Sarpanchs and the Politics of Gram Panchayats

Gram panchayats are the elected village councils that constitute the most local form of elected government in rural India. They are headed by a village council head (the sarpanch) and administered by an unelected civil servant (the secretary). The 73rd constitutional amendment, which reorganized the architecture of rural local institutions in 1993, mandated political reservation in favor of scheduled castes (SCs) for all elected positions in gram panchayats, including the position of sarpanch on which this study focuses. In addition, this amendment required that no position of sarpanch be reserved for the same group for two consecutive elections, hence combining the principle of reservation with a principle of rotation. Local authorities thus determine before each election the number of positions of sarpanchs to be reserved for each disadvantaged category, including the SCs.Footnote ⁶

In Rajasthan, villagers directly elect their sarpanchs, picking among several candidates after short but usually heated campaigns (Vij Reference Vij2010). Sophisticated political maneuvering takes place in the run-up to these elections, and competing candidates have recourse to a variety of licit and illicit campaign strategies to assemble a winning coalition (Vij Reference Vij2010). Given the heterogeneous caste make-up of most Rajasthani villages, these coalitions by definition reach across group boundaries. As a result, the winning candidate is usually elected by the votes of villagers not from his or her own community. Wherever the position of sarpanch is reserved for members of the SCs, only SC individuals can stand for election. Campaigns, however, remain just as competitive, with several SC candidates vying to obtain the votes of villagers from all groups. Thus, although reservation necessarily results in the election of a member of the SCs, SC sarpanchs owe their victory to votes received from other groups.

These local elections are hotly contested in Rajasthan for a simple reason: Substantial advantages come attached to the position of sarpanch. In Rajasthan, gram panchayats perform two important tasks. They select beneficiaries for a number of welfare schemes and decide on the construction and the maintenance of village public goods, such as streetlights, roads, and drains. Members of the gram panchayat, especially the sarpanch, also play an informal role within the village, mediating private conflicts between villagers and serving as brokers between villagers and local authorities.

Although decisions are theoretically the product of deliberations of the village council, the balance of power among the different actors within the gram panchayat—the council members, the sarpanch, and the unelected secretary—varies according to these actors’ respective backgrounds. Wherever the sarpanch is perceived as weak, inexperienced, or uneducated, both the council members and the unelected secretary exert their full influence. As a result, decisions tend to be taken collectively. On the contrary, in many villages in which the sarpanch is perceived as legitimate—because of age, education, and/or caste—he or she exerts greater authority in decision making.

With these nuances in mind, it remains the case that all sarpanchs—even those whose decision-making power tends to be eroded by other actors within the gram panchayat—play a central role in the village. Their signature and their stamp are required on all official documents produced by the gram panchayat, thereby rendering them indispensable to the functioning of the village council and to villagers trying to obtain official documentation.Footnote ⁷ In addition, sarpanchs are the only village-level officials authorized to disburse funds to buy materials for construction, and they are personally in charge of ensuring that public works are implemented.Footnote ⁸ Given this responsibility, sarpanchs are frequently in contact with government officers at the block and at the district level to ensure that funds sanctioned for the village are released. As village council head, they also represent the village in various institutions and assemblies at the block and district level and maintain relations with a host of locally powerful actors (including the police). Last but not least, sarpanchs play an important ceremonial role when they preside over the fortnightly meetings of the gram panchayat or the participative village assemblies (gram sabhasFootnote ⁹ ).

The social status acquired by becoming the top official of the village is not, however, the main reason why these elections are so hotly contested. The direct or indirect monetary returns that sarpanchs derive from their function constitute a more influential factor. Sarpanchs receive a relatively meager monthly salary (INR 3000 as of 2012). A vast majority of them, however, are able to increase their net wealth during their tenure in proportions that appear unrelated to their previous income or to their official income. Sarpanchs’ new connections, as well as their central role in assisting with local development works, almost invariably provide them with the ability to launch a profitable contracting or land acquisition business. The role of sarpanch may also enable them to embezzle public funds,Footnote ¹⁰ especially when they run their own contracting businesses. Regardless of whether their activities are licit or illicit, sarpanchs’ new connections provide them with new opportunities to amass wealth, which ultimately pave a path to social mobility for the official and his or her immediate family (Baviskar and Matthew Reference Baviskar, Baviskar and Mathew2009; Kumar and Rai Reference Kumar and Rai2005).

My observations provide anecdotal evidence supporting this widespread presumption among the villagers I interviewed. The major acquisitions made by sarpanchs and their families during their tenure suggest that they, in most cases, become conspicuously richer. Six of the eight sarpanchs whom I observed had recently acquired their first sport utility vehicle, thus proving correct the popular nickname of “Bolero sarpanchs.”Footnote ¹¹ An overlapping sample of six of the eight sarpanchs had also undertaken major renovation works at their residence.

Changes in Village Life under an SC Sarpanch

Given the role played by sarpanchs in village life, what does change, and what does not change, when a member of the SCs is elected sarpanch?

As suggested by several recent studies (Chauchard N.d.; Dunning and Nilekani Reference Dunning and Nilekani2013), it is first important to note that sarpanchs from traditionally disadvantaged groups, such as the SCs, do not fundamentally overturn the traditional social order. The material gains made by villagers from the SCs, if any, remain minimal, and traditionally dominant groups remain dominant. This is because the sarpanch's ability to aid SC villagers is limited by the local political context. To gain office, SC sarpanchs depend on a multigroup coalition and thus may not wish to make decisions viewed as favorable to their own group. Indeed, SC sarpanchs may even behave as proxies for influential local strongmen.Footnote ¹² The most significant constraint faced by SC sarpanchs may, however, be the institutional structure of the gram panchayat system itself, in which council members may play an outsized role, especially in the face of a sarpanch perceived as weak or “illegitimate,” which is often the case with SC sarpanchs.

In spite of this notably weak impact on the redistribution of resources, the arrival of an SC villager in office nonetheless entails a series of tangible changes in village life. Reservation for a member of the SCs may lead to a major reversal of traditional caste-based roles in the village. Simply put, the election to a position of power of a member of a stigmatized group exposes villagers to sights that would have been unlikely before reservation was implemented. Villagers may observe a member of the SCs seating ceremoniously on a dais, providing his or her opinion or signing off on the council's decisions. As the sarpanchs and their entourage walk through village streets to assess various public works, villagers see members of the SCs on streets on which they otherwise dared not venture. The conjunction of each of these apparently minor changes represents a significant change to the social environment in which villagers live.

The second visible consequence of reservation for a member of the SCs derives from the ability of sarpanchs to accumulate wealth during their tenure. Given this reality, reservation of the office of sarpanch for a member of the SCs often ensures that at least one household within that category will display—often in the most ostentatious manner—ownership of a number of goods that members of these castes typically do not possess.

Third, the ascension of an SC villager to the position of sarpanch increases several forms of contact between some SC and some non-SC villagers. Given the social role played by the sarpanch in the village, reservation first appears to increase the frequency with which non-SC villagers enter the SC hamlet. Villagers from all castes visit the sarpanch in his or her courtyard and in some cases even inside his or her home.Footnote ¹³ Interestingly, reaching the sarpanch's courtyard requires villagers to take roads that they do not usually take and to pass by homes they would usually not go by. In that sense, the relatively frequent visits that elders and household leaders from all castes pay to the sarpanch may subtly redefine the geography of the village: They now routinely visit areas that were previously considered unworthy of a visit. In reverse, the visits made by the sarpanch and his or her entourage in the course of their work across the village give several members of the SCs access to areas of the villages in which they are otherwise rarely seen.

Finally, reservation for an SC sarpanch appears to create a new channel of communication between the village's SC community and formal and informal institutions. Sarpanchs attend numerous meetings at higher levels of government and are typically engaged in a multitude of deals (lawful or not) with various institutional actors. As implied by Dunning and Nilekani (Reference Dunning and Nilekani2013), sarpanchs also frequently develop ties to powerful party officials, who rely on them for grassroots mobilization. In sum, reservation offers the SC community access to a network of relatively powerful individuals—local functionaries, local police commanders, and in some cases even members of state legislatures—who were previously out of reach. In addition to the personal benefits that an SC sarpanch may derive from these important connections, another byproduct of reservation may come in the form of an extra measure of linkage for the SC community (Krishna Reference Krishna2004). Given the crucial importance of these links in the event of a conflict between an SC villager and another person,Footnote ¹⁴ this undeniably constitutes an important change in village life.

The Psychological Impact of SC Sarpanchs: Hypotheses

Given the changes in village life just described, how should we expect reservation for a member of the SCs to influence the psychology of intergroup relations? Whereas an important literature has suggested that a group's access to representation may drive short-term, emotion-driven psychological reactions in majority populations (Coser Reference Coser1956; Horowitz Reference Horowitz1985; Olzak Reference Olzak1990; Petersen Reference Petersen2002), this study focuses on the cognitive effects of descriptive representation. What new group-related information is revealed as a result of reservation for a member of the SCs? Which group-related beliefs may, in turn, be expected to change?

Researchers who have explored the effect of a group's access to political representation on the psychology of members of dominant groups (Beaman et al. Reference Beaman, Chattopadhyay, Duflo, Pande and Topalova2009; Hajnal Reference Hajnal2001; Reference Hajnal2005) have emphasized its potential impact on stereotypes. Access to political representation, these authors argue, provides citizens with updated information regarding the characteristics of politicians from that group. Because this information may be more positive than previously held information, representation may lead citizens to update their negative stereotypes about members of disadvantaged groups, replacing them with more positive ones. In that sense, descriptive representation may reduce statistical discrimination. These intuitions about the ability of descriptive representation to reduce statistical discrimination have relied on two interrelated assumptions: that updated group-related information leads individuals to develop more positive beliefs about members of a disadvantaged group and that this updated information may arise from exposure to an atypical member of that group (i.e., an elected politician).

Decades of studies in social psychology, insisting on the stickiness of stereotypes and their ability to resist new disconfirming information, however, suggest that such assumptions may be very optimistic. Ethnic stereotypes are learned early in life (Dunham and Degner Reference Dunham and J2010). They are deeply ingrained (Allport Reference Allport1954; Devine Reference Devine1989; Fazio et al. Reference Fazio, Jackson, Dunton and Williams1995) and too stable (Fiske Reference Fiske, Fiske, Gilbert and Lindzey1998) to be influenced by new information acquired through exposure to a single member of a disadvantaged group in a stereotype-inconsistent position. A variety of tactics allow members of dominant groups to maintain their stereotypes and to create cognitive consistency (Macrae, Hewstone, and Griffiths Reference Macrae, Hewstone and Griffiths1993). They may overlook individuals who disconfirm their current views or discount inconsistent evidence as an exception to the rule (Macrae, Hewstone, and Griffiths Reference Macrae, Hewstone and Griffiths1993; Weber and Crocker Reference Weber and Crocker1983). Attributional and other interpretative biases (Hewstone Reference Hewstone1994; Maass and Arcuri Reference Maass, Arcuri, Semin and Fiedler1992) may allow them to see the counter-stereotypical individual “in ways that render his or her behavior more in line with expectations, or as being due to temporary, external, situational constraints” (Moreno and Bodenhausen Reference Moreno and Bodenhausen1999, p. 7). Perceivers may also see the counter-stereotypical individual as not reflective of the group as a whole (Allport Reference Allport1954; Kunda and Oleson Reference Kunda and Oleson1997; Weber and Crocker Reference Weber and Crocker1983) – that is, they may “subtype” him or her. Given the intractability of stereotypes and the enduring nature of prejudice described in these works, it is less clear why one should expect a single local official to change the way people think about outgroups. Accordingly, the type of rational belief update—new information automatically leading to a decrease in statistical discrimination—described in Beaman et al. (Reference Beaman, Chattopadhyay, Duflo, Pande and Topalova2009) and hinted at in Mansbridge (Reference Mansbridge1999) remains an unlikely consequence of descriptive representation, especially in the short time span covered by a single political mandate.Footnote ¹⁵

The fact that descriptive representation may not alter stereotypes does not mean, however, that we should expect it to have no impact on the psychology of intergroup relations. Descriptive representation may affect at least two other group-related beliefs that contribute to the persistence of discriminatory behaviors. It could first affect perceived social norms of interaction with members of the SCs. Villagers may, in other words, perceive that other members of dominant groups are more tolerant and less discriminatory in their interactions with members of the SCs under reservation. Reservation entails de facto shifts in social norms of interaction with at least one member of the SCs: When a village is reserved, the SC sarpanch enters a variety of homes within the village, receives the visit of villagers, and builds relationships with non-SC villagers. In a context in which segregation and hostility otherwise prevail, these changes represent a major breakthrough. As villagers living in reserved villages are faced with a number of norm-breaking situations, their overall perception of social norms of interaction may be evolving toward more tolerance and less segregation. Consider how a young upper caste villager critical of reservations sardonically described the new social status of the sarpanch's extended family: “Since he [a SC villager] became sarpanch, people salute them and even invite them [home].”Footnote ¹⁶

These changing perceptions of social norms of interactions should, in turn, matter for interpersonal relations. A rich tradition of scholarship on conformity and social consensus has suggested that, while engaging in social behaviors, individuals often rely more on perceived social norms than on their own personal beliefs (Allport, Reference Allport1954; Asch, Reference Asch and Macoby1958; Cialdini, Kallgren, and Reno Reference Cialdini, Kallgren, Reno and Zanna1991; Crandall and Stangor Reference Crandall, Stangor, Dovidio, Glick and Rudman2005; Kuran, Reference Kuran1995; Stangor, Sechrist, and Jost, Reference Stangor, Sechrist and Jost2001). Building on these intuitions, research has suggested that perceived social norms of interaction with members of disliked groups are more powerful predictors of intergroup behavior than individual attitudes toward that group (e.g., Blanchard et al. Reference Blanchard, Crandall, Brigham and Vaughn1994; Paluck Reference Paluck2009; Paluck and Green 2010; Paluck and Shepherd Reference Paluck and Shepherd2012). If reservation did affect these perceived social norms of interactions, we should thus expect upper caste villagers’ behaviors to evolve as well. As upper caste villagers perceive that social norms relating to interactions with SCs are changing around them, they may adjust their own behaviors.

Second, I argue that reservation could affect villagers’ perceived legal norms of interaction toward members of the SCs. Changes in these beliefs are more political in nature. The legal arsenal criminalizing untouchability-related practices has, over the years, become extensive (Shah et al. Reference Shah, Mander, Thorat, Deshpande and Baviskar2006). This legislation is rarely implemented, either because local authorities do not cooperate or because wronged members of SC communities lack the capacity to take their claim through a fragmented and biased system of local courts. As a result, flagrant untouchability-related behaviors are frequently left unpunished in spite of the existence of extremely repressive laws. When an SC sarpanch assumes office, the possibility that local authorities will use and implement these laws may be perceived as more likely. Because of the sociopolitical context in which they serve, it is unlikely that SC sarpanchs would personally come forward as strong advocates for victims of these practices, and they do not have direct authority over the police. Yet, given the political connections and linkages that SC sarpanchs build through their functions, villagers may perceive that other powerful local actors with whom SC sarpanchs now interact on a regular basis (party leaders, members of state assemblies, etc.) could pressure the police on their behalf. As a consequence, they may be more likely to perceive that otherwise relatively common humiliations could now be considered atrocities under Indian law and that local authorities would be more likely to pay attention to caste-related incidents. Echoing many other upper caste men who explicitly made this connection in reserved villages, one of my interlocutors summarized the issue: “SCs are now protected, so we make sure not to do anything wrong; otherwise a complaint might be very easily recorded against us.”Footnote ¹⁷ As hinted by this statement, these changes should in turn be expected to have an effect on the practice of untouchability. Because the probability of legal sanctions by the authorities in part predicts the behaviors of members of dominant groups, changes in these beliefs should constitute a major predictor of untouchability-related behaviors.

The rest of this article evaluates the impact of reservation on the various beliefs considered in this section and on the readiness of non-SC villagers to engage in untouchability-related practices.

Sampling Villages: Exploiting A Discontinuity

In Rajasthan, electoral officers of each district bureaucracy reserve a fixed number of sarpanch positions before each election according to a rule based on the share of SC population at the panchayat samiti level.Footnote ¹⁸ Namely, the total number of gram panchayats (GPs) reserved within each panchayat samiti at each electoral period is proportional to the share of the SC population in the panchayat samiti area. To ensure rotation across GPs in the implementation of reservation, electoral officers have in addition been instructed to rank GPs according to their proportion of SCs and to go down that list, reserving GPs with increasingly small SC populations at each successive electoral period. Following this rule, illustrated in Table 1, the set of reserved GPs systematically differs from each other across electoral periods in terms of their SC population. Online Appendix B provides additional details on the implementation of this rule in each of the panchayat samitis in which the survey took place.

TABLE 1.

Reservations for the Scheduled Castes in Jalore Panchayat Samiti (Jalore District)

*Based on 2001 Census of India Data

Note: Since members of the SCs constitute less than 20% of the population in Jalore Panchayat Samiti, 5.5 (that is, either 5 or 6) GPs are supposed to be reserved at each electoral period, as is the case here. Assignment to reservation then unfolds in decreasing order, starting with the GPs with the largest SC population share and ending with the GPs with the smallest SC population share.

Given the rules presiding over this rotation, if I randomly sampled GPs within Rajasthan, I would not be able to determine whether the variation in my outcome variables was due to reservation for members of the SCs or to a higher share of SC population within reserved GPs. However, because assignment to reservation depends on this demographic principle, this system enables me to generate a natural experiment on a subset of the data. Namely, I exploited the discontinuity in the implementation of reservation. I selected GPs whose SC population was around—some slightly above, some slightly below—the local threshold for reservation in 2005 (as seen in Table 1, in Jalore this “threshold” was 18.07%). I thus focused on comparisons between two sets of GPs: GPs that were reserved in 2005 (and, given the rules detailed earlier, had never been reserved before) and GPs that were not reserved in 2005 and had never been reserved before.

Practically, to sample the villages, I proceeded in four steps.

1. 1. I first selected four districts.Footnote ¹⁹ They were purposefully chosen to ensure that different parts of the state would be represented in the sample.

2. 2. I then randomly selected four panchayat samitis within each of these four districts.

3. 3. Within each of the 16 targeted panchayat samitis, I listed GPs that had been reserved for SCs in 2005 as well as the 5 “never reserved” GPs with the largest share of SC population.

4. 4. Selecting from this group of GPs, I formed 8 matched pairs of GPs in each district, for a total of 32 pairs and 64 GPs. In each case the survey team targeted the largest village under the authority of the GP. Footnote ²⁰

Each of these pairs contained one reserved GP and one unreserved GP that were geographically proximate.Footnote ²¹ Given limited resources, the pair matching was designed to ensure that the targeted GPs were matched on a number of key observables that could potentially affect my outcome variables.Footnote ²² In addition to being matched on SC population at the GP level, selected pairs were thus also matched on the population of their head villages, which the survey targeted, on their relative distance to a city, on caste make-up (a dichotomous variable indicating whether the majority of the non-SC population of the GP belonged to a single subcaste), and the identity of the dominant (i.e, the most numerous) SC subcaste in the GP.Footnote ²³

Although this pair matching —detailed in Online Appendix C— allowed me to ensure that selected reserved and unreserved GPs were comparable on a number of key variables, my sampling strategy also guaranteed a more general form of balance. Namely, because these pairs were selected around the local thresholds for reservation, my sample was balanced on a large set of covariates drawn from the census of India. Table 2 and Online Appendix D (Table D.1) show that selected reserved and unreserved GPs are statistically indistinguishable on a large number of village characteristics (including their population of SCs), a necessary condition for a valid natural experiment (Dunning Reference Dunning2008).

TABLE 2.

Difference of Means between Reserved and Unreserved Villages, with Standard Errors

Note: All analyses at the GP level and based on data from the 2001 census of India.

Sampling Villagers

In each of the 64 villages visited, 12 respondents were interviewed, for a total N of 768 respondents. The surveys were administered in November 2009 to a sample of 12 non-SC individuals within each village. Respondent sampling followed an “as-random-as-possible” procedure within each of the caste-homogeneous units into which Indian villages are divided. Based on information on patterns of settlement in each village,Footnote ²⁴ each interviewer was assigned to survey members of a specific caste group. After being assigned to a specific group, interviewers were “placed” by supervisors in front of a house at one end of their sections and began by attempting to survey that household. They were instructed to subsequently walk toward the other end of their zone and to interview a member of every n^th house (the n number depended on caste settlement size) along the way. Within each house, interviewers surveyed the “first available male” (in the absence of female interviewers, no female was interviewed). If no male was available, interviewers were instructed to come back to each house twice before they were allowed to target another household. Proceeding in this fashion, the overall response rate was 67.8%.Footnote ²⁵

Eliciting More Honest Responses: The Interview Process

To generate self-reports of untouchability-related beliefs, this study relied on an original MP3-player-based “audio self-administered questionnaire.” In this methodology, respondents reacted to statements made by “villagers like [them] in earlier conversations with the research team” that they heard through the MP3-player's earphones. To note their reactions to those statements, they marked an answer sheet using simple shapes and logos,Footnote ²⁶ a portion of which is reproduced in Figure 1. The answer sheet had as many lines as there were questions in the audio survey, and each line presented respondents with various response choices represented by thumbs. Instructions contained in the instrument detailed what each thumb meant, and a “voice” provided instructions on how and where to react to each statement. Respondents ticked one of the shapes on each line; if they did not know what to answer or refused to respond, they did not tick anything and moved on to the next line. To enable illiterate respondents to identify the line associated with each statement, a logo (in Figure 1, a chair, a tea glass, or a hanger) was mentioned after each statement. Interviewers played a minimal role: After training respondents, they simply pushed the “play” button and sat aside until the audio survey ended. After their return, if respondents requested that a question to be repeated, interviewers used the “skip forward” function. After they completed the audio survey, respondents folded their answer sheets and placed them in a locked box.

FIGURE 1.

The Answer Sheet

This methodology, whose design and implementation are detailed in Online Appendices E and F, was chosen over several alternatives. It was chosen over audio computer interviewing (ACASI), which would likely have attracted unwanted attention and limited our ability to convince older villagers to participate, as determined in a small pretest. This methodology was also chosen over face-to-face interviewing and techniques known to reduce misreporting such as list, survey, and endorsement experiments. The use of recordings had several advantages. First it ensured a perfectly uniform delivery of the survey. It also provided respondents with culturally grounded questions, presented in their local dialect through the reflections of “fellow villagers.” Third, whereas only a limited number of survey or list experiments can be included in an instrument, this methodology allowed me to generate individual-level responses to an extensive series of items, and thus to disentangle various psychological mechanisms. Last but not least, as with other self-administered survey methodologies (Harmon et al. Reference Harmon, Turner, Rogers, Eggleston and Li2009; Tourangeau and Smith Reference Tourangeau and Smith1996), the isolation and the privacy implied by this methodology significantly reduced misreporting.Footnote ²⁷