by Abhijit V. Banerjee, Rohini Pande & Felix Su
Elections are quite possibly India’s proudest achievement. Over the next few weeks, 700 million voters (which is more than four times the number in the next largest election—this week’s contest in Indonesia) will have the option to exercise their franchise, under conditions that are mostly safe and free from direct intimidation. And while some politicians will try to make it otherwise, most people will vote in their own name and only in the own name. After the election the loser will congratulate the winner and promise (however disingenuously) to work together with the winner for the country’s future. And that will be that.
It was not always like this. Ballot-box stuffing was common, as was voter intimidation. Unsurprisingly, the losers would often challenge the result. In some cases—most famously in Kashmir—the election did little to enhance the legitimacy of the government that it brought to power.
In all of this there has been enormous progress over the last decade, thanks to the Election Commission and the willingness of incumbent governments to give it a free hand. For India, where few government services inspire confidence and even fewer have been improving, elections stand out as a remarkable example of state effectiveness.
Given this observation, it is ironic that the Indian voters are not more enthusiastic about voting. In the 2004 national elections only 58% voted. Moreover the trend is not promising: in the 1998 elections 62% voted, while in 1999, 60% voted. In Uttar Pradesh, the biggest state in the country, only 48% showed up at the polls in 2004.
One standard explanation for why citizens do not vote is that they consider it irrelevant to their lives. For example, African Americans in the United States have a substantially lower propensity to vote than the rest of the population. It is often suggested that they believe, rightly or wrongly, that there is little point in privileging one white guy over another, given that both will ignore them after the election.
The problem in India does not seem to be about class or ethnicity: in the 2008 Delhi state elections, the turnout was only slightly lower in polling stations located in slum areas (56%) than in non-slum areas (58%). In the 2007 UP state election, turnout in areas with a higherthan-median fraction of Scheduled Castes roughly mirrored the state average. However, it is not hard to imagine that many voters—cutting across class, caste and religion—might be relatively cynical about the role of the government in their lives, given the manifest evidence of government failure all around them.
A different reason for not voting, which might at first seem very similar, has to do with feeling disempowered. It is not so much that citizens do not want to exercise their franchise, but rather that no one has taken the trouble to make the case to them that they should. The only people who ever ask them to vote are the politicians, who come around during elections, and less often afterwards.
Some evidence for this latter hypothesis is offered by the experience of several NGO campaigns implemented during the recent Uttar Pradesh and Delhi state elections. During the 2007 UP state election, Saarthi, a local NGO, carried out two campaigns in about 400 villages spread across 18 constituencies in the three districts of Lalitpur, Sitapur and Bahraich. In each campaign volunteers from Saarthi spent a day talking to villagers in a series of meetings that culminated in a puppet show. Each campaign delivered a simple and largely uncontroversial message. “Vote for clean candidates, corruption hurts everyone,” said one; “Vote on issues, not on caste,” said the other.
Since the villages where the campaigns took place were picked by lottery, we can infer the effect of the campaigns by comparing villages that were selected for the campaigns with villages that were not chosen. The results are striking, especially given the fact that substantially less than a quarter of each village was exposed to either campaign. Voter turnout in villages not chosen for the Saarthi campaigns was 54%. Villages which received a Saarthi campaign saw a 6 percent increase in turnout. Interestingly, the two campaigns affected different demographic groups in different ways. The caste campaign invigorated male voters: adult male voter registration increased by 6% in villages that received the caste campaign. This was paralleled by an 11% increase in male turnout. The effect among women was much smaller. In contrast, the corruption campaign enthused women, who increased their turnout by 8%. In contrast, male turnout was unaffected by the campaign.
A related exercise was carried out during the 2008 Delhi state election in collaboration with the newspaper Hindustan. Hindustan had the innovative idea of printing on each day leading up to the election information about the candidates running in a certain cluster of constituencies. Each newspaper profile emphasized both the candidate’s background (wealth, education, criminal record) and, for those who had been one, and some measures of his performance as an MLA carefully put together by a Delhi-based NGO, Satark Nagrik Sangathan.
The program that we studied involved distributing in each of 200 slum neighborhoods around Delhi city 400 free copies of the particular issue of the newspaper that contained the information pertaining to those particular neighborhoods. In the days leading up to the day when the newspapers were distributed, a network of eight NGOs conducted a mobilization campaign in those neighborhoods. The campaign involved three days of door-to-door visits in each area, where nonpartisan fieldworkers recited and handed out pamphlets delineating the roles and responsibilities of MLAs
Once again, turnout increased in the randomly chosen polling stations covered by the campaign. The magnitude of the effect was significant --a 4% increase in overall turnout. The effect is smaller than in Uttar Pradesh, which may reflect that urban voters are somewhat better-informed and/or more cynical. But the fact remains that even in a major metro, there exists a set of voters who were willing to queue at the polls because someone had cared enough to communicate with them about the election.
What seems clear is that there are a lot of voters in India—perhaps more in rural areas, perhaps more among the lower castes—who are neither disinterested nor jaded (nor are they, by the way, the shrewd, all–knowing, rural Indian voters from some social scientists’ fantasies). They are open, they are interested, and they are waiting to be included in the grand conversation about Indian democracy that is being carried out (as of now) on their behalf by others.