Why the Poor Vote in India: “If I Don’t Vote, I Am Dead to the State”

Assessments of the Indian state have often pointed to its general failure to improve the conditions of the poor and the very poor. Legions of studies (see, for example, Bardhan 1984, 2010; Kohli 1989; Weiner 1991; Dreze and Sen 1996; Dube 1998; Mehta 2003; and Varshney 2005) have described how the state has failed those most in need.

Lokniti, the Programme for Comparative Democracy of the Delhi-based Centre for the Study of Developing Societies (CSDS), carried out all three surveys. The 2009 State of the Nation survey covered all states except Jammu and Kashmir, Goa, and the North East; the 2004 National Election Study covered all states and union territories.

First, we randomly selected localities from the list of polling booths used by the electoral commission. Next, we confirmed the income and caste profiles of localities and neighborhoods through multiple sources, all independent of one another. When a locality fitted our required profile, we selected it. For neighborhood and locality selection, our income parameters included upper, middle, and lower-income households; and for caste, our parameters included households from the upper, backward (or intermediate), and ex-untouchable castes. In rural and urban settings, we conducted focus groups in nine types of neighborhoods.

The focus groups and follow-up interviews were conducted in the following districts: Meerut, Gautam Budhnagar, and Azamgarh in Uttar Pradesh; Sangli, Vardha, Mumbai, and Nagpur in Maharashtra; and Madurai, Chennai, and Chidambram in Tamil Nadu.

These three large Indian states have very different income levels. Per capita income is highest in Maharashtra, among the lowest in Uttar Pradesh, and average in Tamil Nadu. To be specific, in 2002–2003 in India, per capita income at constant prices was Rs. 11,013 per annum; per capita income was Rs. 15,580 in Maharashtra, Rs. 5,603 in Uttar Pradesh, and Rs. 12,696 in Tamil Nadu. Only Goa’s per capita income was higher than Maharashtra’s (see sampark.chd.nic.in/images/statistics/SDP2005R6.pdf).

Tamil Nadu and Maharashtra have been home to movements demanding social equality, but similar movements have not taken hold in Uttar Pradesh. In Tamil Nadu, where the Dravidian movement existed through the middle half of the twentieth century and where a Dravidian party finally came to power in 1967, there has been a history of mobilization among the marginalized. In Uttar Pradesh, by contrast, the electoral mobilization of the marginalized and the poor is a recent phenomenon (Chandra 2004). We did not conduct research in West Bengal, where a leftist party focusing its attention on the rural poor has been in power since 1967. Banerjee (2007) has found independently that the poor vote for reasons that echo some of our findings, although she sees voting as akin to the performance of a religious ritual.

These three states also differ on the dimensions of interparty competition and history of voter mobilization. Uttar Pradesh is now characterized by multiparty competition, with each of the three major parties representing a social/caste bloc. In Maharashtra, electoral competition is largely between the centrist Congress and its splinter parties and two parties on the right, whereas in Tamil Nadu, at least until the last elections, two regional parties have competed to control the state legislature.

In coding our data, we took the broadest possible view with respect to defining relationships of patronage and exchange: that is, expected or actual benefits could accrue from interpersonal or community ties, or by means of public policy. We also used a broad classification in defining a community, which could comprise an ethnic, a linguistic, a religious, or a class group and could also mean a locality, a village, a city, an electoral district, or a region.

There are party workers in poor localities, but these workers exercise less influence in their respective parties than do party workers in more affluent localities. Poor respondents also regularly lamented the fact that they did not know prominent politicians or government functionaries in their districts.

Ethnographic work by Banerjee (2007) in two villages of communist party governed West Bengal also reports the use of the language of rights by the poor. On the whole though, assertion of a right or citizen affirmation has so far remained an underreported motivation for voting in the literature on participation.

Ethnic, racial, and religious identities, because of their relative permanence, emerge as an effective organizing principle for durable voter coalitions (Wilkinson 2007; Jha et al. 2005; Fearon 1999). Since voters are convinced that voting their coethnics into office offers them the best chance of accessing state resources, they show up on Election Day to vote for coethnic patrons (Horowitz 1985; Chandra 2004; Posner 2005). Scholars working in Latin America have found that parties can motivate turnout by targeting swing voters (Stokes 2005) or core supporters (Nichter 2008) with direct cash transfers before elections.

Explanations based on clientalism and patronage focus more on party–voter relations rather than state–voter relations.

In Weapons of the Weak, Scott (1985) shows that the poor and the marginalized in Malaysian villages do not give in to subjugation. In the absence of opportunity, they resort to passive or hidden forms of resistance against the dominant classes. Subscribing to a framework grounded in religious ideas as well as in traditional norms of reciprocity, they hold the dominant classes accountable for their behavior.

The record of public service delivery varies across Uttar Pradesh, Maharashtra, and Tamil Nadu. Tamil Nadu and Maharashtra provide a wider set of services than Uttar Pradesh, and yet, in the evidence drawn from focus group discussions and interviews, we do not observe any difference in how the poor described their relationship with the state. We do however find that the poor complain about a larger set of issues in Tamil Nadu and Maharashtra than Uttar Pradesh. But this probably reflects the “endowment effect”: citizen expectations of the state are informed by the initial endowment of state provisions (Sunstein 1993). Even when the state provides more services, its relationship with the poor still remains capricious, and irrespective of whichever state they live in, the poor perceive their treatment by the state as arbitrary and neglectful.

Corbridge et al. (2005) confirm this finding quite succinctly, saying that the poor see the state when the state wants to see them.

The poor certainly face the capriciousness of the state in urban areas, but they are frequently also treated with indifference in rural areas. In urban areas, the residential situation of the poor is precarious; tied to this problem are issues of access to the public distribution system and other livelihood issues. The poor are under constant threat of having their settlements uprooted and losing their recognition as inhabitants of the city. The primary fear of the poor in urban areas is that unless they have political protection, politicians and city administrators can easily dispense with their settlements, and their hutments can be destroyed. The residents of slums in cities like Delhi and Mumbai, where there is a large premium on land, often made this connection with remarks like “We have to make ourselves count; otherwise, we will be evicted” or “You think the parties are interested in our welfare? If we are very fortunate, our slum will be legalized.” In rural areas, both farm and nonfarm work pay close to the minimum wage. In many instances, however, the state’s incomplete implementation of the land-redistribution program has left the poor with land entitlements but without control over the actual holdings. The state has generally failed to intervene on their behalf to ensure the transfer of land. It has also failed to provide them with reliable access to public goods like clean drinking water, electricity, schooling, and health care. Nevertheless, the poor citizens we spoke with still expected the state to deliver public goods. For the most part, however, access to the state and to jobs with the state remained out of their reach. For a large section of these poor citizens, issues of livelihood and security were central to their lives, and many said that they would offer bribes to state officials and become clients of the state if they had the resources to do so. They also said that they seldom counted with the state, and that the state was not interested in addressing their issues. When they did make contact with the state for the redress of individual as well as collective problems, these interactions were fraught with hurdles.

Ample evidence for this claim has been provided by reports of the Planning Commission and the Central Vigilance Commission, both governmental bodies.

Over the past six decades, the Indian state has regularly conducted elections for the national parliament and the state assemblies. An independent electoral commission has overseen these elections, and the election results have seldom been challenged. Incumbents have been replaced regularly, and, as a result, the rate of turnover of elected representatives in India is one of the highest among democratic countries. The use of party symbols on paper ballots in the past—and, more recently, on electronic voting machines—has ensured that illiteracy will not prevent poor voters from choosing parties. As another feature of the electoral process in India, updating the voter lists is the responsibility of the Election Commission. This arrangement relieves poor citizens from the burden of self-registration for voting. A variety of steps have also been taken to reduce the threat of coercion of voters. Historically, intimidation of voters has not been entirely absent. In fact, in the first two decades of India’s democratic experience, the landed elite exercised control over the turnout in parts of rural areas (Rudolph and Rudolph 1967), but these practices have been on the decline. The state makes unprecedented security arrangements for elections. In regions known for electoral violence or fraud, elections are staggered over many days to allow for movement of security forces between different areas. In addition, the mass media widely report on electoral malpractice. If irregularities in an election are confirmed, the Election Commission countermands the election results and orders repolling (Lyngdoh 2004). Not surprisingly, the Election Commission enjoys the highest credibility among all state institutions (Linz et al. 2007). With the possible exception of the elections in Jammu and Kashmir, especially in the 1970s and 1980s, and those in some northeastern states, elections in India are largely free and fair, and a large majority of Indian voters see them as such (Rao 2004).

For a discussion of why a sympathetic relationship between a representative and the people he or she represents is central to political representation, see Rehfeld 2006; Urbinati 2000, 2006; Urbinati and Warren 2008.

The locality-level campaign survey was conducted across 138 localities in four states—Bihar, Maharashtra, Uttar Pradesh, and Tamil Nadu—in the months of April and May during the 2009 parliamentary elections. The campaign was observed in these localities in the final week before Election Day when campaigning was at its peak. In Uttar Pradesh, Maharashtra, and Tamil Nadu a total of 45 poor localities were included in the survey. Twenty-one of these were predominantly SC localities while 24 were predominantly upper and backward caste localities.

Harris (2005) reports similar findings. Using a more exhaustive measure of participation akin to that of Brady et al. (1995), he found that the poor were more likely than the non-poor to be engaged in political activity.