Empowering Female Leaders and Voters in Rajasthan, India
- Rural population
- Transparency and accountability
- Women’s/girls’ decision-making
- Gender attitudes and norms
- Diversity quotas
- Gender quotas
In many low- and middle-income democracies, the entrenchment of the economic elite in politics may undermine government effectiveness by limiting the accountability of elected officials. Researchers evaluated a voter information campaign and exploited the random assignment of reservations for women in village councils to measure the impact of information and reduced incumbent advantage on village council elections. Both interventions increased the number of candidates and drove the worst performers out of the race, though in the case of the voter campaign they had short-term costs in terms of officials’ performance post-election.
The effectiveness of democratic governance relies on high quality citizens entering politics and the electorate holding their representatives accountable. In many low- and middle-income democracies, the entrenchment of the economic elite in politics may undermine this process. Incumbents may be reelected despite bad performance due to voters’ lack of information. Additionally, citizens may be reluctant to seek elected office due to the costs associated with entering an election. A growing body of evidence points to the importance of entry costs in affecting candidate selection and of voter awareness for candidate selection.1 However, less is known about how voter information and entry costs influence incumbents’ decisions to re-run and for new-comers to seek office. There is also little evidence about whether reducing incumbency advantages can improve the quality of governance and government effectiveness.
Context of the evaluation
Village councils or Gram Panchayat (GP) are the lowest tier of elected representation in India. They are responsible for the provision of village infrastructure and identifying welfare recipients. In addition, GP administer the Mahatma Gandhi Rural Employment Guarantee Act (MGNREGA), the government’s flagship anti-poverty program, which guarantees 100 days of work per year to rural households at the state-regulated minimum wage. The GP registers potential beneficiaries, plans and executes public works, and ensures wages are paid on time.
In Rajasthan, each council represents an average of 3,000 people across multiple villages. The GP makes decisions by majority vote, with each member elected to represent an individual ward. The head councilor (Sarpanch) is elected at-large by a direct vote of the GP electorate. While the sarpanch does not have veto power, she is the only full-time member of the council and has significant power over final council decisions. Since 1993, each Indian state is constitutionally required to reserve one third of sarpanch seats for women. In Rajasthan, GP to be reserved for women are randomly selected before each election. This evaluation focuses on the group of GP in which sarpanch seats were not reserved for women in the 2010 election. In GP where sarpanch seats were reserved for women in previous elections, the removal of reservations weakens the incumbency advantage because female incumbents often face social and family pressure to not re-run for office. In turn this lowers the costs of entry for other potential candidates.
Details of the intervention
Researchers conducted a randomized evaluation to test the impact of a voter information campaign and reservations for women in the previous electoral cycle on candidate entry, election results, and post-election performance in village council elections in Rajasthan. At the time of the study, two previous election cycles included reservations for women. For the 2010 election, the government randomly assigned half of the GP in the state to be reserved for women regardless of whether they had been reserved in the past. Researchers used this random assignment to analyze the impact of previous reservations in the GP that were not reserved in the 2010 election.
The second intervention was a voter information campaign. Prayatn Sanstha, a local NGO, ran the campaign between one and 20 days before the elections using publicly available data on MGNREGA implementation. Prayatn used calendar posters and theatrical performances to provide information on the roles of village council leaders and leaders’ past performance implementing MGNREGA. Election districts were randomly assigned to receive either a “general” or “gender” information campaign. The general campaign emphasized that citizens should vote based on elected officials’ job performance, rather than voting based on caste or gender. In addition to the components from the general campaign, the gender campaign emphasized that both genders could run in unreserved councils, and that women council heads typically performed as well as men.
In total, Prayatn distributed 130,000 informational calendars and performed 472 shows with an average attendance of 32 percent of the adult population of each village. Using administrative sources and household surveys, researchers collected data on election results, measures of the number and quality of candidates entering each race, the likelihood of incumbents running, and post-election performance as measured by employment generated through MGNREGA.
Results and policy lessons
Researchers found that both reservation for women in the previous electoral cycle and the voter awareness campaign increased the number of candidates and decreased the probability that an incumbent ran and won an election. They found some evidence that newly elected leaders in constituencies with the general information campaign performed worse in terms of employment generated under MGNREGA in the short-term.
Candidate entry: Both the information campaign and previous reservations for women led to an increase in the number of candidates. An average of seven candidates entered each election where the seat had not been previously reserved for women. One additional candidate ran in GP which either had been reserved or exposed to the information campaign. Past reservations for women had a greater effect on the probability of women running—a 48 percent increase—whereas the information campaign increased participation from lower castes and tribal candidates. When seats were no longer reserved, male family members often replaced female incumbents as GP candidates.
Election outcome and accountability: Both interventions decreased the probability that incumbents ran and won elections. Furthermore, they pushed the worst performing incumbents out of the race, and among those who chose to run, it reduced low performing incumbents’ probability of winning. Incumbents who performed one standard deviation below mean performance in implementing MGNREGA were 22 percent less likely to run for reelection in villages with the information campaign, and 34 percent less likely to run in previously reserved villages. The best performing incumbents, however, were as likely to run and win as in comparison villages. In previously reserved villages which received the information campaign, the probability of running increases by 26 percentage points, from a base of 54 percent, for the best performing candidates and their family members.
Post-election performance: In non-reserved councils, the general campaign led to a lower performance implementing MGNREGA and had no statistically significant impact in villages with previously reserved councils. These results indicated some short-run costs of broadening the political base through the information campaign. Sarpanch from lower castes may be disadvantaged at first in dealing with the whole village and the administration to organize MGNREGA work. One should keep in mind however that our analysis only measures the short-run impacts on a single program. Broadening the candidate pool could lead to a process of learning and selection which could lead to better outcomes in the long-run.
Banerjee, Abhijit, Esther Duflo, Clement Imbert and Rohini Pande. “Entry, Exit and Candidate Selection: Evidence from India.” 3ie Draft Grantee Final Report, 2013