Are women better politicians?

Gender quotas have been the policy lever of choice to increase female representation in political bodies. More than 70 countries now maintain gender quotas at the national level, and many others maintain them sub-nationally. A prominent form of these quotas is the reserved-seat quota used in local governments in India, where specific constituencies are designated exclusively for women candidates, and men cannot compete in these reserved-seat races. Descriptively, these quotas have been a success as they automatically improve the number of women in any elected body that has quotas. Therefore, India provides an ideal laboratory to study how quotas affect not just representation, but electoral accountability.

Despite the substantial evidence on these institutions, many questions remain unanswered. In particular, scholars have not investigated the potential for reserved-seat gender quotas to influence electoral accountability. Elections hold politicians accountable through two main channels:

1. Elections discipline incumbents by incentivising them to work in their constituents’ interest.
2. Elections select candidates with better qualities for office.

Since men cannot contest reserved seats, quotas mimic temporary term limits on male politicians. Party gatekeepers who control candidate selection can also complicate matters if they treat female and male incumbents differently, further complicating accountability analyses. Finally, voters may also have taste-based preferences for male and female candidates. Taken together, a reserved-seat quota system can change accountability patterns through three different channels:

1. By altering incentives for male and female incumbents differently, quotas may reduce effort by men.
2. Independent of the electoral mechanism, quotas select more women into office.
3. The presence of party and voter discrimination may entangle both the discipline and selection channels of accountability.

Scholars studying traditional term limits have developed methods to identify these discipline and selection effects separately. They compare, e.g. term-limited and non-term-limited US governors serving in the same term to isolate the discipline effect, and compare governors with identical term-limit status but serving in different terms to identify the selection effect (Alt et al. 2011). When it comes to gender quotas, however, these reduced-form approaches cannot disentangle the channels, because reserved-seat quotas impose temporary term limits only on men.

Evaluating the impact of gender quotas on accountability

In our research (Desai, Karekurve-Ramachandra, and Montero 2024), we study the effects of gender quotas in urban local government in Mumbai, India, on accountability as defined previously. Reserved seats are randomly assigned across 227 electoral wards in the city, providing a natural experiment. Each of the 227 wards in Mumbai elects a representative, called a councillor, together, they form the Brihanmumbai Municipal Corporation (BMC), a body with a total budget of almost US$9 billion. BMC councillors have a hybrid role: they make up the municipal legislature, but are also in charge of basic public service provision within their constituency. Praja, an NGO seeking to improve governance standards in India, conducted surveys of voters in the 227 wards of Mumbai. We use this data to show that voter perceptions of women are higher than for men. But what explains this gap?

The randomisation of quotas not only imposes temporary term limits on men, but it also materialises probabilistically. Hence, a reduced form approach cannot disentangle the three channels. We therefore adopt a structural approach: we write down a theoretical model of electoral accountability and use data from Mumbai to estimate it. In our model, voters hold prior beliefs about candidate types that may vary by gender. Exerting effort is costly for low-type politicians, but costless for the high-type. Effort shifts the distribution of policy outcomes.

Political parties observe politician types, but voters do not. Parties can choose to renominate candidates, and may discriminate on the basis of incumbent type or gender. Voters observe policy outcomes and the renomination decisions, update their beliefs about the incumbent, and decide whether to re-elect the incumbent party’s candidate or elect a challenger. We also allow voters to hold expressive (taste-based) preferences over candidate gender in addition to statistical beliefs about politician quality.

Selection drives female councillors’ superior performance

We find that selection primarily drives female councillors’ superior performance, as they are twice as likely as men to be high-types. Regarding discipline, low-type men exert substantially more effort than low-type women, despite facing probabilistic term limits. This stems from parties’ renomination behaviour: a low-type male incumbent, even accounting for the possibility of being term-limited by reservation, is over three times more likely to be renominated than a low-type female incumbent. Parties renominate women at only slightly lower rates in reserved constituencies than men in unreserved constituencies, but discriminate heavily against female incumbents in unreserved constituencies. Overall, the selection effect dominates, and women deliver better policy outcomes on average.

We find that both voters and parties possess a taste-based preference against women. Furthermore, our structural modelling approach allows us to conduct counterfactual experiments to understand the effects of removing gender quotas. Without quotas, (i) the policy component of voter utility falls almost fourfold, and (ii) the share of women in the BMC falls from 56% to 3%, remarkably similar to the 2% female share in the BMC in the pre-quota period (Barry et al. 2004).

Implications for institutional design

Despite improving policy outcomes, quotas reduce overall voter welfare because voters' strong taste-based preference against women outweighs the policy gains. Importantly, quotas are indispensable for ensuring female representation because even if parties did not discriminate against women, the female share in the BMC would still be at most 35% without quotas due to taste-based voter preference in favour of men. Our counterfactual analyses yield two important implications for gender quota policies. First, two-term random reservations would improve voter welfare by 9–11% relative to the existing baseline while achieving the same descriptive representation goal as the current one-term system – interestingly, a similar proposal was mooted by the prominent political party in Mumbai, the Shiv Sena.[1]1. Before the party's split in 2022.  Second, purely rotating reservations that are currently in use in many other local bodies in India impose harder term limits on men, eliminating their discipline incentives and reducing voter welfare.

References

Alt, J, E Bueno de Mesquita, and S Rose (2011), “Disentangling accountability and competence in elections: Evidence from US term limits,” Journal of Politics, 73: 171–186.

Barry, J, T Honour, and S Palnitkar (2004), “Social movement, action and change: The influence of women’s movements on city government in Mumbai and London,” Gender, Work & Organization, 11: 143–162.

Desai, Z, V Karekurve-Ramachandra, and S Montero (2024). "Are women better politicians? Discrimination, gender quotas, and electoral accountability," Unpublished manuscript.