged.
Historically, women have been underrepresented in the highest offices due to great obstacles, visible and invisible (glass ceilings), to entering politics (Htun Reference Htun2005; Norris and Inglehart Reference Norris and Inglehart2001). However, in the last two decades, the number of women elected to office in Latin America has steadily grown. Researchers have proposed two main explanations for this trend: the adoption of gender quota laws (Escobar-Lemmon and Taylor-Robinson Reference Escobar-Lemmon and Taylor-Robinson2005; Htun Reference Fox and Lawless2005) and more opportunities for women due to economic, social, and cultural development (Htun and Piscopo Reference Htun and Piscopo2014; Norris and Inglehart Reference Norris and Inglehart2001).
Gender quotas demand a percentage of women candidates or legislators, forcing parties to incorporate more women as politicians and thereby reducing the costs that women incur when running for office. Gender quotas have proved to increase women’s legislative representation. Folke et al. (Reference Folke, Rickne and Smith2017) analyzed all national-level legislatures elected from 1945 to 2016 in 12 democracies and found that a gender quota led to a large inflow of women in office. Research on Latin America has estimated that on average, gender quotas have boosted women’s representation by 10 percent (Htun Reference Fox and Lawless2005, 118).
The effect of quotas on the candidacies of first ladies is unclear. Potentially, quotas could benefit first ladies because quotas open up space for more women in politics and party gatekeepers may choose women with family links as candidates (Jalalzai Reference Jalalzai2013, 19). However, some studies have found that women candidates do not benefit more than men from personal connections when gender quotas are in place (Schwindt-Bayer et al Reference Schwindt-Bayer, Vallejo and Cantú2020; Beer and Camp Reference Beer and Ai Camp2016). Similarly, Baturo and Gray (Reference Baturo and Gray2018) have shown that the value of family connections diminishes when societies accept women’s political participation as normal. Furthermore, first ladies may not need quotas to attract parties, since they already have sizable political capital.
Research has argued that women in public office enhance other women’s identification with the political system and their ability to influence it (Burrell Reference Burrell, Thomas and Wilcox2005). Women legislators have been associated with more women running for office and with a broader social acceptance of female politicians (Thames and Williams Reference Thames and Williams2013; Escobar-Lemmon and Taylor-Robinson Reference Escobar-Lemmon and Taylor-Robinson2005). There is also evidence that some women politicians actively pursue women’s votes; for example, presidential candidates in Chile and Brazil (Reyes-Housholder Reference Reyes-Housholder2018). Furthermore, according to Krook and O’Brien (Reference Krook and O’Brien2012), most of women’s appointments as ministers respond to the increasing number of women in the political elite. When deciding whether to compete in elections or not, first ladies may take into account women’s record in both the legislative and executivebranches.
Two other factors are related to women’s political involvement. One is sexism, a phenomenon expressed in mistrust of women as leaders (Fox and Lawless Reference Fox and Lawless2004). Recent efforts to promote women’s access to democratic institutions in Latin America have revealed patriarchal, sexist, and gender violent practices (Freidenberg Reference Freidenberg2017). Escobar-Lemmon and Taylor-Robinson (Reference Escobar-Lemmon and Taylor-Robinson2016) found that women need to become “surrogate males” to increase their chances of being appointed to cabinets in Argentina, Chile, Colombia, Costa Rica, and the US. The level of sexism varies nonetheless across the region. Morgan and Buice (Reference Morgan and Buice2013) examined survey data and showed that in 12 of 19 countries, at least two-thirds of the participants disagreed that men are better political leaders than women—but at least a third of Paraguayans, Panamanians, Colombians, Haitians, Jamaicans, and Dominicans did agree. In sum, women are unlikely to become competitive candidates and to run for office if sexism is widespread.
Second, research suggests that increasing the women’s share in pipeline professions decreases gender disparities in elective office (Fox and Lawless Reference Fox and Lawless2004, 265). The expectation is that improving the educational and professional status of women raises the chances of observing more women in power (Norris and Inglehart Reference Norris and Inglehart2001, 130).
Works on political newcomers may shed light on why politically inexperienced first ladies become candidates. Researchers have identified two main factors that explain why citizens seek to be elected: political ambition and structure of opportunity (Fox and Lawless Reference Fox and Lawless2005; Borchert Reference Borchert2011). Although the motivations to pursue a political career are often a combination of factors (Borchert Reference Borchert2011), the latent variable of political ambition is exhibited when individuals enter their first political contest (Fox and Lawless Reference Fox and Lawless2005; Maestas et al. Reference Maestas and Sarah Fulton2006). This literature suggests that politically ambitious first ladies may tend to seek an elected position before entering the executive branch. Once in the executive, they may be more inclined to gain experience in politics by becoming directly involved in public policies, as Winfield (Reference Winfield and Norris1997, 167) has documented for American first ladies. Arguably, such experience is instrumental to preparing a future political career.
As for the structure of opportunity, favorable circumstances influence the decision to run for office (Fox and Lawless Reference Fox and Lawless2005, 644). Fox and Lawless (Reference Fox and Lawless2004, Reference Fox and Lawless2010) indicate that the support and encouragement from politicians affects women’s decision to run for office. O’Brien and Reyes-Housholder (Reference O’Brien, Reyes-Housholder, Rudy and Elgie2020) claim that candidate supply and political opportunities explain, better than cultural variables and mass-level indicators of gender equality, the increase of women in the executive branch. These works suggest that former first ladies may run for office when they perceive high chances of winning.
Most first ladies are political newcomers, a category that has been associated with new democracies and democracy level. Carreras (Reference Carreras2017, 9) proposes—and finds supportive evidence—that newcomers are more frequent in new democracies because political institutions have not taken root yet and voters are still inexperienced. With respect to democracy level, Corrales (Reference Corrales2008) argues that voters demand new politicians when political liberties decline and finds that newcomers succeed when authoritarianism increases. However, it is unclear if authoritarianism also paves the way for more women candidates because higher levels of democracy favor women’s access to power. For example, Morgan and Buice (Reference Morgan and Buice2013) found a connection between democratic values and gender egalitarianism after examining survey data from 19 Latin American countries.
In addition, newcomers have been linked to macroeconomic conditions. Corrales (Reference Corrales2008) claims that economic crises and accelerated growth increase the electoral demand for less conventional candidates because both junctures generate economic anxiety. In support of his argument, he found that economic growth and high inflation rates favor newcomers in Latin America. In contrast, Carreras (Reference Carreras2017) found that newcomers are less likely to be elected when the economy is growing. His interpretation is that voters turn to newcomers primarily in adverse economic scenarios. Although it is unclear what economic conditions may favor first ladies, they may consider the macroeconomic situation when deciding whether or not to run for office.
Theory: Political Experience and First Chance
We argue that accounts that portray former first ladies who run for national office as delegates of ex-presidents conflate women who do not have electoral experience before becoming first lady and those whodo. Contrasting cases can be illustrative. Since President Leonel Fernández was not allowed to run for another term, Dominicans living abroad drove the candidature of his wife for the vice presidency in 2011, despite Margarita Cedeño’s political inexperience. Her supporters used the slogan “With her we continue with him” (Con ella seguimos con él) as a symbol of continuity of her husband’s policies (Cruz Tejada Reference Cruz Tejada2010). Fernández’s Dominican Liberation Party nominated Cedeño for the vice presidency, and she became the running mate of Danilo Medina. They were elected (2012–16) and reelected (2016–20). Similarly, Xiomara Castro’s only political involvement before running for the Honduran presidency in 2013 was her participation in the 2005 presidential campaign of her husband, Manuel Zelaya, and organizing demonstrations after the coup against him in 2009.
The paths of Cedeño and Castro contrast to those of Hilda “Chiche” Duhalde, Cristina Fernández, and Emilia Alfaro, who were elected politicians before becoming first ladies and experienced a political upgrade after serving in the executive. Duhalde was a federal deputy before her husband, Eduardo Duhalde, became president of Argentina (2002–3). After his term, she was elected deputy and senator. Similarly, Alfaro was a deputy before her husband, Federico Franco, reached the Paraguayan presidency (2012–13), and she became a senator after serving as first lady. Fernández had vast political experience as deputy and senator before becoming first lady (2003–7). In 2007 she became the first woman Argentine president. (Table 4 in the appendix shows the career paths of all former first ladies who have run for office).
We argue that first ladies with elected political experience are more likely to run for office. First ladies who have been elected have already shown their ambition to compete for office, one of the main reasons why individuals become political candidates (Borchert Reference Borchert2011; Fox and Lawless Reference Fox and Lawless2004, Reference Fox and Lawless2005, Reference Fox and Lawless2010). Socialization and practical experience in politics also influence how politicians think about their skills for holding elected positions and thereby affect their motivations to pursue a candidacy (Fox and Lawless Reference Fox and Lawless2005, 653). Arguably, these first ladies have gained confidence in their abilities to have a political career, and they can show that they know how to perform well in elected office. In line with this, research has shown that the strongest candidates for office, other than incumbents, are experienced politicians (Berkmann Reference Berkman1994, 1028).
Our argument aligns with the observation that politically experienced first ladies seem to act strategically in the executive branch. As first ladies, they tend to pursue activities that portray them as competent decisionmakers and attractive personalities. They often give interviews, get involved in public policies, participate in charities, and engage in political meetings. For example, the press and some analysts informally called Nicaragua’s first lady Rosario Murillo “copresident” for her involvement on domestic and international affairs (Brandoli Reference Brandoli2016). In 2016, Murillo formalized her influence by running as vice president alongside her husband. They were elected in November of that year. To sum up, first ladies with electoral experience are politicians who, after leaving the executive, are in a favorable position to lead another campaign.
To be sure, our argument does not imply that citizens differentiate between politically inexperienced and experienced first ladies when casting a vote. Additionally, the argument does not imply that the previous political experience (or lack thereof) of a former first lady reveals meaningful information about her personal relationship to an ex-president or the ideological distance between them.
Furthermore, we do not claim that previous elected experience and delegate status are necessarily mutually exclusive categories. In theory, a politically experienced first lady could agree to act as a delegate of a former president while an inexperienced woman may act independently. However, our argument does claim that the behavior of politically experienced first ladies will tend to be distinct from inexperienced ones because they have already demonstrated political ambition. This ambition, added to the practical knowledge of leading a successful campaign, makes them more likely to run for office. In contrast, the conventional view does not differentiate among former first ladies who run for office: they are all assumed to simply represent a male politician. In sum, we propose
H1 Former first ladies are more likely to run for national office when they have experience as elected politicians.
An implication of the conventional wisdom is that former first ladies will try to capitalize on the popularity that their associates enjoyed as presidents. If such a view is correct, then former first ladies should run at the first chance they have after leaving an executive led by a popular politician. However, our expectation is different. Since we regard politically experienced first ladies as independent politicians, we expect them to run at the first opportunity they have after leaving the executive. Running early offers first ladies an “incumbency advantage,” due to their access to public resources, media coverage, voters’ attention, and influential networks. These advantages decrease or are lost if first ladies wait until later elections to run for office. This expectation is aligned with the scholarly literature that suggests that women politicians often need to take opportunities as soon as they can, due to limited future opportunities (Arana Araya et al. Reference Arana Araya, Hughes and Pérez-Liñán2020).
It could be argued that experienced first ladies would wait for more favorable timing before running for office if they were associated with an unpopular leader. However, we argue that in most cases, the advantages of becoming a candidate early offset the costs of being linked to an unpopular president. Experienced first ladies have their own political capital, which can shield them from negative associations. That explains why Patricia Alfaro of Paraguay successfully ran for the Senate in 2013 even though her husband, Federico Franco, had a popularity rate of 27 percent in his last year in power. In addition, first ladies who distance themselves from an ex-president may run a campaign critical of them, as Susana Higuchi of Peru did in 2000. Therefore:
H2 Politically experienced former first ladies are more likely to run for national office as soon as they leave the executive branch.
Empirical Analysis
The sample covers 1999–2016 because 1999 was the first year in which a former first lady ran for office in Latin America: María Julia Pou won a seat in the Uruguayan Senate. We will also reexamine the theory since 1990, the year women became more competitive as candidates with the election of Violeta Chamorro as the first woman president in the region. Interestingly, the same first ladies are included in both samples because those who left power since 1990 were still alive in at least one of the elections that took place in 1999 or afterward.
We study legislative and presidential elections because former first ladies have actively run for both branches (12 times for the executive, 14 for Congress) and we cannot identify a priori the branch for which former first ladies will run. Some, such as Cristina Fernández and Keiko Fujimori of Perú, have run for both branches. However, since presidential and congressional elections follow different electoral rules and have their distinct logics, we will conduct a robustness check just focusing on the most common type of candidacy (i.e., congressional). Subnational elections are not included because they have failed to attract former first ladies: only Marta Larraechea of Chile and Marisabel Rodríguez of Venezuela ran for mayor and lost.
The data for the 90 first ladies who were available to run for office in the 94 elections that took place from 1999 to 2016 come from a dataset built using multiple sources. These include the presidents’ and first ladies’ websites, first ladies’ public speeches, biographies, published interviews, and media outlets (sources and information about first ladies included are in the online appendix).
Since our hypotheses refer to individual predictors, the unit of analysis is first lady–election year. We exclude from the sample the first lady–election years in which there was only a presidential election and the first lady was constitutionally forbidden to run. This applies to Raquel Blandón (Guatemala, 1991), Elizabeth Aguirre de Calderón (El Salvador, 1999), Lourdes Rodríguez de Flores (El Salvador 2004), Vanda Pignato (El Salvador 2014), and María Gabriela Chávez (Venezuela 2013). We also do not include the six women who could not run for office because they held a foreign nationality: Virginia Gillum (Bolivia), Josefina Villalobos (Ecuador), Bessy Watson (Honduras), Mary Flake (Honduras), Aguas Ocañas (Honduras), and Eliane Karp (Peru). Finally, because our focus is the behavior of former first ladies, we do not include the two candidacies of first ladies that occurred in midterm elections (Cristina Fernández in 2005 and Cilia Flores of Venezuela in 2016).
Our dataset allows us to identify relevant characteristics of the 90 former first ladies’ trajectories. For example, 68 of them had hands-on experience in public policy, 7 had experience as popularly elected politicians before reaching the executive, and 7 had experience as appointed politicians. The candidates available per country from 1999 to 2016 varied between 2 (Dominican Republic) and 8 (Guatemala).
The dependent variable Attempt captures whether a former first lady ran for the presidency, vice presidency, or Congress (coded as 1) or not (zero). Noticeably, Attempt is indifferent as to whether first ladies were elected or not and to who runs more than once. Previously Elected captures the first hypothesis. It takes the value of 1 when a candidate held a popularly elected position before becoming a first lady and zero otherwise. Vice presidents, governors, senators, representatives, subnational legislators, members of constitutional assemblies, mayors, and city counselors were coded as holding an elected position. To test the second hypothesis, we will interact Previously Elected with First Chance, a dichotomous variable that captures when a first lady becomes a candidate in the first election after leaving the executive.
We control for first ladies’ appointed (i.e., not popularly elected) previous political experience. Previously Appointed takes the value of 1 when a first lady had experience as an appointed politician and zero otherwise. Ministers, vice ministers, ambassadors, consuls, and non–popularly elected party leaders are included. As a robustness check, we also control for Politician. This dichotomous variable identifies all first ladies with elected or appointed previous political experience.
According to the conventional view, the popularity of a male president should indicate how competitive his first lady can be as a candidate. Popularity is taken from the Executive Approval Project (Carlin et al. Reference Carlin, Hartlyn, Hellwig, Love, Martínez Gallardo and Singer2016) and captures the average popular support that former first ladies’ presidents had in their last year in office.
The literature proposes that women are more likely to be elected when there are gender quotas implemented and the number of women politicians is higher. Quota Adopted receives the value of 1 when a gender quota law has been implemented to elect legislators, and zero otherwise. Given that some quotas may not work as planned, as a robustness check we use Effective Quota. This variable is coded as 1 if the quota reaches 10 percent of de facto threshold (for either candidate or reserved seat quotas) and noncompliance is punished, or there are clear rules about the rank order of candidates. Women in Congress measures the percentage of women in the single or lower house for any given year. The three variables are taken from Hughes et al. (Reference Hughes, Paxton, Clayton and Zetterberg2017).
We also control for factors that limit women’s political advancement. Sexism measures the national level of sexism. It comes from the World Values Survey’s 2015 question, “On the whole, men make better political leaders than women do.” This variable ranges between zero (strongly disagree with the statement) and 4 (strongly agree) and represents the average scores for each country for the year the survey was conducted. Since the WVS does not include all countries, to fill out the missing values, we followed Arana et al. (Reference Arana Araya, Hughes and Pérez-Liñán2020) and used a latent growth curve model that estimated the predicted values of Sexism. Since Sexism is correlated with the women’s level of education, we used as predictors of the variable the women’s percentage of completed primary, secondary, and tertiary schooling (Barro and Lee Reference Barro and Lee2013). Sexism also correlates with year (–0.55). Therefore, we added as a predictor a time counter with a random slope that captured the linear trend by country.
As a robustness check for sexism, we use Political Discussion. This variable is taken from the V-Dem Project (Coppedge et al. Reference Coppedge, Gerring, Henrik Knutsen, Lindberg, Teorell, Altman and Bernhard2020) and captures whether women can openly discuss political issues in private homes and in public spaces.
Education captures women’s ability to reach positions of power using the average number of years of schooling for women in each country. It is taken from the Educational Attainment Dataset (Barro and Lee Reference Barro and Lee2013), which has five-year interval data for the sample (the post-2010 data are based on projections).
Research about political newcomers suggests that the age and level of democracy may relate to the opportunities for women in politics. Therefore, Democracy Age takes the number of years a country has been a democracy. The data are taken from Polity IV (Marshall et al Reference Marshall, Gurr and Jaggers2017). Democracy Level differentiates between semidemocracies and democracies. The data come from Mainwaring and Pérez-Liñán Reference Mainwaring and Pérez-Liñán2013. Scholars in this literature also propose that economic changes offer opportunities for nontraditional candidates. We capture this argument with annualized measures of GDP per capita growth (GDP Growth) and inflation, taken from the World Bank Indicators (2018). Due to the high variation in Inflation, we use its natural logarithm.
In addition, some women may have had the ambition to be elected to office but did not have the chance to become candidates before becoming first ladies. These women had nonetheless the opportunity to show their political commitment by getting involved in public policy once in the executive. Therefore, Public Policy takes the value of 1 when a former first lady had hands-on experience in public policy in the executive branch, and zero otherwise. This experience is documented in the online appendix.
Results
Given that the dependent variable reflects the attempts of first ladies to get elected, we estimate discrete time duration models. Using this technique, once a first lady runs for office, she becomes censored and drops from the sample.
Table 1 presents six probit models. The baseline model (1.1) includes the independent variables Previously Elected and First Chance and controls that capture conventional wisdom arguments. In 1.2, we interact the independent variables to test H2. As a further test of the conventional wisdom, in 1.3 we examine whether former first ladies are more likely to run for office at the first chance they have, conditional on the outgoing president’s approval rates. In models 1.4 and 1.5, we retest the argument using variations of the political experience of former first ladies. Model 1.4 uses Previously Appointed while model 1.5 adds Politician. Finally, model 1.6 clusters standard errors by country to account for unobserved country-level idiosyncrasies and error correlation.
Former First Ladies Who Run for Office
+ p < 0.1, * p < 0.05, ** p < 0.01.
Entries are probit coefficients (standard errors).
The results in table 1 support H1 across all models: former first ladies who have previously held an elected position are more likely to run for office. The result does not hold when Previously Elected is replaced by Previously Appointed (1.4) but does hold when Politician replaces Previously Elected (in 1.5). Since Politician includes elected and appointed previous political experience, the results suggest that the statistical significance of this variable is explained by the elected experience. In support of H2, model 1.2 shows that first ladies who have electoral political experience are also more likely to become candidates the first chance they have after leaving the executive.
Popularity is not statistically significant across any model, and model 1.3 shows that women associated with popular presidents are not more likely to run for office as soon as they can. These results strongly suggest that the popularity of presidents is unrelated to the candidacies of former first ladies, as the conventional wisdom proposes.
Revealingly, the variables that capture arguments from the literature about women in politics and political newcomers are not statistically significant across most models, reinforcing the notions that first ladies are a unique type of candidate and that the scholarly literature has not addressed their motivations to run for office. The exceptions are Quota Adopted in 1.3, Women in Congress in 1.4, and Sexism, Education, and Inflation in 1.6. The somewhat divergent results in 1.6 suggest some differences across countries.
Curiously, the only control variable that is statistically significant across all models (except 1.6) is Public Policy. This gives insight about the first ladies’ engagement in politics and reinforces the proposition that the trajectory of first ladies is central to understanding their candidacies.
We estimated predicted probabilities to gain a better understanding of the substantive impact of Previously Elected and the interaction Previously Elected*First Chance. The lefthand graph in figure 1 shows that when first ladies do not have experience as elected politicians, the predicted probability that they will run for office after leaving the executive is only 0.2 percent, but rises to 70 percent when they have such experience. The righthand graph shows that the probability that previously elected first ladies will run for office the first chance they have is 86 percent, and 16 percent otherwise. This strongly suggests that they decide to become candidates as soon as they leave the executive. In contrast, the chances that politically inexperienced former first ladies will run for office are almost unaffected by whether they run in the first election after leaving the executive power (4 percent of probabilities) or not (5 percent).
Political Experience and Running for Office.
In table 2 we retest the arguments using five alternative samples. In 2.1, we exclude elections under semidemocratic governments because the motivations that first ladies have to run for office may vary across regime types. It could be argued that former first ladies are unlikely to become candidates if they are too old, have not shown interest in running for office on multiple occasions, or served for too short a period to use their position in the executive to support a political career. To account for these limitations, 2.2 excludes women who became first ladies at 70 years old or older, reached the age of 70 without running for office, did not run after 7 elections as former first ladies, or served in the executive for less than a year. As a consequence, 14 women were dropped from the sample.
Alternative Samples
+ p < 0.1, * p < 0.05, ** p < 0.01. Entries are probit coefficients (standard errors).
Model 2.3 excludes the seven daughters and three sisters of presidents who served as first ladies because they may not have had electoral ambitions, since they did not choose to be associated with a politician. Model 2.4 uses an extended sample and includes former first ladies since 1990. And model 2.5 includes only legislative elections because the reasons that lure first ladies to run for office may differ across branches of government. The observations in which first ladies run for the executive power are excluded from 2.5 because they could not run simultaneously for Congress.
The results in table 2 are categorical: they hold across samples. Perhaps the most notable change is the increased effect size of Previously Elected when the sample only includes competitive (2.1) and congressional elections (2.5). Arguably, this occurs because most candidacies occurred in democracies and were for Congress.
In table 3 we conduct eight robustness checks using alternative specifications. Model 3.1 includes a dichotomous variable that identifies countries that, at some point, constitutionally forbade relatives of the acting president to run for the presidency: Bolivia, Brazil, Costa Rica, Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, Panama, Paraguay, Peru, and Venezuela (we did not find any restriction on running for Congress). In 3.2, we retest the argument that women politicians encourage other women to enter politics, adding Diffusion. This variable takes the value of 1 when a former first lady has run for office in the same country and zero otherwise. In 3.3, we add Populist, a dichotomous variable that captures whether a former first lady accompanied a populist leader in office, as identified by Kyle and Gultchin (Reference Kyle and Gultchin2018). In 3.4, we include Leftist, a variable that takes the value of 1 when a leftist government is in power and zero otherwise. The aim is to test whether leftist leaders open up more opportunities for women, as Escobar-Lemmon and Taylor-Robinson (Reference Escobar-Lemmon and Taylor-Robinson2005) suggest (although the study of Funk et al. [Reference Funk, Hinojosa and Piscopo2017] finds that left parties do not necessarily increase women’s political representation). We take this variable from Arana Araya (Reference Arana Araya2017).
Alternative Specifications
+ p < 0.1, * p < 0.05, ** p < 0.01. Entries are probit coefficients (standard errors).
++ This variable represents the coefficients (standard errors) for Restrictions, Diffusion, Populist, Leftist, GDP per capita, Political discussion, Region, Effective quota, in models 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, and 3.7, respectively.
In 3.5, we include the level of GDP per capita—taken from the World Bank (2018)—as an additional test of the argument that the economic context affects political opportunities for women. In 3.6, we reexamine whether gender inequality is associated with the candidacies of former first ladies, including Political Discussion. Region (model 3.7) controls for a potential geographic concentration of the candidacies. It takes the value of 1 when a former first lady is from South America and zero otherwise. Model 3.8 retests the argument that electoral quotas may be related to the first ladies’ candidacies using Effective Quota. Revealingly, none of the new variables included became statistically significant.
Conclusions
The conventional view implies that the political trajectory of first ladies is irrelevant to understanding their subsequent political behavior. It also entails the notion that women who served with popular presidents should be more likely to run for office, especially as soon as they leave the executive, to capitalize on such popularity. However, we found no support for these implications. In contrast, we found strong support for our argument that electorally experienced first ladies who run for office should be regarded as politicians who use the role to become competitive candidates. Our results show that there is a 70 percent chance that a first lady with elected political experience will run for national office, while the likelihood is less than 1 percent otherwise. Also as hypothesized, we found that first ladies with electoral experience take advantage of their first chance to compete in elections, irrespective of the presidents’ popularity. Furthermore, the findings show that those who engage in public policy as first lady are more likely to run for office afterward.
The results suggest that the conventional view is based on a prejudice that gives no leverage to understanding when former first ladies will run for office. However, differentiating between those who have elected experience and those who do not allows for making accurate predictions of which and when former first ladies will run for office.
We believe that the candidacies of politically experienced former first ladies are likely to promote women’s representation by helping to balance the gender disparity in positions of political power. Furthermore, through their engagement in the public debate, they are likely to promote the entrance of more women into the electoral arena. Although they are a fraction of women legislators, some have successfully run for the presidency and vice presidency, which allows them to exercise considerable political influence.
The election of former first ladies to national office is likely to keep growing. The pool of potential candidates keeps expanding, and they are highly successful. The trend is clear: 15 of the 26 candidacies that we studied occurred in the last 6 years of our sample. Since then, 5 women have aimed for the presidency: Margarita Zavala of Mexico in 2018 (she quit before the election), Sandra Torres in 2019, and Xiomara Castro, Keiko Fujimori, and Cristiana Chamorro in 2021 (although ultimately Chamorro could not run because the Nicaraguan government disqualified and arrested her). Former first ladies have also been active in the vice presidency. Lucía Topolansky was appointed vice president of Uruguay in 2017 but quit in 2020 to become s senator. Cristina Fernández became vice president in 2019; Margarita Cedeño failed in her attempt to be reelected as vice president in 2020; and Rosario Murillo was reelected as vice president in the controversial 2021 general elections. The trend may also add first ladies at the subnational level, something that has already happened in Argentina (Martin Reference Martin2018).
This study creates avenues for research. Qualitative research can shed light on how political couples who reach the presidency work to advance their careers as a team. As noted by Joignant (Reference Joignant2014, 29), researchers have failed to examine the potential transfer of political capital in a married couple. To open up the black box of what first ladies do in the executive power, scholars could examine their involvement in public policies and the consequences of such engagement. Furthermore, the effect that former first ladies who win elections have on the circulation, competition, and integration of elites deserves more attention (Arana Araya Reference Arana Araya and Farazmand2018). Having even larger political families in the region may constrain political competition. However, if experienced first ladies act independently and promote women’s representation, their contribution to diversifying the political elite could offset their effect on increasing the influence of political families.
Supporting Information
Additional supporting materials may be found with the online version of this article at the publisher’s website: Appendix.
