QUOTA RULES: INTERSECTION OF COMMUNITY AND GENDER

Quotas in the form of reserved seats for disadvantaged castes and communities in India have existed since the latter end of British rule.Footnote ³ Recognizing that strong social bias against lower castes, alongside their historic economic and educational deprivation, would make candidates from these communities less politically competitive after independence, the drafters of the Indian constitution implemented quotas for these groups in the national parliament and across state assemblies starting in 1950 (Jensenius Reference Jensenius2016). Specifically, this institution designated a set number of reserved seats to members from castes classified either as Scheduled Castes (SCs) or indigenous tribes classified as Scheduled Tribes (STs), which were proportional to their share of total population within each state. For a caste to qualify for SC status (approximately 450 castes today), they generally must have been historically treated as “untouchable” by the rest of Hindu society, thus facing significant prejudice due to their caste origin (Chandra Reference Chandra2000, 27). Qualification for ST status is similar but requires a lineal connection to an indigenous community.

In 1979, the Mandal Commission convened by parliament recommended the expansion of these reservations to a newly established classification, Other Backward Classes (OBCs). In contrast to SC and ST communities, which qualify for reservation based on a history of pervasive societal bias against their particular caste or community, OBC communities qualify for reservation based on social, educational, and economic indicators that lag significantly behind state averages (GOI 1981). Notably, some non-Hindu minorities, such as Muslim communities, qualify for OBC status at the local level. Given the broader qualifications, OBCs have always formed the largest block of reservations and thus are a significant electoral consideration. The specific communities that qualify for reservations and the specific benefits reservation status confers differ across states and can change over time. Generally, however, positions are reserved in elected governance institutions at the state and local levels, as well as for civil service positions in governmental agencies and student slots in public education institutions for individuals from a community who qualify for either OBC, SC, or ST status (Jensenius Reference Jensenius2016). Only individuals from these communities can fill these positions. Numbers differ by state depending on population proportions, but on average the distribution of reservations is approximately 27% for OBCs, 16% for SCs, and 8% for STs. All other communities are generally referred to as “General” (GEN) or “forward castes” and receive no reservation benefits.Footnote ⁴

In 1992, state governments were significantly decentralized across India. Each state was required to conduct elections at the local level and to devolve considerable powers of expenditure to local governance bodies (Bhavnani Reference Bhavnani2009). These local bodies are divided across each state into districts or zila parishads, under which the municipal bodies govern urban centers, whereas rural India is further subdivided into blocks or panchayat samitis and again into village councils or gram panchayats. Across these elected bodies, gender quotas were instituted in the form of reserved seats for women from all communities across India; however, gender reservations are not as extensive as the community reservations. Positions for women are only reserved in elected governance institutions at the local level. Therefore, at the local level, these gender reservations cut across community reservations. Across all seats, whether designated for reserved communities (i.e., OBCs, SCs, and STs) or open seats that are unreserved (i.e., GEN), 33% must be reserved for a woman from that community (i.e., 33% of all SC seats must be reserved for SC women) (Jensenius Reference Jensenius2016), with several states voluntarily going up to 50% gender reserved seats (Kumar and Prakash Reference Kumar and Prakash2012). When combined with community reservations, this establishes a complex system in which every electoral ward can hold one of eight labels, which determines who is able to run in that seat. Table 1 lists each of these eight labels, their acronym, who qualifies to run in a seat assigned that label, and the proportion of seats assigned that label across the four municipal corporations in the 2017 elections.

Table 1. Reservation Qualifications Proportion of Seats Across Municipal Corporations, 2017

To determine the number of seats allocated to each label within an elected body, reservations are essentially layered. The first layer is strictly community reservations (i.e., GEN, OBC, SC, ST), which are allocated based on each community's proportion of total population across all wards of the elected body. This process establishes the number of labels from each community that will be assigned across wards. This number of community labels is then placed into a lottery and randomly drawn and assigned to a ward,Footnote ⁵ which will then be assigned that label and the accompanying community reservation for the next election. The second layer then attaches a gender reservation randomly to 33% of all seats allocated to each community, again through a lottery. This layering process ensures that women from each community receive a minimum of 33% representation; however, the siloes created by this process also establish an inflexible system that can deter political development.

To make the process of assigning reservations clearer, I use the empirical case of the Kanpur Municipal Corporation (KMC), the primary governing institution of Kanpur city in Uttar Pradesh, as an example. For the 2017 KMC elections,Footnote ⁶ the KMC had 110 total seats across 110 single-member wards. Members of these seats are called parshads or councilors. Across the 110 parshad wards, 21 seats were allocated as reserved for OBCs, 14 for SCs, and none for STs, in proportion to their respective populations in Kanpur.Footnote ⁷ The remaining 75 GEN seats were open to all candidates. Of each community grouping of seats (i.e., GEN, OBC, and SC), 33% were then reserved for women from that community (i.e., 25 of the 75 GEN seats were reserved for GEN women, or any woman from any community). These eight community- and gender-reservation labels were then assigned across all 110 wards.

STAY IN YOUR LANE: LACK OF COMPETITION OUTSIDE RESERVED SEATS

Using electoral data from the four municipal corporations for the 2017 elections only, this first analysis examines all candidates across 390 wards. In this electoral year, 3,784 candidates competed across all communities.Footnote ⁸ Overall, men and women from these communities competed in the 2017 elections, and a large proportion of candidates (30%) competed outside their reservation against candidates from different communities and different genders.Footnote ⁹ Candidates competing outside their reservation could have been either men or women from a reserved-community (i.e., OBC, SC, or ST candidates) competing for a GEN seat, women from a reserved community competing for an open-community or gender-reserved GEN women's seat, or any woman competing for any reserved-community or open-gender seat (i.e., OBC, SC, and ST seats). Notably, no GEN men could compete outside a reservation because they have no reservation. Each of these candidates were competing in a much more unrestricted field; thus, they faced more competition than if they had only competed in the reserved seats for which they were qualified.

Promisingly, we do not see an electoral environment in which candidates from reserved communities and women are blatantly discouraged from competing outside their specific community reservation. As discussed, the community quota system was designed to make candidates from marginalized communities more electable across the general electorate, not just within their communities (Jensenius Reference Jensenius2017). Following the description of the reservation assignment lottery previously, the wards with the highest proportions of voters from reserved communities are more likely to be reserved for those specific communities. Therefore, when candidates from reserved communities compete outside their reservation in open wards, they must appeal to a larger number of voters from other communities. Furthermore, women from OBC and SC communities particularly competed outside their community-reserved seats (60% and 46%, respectively), often in greater proportions than men from their communities (49% and 34%, respectively).

When these data were nuanced to analyze how women competed outside their reservation, however, significantly fewer women competed outside the gender reservation. Most women who competed outside a reservation left their community reservations but remained in gender-reserved seats.Footnote ¹⁰ Overall, compared with the proportion of total women who competed outside their reservation (500 women, 38%), these numbers are notably smaller. Of 1,330 total female candidates, only 183 (14%) competed for open-gender seats; thus, 86% competed exclusively for gender-reserved seats, against women only. Most of the 500 women who competed outside their reservation were from reserved-communities who left their community reservation to compete for the open-community or gender-reserved GEN women's seats, but, again, only against other women. Furthermore, the proportion of candidates who were women in the open-gender races was never higher than 8% and was often much lower. Therefore, the overwhelming majority of competition among male candidates was fellow men, and among female candidates it was fellow women.Footnote ¹¹

Certainly, low numbers of women competing in open-gender wards are problematic for a gender-reservation system that has been in place since 1992, through five election cycles. Similar to community reservations (Jensenius Reference Jensenius2017), the literature generally considers the establishment of a foundation of political representation and experience that encourages female candidates to compete against men to be a primary goal of gender quotas (Deininger et al. Reference Deininger, Jin, Nagarajan and Fang2011). However, the achievement of this goal would entail women competing against men in open-gender elections. As these data show, women in these contests are not moving beyond gender-reserved seats in large numbers. More importantly for quota effectiveness in the long-term, women are not gaining experience competing against men; they are not demonstrating that women can beat men in open elections on a broad scale; and they are not habituating male candidates and voters overall to intergender competition. Notably, Bhavnani's (Reference Bhavnani2009) analysis of Mumbai elections found that, in open-gender seats, higher numbers of female candidates increased the odds that a woman would win. Therefore, the shortage of female candidates in open-gender elections is likely the primary reason why, across these municipal corporations in 2017, only 11 women (4% of all winners), competed for one of the 259 open-gender seats and won.Footnote ¹²

As discussed, the literature is optimistic that, over time, quotas will enable women to compete in open-gender wards by habituating parties and voters to female candidates (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018; Bhavnani Reference Bhavnani2009). Ideally, by requiring parties to nominate women and voters to vote for women when a ward is gender reserved, those parties and those voters will be more accepting of female candidates in subsequent elections. Corresponding with Bhavnani's findings in Mumbai, across these municipal corporations in 2017, all women who won in open-gender wards did so in wards that had previous experience with a woman candidate. In 2012, the ward was either gender reserved or a female candidate won there in an open-gender contest. Furthermore, as Table 2 demonstrates, the success of women in open-gender contests in 2017 differed considerable depending on whether that ward had previously been gender reserved in 2012 (column A) or open gender (column B). Wards that were previously gender reserved were more likely to elect a woman in 2017 to a statistically significant degree. The correlation is more pronounced when the comparison is expanded to all wards that had a woman parshad in 2012 (column C), whether gender reserved or open, with wards who had a male parshad in 2012 (column D).Footnote ¹³ Both comparisons demonstrate that 2017 open-gender wards that had elected a woman in the 2012 election were more likely to elect a woman in the 2017 election.

Table 2. Next-Election Effects on Women Winning Open-Gender Seats, 2012–2017

Note: *significant at 10%, **significant at 5%, ***significant at 1% using one-sided difference in proportions and means tests.

PARTY RESTRICTIONS

The primary movers in restricting or encouraging women to run outside the gender reservation are the political parties. Party nomination is the primary avenue candidates follow to gain office across these municipal corporations because they provide the valuable branding, resources, expertise, and voter blocs that new political entrants lack (Jensenius Reference Jensenius2017). For municipal elections, parties generally seek out locals with strong social reputations in their ward and then assist them with the support needed to run a successful campaign (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018). Only 16% of candidates across these municipal corporations were elected as independents without a party affiliation in 2017, which is similar to trends observed elsewhere in India (Chathukulam and John Reference Chathukulam and John2000).Footnote ¹⁴

In terms of candidates nominated to compete in municipal corporation elections, gender inequalities are stark. Across these municipal corporations, only 36% of nominated candidates were women among the top 10 performing parties.Footnote ¹⁵ Notably, this trend is common across most parties, whether national or regional. Of the two dominant national parties, the Bharatiya Janata Party (BJP) and Congress (INC), Congress had a slightly better ratio of male to female candidates, with approximately 67% more male candidates, whereas the BJP approached 78% more. The smaller regional parties were generally even more skewed, with some having more than two to three times as many male candidates nominated as women. An interesting exception was the small regional Bahujan Mukti Party (BMP), which actually fielded 36% more female candidates than men (albeit only 18 total). Those competing outside a party banner in the independent ranks were primarily men as well, by a nearly two-to-one ratio. These findings demonstrate low numbers of female candidates overall and suggest that few women are being nominated outside the gender-reserved seats. Again, the percentage of women nominated by parties (36%) was not much higher than the percentage of reserved seats (33%).

That number drops further when considering women nominated by parties to compete in open-gender wards, to only 11%. Most women were nominated to compete only in gender-reserved wards. Of the few women competing against men in open-gender elections (n = 116), only 56% were nominated by a major party (n = 65); 44% did so as an independent candidate.Footnote ¹⁶ Low nomination numbers for women competing in municipal corporation elections have also been noted in the literature. In Mumbai, Bhavnani (Reference Bhavnani2009) found that only 7% of female candidates received party nominations in open-gender wards, which is comparable to the 11% observed here. Still, the source of bias is likely quite nuanced, and several additional factors must be considered.

Overall, these skewed numbers indicate a probable gender bias within the party hierarchy that controls the candidate selection process. A small number of key party members, usually men, control the nomination process for positions within the parties in India from the national level (Jensenius Reference Jensenius, Franceschet, Krook and Tan2018) to the state level (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018) to the local level (Anitha et al. Reference Anitha, Bharadwaj, Jayakumari, Krishnan, Nisha and Praveena2008; Chathukulam and John Reference Chathukulam and John2000). Therefore, the perceptions of a small number of people can exert disproportionate influence over the entire process. Nominations are granted based on candidate electability, which is generally focused on ethnicity, community, caste, funding, and “muscle power,” or workers on the ground who can contribute to the campaign (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018; Singh Reference Singh and Saxena2003). If a candidate lacks significantly in any of these categories, they are likely to be discounted (Anitha et al. Reference Anitha, Bharadwaj, Jayakumari, Krishnan, Nisha and Praveena2008; Chathukulam and John Reference Chathukulam and John2000). On these terms, any perception by these gatekeepers within the party that female candidates are electorally weaker than male candidates results in lower numbers of women nominated (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018). Although women can qualify just as easily as their male counterparts for community-based seats, they are generally much weaker in terms of funding and “muscle power.” Female candidates are usually very reliant on their families and male benefactors for both funding and supporters on the ground (Anitha et al. Reference Anitha, Bharadwaj, Jayakumari, Krishnan, Nisha and Praveena2008; Chathukulam and John Reference Chathukulam and John2000; Jensenius Reference Jensenius, Franceschet, Krook and Tan2018).

Although parties are nominating very few women to run against men, parties do not appear to be completely ignoring competitive female candidates. Of the 11 women who won against men in open-gender elections in 2017, eight were nominated by a party. As discussed previously, gender reservations should encourage both party leaders and voters to learn about the ability of female candidates to win elections and make communities both more comfortable nominating and voting for women, even when they are competing against men (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018; Bhavnani Reference Bhavnani2009). Promisingly, Bhavnani found that parties in Mumbai were approximately five times more likely to nominate a woman to run in an open-gender ward if that ward had been previously reserved for a woman. However, as Table 3 illustrates, the analysis of these four municipal corporations in Uttar Pradesh did not show a comparable effect from gender reservations alone. Of the 67 women nominated by a party to compete against men in an open-gender ward in the 2017 elections, 49% were nominated to ward areas that were reserved for a woman in 2012 (column A), whereas 51% were nominated in open-gender wards (column B), a difference that is not statistically significant. Nonetheless, if the analysis is expanded to wards that were represented by a woman in 2012 (column C), whether gender-reserved or not, then a majority of women (57%) who were nominated to run in an open-gender ward in 2017 were nominated in a ward that had a woman parshad in 2012, a difference that is statistically significant from the wards that were represented by men in 2012 (column D). However, this difference is not as dramatic as that reported by Bhavnani.

Table 3. Next-Election Effects on Party Nomination of Women in Open-Gender Wards, 2012–2017

Note: *significant at 10%, **significant at 5%, ***significant at 1% using one-sided difference in proportions tests.

Further reinforcing the likelihood that parties are not shutting out strong female candidates is the shortage of women in the independent ranks. If competitive women are snubbed by the party for their ward, they still have the option to run as an independent, but they do not appear to be doing so in considerable numbers. However, winning without party support is very difficult, given the low success rate of independents. Notably, the parties themselves are primarily responsible for finding and developing candidates. As discussed, these male-dominated party hierarchies are often resistant to developing strong female candidates who may pose a challenge to themselves in the future (Anitha et al. Reference Anitha, Bharadwaj, Jayakumari, Krishnan, Nisha and Praveena2008; Chathukulam and John, Reference Chathukulam and John2000) and likely therefore do not prioritize the development of women within the party (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018).

Overall, the analysis shown in Table 3 demonstrates that although gender reservations alone may not encourage parties to nominate women in open-gender contests, parties are approximately one-third more likely to nominate women in open-gender wards that have had previous experience with a women parshad. Thus, they are theoretically more likely to vote for a female candidate in the current election. Furthermore, these results also demonstrate that experience with a woman parshad does not discourage parties from nominating female candidates in the future.

These findings create concerns regarding the short-term effectiveness of the gender reservation as a foundation for rapid female candidate growth. If female candidates are unable to broadly use the reservation to become considerably more competitive for party nominations in local elections without a gender reservation, then it is difficult to foresee how women will progress into state- and national-level elections in the near future, where no gender reservations exist. Again, despite the implementation of the gender reservation at the local level in 1992, the number of women in the national parliament and across state assemblies has practically remained stagnant over the past two decades, hovering around 11% at the national level and often lower across state assemblies (Jensenius Reference Jensenius, Franceschet, Krook and Tan2018). Across the 38 combined state assembly seats of the four cities used in this analysis, there were only seven women (18%) in 2018. Even more concerning is the observed decrease in the number of new woman candidates competing for state legislatures across Indian states with entrenched gender biases, despite the establishment of gender reservations at the local level that could serve as feeders (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018).

The causality dilemma inherent in these findings is present in the political gender-equality discussion across political systems. Parties will not nominate women if they are perceived as weaker than their male counterparts; however, candidates competing in political environments in which the party is strong need party support to become competitive candidates. Therefore, without parties taking political risks on female candidates, it is difficult for women to become politically strong enough to receive a nomination. Gender-quota systems essentially force parties to take this risk on female candidates by requiring them to nominate women. However, in quota reservation systems such as India's, parties are only forced to nominate women to gender-reserved seats and are never required to nominate women to compete against men. Proposed solutions to this dilemma in other political systems have been candidate quotas, which require parties to nominate a percentage of women across all seats. This system encourages parties to develop strong female candidates who can win against both male and female candidates, as they could potentially face both. However, this system does not create a minimum level of women in an elected body because potentially every woman nominated could lose to a man, particularly if the party protects its male candidates by shunting all its female candidates into uncompetitive wards where it expects to lose anyway (Schwindt-Bayer Reference Schwindt-Bayer2009).

LACK OF CONTINUITY

In Mumbai, Bhavnani (Reference Bhavnani2009) found that most of the women who won in open-gender wards were incumbents brought into politics through a past gender reservation, possibly demonstrating that quotas can help build a cohort of incumbent women who can develop the political experience to challenge men at the state or national levels. However, an analysis of these four municipal corporations shows that incumbency is very low for both women and men.Footnote ¹⁷ Proportionally more male incumbents returned to run again (10%) than did female incumbents (7%). Furthermore, more men returned and won again as an incumbent (5%) than did women (2%). Few men or women parshads return as incumbents in general; however, the distinct advantage gender reservations provide does not encourage or enable women to run again at a degree comparable to that of men. These findings align with those of Bhalotra, Clots-Figueras, and Iyer (Reference Bhalotra, Clots-Figueras and Iyer2018) across state legislatures in India; they also found that women were less likely to run again as incumbents.

Several structural factors likely restrict the ability of the gender reservation to encourage incumbent female parshads to return: institutional disincentives, incumbency disadvantages, and reservation disqualification. First, past qualitative work with parshads in Jaipur described both men and women who reported not planning to seek re-election due to low pay, too much work, and frustration with intransigent bureaucracy (Turnbull Reference Turnbull2018). Parshads reported receiving a minimal expense allowance that was far below a middle-class salary, which did not justify complaints of frustrating working conditions and long hours. Reports of corruption opportunities posit that this allowance could be padded, but first-time parshads, particularly inexperienced women, likely face difficulties gaining access to these revenue streams.

Incumbency may pose a disadvantage to re-election as well. Only 9% of parshads overall sought re-election in a subsequent election, and only 4% were successful. This could indicate that re-election is extremely competitive, which further disincentivizes parshads from attempting to return. Across the country, previous research has found that incumbency can actually confer a disadvantage on candidates in India.Footnote ¹⁸ Bhavnani (Reference Bhavnani2009) found an incumbency disadvantage for both men and women in Mumbai municipal elections, and larger cross-state studies have found similar disadvantages across national and state elections (Linden Reference Linden2004; Uppal Reference Uppal2009).

Finally, perhaps the most important consideration for any parshad considering running again in the next election is disqualification by the reservation lottery. If the reservation for a parshad's ward changes in the next election to one they do not qualify for, they are unable to run again in that ward and are displaced. Of the 390 parshads who won in the 2012 elections, around 77% faced a reservation change in their ward for the next election. Notably, 45% of parshads overall faced a displacing ward change that essentially blocked them from running again in that ward because their ward was assigned a reservation in the lottery that they did not qualify for. Unsurprisingly, more men (59%) than women (21%) were displaced.Footnote ¹⁹ Men faced the dual possibility of a community and/or a gender reservation blocking them from running again. However, community reservation lines are rigid, regardless of gender reservations. Just as men cannot compete across community-reservation lines, neither can women. All of the women who were blocked from running again in their ward did not fit the community-reservation criteria for the newly designated ward reservation because they were either GEN women who could not run in a reserved-community ward, or OBC, SC, or ST women whose wards had been reserved for a community other than their own. This is a notable downside to gender-reservations cutting across community reservations: 28% of GEN women, 18% of OBC women, and 20% of SC women were unable to run again in their respective wards due to reservation displacement.

Delving deeper into the various reservation possibilities, the probability of being displaced by a reservation differs significantly across groups. At proportions based on the 2011 census (until they are adjusted with the 2020 census), Table 4 lists how many seats were available and how many seats were blocked for each category of candidate across these municipal corporations. If a candidate's ward was reserved for any of the reservations off-limits to them, they could not compete in that ward. Given that 54% of all seats are reserved for someone, several communities and genders were blocked from competing in a majority of seats.

Table 4. Blocked by Reservation

As expected, GEN men face the highest number of blocked seats at 54% because they qualify for no reservations. Notably, men from the SCs, the disadvantaged communities specifically for whom reservations were initially designed to increase political representation, are close at 47% off-limits. The candidates with the most options by far are OBC women. OBC candidates overall have the largest block of reserved seats, of which a one-third are reserved for OBC women. In addition, OBC women can compete for both GEN women's and GEN seats. Due to a smaller number of seats reserved for SC candidates, SC women can compete for a supermajority of seats, but not as many as OBC women. This comparison illustrates a problematic side effect of community-exclusive reservations established in proportion to the population. The less populous communities (i.e., SC) are blocked from the large number of seats reserved for the more populous communities (i.e., OBC). Although GEN men face the greatest chance of being blocked, the reservation system is still likely to block the disadvantaged communities it was designed to support. The SC communities tend to be significantly more economically and socially disadvantaged compared to OBC communities, but under this system they have fewer electoral options. Although reservations create a foundation of representation, to a degree they can also create an artificial ceiling that favors the more populous, more advantaged communities and constrains the smaller ones. This is particularly problematic given the likelihood that small minorities are already more likely to be excluded from key decision-making and resource-allocation processes.

For female parshads specifically, the reservation system institutes an additional consideration: not only do community reservations change, but gender reservations do as well. In fact, in contrast to community reservations, which can be maintained over subsequent elections,Footnote ²⁰ many states mandate that a ward cannot be reserved for a woman in consecutive elections (Chathukulam and John Reference Chathukulam and John2000; Ghosh Reference Ghosh2001), likely to ensure that men from these wards are not permanently excluded from office (Bhavnani Reference Bhavnani2009). This is the case in Uttar Pradesh. Therefore, across these four municipal corporations, gender reservations were randomly assigned in 2017, but only to wards that were not gender reserved for the previous election in 2012.Footnote ²¹ Electoral rules did allow candidates to compete outside the ward they reside in, so even with a reservation change in their home ward a candidate could run on the ballot of another ward open to them. However, other literature has described voter resistance to candidates who “parachute” in from outside wards (Bhalotra, Clots-Figueras, and Iyer Reference Bhalotra, Clots-Figueras and Iyer2018; Turnbull Reference Turnbull2018). This situation presents a significant challenge to parshads displaced by a new reservation. As Table 5 illustrates, parshads who were displaced were less likely to run again or win again than those not displaced, a statistically significant difference. Of all the candidates displaced, only 6% ran again, and none won, whereas 19% of the candidates not displaced ran again and 8% won.

Table 5. Effect of Displacement Likelihood to Run Again and Win Again, 2012–2017

Note: *significant at 10%, **significant at 5%, ***significant at 1% using one-sided difference in proportions tests.

If the analysis is narrowed down to women specifically, the effect is less noticeable (Table 5, bottom two rows). Of the women displaced by a reservation change, 13% went on to run again in the next election; of the women not displaced by a reservation change, only 18% went on to run again in the next election. The difference is not statistically significant. Still, the likelihood that an incumbent woman parshad would win again was, to a degree, statistically different if she was not displaced. Seven women (6%) who were not displaced won, whereas no displaced women won again. These findings also demonstrate that it is more difficult for a woman parshad displaced by a reservation change to engage competitively in the next election than it is for men.

General men are undeniably the most numerically constrained by the reservation system, with over half of all seats off limits to them. Female parshads are subject to the reservation lottery as well but have more options in subsequent elections than their male counterparts; they can still run in the gender-reserved wards that are off limits to men. Still, if women rarely compete for open-gender seats, either because they choose not to or are unable to, then nearly two-thirds of all seats are normatively off limits to them. The overlap between gender and community reservations constrains this number further because women cannot run in community-reserved seats they do not qualify for. For example, if an SC woman in Kanpur was displaced by a reservation change, but she was only able or willing to compete for a gender-reserved seat in 2017, of 110 seats, only 30 seats were truly open to her (25 GEN women's and 5 SC women's seats). Combined, these institutional constraints and normative practices discourage incumbency for both women and men. However, these institutional side-effects restrict the positive impact the quota institution can deliver to the marginalized groups it was designed to support, such as women. The value of bringing women incumbents back for subsequent terms was demonstrated by Beaman et al. (Reference Beaman, Chattopadhyay, Duflo, Pande and Topalova2009), who in rural India found substantial increases in the constituent perception of how successful their women leaders were after two consecutive terms only; one term was rarely enough. Therefore, if women are brought in by reservation for only one term, the ability of the gender reservation to both alter perceptions of female political ability and the development of experienced women to make inroads at the state and national levels is likely diminished.