Representation from Below: How Women’s Grassroots Party Activism Promotes Equal Political Participation

Abstract

Extensive research investigates the impact of descriptive representation on women’s political participation; yet, the underlying mechanisms remain unclear. This article develops a novel theory of descriptive representation, arguing that women politicians mobilize women’s political participation by recruiting women as grassroots party activists. Evidence from a citizen survey and the natural experiment of gender quotas in India confirm that women politicians are more likely to recruit women party activists, and citizens report greater contact with them in reserved constituencies during elections. Furthermore, with women party activists at the helm, electoral campaigns are more likely to contact women, and activist contact is positively associated with political knowledge and participation. Evidence from representative surveys of politicians and party activists and fieldwork in campaigns, further support the theory. The findings highlight the pivotal role of women’s party activism in shaping women’s political behavior, especially in contexts with pervasive clientelism and persistent gender unequal norms.

INTRODUCTION

Vast gender gaps in political knowledge and political participation persist across the world (Barnes and Burchard 2013; Desposato and Norrander 2009; Robinson and Gottlieb 2021). As of 2022, only one-quarter of the legislators worldwide are women and women are under-represented across party hierarchies (O’Brien 2015). These gender inequalities raise normative concerns, while leaving unrealized the full potential of women’s political participation such as greater democratic legitimacy (Clayton, O’Brien, and Piscopo 2019) and enhanced substantive representation (Chattopadhyay and Duflo 2004). Existing research suggests that descriptive representation offers a key means of increasing women citizen’s political knowledge and participation by changing attitudes, norms, and perceptions. ¹ However, it is unclear how these effects occur in settings where unequal gender norms remain entrenched and continue to impose high costs on women’s political participation (Jayachandran 2015).

This article offers a novel theory of descriptive representation, suggesting that women politicians need not change deeply entrenched beliefs to mobilize women into politics. Instead, women politicians can leverage their position in the electoral hierarchy to build political infrastructure that eases the proximate constraints on women’s political participation. Women do so by transforming grassroots party organization to recruit women as party activists—party agents who mobilize citizens on behalf of politicians. ² Women in local politics are well placed in the party hierarchy to recruit women to become party activists. They can simultaneously lower household and party gatekeeping barriers that impede status-quo party recruitment and keep activist roles outside of women’s reach. This institutional change puts women party activists at the helm of ground campaigns, which increases the likelihood that ground campaigns will canvass men and women more equally, shrinking the gender gap in partisan contact (citizen–activist interaction). Men and women who receive this partisan contact are, in turn, more likely to know more about politics and are more likely to participate in politics.

Women politicians can impact citizen’s political engagement through several mechanisms. ³ This article focuses on party activism because of the critical role party activists play in shaping citizen’s political behavior and in mediating access to the state in clientelistic settings both during and between elections (Auerbach and Thachil 2018; Brierley and Nathan 2020). ⁴ India is a case in point. To access state benefits, more than a fifth and a third of citizens in rural and urban India turn to party activists (Auerbach and Kruks-Wisner 2020). In every national election, approximately five hundred million citizens are contacted by party activists, with a similar or greater intensity of activist contact in India’s state and local elections. ⁵ Crucially, activism is a critical pathway for political entry. Becoming a party activist is the first step to becoming a politician, especially for those who are from non-elite backgrounds (Goyal Forthcoming). Studies of party organization and party building highlight the importance of grassroots party presence for party resilience (Chhibber, Jensenius, and Suryanarayan 2014). Yet, men dominate as party activists, and the cause and consequences of the vast gender gaps in party activism remain unknown.

This article provides the first evidence establishing the link between descriptive representation and women’s grassroots activism using the natural experiment of gender quotas in the municipal council of India’s capital city, Delhi. An as-if random protocol reserves half of all constituencies for women to contest elections, allowing for the identification of the effect of mandated representation while overcoming selection bias. I use a representative citizen survey with a novel measure that asks about party activists’ gender and contact during local electoral campaigns, confirming the influence of women’s recruitment of women activists on citizen–party interaction. Data from surveys with local politicians and party activists in Delhi, and shadowing of ground campaigns in diverse sites and elections in India, corroborate support for the theory and increase external validity of the findings.

The article’s theory is well supported by various data sources. Men and women in reserved constituencies are approximately 2.4 times as likely to be contacted by women party activists relative to those in non-reserved constituencies. Results underscore that women politicians have the same canvassing capacity as men. Furthermore, party activists reach out to citizens more equitably in reserved constituencies and the gender gap in partisan contact—which is 20 percentage points in non-reserved constituencies—halves in reserved constituencies. Additionally, receiving partisan contact is positively associated with citizens’ political knowledge and participation in multiple political domains, as confirmed through a qualitatively informed fixed effects strategy and matching estimates. Qualitative evidence strengthens support for theory by opening the black box of grassroots party recruitment and local campaigns, showing who becomes an activist and how door-to-door local campaigns are organized.

This article introduces the link between descriptive representation and grassroots party activism and shows its implications for partisan mobilization and political participation. These findings advance our understanding of descriptive representation, suggesting that the spillovers of descriptive representation are more fully understood by focusing on what women do when they are in politics, not only on what women symbolize. The article theorizes “party activism” as a key mechanism that moderates how descriptive representation influences political participation and policy outcomes in developing countries (Clayton 2021), and has direct implications for women’s political and claim-making behavior. The article extends research on women’s representation and gender quotas by illustrating how women’s influence can reach beyond policy goals to build inclusive party organization that are the bedrock of democracy (Bermeo and Yashar 2016). It also contributes to research on gender and development by showing that descriptive representation advances political participation in deeply patriarchal societies with entrenched gender unequal norms (Jayachandran 2015). Although policy interventions can lower supply-side or household gatekeeping barriers to women’s political participation (Gottlieb 2016), the findings show how descriptive representation via party activism can simultaneously lower barriers imposed by multiple gatekeepers. This finding holds implications for interventions aimed at advancing equality in multiple forms of political participation.

This article contributes to distinct bodies of scholarship approaching political selection and recruitment. Comparative scholarship on candidate selection underscores the greater importance of party recruitment for women and shows how male party gatekeepers fail to recruit women candidates (Krook 2010). This article improves our understanding of party recruitment by opening up the black box of gender and party recruitment at the activist level, and offering an in-depth account of how party gatekeepers influence grassroots activist recruitment and the outreach of electoral campaigns. The literature on clientelism has also studied how brokers emerge in grassroots politics (Auerbach and Thachil 2018; Brierley and Nathan 2020; Wantchekon 2003), but has largely overlooked the role of identity. This article contributes to our understanding of clientelism by theorizing how descriptive representation diversifies party networks from the bottoms-up and showing that women can gain a competitive edge in clientelism.

HOW DESCRIPTIVE REPRESENTATION INCREASES WOMEN’S POLITICAL PARTICIPATION?

The relationship between descriptive representation and political participation has received significant attention in the gender and politics scholarship and in public discourse. In developing countries, where stark gender gaps in politics exist, this relationship holds significant promise in the wake of gender-quota reforms. The near global implementation of gender-quota reforms means women have entered political office in unprecedented numbers, despite the persistence of economic underdevelopment and unequal gender norms (Tripp and Kang 2008).

Scholarship investigating the effect of women’s presence in politics on women’s political participation in the Global South has predominantly found a positive relationship between the two, particularly in India. Exploiting the randomized implementation of gender quotas in rural India, studies have provided causal evidence for this relationship in a deeply patriarchal society (Chattopadhyay and Duflo 2004; Deininger et al. 2015). Cross-national research further extends these findings (Barnes and Burchard 2013). However, despite publication bias against statistically weaker findings, there are increasing concerns about null and negative findings (Beaman et al. 2009; Goyal 2020; Liu 2018; Morgan and Buice 2013).

In addition to mixed findings about whether descriptive representation affects women’s political participation, how or why there may be an effect is unclear. Building on representation theory (Mansbridge 1999), scholars have argued that descriptive representation demonstrates to women that politics is not solely a man’s domain; in turn, these signals lower the internal barriers holding women back from political participation and increase women’s political efficacy and subsequently improving their political participation. Women can act as role models (Wolbrecht and Campbell 2007) and change stereotypes about politics and women’s role in politics (Beaman et al. 2009; 2012). According to this symbolic theory of representation, attitudinal changes universally increase women’s political participation.

This logic rests on two premises, both of which have limited evidence to support them. The first premise, that having more women in leadership positions, will improve women’s attitudes toward politics or boost their political efficacy is not strongly supported by empirical evidence. Research in developing countries highlights two key hurdles: slow moving and restrictive gender norms and women’s lower political knowledge. Beaman et al. (2009) find no short- or long-term attitudinal effects of mandated representation among women constituents in India. Instead, they find gender stereotypes strengthen among women constituents in the short term, with an explicit “backlash” occurring among citizens in the long run (1532); the backlash persists despite implicit same-gender leader preference. Highlighting the regressive nature of gender norms in Asia, Liu (2018) calls symbolic effects into question and shows that female political leaders generate a backlash effect on women’s political engagement where gender norms are regressive. Goyal (2020) provides experimental evidence for backlash against women politicians in India. She conducts a visual experiment showing that the change in women’s political efficacy is negative when they are exposed to the photograph of their female representative relative to women who are exposed to a male representative’s photograph. Investigating a development program in Mali, Gottlieb (2016) finds that in place where norms against women’s public roles remain entrenched, non-partisan interventions that are aimed at improving women’s political participation can backfire as women self-impose limits to future civic participation and men erect new barriers as a form of backlash.

Research also shows changing internal attitudes—the second premise—is insufficient for increasing political participation in contexts with highly unequal gender norms and systemic barriers to women’s political participation. Using data across 18 sub-Saharan African countries, Coffe and Bolzendahl (2011) conclude that large gender gaps in political participation cannot be explained by gender differences in political attitudes. Iyer and Mani (2019) find that neither internal nor external political efficacy measures are a significant predictor of political participation in rural India. Thus, growing evidence suggests that attitudinal effects either do not materialize or are insufficient in promoting women’s political participation.

This article offers an alternative theory of descriptive representation which suggests that women politicians need not change entrenched gender norms or internal beliefs to mobilize women. Instead, descriptive representation enables women to ease proximate constraints on women’s political participation through the channel of women’s grassroots party activism. This theory also suggests that the positive effects of descriptive representation are not universal. They also not subject to a ceiling or untenable in the short term. Instead, they are conditional on receiving effective party activist contact, which is one explanation for the mixed findings in the literature. ⁶ Aggregate null results, therefore, do not mean that descriptive representation is inconsequential for women’s political participation, but rather that effects are heterogeneous and not universal, and there is potential for moderating variables as discussed in Clayton (2021). The theory presented in this article is one step in this direction.

Grassroots Party Activism-Based Theory of Descriptive Representation

Some of the most persistent gender inequalities worldwide are in political party positions, with male-dominant parties perpetuating inefficiencies that keep women out of politics. In developing countries, grassroots party activism constitutes a crucial political channel through which ordinary citizens become politicians and engage the state. While fewer women are activists, the gender gap in party activism has received less scholarly attention. Primary data from recent studies highlight gender gaps in more formalized forms of party activism (Auerbach and Thachil 2018; Brierley and Nathan 2020; Goyal and Sells 2021), but this only scratches the tip of the iceberg. The majority of party activists are only informally affiliated with political parties and comprise the lowest hierarchy in the party organization. Such informal party agents are not mentioned in formal part lists, even when they have remained active in the party for several years. ⁷

Research on gender gaps in party activism is also sparse; however, research on women’s participation in electoral politics suggests that such supply-side factors as household barriers (Chhibber 2002), domestic responsibilities (Iversen and Rosenbluth 2010), and lack of ambition (Lawless 2015) can pose barriers to women becoming party activists. For instance, conducting ethnographic research on women party workers in India, Bedi (2016, 185) documents that politically active women in intergenerational and patrilocal households have to negotiate their domestic chores with other women family members in order to get approval to participate in party politics. Because activism brings various advantages to the family and the local community, it builds relational support for women’s political work. Women whose families are highly conservative or women who are unable to reduce their domestic burden, such as unmarried women or women with young children, are less likely to become party activists. In addition, the gendered nature of political socialization can depress women’s political ambition and self-esteem, making their activism less likely. Indeed, a sizeable minority of women activists that I interviewed mentioned that prior to becoming party activists they had “no politics in their life.” In contrast, men activists were more likely to mention being active in youth or student politics and were far more likely to become an activist before getting married.

In addition to these supply-side barriers, demand-side barriers further depress the recruitment of women as party activists. Male-dominant parties can gatekeep women out of political positions, although party recruitment is more important for women’s political selection and is considered the key reason for women’s political under-representation in a range of contexts (Krook 2010; Lawless 2015). At the activist level, the gendered division of party activities can direct women to non-electoral work which is less valued by party leaders (Daby 2021). This can reinforce women’s lack of trust in the sponsorship of male party leaders, and consequently make male party leaders less successful in recruiting women activists (Gulzar et al. 2023; Preece, Stoddard, and Fisher 2016). Male-dominant parties and recruitment networks may also be unable to access social networks and spaces where women are active and autonomous participants, thereby ignoring women talent that can be incorporated into party politics.

How Women Recruit Women as Party Activists

Descriptive representation offers a resolution—it eases the household- and party-side barriers hindering women from becoming party activists. Women politicians have two comparative advantages in recruiting women party activists. They can gain approval for women’s party activism from household gatekeepers, lowering supply-side barriers. By becoming one of the gatekeepers to party activist recruitment, they also lower demand-side barriers that exist in male heavy party organizations.

Responding to household gatekeepers, women politicians actively and effectively persuade other women to become party activists. Evidence shows that women are both more persuasive and more co-operative in women-majority groups (Díaz-Martin et al. 2022). Women can convince family gatekeepers in the household to allow women to participate in politics, and families are less reluctant when women’s political participation is with other women (Cheema et al. 2023; Prillaman 2021). The presence of other women in public and political spaces can lower women politicians’ and women party activists’ concerns about their reputation and safety. Becoming a party activist also has instrumental benefits. In patronage settings where even marginal political access helps get things done, having social and political connections to local party leaders is extremely valuable. These benefits accrue to women and their families and may further lower a family’s reluctance to a woman’s political participation. However, only lowering supply-side barriers remains insufficient for women’s participation, as women continue to face hurdles accessing opportunities inside male-dominant party organizations.

Responding to party-side barriers, women politicians are well positioned to identify and recruit women from networks and spaces where women are relatively more numerous and active. For instance, my field research shows that in India, women politicians recruit women activists from healthcare centers (anganwadi), school management committees, tuition centers, self-help groups, temples, and NGOs. Tapping into this larger pool of potential recruits lowers the costs of recruiting women activists. Recruiting women activists complies with women’s social roles which lowers reputation costs, as women politicians are likely to be targets of violence and harassment if they are seen with strange men (Krook 2017). Women can use their better understanding of women’s preferences and utilize their party networks to assign women to more meaningful activities where women activists can claim credit for their labor and develop stronger grassroots networks. The recognition that women activists’ labor can translate into material and political opportunities can motivate women activists to join and work hard for the party (Bedi 2016). Meanwhile, women politicians not only benefit from the support of their loyal and capable women activists during party nominations, it is also easier for women to claim credit for recruiting women to build the party’s grassroots organization and for providing party leaders with valuable infrastructure to access the latent women’s vote. Therefore, such recruitment becomes a win–win strategy for women politicians (Goyal and Sells 2021).

This argument suggests that women politicians are more likely than men politicians to recruit women as activists. In the context of electoral campaigns, women politicians can achieve this change in two ways. First, they can increase the number of informal women activists but keeping the number of men activists the same (pathway one). They can also achieve this change by offering more informal roles to women over men activists (pathway two). Both pathways are consistent with the theory. In either pathway, women are more likely than men to recruit women activists, and the gender gap in party activism is lowered. I summarize this hypothesis below.

H1: Women politicians are more likely than men politicians to recruit women party activists.

Implications for the Gender Gap in Partisan Mobilization

Descriptive representation decreases both party recruitment and political mobilization inefficiency. In contexts with restrictive gender norms, which also often have gender-segregated spaces, party organizations dominated by men are ineffective or uneven mobilizers. Such patriarchal settings restrict mixed-sex interactions initiated by men and have fewer women present in public spaces, preventing men from accessing women constituents. Even in industrialized contexts where contact is pre-planned and aimed at men and women equally, the canvasser’s visible traits, such as gender and ethnicity, influence the likelihood and substance of contact. Monitoring male activists’ attempts to contact women does not minimize this principal-agent problem (Enos and Hersh 2015). Lack of women in public spaces can make these tendencies worse. In other words, I argue that there is likely to be a gender gap in partisan contact in such settings.

Ground campaigns that have a more balanced gender composition of party activists, are more likely to be efficient and equitable in their outreach and policy content and can eliminate or lower this gender gap. Women and mixed-gender party activist groups can access men in public spaces. For example, Bedi (2016) notes that women party activists in India can “enter kitchens,” unlike men who are confined to the doorsteps. Darwin (2017) finds that women brokers in Indonesia organized meetings with both men and women not only in prayer halls or community-level religious groups, but also in spaces that are traditionally male spaces such as coffee shops. In the United States, Carpenter and Moore (2014) find that women activists have a higher outreach and are more effective in persuading citizens. In other words, women are more likely than men activists to reach out to men and women more equitably.

H2: Constituencies with women politicians are more likely to lower the gender gap in party activist campaign contact.

Implications for Citizen’s Political Participation

Women play a role in promoting gender parity in political participation by conducting more gender-equal partisan outreach. Partisan contact serves as a powerful mobilizer for several reasons. Field experiments in American politics show door-to-door contact—partisan and non-partisan—effectively increases citizen’s political participation (Alvarez, Hopkins, and Sinclair 2010; Green and Gerber 2019). Contact with either men or women party activists allows constituents to receive direct information about political events. Moreover, unlike scripted forms of contact that are a hallmark of non-partisan contact, because party activists tailor their conversation to the recipient, such conversations can boost women’s interest in politics and increase their political knowledge, which can in turn increase political participation. ⁸ Research randomizing partisan contact is sparse (Alvarez, Hopkins, and Sinclair 2010 is an exception in the US), and to the best of my knowledge, Cruz (2023) is the only study that uses a field experiment showing that partisan contact persistently increased political engagement in mayoral elections in the Philippines. However, experimental evidence supports the claim that partisan content shapes political behavior (Gulzar et al. 2023; Wantchekon 2003).

Partisan contact provides additional benefits for women’s mobilization. Because parties benefit equally from men’s and women’s electoral participation, party activists can persuade household heads to let women participate in electoral politics, making it easier for women to turn out to vote. Even in cases where the contact is targeted to women, the household benefits from establishing an informal political connection; this relationship can help poor and lower-income families to access state benefits and avoid episodes of state repression. Partisan contact aimed at women signals to fathers, husbands, and mothers-in-law that politicians observe and attach electoral value to women’s political participation. This increase in the instrumental value of women’s political participation makes it more likely that women turn out to vote. ⁹

Partisan contact during elections can have a multiplier effect on non-electoral political participation through spillover effects. Such spillover effects will be stronger if activist contact is reinforced between elections, as it is likely that those who are contacted during elections remain connected through neighborhood meetings, phones, or WhatsApp groups. Party activists may also conduct door-to-door campaigns in between elections to mobilize citizens for public protests or for religious or political events, but less is known about this outreach and its effects on citizen’s non-electoral participation. ¹⁰

H3: Citizens who receive door-to-door contact during elections are more likely to have higher political knowledge, and are more likely to participate in politics during and in between elections.

CONTEXT AND RESEARCH DESIGN

The article focuses on Delhi, India’s capital city with a population of 23 million, as its empirical site. The positive association between descriptive representation and women’s political engagement is more likely to be observed in Delhi, as women have held top political positions and are well represented in municipal politics with a 50% reserved-seats gender quota. Additionally, Delhi’s urban residents have greater access to information, mobility, and decision-making power compared to rural regions in India, reducing external limitations on women and potentially strengthening the effect of descriptive representation on political engagement.

The Natural Experiment of Randomized Gender Quotas

The randomized reserved-seats gender quotas in the Municipal Corporation of Delhi (MCD) allow for causal inference. India’s gender reservation policy mandates that only women can contest elections in constituencies reserved for women, which are selected as-if-randomly. Unreserved constituencies are open to both men and women. The MCD has 272 wards that each elect a single councilor every 5 years through plurality rules, as shown in Figure 1. In 2017, half of the municipal constituencies were reserved for women through an as-if random process that involves reserving every second constituency from a serially ordered list of municipal constituencies; these serial numbers are in-turn as-if-randomly assigned. Many studies have used this design for the purpose of causal inference (Chattopadhyay and Duflo 2004). Section A.1 of the Supplementary Material provides balance tests supporting the design’s internal validity.

Figure 1. Gender Reservation in Delhi’s MCD 2017

Data

The study surveyed 1,664 low- to middle-income residents in 51 neighborhoods and 183 polling stations across 17 randomly selected municipal constituencies in Delhi, 9 in reserved wards and 8 in non-reserved wards, approximately 18 months after the 2017 municipal elections. This is among the few surveys to highlight measure political participation in an urban Indian setting. Section A.2 of the Supplementary Material details the sampling procedure and ethical considerations. Table 1 presents summary statistics showing gender gaps in resources, mobility, and political knowledge and participation, highlighting the existence of restrictive gender norms in Delhi, even with universal access to mobile phones. Women are less likely than men to independently own a mobile phone, require permission to attend political events, and are less likely to discuss politics with family and friends. It is striking that gender gaps in Delhi are larger than in rural parts of Uttar Pradesh (India’s poorest state) (Iyer and Mani 2019).

Table 1. Gender Gaps in Politics

Note: For all variables, the data source is the first wave of the panel citizen survey ( $ N=1,664 $ ), except for the second variable and the third variable where the data source is the second wave ( $ N=1,304 $ , re-interview rate $ = $ 78%). All variables are dummies otherwise indicated and missing values are coded as missing. The difference column reports the t-test of differences in means between men and women. Standard errors are clustered at the constituency level. *** $ p<0.01 $ , ** $ p<0.05 $ , * $ p<0.10 $ .

The citizen survey includes measures of activist contact. The survey asks: “In the last municipal elections in 2017, did any party activist visit you personally?” Party activists contacted close to 7 million—60% of Delhi’s electorate in a span of 1 month. Yet, contact is not equal; women are 11.7 percentage points less likely than men to be contacted by party activists. No election survey has collected data on party activist gender. This survey introduces a new measure, asking respondents: “Do you remember whether the party worker(s) were mostly men or women or a group of men and women party activists?” ¹¹

Although marginally less interested in politics than men, women believe both that their vote is equally important and that they can influence MCD politics, echoing findings that the gender gap in political efficacy is less pronounced in the Global South. Women feel more capable of contesting elections than men. Yet, women remain less likely than men to know about and participate in politics. Only 11% of women could name their local representative, whereas a mere 2.9% know about gender quotas. Even among men, the knowledge of gender quotas is low. I add and average these measures to create the political knowledge index.

Gender gaps are the smallest on registering a complaint, which citizens can do from home, and, as expected, on electoral participation. Gender gaps in turnout have shrunk in India since 1990s. Administrative data also confirm that the gender gap in turnout is marginal in Delhi’s local elections (less than 2 percentage points). However, women are less likely than men to participate in political events and protests related to women’s and environmental issues. Note that non-electoral participation is self-reported and the gender gap may reflect gendered measurement bias, especially if men over-report and women under-report their political participation. Therefore, it is important to interpret this gender gap with this caveat. I add and average the civic measures to create the non-electoral participation index.

Empirical Strategy

Because vast majority of women only run for office in reserved constituencies, I follow convention (as in Beaman et al. 2009) to report reduced form effects of mandated female candidacy in 2017. I run the following regressions:

(1) $$ \begin{array}{rl}{Y}_{ij}={\beta}_0+{\beta}_1{R}_j+{\epsilon}_1,& \end{array} $$

(2) $$ \begin{array}{rl}{Y}_{ij}={\beta}_2+{\beta}_3{R}_j+{\beta}_4{G}_{ij}+{\beta}_5{R}_j{G}_{ij}+{\epsilon}_1,& \end{array} $$

where $ {Y}_{ij} $ is the outcome of interest for respondent i in constituency j, $ {R}_j $ is the treatment and refers to the j constituency’s reservation status, and $ {G}_{ij} $ is the gender of the ith respondent in constituency j. The gender gap for an outcome is estimated by running a regression which includes an interaction between gender and reservation status of respondent i in constituency j as in Equation 2.

Technically, the treatment is the assignment of constituencies to gender reservations within a setting where quotas have been long introduced (quota shock is held constant). Note that the control group is not reserved in 2017 and the treatment group is reserved in 2017. Because there have been reservations in the past (since 1997), and these are orthogonal to reservations in the future, this comparison only provides contemporary effects of reservations and averages out over historical effects. Section A.1 of the Supplementary Material overviews the history of Delhi’s reservations. Qualitative evidence suggests that the effects of reservations on party activism persist and strengthen over time which means that causal estimates that I present here are under-estimates.

Substantively, the treatment is the assignment of an equal number of women as major party candidates as men. First, all major parties nominate candidates in all 272 constituencies, and out of 134 non-reserved districts, only five women ran as major-party nominees. Independent or small-party candidates are rarely successful. Second, because gender quotas are not rotated, but newly randomized and released only a few weeks prior to elections, candidates or parties cannot use past reservation status to strategically organize gendered recruiting efforts, as underscored by interviews. Candidates and incumbents cultivate and signal their grassroots support between elections, hoping to be rewarded with a party ticket at election time, which enables them to expand and engage their activist base during elections. Winning politicians strengthen their support base. It is important to note that like in most countries, electoral campaign period, and therefore the most intense activist recruitment period in India, lasts between 4 and 6 weeks. ¹² The vast majority of party activists and candidates are locally embedded in their constituencies and enter politics during campaigns. Sixty percent of residents report contact with party activists during elections, highlighting the concentration of activist–citizen interaction in this brief period. Evidence shows that campaigns conducted 3–4 weeks before elections significantly enhance voter turnout (Wantchekon 2003).

The estimates of interest are $ {\beta}_1 $ and $ {\beta}_5 $ , and the results do not include any controls for transparency. Section A.5.1 of the Supplementary Material shows that the results are robust to covariate adjustment. The results report robust standard errors clustered at the constituency level and are implemented in R. The clustered design is sufficiently powered (80%) to detect treatment effects as small as 6–10.3 percentage points with binary outcomes, assuming an intra-class correlation coefficient observed in the data that range from 0.001 to 0.03. I also report test statistics from wild cluster bootstrap procedures and a two-sided randomization inference test of zero treatment effects, which provides inference with correct size regardless of cluster or sample size.

MAIN RESULTS

Women’s Grassroots Party Activism

Table 2 presents the results corroborating support for H1. Column 1 constant indicates that 6.4% of non-reserved constituents are contacted by women activists. In stark contrast, citizens in reserved seats report approximately 2.4 times more contact with women activists than those in non-reserved constituencies. In other words, more than 15% of citizens in reserved constituencies reported being approached by women activists, indicating a fundamental change in the nature of who conducts ground campaign; these results lend strong support to the first argument.

Table 2. Women Party Activists Are More Likely to Contact Citizens in Reserved Constituencies

Note: The interaction is between reserved 2017 and women respondent. Standard errors are clustered at the constituency level. *** $ p<0.01 $ , ** $ p<0.05 $ , * $ p<0.10 $ .

Column 2 reports the interaction of gender and reservation status, showing that both men and women are equally more likely to be contacted by women activists in reserved constituencies, indicating women activists canvass men as effectively as they canvass women. Women activists may find it easier to access men in public spaces or may work harder to access men. These results also suggest that women politicians are less likely to be using a gendered campaign strategy—that is, using men activists to contact men voters and women activists to contact women. If that were the case, one would expect the results to show women being contacted by women activists but not men.

Columns 3 and 4 report contact by men activists. In reserved constituencies, citizens have a substantially lower likelihood of being contacted by men. Precisely, 21.2% citizens report contact with men activists in non-reserved constituencies. However, in reserved constituencies, less than half as much, that is, only 8.5% report contact by men activists. It is noteworthy, that although fewer all-men activists are making contact in reserved constituencies, women citizens in reserved constituencies relative to men citizens in non-reserved constituencies are 7.7 percentage points more likely to be contacted by men activists. This finding shows that all-male activist groups that organize under the leadership of a female politician are better able to reach female voters. Men activists may feel forced to exert greater effort to reach out to women voters, or women leaders may direct contact at women voters. Women leaders may also select and recruit male activists who are capable of reaching women voters.

In contrast, columns 5 and 6 demonstrate a slight increase in contact by mixed-gender groups of activists, although it is statistically insignificant. Collectively, these results suggest that more all-women activist groups, and fewer all-men activist groups are mobilizing citizens where women are candidates. Moreover, in reserved constituencies, and therefore under a woman’s leadership, these fewer-all men groups conduct more outreach toward women relative to when these groups are active in non-reserved constituencies (under a man politician’s leadership).

Partisan Contact in Ground Campaigns

Table 3 reports that citizens in reserved constituencies receive a level of partisan contact equal to that in non-reserved constituencies. The point estimate is close to zero and is statistically insignificant. This indicates women and men have roughly the same canvassing capacity. However, the politician survey conducted in this study reveals the perception of women politicians as weak mobilizers is widespread among male peers. In the light of this bias against women, this finding offers an important policy implication: providing information about women’s door-to-door outreach effectiveness can change male misperceptions of women as weak campaigners, therefore lowering statistical bias. Such misperceptions hinder women’s abilities to secure party nominations and to deliver public services.

Table 3. Women’s Door-to-Door Campaigns Contact Citizens as Effectively as Men’s Campaigns but More Equitably

Note: Standard errors are clustered at the constituency level. *** $ p<0.01 $ , ** $ p<0.05 $ , * $ p<0.10 $ .

In addition to conducting as much contact as men, women politicians halve the inequality in partisan contact, lending support to the second hypothesis. Column 2 shows that women in non-reserved constituencies are 15.7% less likely to be contacted by party activists than men. However, in reserved constituencies, this gender gap lowers by 7.7%. Each constituency has an electorate of about 20,000 women voters; thus, 1,540 additional women get contacted in reserved constituencies. This remarkable change is the effect of only one election cycle of reserved seats in 2017—the fifth cycle since the quota’s implementation. It is plausible that the effects are stronger with more contiguous years of reservations. Nevertheless, they indicate that reservations have an effect even in one election cycle (reserved in 2017 vs. not-reserved in 2017), even after the initial gains of quota adoptions have been realized, and accounting for the possibility that these effects average out over historical persistence.

The results in Tables 2 and 3 suggest that overall contact has remained intact, and that citizen contact with men activists has decreased in the same amount that citizen contact with women activists and with mixed groups has increased in reserved constituencies. What has changed is that in the status quo (under men’s representation), that is, non-reserved constituencies, men were more likely than women to be contacted by party activists. Under reserved constituencies, this imbalance or gender gap in party activist contact is decreased. So overall men citizens are receiving less contact, but it is because they were favored in the status quo. On net, it is not men citizens who are now worse off, but instead women, who were being ignored by party activists, are now better off.

Together the results in Tables 2 and 3 also offer preliminary insights as to whether women have increased the number of party activists (pathway one), or whether they have substituted men with women activists in their campaigns (pathway two), at least in the context of both formal and informal party activist recruitment observed in ground campaigns. Because overall contact has not increased (Table 3), and assuming that more party activists will increase contact to at least some extent (however minor), it means that women do not have bigger campaign organizations per se but have more women in their campaigns. Contact with men activists indeed has been substituted with women and mixed groups of activists (Table 2). Data from ground campaigns also lend support to the second pathway. It suggests that women do not necessarily have bigger campaign organizations than men but have more women in them.

Mobilizing Citizen’s Political Participation

The final premise suggests that receiving partisan contact is likely to increase citizens’ political participation. An ideal experiment to investigate this would cross-randomize real-time campaign manager outreach strategies in reserved and non-reserved constituencies. Lacking this ideal experiment, I use observational data to provide a preliminary test of whether partisan contact correlates with increase in political knowledge and participation.

Fieldwork informs my empirical strategy. My fieldwork suggests that candidates focus their contact efforts on their stronghold neighborhoods, considering factors such as caste, religion, and political affiliation. However, within these neighborhoods, they maximize outreach and are less selective in whom they engage with. To mitigate endogeneity bias stemming from candidate selection, I introduce a fixed effect at the highly localized, within-constituency neighborhood level. Within a neighborhood, there is an element of randomness in activist contact due to the densely populated and busy nature, as well as the lack of individual data for screening. This strategy, therefore, provides a robust empirical test of the correlation between contact and political engagement. Additionally, I include controls for individual-level variables, such as education, marital status, caste, religion, age, migrant status, home ownership, and employment, to account for potential correlations with being contacted. Section A.5.2 of the Supplementary Material presents results with and without controls, demonstrating bias reduction, and robustness is confirmed through two matching techniques: Mahalanobis distance and full propensity score matching using the MatchIt package in R.

Figure 2 demonstrates significant correlations between partisan contact and political knowledge, electoral participation, and civic engagement in both reserved and non-reserved constituencies, largely supporting the third hypothesis. The exception is non-reserved constituencies, where contact has limited impact on women’s political knowledge. While the point estimate for voting in non-reserved constituencies is substantively significant, it becomes marginally insignificant after controlling for numerous covariates. Fieldwork observations suggest that the nature of contact targeting women differs between reserved and non-reserved constituencies. Women may be encouraged to participate in politics and vote but may not benefit as significantly in terms of political knowledge in non-reserved constituencies. Additionally, preliminary analysis indicates weak to no differences in the effects of contact between women activists and men activists on political knowledge and participation. Please refer to Supplementary Tables A11 and A12 for these results. However, given the correlational nature of the analysis, these findings are preliminary and should be interpreted cautiously.

Figure 2. Campaign Contact Is Correlated with Citizen’s Political Engagement

Note: The circle represents the point estimates of the dependent variable which is any partisan contact. The bars represent 95% (thinner) and 90% (thicker) confidence intervals. Robust standard errors are reported. Section A.5.2 of the Supplementary Material presents tabular results.

Reduced-Form Effects of Reservations on Citizen’s Political Engagement

In this subsection, I present additional results consistent with this article’s theory: weak reduced-form effects of reservations on political knowledge and participation. The theory suggests that the effects of descriptive representation are conditional on women being mobilized. Supplementary Table A15 reports that additional results show the weak reduced-form effects of descriptive representation on women’s (and men’s) political knowledge and participation, which strengthen support for the theory. Reservations do not universally increase political knowledge and political participation. In settings with gender-restrictive norms such as India reduced-form results may average out backlash against women politicians. Indeed, public support for gender egalitarianism is weak and remains far below the world average in India; it has also declined since the implementation of gender quotas in 1990s, further raising concerns of a backlash. ¹³ Consequently, the effect of reservations on political behavior will be difficult to detect in studies that report reduced-form effects without accounting for (countervailing) mechanisms.

QUALITATIVE EVIDENCE

Evidence from the Politician Survey

To complement insights from the citizen survey, a politician survey was conducted in 2019–20, with 92 incumbent municipal politicians in Delhi. The survey is representative of municipal incumbents and has a response rate of 33% (see Supplementary Figure A5). Using data from this survey, I find that women politicians report having more women party activists in their networks. Asking politicians to name their top party activists, I find that 37% of women incumbents named a woman party activist compared to 16% men. The survey also asked politicians “Were your ground campaigns in the last MCD election comprised of mostly men or mostly women party activists?” Enumerators had the option to code “gender-balanced” on their screens, but it was not explicitly mentioned to avoid signaling cheap talk. In line with the theory, Figure 3 confirms that women politicians are much more likely than men to report having gender-balanced campaigns. Moreover, women politicians who are independent of their husbands are more likely to have balance as compared to those men who have captured quota positions by fielding their wives as candidates. These findings suggest that the male capture of gender reserved seats remains a barrier to realizing the full benefits of descriptive representation. A representative politician survey conducted in one of the most rural and poor Indian states, Bihar, which includes the same question finds similar results. See Supplementary Figure A13. However, it also shows no negative impact of male capture of reserved seats as far as use of women in campaigns is concerned.

Figure 3. Women Politicians Report More Women Activists in Their Ground Campaigns

Note: The figure plots the response of 82 politicians who report their campaigns either balanced or mostly women, or mostly men, for three categories of politicians: women politicians ( $ N=30 $ ), husbands of women politicians who have captured quota seats and have answered the survey ( $ N=12 $ ), and men politicians ( $ N=40 $ ). Ten politicians did not respond to the question.

How do women recruit women party activists? Fieldwork interviews highlight that women politicians and aspiring candidates actively knock on doors, visit women-majority spaces, and convince women to partake in party politics. During fieldwork interviews, a woman politician articulated a clear expression of women’s demand-side push, configuring the household is both as a politically and materially confining space for women: “Women do not even come outside of the house. I took the women out of the four walls of the house.”

Women politicians were vocal about how they bring women into party activism. In contrast, men politicians were silent, and none of them mentioned explicitly exerting effort to recruit women activists. It is also noteworthy that no women politician credited “seeing” other women in politics as the reason for entering politics. Instead, women shared how someone “brought” or “connected” them into politics, for example, a political candidate asked them to canvass door-to-door during a campaign or they undertook a task that eventually paved the way for them to become a formal party official. For example, Neelam (all identifying information has been changed) started her activist career when a women candidate for a major party asked her to participate in the candidate’s campaign in the run-up to the local elections. Today, Neelam is a municipal politician herself and richly demonstrates the entire cycle of how citizens first become activists, mainly during elections, and then candidates:

I joined politics when Mrs. Rama contested elections and asked for my help. But when I was working for Mrs. Rama’s campaign, I did not do so thinking that I will join politics. Rama ji and the then Member of legislative assembly, Mr. Shama saw my capacity to work and the style of work, and they said that they will make me the President of women’s party wing of this District because there was no organization at that time. At that time, women did not like joining politics; and their family members disliked it as well, but when I came here and was made the President of women’s party wing, I got the educated women out of their houses. I made a good president [of women’s party wing.] ¹⁴

Many women party activists have built extensive experience in problem-solving in their constituencies, and women politicians use their networks and access to spaces where women are more active such as healthcare centers, self-help groups, temples, NGOs, and neighborhood or school management committees to recruit active women talent. Aama describes her journey,

I do not have any political association in the family at all; I just had a passion for social work. If any work was not being done by the government officials, I would take that on. Earlier I used to live in Nagarwasi, and there was a sewage line related work which was done four times in the course of a year. I got that work done along with the RWA. Shiela used to be the municipal councilor back then, and she appointed me as the Mahila Morcha Adhyaksh in the area [District President of women’s party wing]. ¹⁵

Evidence from Ground Campaigns

The gendered dynamics of grassroots activism are most clear in ground campaigns during door-to-door candidate canvassing. Based on observations made during several years of fieldwork in multiple sites and during both quota and non-quota elections in India, I present a stylized sketch of a ground campaign. ¹⁶ Door-to-door campaigns are candidate-centric and begin 4–6 weeks ahead of elections. Aspiring political candidates rely on the grassroots activist following they have developed over the years as well as the prestige of their party nomination to expand their activist following in the run-up to the elections. They understand their constituency’s socioeconomic backgrounds—class, caste, religion, and partisan—support at the neighborhood level and target their personal campaigns most intensely at neighborhoods that are electoral strongholds. They are also more likely to reach out to the same neighborhoods where they have campaigned previously.

In a door-to-door campaign, a group of 5–15 party activists, many times without the candidate, walks in a pre-selected neighborhood. Activists carry megaphones or play loud music from a vehicle to attract crowds. They then break into smaller groups, knocking on doors and talking to people gathered in public spaces or around small tea stalls or shops in residential areas or those who are standing on their porches or are out on their balconies.

During these campaigns, party activists engage in unscripted conversations with citizens and provide information about the candidate’s sociopolitical background, policy platform, and party affiliation. They tailor their conversations, emphasizing the aspects of the candidate profile—language, caste, or candidate’s family background—they believe may resonate with the voter or discuss how the candidate plans to address the dominant service issue facing the neighborhood. These conversations may be short, lasting only 3–5 minutes—or can become more in depth, lasting 15–20 minutes or longer. Party activists distribute calendars or pamphlets and save citizen’s personal information such as mobile or WhatsApp numbers, which they use to send reminders and invitations.

When men activists conduct these campaigns, they mobilize men because men dominate public spaces. In contrast, women’s campaigns featuring women activists find it easier to access women voters. For example, consider the following observations made during the campaigns of Delhi’s state elections in 2020:

Woman candidate was constantly surrounded by four to five women from the middle-class strata, who were all activists for the party. She visited around 80–100 homes. Woman candidate would arrive at a house, and a woman constituent would wrap a red shawl around the candidate, along with handing her a marigold garland. The shawl and garland was distributed by one or two women activists who would walk ahead of the contingent, hand over a garland or a shawl to the closest woman, who would wrap it around. Woman candidate interacted mostly with women, asking them about the most recent electricity bill and telling them to vote for her. Men hardly ever interacted with her, and if ever, they seemed wary of it. Roughly 85 percent of the people the women candidate and activist met were women.

In contrast, the campaign of a man candidate from the same party, comprised mainly of young men activists:

One of the most noticeable aspects of the campaign was the fact that only three women were a part of the same, out of around 120 men, who belonged to various age groups.

According to the male campaign manager, who is the campaign manager of male candidate, the candidate had 12 back to back open meetings scheduled for the whole day, in various areas of the constituency. In each of these street meetings, a stage was set up where dancers would perform on folk/Bollywood songs. Followed by this dance performance, the party activists would address the crowd for about 30 minutes before the candidate’s arrival. There were eight men activists, who took turns to talk to the crowd. They talked about the party’s manifesto and raised some common points. One major thing that was being talked about was water. While some women were present with their children, the crowd that gathered to listen was dominantly men who had been either contacted by the male activists or had assembled at the venue after hearing the noise from the songs.

Evidence from Interviews with Party Activists

Fieldwork and interviews with 1,243 formal municipal constituency-level party activists in Delhi also reveal similarities and differences between parties and the type of women who become party activists. To the best of my knowledge, this is the first multi-party survey and the first survey of both men and women party activists conducted in any developing country. Supplementary Figure A6 provides the distribution of party activists across parties and gender. Most women party activists are married, middle-aged, and socially active in their local communities and associations. Majority of them join politics after getting married, unlike men many of whom enter politics before marriage. This implies that women have less experience in party politics than men despite being of a similar age. Fieldwork showed that there is also a noticeable difference in the locations where men and women party activists conduct their political activities. Men usually use their shops or public land, whereas women typically operate from their homes or NGO offices. This discrepancy highlights how gender-based resource access disparities shape the places where women engage in party activities and the limited visibility of their party activities.

The survey reveals three consistent findings across parties. First, the majority of party activists (over 80%) indicate that party recruitment is the primary factor that connects them to the party organization (Supplementary Figure A7). Second, women activists are more likely than men to report being recruited by women involved in local politics (Supplementary Figure A8). Qualitatively, constituencies that lack women politicians or have low levels of women’s electoral participation have less formalized and less active women’s party wings. Finally, women party activists are no more likely to have family members in politics than men in all three parties, suggesting that women who are recruited are not mere tokens or wives of men party members (Supplementary Figure A10).

The survey also highlights the differences in the quantity, quality, and characteristics of women party activists recruited across three major political parties. Echoing studies on party organization in India (Chhibber, Jensenius, and Suryanarayan 2014), the BJP has the most organized and active network of women party activists, followed by the AAP and then the INC. BJP party women come from families that own small- to medium-sized enterprises. They are not only socially highly active in religious or neighborhood communities, but they are also active in planning social activities with other women party members, centered mainly but not exclusively on religion. In contrast, AAP party women are more involved in local problem-solving, particularly on education- or water-related concerns, but have fewer ties to other party members. AAP party women are also more likely to come from poor to lower-middle class backgrounds. Finally, the INC has least organized network of women activists. INC party women are wealthier but are the least engaged in party activities.

The survey underscores strong backlash and reluctance against women’s political participation. Party activists highlighted how the “patriarchal society” remains vehemently against women in politics. Household constraints and regressive gender norms were the top two reasons cited by more than 30% of men and women activists as strong barriers against women’s party activism. Furthermore, only 60 out of 550 activists mentioned experiencing role-model effects (Supplementary Figure A9). Majority of these mentioned national male leaders as role models. It is worth noting that no one mentioned local women (or men) leaders as role models. Even among this politically active group where role model effects are most likely, there was no evidence for the same. Instead, women party activists highlighted that their presence in politics is not welcomed as manifest in the high levels of harassment and violence they persistently experience when they conduct campaigns. They reported using persuasion, information, and service provision which helps them to allay concerns and fears that citizens have against women’s political leadership.

Together, the qualitative data from various sources (politicians, activists, and campaigns) and measurement techniques (direct observations and interviews) also allay the concern that the gender composition of campaigns has not changed, but instead men and women activists are equally present in reserved and non-reserved seats, but women are more effective under women’s leadership.

CONCLUSION

This article advances grassroots party activism-based theory—a unified explanation—that links descriptive representation with women’s political participation in partisan, civic, and electoral politics. Although the theory could apply in many settings, it is especially relevant in developing contexts where party activists play a key role in problem-solving and electoral mobilization, and is a stronger explanation in settings where regressive gender norms remain entrenched, hindering attitudinal and symbolic effects.

The argument offers potential explanations for key puzzles relating to gender and politics in the Global South. First, it can explain why top-level women leaders have done little to improve women’s representation in parliaments. The theory suggests that, in the absence of women’s presence in grassroots party networks, high-level women leaders remain beholden to male-dominant party organization. Paradoxically, women’s “representation from below”—not only at the top—is key to laying the groundwork essential for political equality inside parties. This is also an important scope condition of the theory. The argument also offers an explanation why India’s gender gap in electoral turnout has closed in the last two decades. Women party activists have not only put women’s issues on party platforms and manifestos, but also carved women as an electoral force parties must reckon with. The global trend toward decentralization and the increasing presence of women in local governments indicate that this explanation for the reduction of the gender gap in voter turnout is likely to have broader applicability.

Crucially, these findings help to clarify how women politicians deliver substantive representation in settings where women politicians are less qualified and less experienced (Chattopadhyay and Duflo 2004), have weaker political networks (Grossman, Garcia-Hernandez, and Michelitch 2022), and are discriminated against by bureaucrats (Purohit 2023). Mobilizing the support of women party activists, women politicians can use grassroots tactics to pressure low-level bureaucrats to get things done and can rely on women activists to respond to vulnerable and poor citizens.

This article hopes to spark a research agenda on women’s party activism in developing contexts. Men party activists are influential figures in distributive politics (Auerbach and Thachil 2018). Thanks to women in local politics, women party activists have entered this political milieu. Understanding the quality, limitations, and the intersectional nature of women’s party activism is imperative for our understanding of not only women’s political behavior but also broader forces of democracy, party building, and development.

The findings of this article challenge several widely held beliefs. They dispel the notion that women politicians must alter attitudes and perceptions to make progress in politics. They refute the idea that women are incapable of engaging in clientelism and reject the perception of women as mere tokens who act solely on the command of their husbands. Instead, women have taken a step forward by transforming political party institutions from the grassroots and leveraging women’s unrealized political agency.