ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Political Reservations and Service Delivery in Village Government

The impact of reservations for the Scheduled Castes as village heads (pradhans) is examined on school and governance outcomes. In general, SC reservation is not associated with improved outcomes. The impact differs spatially, based on past history of landlord or non-landlord control under British colonial rule. In historically landlord areas with greater presence of dominant classes and already worse school quality and governance, reservations are associated with no change in outcomes. In the non-landlord areas, reservations are associated with worse outcomes. The findings can be attributed to negative perceptions, discrimination and domination faced by the SCs. For effective formal policy empowerment of the SCs, the attitudes, beliefs and perceptions that dictate the informal rules of individual and social group interactions need to be addressed.

The authors would like to thank the Asha Trust and the team of college students that participated in the data collection for this paper. They thank seminar participants and colleagues for helpful comments on an earlier version of this paper. The findings, interpretations and conclusions expressed here are those of the authors alone, and do not represent the views of the World Bank, its executive directors, or the countries they represent.

This paper looks at the impact of political affirmative action in the form of reservations for the Scheduled Castes (SCs) in the office of the gram pradhan or head of the local village government. We examine the impact of the above on local governance and school outcomes in Uttar Pradesh (UP). We use primary data from a survey of public primary schools in 130 gram panchayats in UP.1 In unannounced visits to schools, information was gathered on teacher effort, student performance, fee charged, and scholarship reaching students.

The SCs have historically been at the bottom of India’s caste hierarchy. After independence, although an end was put to the legal recognition of the rules of caste, the social behaviour of the high castes is still largely governed by the norms of the caste system (Thorat 2002). The 73rd amendment to the Indian Constitution in 1993 made it mandatory for states to reserve a fraction of the seats for the village council head for individuals from historically disadvantaged groups, namely the SCs, Scheduled Tribes (STs), and women. Simultaneously, the devolution of power to village governments also began.

In UP, the reservation in pradhan seats for the SC candidates is linked to the proportion of the SC population in the gram panchayat’s total population. This paper uses the non-linearity in the reservation rule and census population data to identify the impact of reservation on school and governance outcomes. Our data allow us to look at the quality of public service, unlike most empirical studies on governance that look at the physical presence of public goods.

What, then, can we expect from the political affirmative action in village governments for the SCs? The findings suggest that such affirmative action for a stigmatised group might not help reduce its social exclusion and can actually lower the quality of public services that the group depends on. Affirmative action is necessary, but it might not be sufficient to improve the outcomes in public services that the disadvantaged group depends on.

This paper’s evaluation of the impact of political affirmative action for the SCs focuses on the impact on public schools because most SC families depend on public schools, whereas most high caste families do not. The survey for this study took place in 2002–04, in the second election term after the SC affirmative action came into effect in 1995. Following the 73rd amendment, the oversight and control of village schools began to be devolved to village governments. However, it is well-recognised that high teacher absenteeism is common in many states, particularly in north India. Even when teachers are present in school, there is little active teaching resulting in dysfunctional schools.2 Although local governments are responsible for overseeing the provision of public services, in reality, village governance is weak and oversight does not occur in states such as UP (Pandey et al 2010).

Our findings show that by almost all outcomes, SC reservation is associated with a lower quality of public schools and worse local governance. Our explanation is twofold: (i) SC pradhans face intimidation and discrimination from dominant interest groups that results in worse governance, and (ii) teachers, who mostly belong to the higher castes, work less(er) when the pradhan is SC. Even though the formal rules of policy have changed in favour of the disadvantaged social groups, our findings suggest that the informal norms of the caste system do not permit these groups to benefit from such change, at least in the short run.

In contrast, Rohini Pande (2003) finds that mandated representation for the SCs and STs in state legislatures in India improved targeted transfers to these groups. It is important to note, however, that caste dynamics are likely to be different and stronger in the local democracy of a village, where all individuals living in a gram panchayat likely know each other and interact daily, compared to state-level democracy. Another study (Chattopadhyay and Duflo 2004), using the fact that gram panchayats were randomly selected for reservation, analysed the impact of political reservation for women on investments in local public goods in the gram panchayats of West Bengal. It found that women leaders invested more in public goods that were directly relevant to the needs of rural women, drinking water, fuel, and roads. The results from the West Bengal study are not directly comparable to the present study because the former analyses the impact of affirmative action for women while this study looks at the impact of such action for the SCs. The extent of empowerment and inclusion of the two groups before and after a change in policy may differ depending on the informal norms and perceptions among other factors. Furthermore, the two studies look at different public good outcomes that are not comparable.

Other studies that examine the impact of caste echo our results. In a study of higher education in UP, the SC students reported facing negative and hostile attitudes from fellow students and teachers (Pandey and Pandey 2018). Most high castes as well as the SC students believed that students who belong to the reserved caste category (affirmative action) had inferior abilities and those belonging to the general (or high) castes had superior abilities. In rural UP, Hoff and Pandey (2014) found no caste gap in the performance of school students where caste identity was anonymous. When caste identity was made salient, the performance of SC students dropped and a significant caste gap emerged in favour of high-caste students.

Policies need to recognise the prevalent negative attitudes, perceptions and stereotypes of disadvantaged groups, such as the SCs, and how these hold back individuals and groups. To see that a disadvantaged group benefits from affirmative action, policies can go further and attempt to shift mental perceptions pervasive in society that dictate the informal rules of how individuals and social groups interact. Mental models, attitudes, beliefs, and perceptions influence individuals and group behaviour (Hoff 2015, 2016), and these are likely to affect the impact of a formal rule empowering individuals from weaker groups.

Data Description

Data was collected in a primary survey of 130 gram panchayats across 26 districts in UP. The districts were chosen in the following way. One set of districts is situated along both sides of the length of the border that separated the land revenue system of the landlord-based Oudh (or Awadh) Province (landlord-based) from the non-landlord-based North West Province (NWP) (non-landlord-based NWP) under the British rule. A second set of districts is situated in the NWP, on each side of the border separating the land revenue systems that were either landlord-based (landlord-based NWP) or non-landlord-based (non-landlord-based NWP) (Figure 1).3

One block was randomly chosen in each district, and five gram panchayats per block were randomly selected. A block is an administrative unit between a district and gram panchayat. In each gram panchayat, one randomly selected primary public (government) school was studied August 2002 to February 2003. In each school, a sample of 10 children from Class 4 was randomly chosen. All teachers providing primary-grade instruction were in the sample.

Researchers made seven unannounced visits to schools during this period. At each visit, researchers noted the presence or absence of a teacher and their activity upon the team’s arrival. Two measures of teacher effort—activity and attendance—were constructed for a teacher, averaged over the visits. Attendance is the fraction of survey visits during which the teacher was present in the school. Activity is the fraction of survey visits during which the teacher was engaged in teaching activity, defined broadly to include teaching, writing on the board, supervising written work, and keeping order in the classroom. Non-teaching activities include sleeping in school, talking with someone outside the classroom, doing non-teaching tasks, and being absent. Sample children were tested on basic competencies in language (Hindi) and mathematics based on the beginning pages of Class 4 textbooks. Language competencies included vocabulary and sentence writing, and mathematics competencies comprised addition, subtraction, multiplication, and division. Test score is the sum of score across all items for a student. School characteristics and family background data from students and teachers were collected. Information was collected on the gram panchayat’s local government: pradhan characteristics, whether the Pradhan seat was reserved for SC, whether the meetings of the gram panchayat council and it’s village education committee (VEC) and health committee occurred.

One year later, as part of a separate study limited to the first set of districts (those bordering landlord-based Oudh province and non-landlord-based NWP), information was gathered on the fee charged by teachers and the scholarship received by students in 2003–04 for all children attending a public primary school in sample households. This survey included a random sample of 10 households in every gram panchayat with at least one child going to a public school. The mandated fee was a uniform and minimal amount in the state.4 Every SC student in primary school was entitled to an annual scholarship of `300 at the time. Excess fee variable is taken as the difference between the fees, charged to a student minus the mandated fee, while scholarship variable is the actual amount received by a student.

Table 1 presents the summary of variables. On average, 21% of teachers were absent during a survey visit and only about 60% were engaged in teaching activity. The average student test score on the literacy and numeracy test was 38% of the maximum score. The amount of scholarship reaching the SC students was `220, 73% of the mandated amount. Students paid `29 in excess fee charged by teachers, at least two and half times more compared to the mandated fee, and roughly half the daily wage of unskilled labour at the time.

SC Reservation and Expected Outcomes

Creating political reservation is justified because the SCs are under-represented in various levels of government otherwise. Low SC participation in the absence of reservation can stem from the intimidation and domination of the upper castes, discrimination, and higher opportunity costs of time. In a study of SC pradhans in UP, 71% reported standing for elections only because the seat was reserved for SC (Sahbhagi Shikshan Kendra 2001).

Following the 73rd amendment, the process of devolution of control over public services to the village government also began. In UP, the VEC is headed by the pradhan. Other members of the VEC include the head teacher of the village public primary school and the parents of children studying in the school. The pradhan and the head teacher jointly operate the school account, which receives the scholarship money for the SC students and other school funds. The VEC hires contract teachers in public schools and decides whether to renew their annual contract. It can inspect schools, complain and recommend action, if dissatisfied with the teachers or schools.

Reservations will affect policy outcomes if SC and non-SC candidates have different policy preferences over public goods. In the sample gram panchayats, the SC children are over-represented and upper-caste children under-represented in public school enrolment (42% SC, 15% upper caste) relative to their population shares (24% SC, 27% upper caste). Thus, the SCs have a higher stake in the quality of public schools compared to the upper castes.5 Other factors determining the effect of reservations include the extent of capture by dominant groups and the ability of the SC pradhan to implement policies. When local governments are subject to capture, the reservations for non-elite candidates enable voters to counter the lobbying by elites. But if reservation worsens the pool of candidates or candidates face intimidation and discrimination from dominant groups, outcomes can worsen. A priori, the net effect of political reservations is ambiguous.

Within a block, the proportion of pradhan seats reserved for the SCs equals the proportion of the SC population (say x fraction) in the total block population. All gram panchayats in a block are listed in the descending order of their proportions of SC population. Every five years, the length of an election term, x fraction of gram panchayats is chosen from this list going from top to bottom. So reservation rotates every five years. In the first term, the first x fraction of gram panchayats is chosen, and in the second term, the next x fraction is chosen and so on. During the survey, gram panchayats were in their second election term following the 73rd amendment. The population figures used to implement the reservation policy were based on the 1991 decennial census estimates, which were the latest estimates available.

We used the gram panchayat-level census population and reservation data to verify whether the reservation policy was followed in practice in the surveyed blocks in the sample districts. The rule used to determine the SC reservation was followed in the first election in 1995. Figure 2 illustrates the rule in four of the surveyed districts. In the second election in 2000, however, the rotation of gram panchayats did not take place due to political reasons at the state level.Reservation from the first election term, gram panchayats with SC reservation and those without SC reservation, continued as such in the second term.7

Identification: In identifying the causal impact of reservations on public service outcomes, the issue is that reservations are a function of an observed covariate, namely the fraction of the SC population in a gram panchayat, potentially related to service delivery. In models of democracy, the population proportions of different interest groups affects public goods allocation (Bardhan and Mookherjee 2002).The reservation variable, whether the pradhan’s seat is reserved for the SCs, is a dummy variable and, as Figure 2 illustrates, is a discontinuous function of the fraction of SC population in the gram panchayat. This feature of the reservation rule is used to identify the impact of reservations.

Identification strategy one: In ordinary least squares (OLS) regression, we control for any smooth effects of the variable related to the SC reservation, namely linear and quadratic controls for the fraction of the SC population in 1991.We also include other controls for gram panchayat population, such as the linear and quadratic terms of population size, density and fraction of the Other Backward Class (OBC) population. The strict fixed effects are included since sample gram panchayats in a district belong to the same block, and reservation is made within a block. Since discontinuity in the relationship between the fraction of the SC population and the reservation variable is behind the identification, the sample is restricted to gram panchayats with a fraction of the SC population in a range close to the point of discontinuity. The sample is a +.15/-.15 discontinuity sample around the point of discontinuity in terms of the fraction of the SC population.10 We also use a +.10/-.10 discontinuity sample.11

Identification strategy two: Using the same controls and discontinuity samples as in OLS, instrumental variable (IV) regressions are run. The discontinuity, that is, the reservation cut-off points are used to create instruments for the reservation variable. Specifically, the instrument is the gram panchayat’s distance in terms of the fraction of SC population from discontinuity point where the distance variable is switched on for the block, that is, the distance of the gram panchayat from discontinuity point is interacted with a dummy variable for the block.

Reservation and Associated Outcomes

Teacher effort: In linear regression of (1), two measures of teacher effort—attendance and activity—are used as dependent variables for teacher i in gram panchayat and district k. The variable reserved for SC is a dummy variable with value 1 if gram panchayat j in district k has pradhan’s seat reserved for SC, 0 otherwise. Zijk consists of a set of district dummies, enrolment size in school in gram panchayat j, controls for ith teacher caste, gender, education, distance of teacher’s house from school, controls for jth gram panchayat whether the gram panchayat has electricity, phone, distance from nearest pucca road, literacy rate, population size, and density and shares of different caste groups. The effect of SC reservation is allowed to differ in landlord and non-landlord areas. Given the extent of capture is different in the two areas as shown in Pandey (2010), it is likely that the effect of reservation for a disadvantaged group differs. The variable landlord is 1 if gram panchayat j in district k
belongs to a landlord district, 0 otherwise.

Effortijk= a + b. Zijk + c.reserved for SCjk+ d.reserved for SCjk*landlordk + uijk … (1)

The results from the OLS and IV regressions of (1) are in Tables 2A and 2B, respectively. Because they are similar, although standard errors are higher in IV as expected, the OLS estimation results are discussed. The SC reservation is associated with lower teacher effort, but this effect is signi­ficant only in non-landlord areas. In these areas, teacher attendance is 46% lower and activity is 53% lower in the SC-reserved gram panchayats compared to gram panchayats with no such reservation. Among the SC-reserved gram
panchayats, attendance and activity are about 42% and 47% higher, respectively, in landlord districts compared to non-landlord districts (Table 2A, columns 1–2). As a result, the net effect of reservation on teacher effort is insignificant in the landlord villages and negative and significant in the non-landlord villages.12 Results are robust in +.10/-.10 discontinuity sample. The coefficients of interest are larger in magnitude and stay significant (Table 2A, columns 3–4).


Are these results capturing the effect of unobserved differences across the SC-reserved and unreserved gram panchayats? It is possible that lower teacher effort in the SC-reserved gram panchayats reflects differences in the unobserved characteristics between reserved and unreserved gram panchayats. However, teachers and gram panchayats look similar on a number of observed characteristics (Table 5, p 53).

Fees charged and scholarships recieved: Regressions similar to (1) were run using fee charged and scholarship received
as dependent variables for student i in gram panchayat j and district k. In the fee regression (2), Zijk includes district dummies and controls for gram panchayat as in (1). SCijk is a caste dummy if student i in gram panchayat j and district k is SC. The scholarship regression is same as (2), except the sample is restricted to the SC students, so the student variables SCijk and SCijk*landlordk are dropped from the equation.

Excess Feejk= a + b.Zijk+ c. reserved for SCjk+ d. reserved for SCjk*landlordk+ e. SCijk+ f. SCijk.landlordk +uijk … (2)


The results from the OLS and IV estimations are similar and in Tables 3A and 3B, respectively. The results from the OLS regression are discussed. The SC-reserved gram panchayats have significantly higher (by 70%) fee charged compared to the gram panchayats with no SC reservation. Note that the fee charged in reserved gram panchayats in landlord areas is not significantly different from the fee charged in reserved gram panchayats in non-landlord areas (Table 3A, Panel A, colu­mns 1–2). The results for scholarship paid are similar but pronounced in +.10/-.10 sample (Table 3A, Panel B, columns 1–2). The net effect of reservation is significantly lower scholarship money (by 52%) received in SC-reserved gram panchayats compared to panchayats with no such reservation in this sample (Table 3A, Panel B, column 2).


Student test score: The regression specification is similar to (1) except that the dependent variable is the student test score for student i in gram panchayat and district k. The controls for teacher characteristics in Zijk in (1) are replaced by the student characteristics, dummy variables for caste, gender, and mother’s and father’s education.

The OLS and IV estimation results are in Tables 4A and 4B (p 54), respectively. In +.15/-.15 sample, the net effect of the SC reservation on test scores is negative and weakly significant (value is .07). In +.10/-.10 sample, the net effect of the SC reservation on the test score is negative and significant in non-landlord areas only where the SC reservation is associated with 37% lower score (Table 4A, columns 1–2). The latter result stays robust after controlling for physical quality of schools (Table 4A, columns 3–4).

Is caste a proxy for other characteristics? Because caste can be a proxy for the pradhans’ socio-economic characteristics, the impact of reservation may be reflecting the effect of such characteristics. The SC pradhans in the sample have lower levels of schooling and political experience, which can adversely affect their ability to govern and implement their desired policies. But controlling for the pradhans’ characteristics, schooling, previous experience and age, the results stay mostly unchanged (Table 2A, columns 9–10; Table 3A, Panels A–B, columns 3–4; Table 4A, columns 5–6).


Figures 3–5 (pp 54–55) present a comparison of the average outcomes in the SC reserved and non-reserved gram panchayats in landlord and non-landlord areas. The results in the preceding section show that these differences in the average outcomes between the SC reserved and non-reserved gram panchayats stay robust in regression estimations. For almost all teacher and student outcome variables, the net effect of reservation is either insignificant or small and negative in landlord areas and unambiguously negative in non-landlord areas. It is in the historically non-landlord areas that the SC-reserved gram panchayats are associated with worse school outcomes and greater corruption or leakage of school funds compared to the non-reserved gram panchayats.

Our explanation of the results is twofold: one, worse governance since the SC pradhan is likely to face intimidation, domination and discrimination from high caste groups, and two, that teachers who are mostly higher caste and class work less when the pradhan is SC. In areas with a history of landlord control, the extent of elite domination is larger and local governments and schools are already dysfunctional to a greater degree as shown in Pandey (2010). Although the SC reservation is not associated with worse school outcomes, neither is it able to reverse greater leakage of funds and poorer school outcomes in these areas. While school outcomes are not associated with the SC reservation in landlord areas, the outcomes are significantly and negatively associated with the SC reservation in non-landlord areas.

Worse governance: A village government is dysfunctional to a greater degree in the SC-reserved gram panchayats compared to those not reserved for the SCs, particularly in non-landlord areas (Table 6). The gram panchayat council was significantly less likely to have met in the SC-reserved gram panchayats in the non-landlord areas while there was no significant difference in the landlord areas.13 The education and health committees of the gram panchayat government were significantly less likely to be functional in these gram panchayats. Worse governance in the SC-reserved gram panchayats is possibly due to the intimidation and dependence on the higher castes. In Sahbhagi Shikshan Kendra’s (2001) report, 60% of the SC pradhans reported facing physical violence, threats and manipulation of votes during their election, and 50% reported that gram panchayat council members do not cooperate. Further, the lower castes with minimal landownership often depend on upper-caste landed households for employment on their fields (Lieten 1997; Lieten and Srivastava 1999).

Influential teachers and SC pradhans: Teacher absenteeism is facilitated by the presence of a significant nexus between the local village elites and teachers, who often share a common caste and class background. In the sample, 83% of the teachers were of castes higher than the SCs while just 17% were SC. On average, the teachers owned (2.44 acres) 35% more land compared to the SC pradhans (1.81 acres). Figures 3–4 compare the average teacher effort by teacher caste and landholding between the reserved and unreserved gram panchayats in the discontinuity sample. In landlord areas, there was no significant difference in effort between the SC reserved and non-SC reserved gram panchayats. In non-landlord areas, the high-caste teachers and teachers with above-median landholding worked significantly less(er) in the SC reserved gram panchayats compared to the non-SC reserved gram panchayats (p values based on Mann-Whitney test are below 0.06). Note that the SC teachers worked no less in the SC reserved gram panchayats in these areas. In fact, they worked harder in the SC reserved gram panchayats, but this difference between the SC reserved and non-reserved gram panchayats is insignificant likely due to the small sample size of SC teachers.


This paper finds that affirmative action in the form of political representation of low caste (SC) individuals as heads of village governments is generally not associated with better school or governance outcomes. Reservation is associated with either worse outcomes in areas that otherwise had relatively functional governance, or no change or small decline in outcomes in areas with already worse governance. Formal policy empowerment of a non-elite group, such as the SCs, is unlikely to benefit them at least in the short run, unless policy attempts to alter the informal rules at play in local power structures of class and caste.

The limitations of the present study are that it is the only study so far that attempts to measure the impact of the SC reservation in local government on public service outcomes, and it is only able to measure the association between reservation and outcomes using a discontinuity sample. Another limitation is that the data is from the second election cycle, whereas currently villages are in the fifth election cycle following the 73rd amendment. Further work is needed with a larger sample of villages, broader set of public services’ outcomes and recent data to further the understanding of the impact of SC reservation on local governance and public service outcomes.

Recent news reports of atrocities targeting SC individuals in India, particularly rural UP, highlight the extent to which the SCs lack empowerment and voice. The incidents indicate that speaking against an unlawful act often has huge social, economic and personal costs for SC individuals and groups. Likewise, it is very plausible that the SC pradhans face difficulty implementing their desired policy preferences. Political reservation is necessary but may not be sufficient to empower village leaders to govern without intimidation and fear. Policies need to be sensitive to the power dynamics at play bet­ween individuals and groups of different castes living locally. To benefit groups like the SCs from affirmative action, the negative perceptions and attitudes that dictate individual and social group interactions and behaviours need to be addressed. These come in the way of formal policy empowering weaker groups. Policymakers and researchers need to examine additional app­roaches, such as behavioural interventions, to make local political reservation effective.


1 A gram panchayat is an administrative unit between a block and a revenue village. It comprises of two to three revenue villages on average and the local village-level government is formed at the gram panchayat level. We use the terms village and gram panchayat interchangeably in the paper, although we always mean the latter.

2 Educational achievements are particularly low in north Indian states, and lower in rural areas and for disadvantaged groups like SCs.

3 The Oudh area and most of the NWP had a temporary revenue settlement, where the revenue was subject to revision every few years, except the easternmost part of the NWP, which was permanently settled. Thus, the entire sample of the first set of villages, whether landlord-based or otherwise, had a temporary settlement.

4 During the survey, for SC students, the annual fee was `1 in Classes 1–3, and `2 in Classes 4–5. For non-SC students, the annual fee was `1 in Class 1, `11 in Classes 2–3, and `12 in Classes 4–5.

5 Dreze and Saran (1995: 210) reported that when the SCs constitute the dominant group in the population, their educational achievements are comparable to those of the upper castes. Since the SCs have a lower literacy rate than the rest of the population, their returns from education are expected to be higher. In UP, at the time of this study, literacy rates for the SCs were 39% and 9% for males and females, respectively, compared to the corresponding state figures of 52% and 19%, respectively.

6 This is true for all sample districts, except two where rotation was followed as per the rule, again due to political reasons. The results reported in the following section continue to hold if we allow the impact of reservation to differ in these two districts. Nevertheless, note that the district fixed effects are included in regressions.

7 The rotation rule was followed in the third election term, which fell after the survey period.

8 In addition, population proportions can have a general equilibrium effect on public goods’ allocation by influencing the prices and returns from allocating different public goods.

9 This sort of identification is an application of the idea behind D T Campbell’s (1969) regression discontinuity design.

10 The discontinuity sample is chosen block by block, since reservations are applied within the block.

11 Since the reservation variable is a function of the fraction of the SC population with slope zero in the smooth parts (being a dummy variable), we do not perform a robustness check by including a piece-wise linear trend of the fraction of SC population since the reservation variable does not have a linear trend.

12 As a robustness check, we allow the effect of reservation to be different in the two landlord areas: Oudh and permanent settlement districts. Considering only reserved gram panchayats, teacher effort is significantly higher in the Oudh and permanent settlement districts than in the non-landlord districts. The net effect of reservation is significant and negative only in non-landlord villages.

13 The gram panchayat government operates via the gram panchayat council consisting of the pradhan and elected council members.


Bardhan, P and D Mookherjee (2002): “Relative Capture of Local and Central Governments: An Essay in the Political Economy of Decentralization,” Working Paper, University of California, Berkeley: Center for International and Development Economics Research, pp 1–44.

Campbell, D T (1969): “Reforms as Experiments,” American Psychologist, Vol 24, No 4, pp 409–29.

Chattopadhayay, R and E Duflo (2004): “Women as Policy Makers: Evidence from a Randomized Policy Experiment in India,” Econometrica, Vol 72, No 5, pp 1409–43.

Dreze, J and M Saran (1995): “Primary Education and Economic Development in China and India: Overview and Two Case Studies,” Choice, Welfare and Development, K Basu, P Pattanaik and K Suzumura (eds), Oxford: Clarendon Press, pp 182–241.

Hoff, K (2015): “Behavioral Economics and Social Exclusion. Can Interventions Overcome Prejudice?” Policy Research Working Paper, 7198, World Bank.

— (2016): “Behavioral Economics and Social Exclusion. Can Interventions Overcome Prejudice?” Inequality and Growth: Patterns and Policy, Volume 1: Concepts and Analysis, K Basu and J E Stiglitz (eds), New York: Palgrave Macmillan, pp 6a-2–6a-31.

Hoff, K and P Pandey (2014): “Making up People: The Effect of Identity on Preferences and Performance in Modernizing Society,” Journal of Development Economics, Vol 106, pp 118–31.

Lieten, G K (1997): “Panchayats in Western Uttar Pradesh: Namesake Members,” Economic & Political Weekly, Vol 31, No 39, pp 2700–05.

Lieten, G K and R Srivastava (1999): Unequal Partners: Power Relations, Devolution and Development in Uttar Pradesh, New Delhi: Sage.

Pande, R (2003): “Can Mandated Political Representation Increase Policy Influence for Disadvantaged Minorities? Theory and Evidence from India,” American Economic Review, Vol 93, No 4, pp 1132–51.

Pandey, P, S Goyal and V Sundararaman (2010): “Public Participation, Teacher Accountability and School Outcomes in Three States,” Economic & Political Weekly, Vol 45, No 24, pp 75–83.

Pandey, P (2010): “Service Delivery and Corruption in Public Services: How Does History Matter?” American Economic Journal: Applied Economics, Vol 2, No 3pp 190–204.

Pandey, P and S Pandey (2018): “Survey at an IIT Campus Shows How Caste Affects Students’ Perceptions,” Economic & Political Weekly, Vol 53, No 9,­ste-Affects-Students-Perceptions.

Sahbhagi Sikhshan Kendra (2001): “A Study of Dalit Leadership in Panchayats of Uttar Pradesh” (report), Sitapur Road, Lucknow.

Thorat, S (2002): “Oppression and Denial: Dalit Discrimination in the 1990s,” Economic & Political Weekly, Vol 37, No 6, pp 572–78.


Updated On : 11th Jul, 2022
Back to Top