Laura Camfield Uma Vennam

Security, productivity and social equity? A community-level exploration of tensions within the Mahatma Gandhi National Rural Employment Guarantee Scheme Laura Camfield Uma Vennam International Conference: Social Protection for Social Justice Institute of Development Studies, UK 13 15 April 2011 1

Laura Camfield, University of East Anglia l.camfield@uea.ac.uk Uma Vennam, Sri Padmavathi Mahila Visvavidyalayam Security, productivity and social equity? A community-level exploration of tensions within the Mahatma Gandhi National Rural Employment Guarantee Scheme Abstract The Mahatma Gandhi National Rural Employment Guarantee Scheme (NREGS) was rolled out across India in 2008. While in Andhra Pradesh NREGS has received predominantly positive evaluations (e.g. Reddy et al., 2010; Uppal, 2009), the paper argues that an evaluation from a wellbeing perspective which engaged with its impact on relational and subjective dimensions of participants wellbeing might look very different. The paper uses individual and group interview and survey data from adults and children aged 15-16 participating in Young Lives, a longitudinal study of children growing up in poverty. It finds that participation in NREGS is high and largely pro-poor; however, those who can only work in the scheme because they are landless have not benefitted as much as expected. The main beneficiaries in the three villages sampled for the qualitative research have been individual farmers, often from higher castes, and to some extent administering officials. So while interview and survey data demonstrate beneficial effects, the systemic mismanagement described in the paper is having a corrosive effect on trust and social relationships. This unintended consequence threatens the sustainability of the scheme and its potential to reduce socio-economic inequalities and vulnerability across the life course. 1. Introduction The paper briefly describes the Mahatma Gandhi National Rural Employment Guarantee Scheme (NREGS) and explores the extent to which it is meeting its stated objectives of enhancing livelihood security, generating productive assets, protecting the environment, empowering rural women, reducing rural-urban migration, enabling asset accumulation, and fostering social equity (Government of India in Porter, 2010). It focuses on the final 1

objective of fostering social equity and argues that from a wellbeing perspective NREGS s failure to realise this is an important problem. The importance of social equity is illustrated by evidence from other studies, predominantly from Latin America (e.g. Adato, 2000), which document the effects of social protection on social cohesion. The paper briefly describes how the data was collected and presents findings addressing social equity and social cohesion in relation to i) positive and negative impacts of NREGS, ii) the extent to which NREGS participants are aware of their entitlements and can hold officials to account, and iii) corrupt practices. It concludes by identifying questions for policymakers. NREGS is based on the Maharashtra Employment Guarantee Scheme promoted by Jean Dreze and Amartya Sen (1989). It departs from previous employment guarantee schemes by banning the use of 'labour displacing machinery' and local contractors who have historically exploited the rural poor. NREGS promises to provide good quality work with a minimum wage and good working conditions, including drinking water, healthcare and childcare if more than five children aged under five are on site (Ambasta et al., 2008). However, the promise of decent working conditions is not always fulfilled as according to Poorest Areas Civil Society Program (PACS, 2008) worksite facilities such as drinking water, shade, first aid, and a crèche were only provided in half of the surveyed panchayats. Reddy et al. (2010) confirm that in Andhra Pradesh only water was consistently provided, and Young Lives survey data (2009) suggests that childcare was available in less than a quarter of worksites, which meant that women either did not participate, took siblings out of school to care for babies at home or onsite, or left babies with preschool children (Sudarshan et al., 2010). NREGS acts as informal social insurance as once a rural household and its members have registered for a job card it guarantees up to 100 days unskilled manual labour for every adult available to work on that day. By guaranteeing work as an employer of last resort the government is effectively providing a minimum wage. The NREGS rate was intentionally set low to avoid early capture of the scheme by the non-poor, which has been the case with other anti-poverty initiatives. However, in Andhra Pradesh the rate is 2

considerably higher than local wages (Jha et al., 2009). Work must be provided within 5km of the household at a rate on or above the minimum wage of Rs 100 per day (Rs 121 from January 2011), although the amount paid varies according to the amount of work done. Types of work include breaking/ moving stones, levelling the ground, checking and repairing dams and irrigation channels, desilting tanks, digging pits around saplings and providing piped water. Work is typically provided from 7am to 12 or 1pm, which is less time than it would take to earn a similar amount as a daily labourer. If work is not made available within 15 days the household is entitled to unemployment benefit, although this is rarely offered (in 2009 less than five per cent of Young Lives sample reported receiving it). NREGS pays the same rate for everyone, irrespective of age, caste, or gender (Johnson and Tannirkulam, 2009). When NREGS started there were widespread predictions that corruption and poor service delivery mechanisms would reduce its impact to such an extent that one of India s most well known economists infamously suggested that the likelihood of money reaching the poor would be higher if we simply drop the money by helicopter or gas balloon into rural areas than route it through employment programs' (Aiyar and Samji, 2009: 5). NREGS aimed to combat this perception with decentralized planning and implementation (at least fifty per cent of works were organised by the Gram Panchayat), proactive disclosure (e.g. display of info on Gram Panchayat notice boards and online), leakproof wage payment system (weekly payment of wages through individual accounts in post offices) and bi-annual social audits. Aiyar and Samji (2009) report that as a result of the social audits over 500 field assistants and ten technical assistants have been dismissed and Rs 60 Lakh worth of embezzled funds has been returned. The audit also raises awareness of workers entitlements such as the 100 day guarantee and provision of shade and water at the worksite. In Andhra Pradesh social audits are regularly conducted and Young Lives data survey (2009) suggests nearly a quarter of villagers participate. However, some studies suggest that they are conducted in a way that doesn t enable genuine participation and by concentrating on implementation issues cannot address more important concerns such as the exclusion of certain social groups or 3

the quality of the work (Gopal, 2009; PACS, 2008). Concerns relating to the quality of the work are partly due to a delay in setting up Technical Support Groups (panels of engineers) to cost and evaluate works and a shortage of junior engineers (PACS, 2008). The lack of engineers led to delays in wage payments as there was no-one to 'sign-off' on work done and meant that work was often poor quality (Ambasta et al. 2008; Mehotra 2008). Lack of staff limited attempts to involve villagers ( social mobilisation ) and tailor programmes of work to local needs - public works were selected based on what had been done before and could be easily measured. Andhra Pradesh Many studies, e.g. Centre for Budget and Governance Accountability (CBGA, 2006, in Dreze, 2006), have placed Andhra Pradesh at the forefront in implementing NREGS due to the high level of population awareness, relative to other states such as Jharkhand, and technological innovations. Reddy et al. (2010) concur with this assessment, but warn that despite the commitment of the political leadership the local Panchayati Raj Institutions are weak which reduces their ability to mobilise villagers in planning and monitoring the work (a core function identified by Moore and Jadhav, 2006). According to Reddy et al. (2010) the indicators for the scheme in March 2010 were all positive: 4.1 million works had been taken up in Andhra Pradesh and 45 per cent of them completed. One hundred per cent of households have job cards, 71 per cent requested work and more than ninety per cent received work within fifteen days. The average number of days worked per household was seventy four, and more than half of households claimed between seventy five and one hundred days. Workers were from the most marginal social and economic groups: 45 per cent were Scheduled Caste, 5 per cent Scheduled Tribe and 51 per cent female (this is confirmed by Young Lives 2009 data where over 80 per cent of Scheduled Caste or Scheduled Tribe households participate and 77 per cent of these have worked in the past twelve months). Reddy et al. (2010) estimate NREGS represents a 9.6% contribution to household income and observe a rise in local wages and reduction in ageand gender-based differentials, with some landlords paying wages in advance to ensure a supply of labour. 4

Despite NREGS strong performance in Andhra Pradesh, its success in meeting its objectives has been mixed. While overall livelihood security has increased (Kareemulla et al., 2009), certain groups have been excluded or excluded themselves, for example, single women, elderly, physically impaired, female-headed households (Holmes et al., 2011; Porter, 2010; Sainath, 2007; Sudarshan, 2010; Young Lives qualitative data, Poompuhar). Although many workers enjoy the experience of working in groups, the fact that these groups are often caste or gender-based creates potentially discriminatory dynamics, for example, risky works being given to Scheduled Caste groups. Non-poor labourers may have benefitted more as they can better withstand delays in payment and variability in wages associated with piece-work (the variability is exacerbated by the use of a Schedule of Rates designed for a system of contractors, who do not pay minimum wages, and machines, Ambasta et al., 2008; Imai, 2007; Mehotra, 2008; Scandizzo et al., 2009). Non-poor households are also more likely to have information about the scheme, acquired through social networks such as relatives or self-help groups, mobile phones, public meetings, and television, and are more able to successfully argue for their rights within the scheme (Shankar et al. 2010; Uppal 2009). The generation of productive assets, which is part of the government s commitment to make the scheme self-sustaining (Mehotra, 2008), has been relatively successful. However, benefits have mostly accrued to individual farmers (Holmes et al., 2011; Imai, 2007), especially in Andhra Pradesh, where 45 per cent of NREGS work is land development (Reddy et al., 2010). Other NREGS works focus on protecting the environment, for example, by increasing the water table. However, the benefits from these are not yet visible and there is concern that the benefits may be reduced by lack of expertise, inputs and technical supervision (Imai, 2007; PACS, 2008). For example, some of the water conservation projects in Andhra Pradesh were unsuccessful as communities had little experience of water conservation, there was no budget for their maintenance, they were difficult to build within the limit of 40 per cent for material costs, and they needed to be combined with soil conservation measures, which did not always happen (PACS, 2008). 5

In relation to women s empowerment, while the scheme can claim to have increased the bargaining power of rural women relative to landlords the Sarpanch in Patna describes how for weeding [the landlords] offered Rs 100 to women. [...] The women say that they are getting Rs 100 in government work even [though] they don t do anything. They ask how much can you give us, if we come to your work? women have been less successful in arguing with NREGS officials for entitlements such as childcare and challenging incorrect payments (see later section on holding officials to account). The scheme has also reduced distress migration, which has apparently increased the likelihood of rural children in drought-affected communities remaining in school. However, male migration for higher wages continues, partly due to a perception of NREGS work as sporadic and of short duration. This partly arises from a history of NREGS projects being delayed while awaiting approval or investigation of an irregularity, or postponed to accommodate peak agricultural periods after requests from landowners. Sjoblom and Farrington (2008: 4) argue that the success of NREGS in reducing rural-urban migration may be in tension with NREGS human capital objective as easy availability of work under NREGA may [...] reduce the potential pace of economic transformation, and lower the prospects of workers gaining new skills. In relation to asset accumulation, while there is little evidence of saving, Young Lives qualitative data shows increased household food consumption and accumulation of consumer goods. According to the Meti in Patna, tribal people now have money and like to eat whatever they want. If they have a desire to eat meat [they] readily go for that. Nonetheless, the survey data in table 1 suggests the amounts earned are relatively small, although obviously of great benefit to poor households. This may explain why the girls focus group in Patna described the scheme as providing small helps only. 6

Table 1: NREGS income by ethnicity/caste (Young Lives, younger cohort household survey, 2009) Child's ethnic group n household Median income from NREGs (household, Rs) N individual Median no: of days worked (individual) Median income from NREGs (individual, Rs) Scheduled 224 3160 463 15.2 1570 Caste Scheduled 207 3650 396 19.4 1750 Tribe Backward 432 3897 961 14.8 2219 Caste Others 81 3507 193 8.8. 1569 Total 944 3521 2014 15.2 1848 The final, and in our view, most important objective of fostering social equity is the one that presents most challenges. While the high participation of Scheduled Tribes, Scheduled Castes and women indicates equality of opportunity, it doesn t signal equality of outcome. In fact Young Lives and other data suggests that the majority of benefits are accruing to farmers (Jha et al., 2009), workers with connections (Uppal, 2009), information (Shankar et al., 2009), alternative income (Imai, 2007; Scandizzo, 2009), larger households, NREGS administrators and their relatives, and members of workgroups who share profits from dubious practices. Before presenting Young Lives data on the impact of NREGS on social equity and social cohesion, we discuss why social equity matters, drawing on the wellbeing approach set forward in McGregor et al. (2007), Deneulin and McGregor (2009) and White (2010). Building on McGregor et al. (2007), White (2010: 162) defines wellbeing as the interaction of the following dimensions: i) material (assets, welfare, and standards of living), ii) relational, encompassing both the social (social relations and access to public goods) and the human (capabilities, attitudes to life, and personal relationships), and iii) subjective (people s perceptions of their (material, social, and human) positions; cultural 7

values, ideologies, and beliefs). She places the subjective as the apex of the triangle (fig. 1) for three reasons: to illustrate the interdependence of the three dimensions, to show how wellbeing emerges from the interplay of subjective and objective, and to ensure that subjective and objective are not separated as this would obscure the cultural grounding of material welfare or standards of living (ibid). Fig. 1: Dimensions of wellbeing (White 2010: 162) White argues that while the novelty of a wellbeing approach lies in its inclusion of the subjective, grounding the human and social in the relational also gives them a distinctive character (ibid). This distinctive character comes from reducing the tendency of subjective approaches towards individualism or a naively positive understanding of social relations by acknowledging the importance of social structure and power relations, even at the household level. The potential tension between the subjective, the relational, and wider social structures is explored by Deneulin and McGregor (2009) who highlight the challenge of choosing, building and living the most adequate social arrangements that will make good living together possible (p18) and the related need for conceptual frameworks and institutional arrangements for negotiating socially coherent wellbeing outcomes and strategies (p19). These approaches provide an alternative basis for 8

evaluating NREGS, namely does it produce good subjective and relational outcomes as well as material ones, and can it be seen as an institutional arrangement that enables socially coherent and cohesive strategies to achieve wellbeing. An example of this is McGregor (2010: 19) use of a three-dimensional wellbeing framework to evaluate community-responsive social protection in Zambia in relation to its effects on women and children. He argues that social protection is social because it is about society collectively protecting the weaker and more vulnerable members of society from harm [... and] maintain[ing] the social fabric of societies. This involves acknowledging that the wellbeing of societies (and economies) contributes to the well-being of individuals and vice versa (ibid). McGregor s (2010) proposition that researchers and practitioners engage with the effects of social protection on community cohesion is an important insight. However, this has previously only been studied in relation to the effects of conditionalities (Adato, 2000, 2007; Streuli, 2010) and targeting (Ellis, 2008; Slater and Farrington, 2009). For example, Ellis makes the point that the sentiment we are all poor here accurately reflects the very small differences in personal and family circumstances separating everyone falling within the bottom 50-60 per cent of per capita consumption (Ellis, 2008:i). As NREGS is a self-targeted scheme that guarantees access to work under Indian law, this literature might seem irrelevant. However, we have attempted to show that participants in NREGS also experience subtle exclusions (for example, of single women or women with young children) and unequal division of benefits, and the way the scheme is managed (discussed in the latter part of the paper) has damaging effects on social relationships. Adato s (2000) evaluation of PROGRESA in Mexico notes the benefits from increased resource flows and improvements in health and education services, including access to information. However, she argues that the creation of new social categories of beneficiary and non-beneficiary that do not map to local understandings of who is poor has affected community social relationships and solidarity which in turn affects the ability of households to translate the economic benefits of PROGRESA into movements out of 9

poverty. The damage this caused was shown by "social tensions, involving resentment, envy or gossip" (ibid: 19), typically when beneficiaries went to collect their benefits, envious exchanges between children of beneficiaries and non-beneficiaries, and less frequently by direct confrontations. More worryingly, non-beneficiaries often refused to work with beneficiaries in communal activities such as cleaning the community or school, saying that they were not being paid or recognised as the beneficiaries were. A promotora from Michoacán2 explained how [non-beneficiaries] feel sad, because they don't have that program, and they are not with the others who are sweeping in groups, all of them united. Because it is very pretty to be all united, all sweeping together, and even having fun, because sometimes we are all full of dust (ibid: 15). While PROGRESA has strengthened social relationships between beneficiary women, potentially building new forms of social capital, the creation of a group of PROGRESA women who participate in separate activities has also reinforced social divisions and may therefore undermine PROGRESA s goal of strengthening social capital at the community level. Adato s (2007) evaluation of Red de Protección Social in Nicaragua found similarly that few people understood the basis for household targeting or perceived the economic differences targeting focused on. This created considerable resentment which nonbeneficiaries expressed through concern over the impact on their children: one day my son told me that a boy (he didn t say his name) told him look, I have a new back pack and you don t, and he started showing him all the new things he had in his back pack. (ibid: 18). The strength of feeling around this particular issue was demonstrated by two of the six communities having a collection in which beneficiary families were asked to contribute some funds for the purchase of school supplies for non-beneficiary families. Finally, Streuli s (2010) qualitative study of Juntos in rural Peru identified similar problems due to targeting errors and lack of understanding of the criteria. She noted that the greatest tensions between beneficiaries and non-beneficiaries occurred in the community most affected by political violence and displacement where relationships between the State and the citizens were still fragile. In relation to Juntos she argues that the adverse effects on community relationships are most damaging for children as it is 10

through these social worlds that children feel and experience well-being" (ibid: 3). However, Streuli (2010: 14) makes a more general point in relation to social protection which is that it is "more concerned providing economic support against economic risks and shocks than with other social risks and non-economic vulnerability, such as social exclusion, discrimination, and violations of human rights [...even though] poverty is about social exclusion and deprivation as well as economic hardship". In the following section we briefly summarise how the data from Young Lives used in this paper was collected before looking at the extent to which these findings from Latin America, which in the case of Streuli (2010) were collected directly from children, reflect a fundamental problem with social protection schemes lack of sensitivity to social relations, exemplified by NREGS. 2. Methodology The survey data used in this paper comes from the second and third rounds of Young Lives data collection in 2006 and 2009. Qualitative data was collected in March 2010 through group and individual interviews/ activities with children aged fifteen to sixteen participating in Young Lives and their peers and key informants such as the NREGS Field Assistant, Meti and the Sarpanch. The fieldwork took place in three rural communities which have been part of Young Lives data collection since 2002 and were used in longitudinal qualitative research in 2007, 2008 and autumn 2010. All names of communities and respondents have been replaced by pseudonyms. Katur Katur is a near-rural village in Anantapur district, forty km from one of the state s major railway junctions and accessible by an all-weather road. While the majority of the population is Scheduled or Backward Caste, there are five Forward Caste households who are the main landowners. The main occupations are agriculture, animal husbandry and wage labour. The area is drought-prone and dependent on migration and now NREGS during the dry season. 11

Poompuhar Poompuhar is a near-rural village in Mahabubnagar district which is accessible by an allweather road and public and private transport. The majority of the population is Backward Caste, although there is a small Scheduled Caste colony at the entrance to the village. The main occupations are agriculture, predominantly cotton, animal husbandry and wage labour. Seasonal migration is common. Patna Patna is a remote tribal village in Srikakulam district without an all-weather access road. While the majority of the population is Scheduled Tribes, the community is dominated by a small number of Forward and Backward Caste households. The main occupations are agriculture, horticulture (e.g. growing mangos and cashews), and collecting non-timber forest produce. Additional employment opportunities are provided by the Integrated Tribal Development Agency. The community experiences frequent cyclones and roads can be cut off by water flowing from the hills. Table 2: Respondents who provided qualitative data used in this paper Patna Katur Poompuhar Group interviews with children 1 (girls), total n=5 2 (girls & boys), total n=10 2 (girls & boys), total n=10 Individual 4 girls, 1 boy 4 girls, 4 boys 3 girls interviews with children Key informant interviews Meti Sarpanch Panchayat Secretary Assistant Project Director Field Assistant Sarpanch Technical Assistant Meti Sarpanch 12

3. Social equity and social cohesion: evidence from Young Lives The first section of the paper presented the background to NREGS in Andhra Pradesh and Young Lives communities. This section looks specifically at social equity and social cohesion in relation to the positive and negative impacts of NREGS, the extent to which participants are aware of their entitlements and can hold officials to account, and corrupt practices. One of the positive impacts identified by officials and participants is increased solidarity between members of workgroups. Triveni, a girl from Katur, describes how the way NREGS work is structured generates a feeling of solidarity as [in other work] support is not given if someone falls behind as the one who supported may fall behind [...] Here all of them work equally and go back home at the same time. The the Assistant Project Director in Patna explains that for this reason the work has helped to grow collectiveness [...] if they have a problem they all respond. For some issues [due to work in NREGS] the money is ready; so if they have a problem in a financial or social aspect they respond very well (this was also the case with beneficiaries in PROGRESA, although the support did not extend to non-beneficiaries, Adato, 2000). But benefits for some participants exist alongside forms of exclusion and adverse incorporation for others. For example, due to the low percentage of worksites with childcare (less than 25 per cent, Young Lives 2009), 25 per cent of women with children under 5 said they were unable to participate due to lack of childcare. 3 per cent of rural households reported that single female household members were refused employment and this was also observed by fieldworkers in Poompuhar. Key informants suggested that delays in payment and uncertainties over the amount of payment meant that few femaleheaded households participated, although they continued to be overrepresented in other forms of daily labour. Children reported pressure from parents to attend on the first day of the week (for example, Mamatha from Katur) or if their parents became sick during a work period as otherwise they would not be paid. 13

The work is physically demanding, even though women and children are usually allocated lighter jobs: when they remove mud they don't check whether there are any children or not; for children and adults they put equally. Then it will be very heavy, madam. Some of it used to fall down from the basket [...] they say "can't you at least take the basket well?" Then we have to carry it well and put it [down] and come [back] (Triveni, Katur) For this reason NREGS work is hard to combine with schooling, even though the metis are relatively sensitive to the needs of students (for example, letting them go early in Katur to avoid them missing college). As according to NREGS regulations children are not supposed to work, even though this was the case in all three communities, they have little protection from exploitation, for example, being charged Rs 500 for a job card in Katur. There are some restrictions on elderly people participating, for example, people over 50 are excluded in Patna because "they are aged people [who] can't do work (Meti). However, these are selectively applied, so in Katur the relatives of the meti are included, even though they are older than others who were rejected and are so infirm that they don t even get up in the field (Triveni). Patna also limits participation to people who are "physically sound, an observation that may explain Porter s (2010) finding that the highest coverage of NREGS within the Young Lives sample was in the second poorest expenditure quintile rather than the poorest. However, it provides some employment to people with disabilities as water carriers. As described earlier, one of the main concerns for poorer participants is the difficulty of predicting how much will be earned due to piecework. For example, in Poompuhar Vasudha s father earned Rs 500-600 for one fifteen day period and Rs 300 for another. The reason for this is that errors in measurement and marking whether people are present or absent are compounded by the use of an inappropriate Schedule of Rates which makes it almost impossible for participants to earn the minimum daily wage of Rs 121. Delays 14

in payment also reduce the appeal to the poor as after working, they have to buy household needs, but they are paying once in fifteen days or even one month also" which means that they need to take loans and are getting "cheated again from outside (Sarpanch, Patna). Despite the scheme s commitment to providing good working conditions, in some communities [the metis] do not provide any facilities like water or for care of children. Nor do these people ask (brother of the Sarpanch, Poompuhar). In Poompuhar workers are also responsible for their own medical treatment, even in one case after an injury with a crow-bar which was clearly work-related. Lack of awareness of entitlements extends to participation in the scheme itself as 70% of those without a job card in Patna, a remote tribal community, believed that they were not eligible or did not know how to apply (Young Lives data, 2009). While participation in social audit is relatively high (table 3), particularly among NREGS target groups of Scheduled Caste and Scheduled Tribe, there is some doubt as to whether these formal mechanisms can address either day-to-day or more profound problems with the scheme (Gopal, 2009; PACS, 2008). There is great variation in awareness and participation in social audit between communities: compare Katur where 75 per cent were aware of social audit being conducted and 63 per cent participated and Poompuhar where 20 per cent were aware of social audit being conducted and 8 per cent participated. The example of Patna where 75 per cent were aware of social audit being conducted, but only 23 per cent participated suggests that knowing about these mechanisms and being able to use them are different things. Table 3: Awareness and participation in social audit by caste group, Young Lives 2009 Child's ethnic or caste group % aware of social audit being conducted in their community % participated in social audit in their community Scheduled Caste 53.2 31.9 Scheduled Tribe 57.2 22.4 Backward Caste 42.1 20.5 Other 40 21.8 Total 46.8 23.3 15

Reddy et al. (2010: 66) has identified a potential problem with NREGS intention to be self-monitoring which is that these very characteristics of the participating households [i.e. that they are predominantly poor and marginalised] may also create the conditions for misuse and neglect of processes, and therefore call for effective institutional and governance structures. Moore and Jadhav (2006: 1293) similarly note that a programme like [NR]EGS is likely to remain in place and on track only if potential jobseekers are organised and mobilised to put continual pressure on the administration as otherwise the disparate voices of the poor will be drowned out by more powerful and cohesive stakeholders such as landlords and officials. This proposition was confirmed by Shankar et al. (2010) who observed that the corollary of the low elite capture in Rajasthan (relative to Andhra Pradesh and Maharashtra), was that corruption at the implementation level was higher because there was little effective monitoring. While Reddy et al. (2010) reported higher administrative standards in Andhra Pradesh (for example, completed and signed muster rolls available for inspection in almost every site), they noted that in a third of cases workers' job cards were incomplete which makes it hard for them to contest discrepancies in payment. This relates to the experience of Reddy, a boy from Katur, who said that the Field Assistant in Katur expects each work group to pay him a weekly 'retainer' of Rs 100 to ensure his records are correct or "he will not calculate properly in meters. And he will not write our bills clearly and somehow or the other he will cause us trouble, madam. A similar problem occurred in Poompuhar where the power of the meti, in the absence of supervision from the field assistant, meant that "whatever the meti writes is the law" (Sarpanch s brother). There is a similar lack of supervision at higher levels - the officers also come in jeeps [...] they just ask one or two persons if the work is going on well. The moment they say yes, these officers go away" (Sarpanch, Poompuhar) and as Reddy et al. (2010) described this extends to the limited nature of the Sarpanch s role. Participants therefore face Hirschman s (1970) choice between exit and loyalty as "if you complain [the officials] say if you don't like this then stop coming " (Sarada and Sahithi, Poompuhar), and if you protest you are not believed (Triveni, Katur). Revanth, a boy from Katur, also suggested that the social 16

impact of protesting was too great: "if we report to higher officials they will dismiss this person. He has a daughter, a son, and all the family members. While NREGS has not followed the apocalyptic trajectory proposed by Moore and Jadav (2006: 1293) involving descent into some mixture of gross corruption, exploitation for narrow political patronage, and loss of reputation and support, numerous examples of petty corruption and more fundamental problems have been identified in the literature (Gaiha, 2005; Mehotra, 2008; PACS, 2008). These problems are particularly acute in remote areas where Scheduled Tribes live, as the example of Patna shows, due to the challenges of NREGS implementation and limited opportunities for financial leakage compared to other programmes. Some of the dubious practices observed in the three communities included warning workgroups of supervisory visits by mobile phone so they can conceal working children ( the children would join the shepherds grazing near the workplace until it was safe to return, Reddy and Govindh, Katur) or increase the number of workers for the purpose of showing a greater number of workers to the checking officers (Revanth, Katur). People were also marked present when they were absent, or present but not working, either because the worker had bribed the meti, the meti was taking the share of the ghost worker ( meti selects their favorite workers to work for them and add five more names to the list and sits and earns free wages, Sarpanch, Poompuhar), or the meti was sharing the profits with the rest of the work group. Field assistants were accused of inflating the measurement of work done and taking the profit, or asking for a retainer from workers to measure accurately. Finally, both participants and fieldworkers (who have been conducting qualitative research in these communities since 2007) observed relationships of patronage between field assistants and farmers who wanted work done on their lands. They noted that poor households in the qualitative sample who had had their land developed often had relationships with local NREGS staff or other powerful people, or had questioned the fairness of the selection process which led to their inclusion to stop them protesting. 17

4. Conclusion Among rural households in Young Lives pro-poor sample membership of and participation in NREGS is relatively high: 68 per cent of households have worked in the scheme during the past twelve months, and the figure rises to 77 per cent for Scheduled Caste and 80 per cent for Scheduled Tribes. However, the median household income from NREGS during the past twelve months is only Rs 3,539 (the individual income is Rs 1,848) and the median number of days worked is just fifteen. This suggests that the impact of NREGS may be relatively limited (we discuss below why this might be); although the marginal value of this cash income to poor households is likely to be considerable. If we evaluate the performance of NREGS against its stated objectives, then it has been least successful in fostering social equity. Social equity extends beyond parity in participation for households from different castes, ethnicities, and socio-economic groups to include parity in the benefits received and the extent to which these can be translated into wellbeing outcomes. The paper argues that equality of outcomes has not been achieved and particular types of participant and administering officials have benefitted more than others. The scheme has actually increased inequality within participating communities by providing greater benefits to landlords, larger households, those administering the scheme, and those within work groups who benefit from dubious practices. The perception of unfairness has led to growing mistrust and resentment, not only in relation to the scheme and its administrators, but also within the community as a whole. The paper discusses some of the reasons why fostering social equity might have proved challenging. For example, lack of knowledge among respondents in relation to eligibility for NREGS and entitlements, particularly in the tribal community, and limited confidence in their ability to get these entitlements from NREGS officials. It suggests that the concentration on the development of private lands (45 per cent of works in Andhra Pradesh), which are typically owned by less poor households, has increased these disparities. The authors note that where poor households in the qualitative sample were 18

included this was usually due to relationships with local NREGS staff or other powerful actors. Finally, inefficiency, mismanagement and pressure from local landlords and politicians has combined to delay the authorisation of schemes, slow their progress, reduce workers wages, and occasionally delay payment; all of which contribute to reducing the benefits of NREGS to its target group low-caste landless labourers. While NREGS has the potential to increase the assets of participating households, the households capacity to convert these increases into sustainable livelihoods is reduced by the practices of the institutions responsible for its implementation. Evidence from Latin America (e.g. Adato 2000, Mexico; Adato 2007, Nicaragua) suggest that even when, as in Nicaragua, social protection schemes are loosely targeted so the majority of households in a community benefit (as is the case with NREGS), there is still resentment among those whose households have either not benefitted or not benefitted as much as others. This resentment can be fuelled by inequities that seem relatively minor, for example, differential access to work for Triveni s grandmother ("They are allowing older people than her, madam. They don t even get up in the field. And my grandmother says do I work worse than them? ). Examples of this sort illustrate the value of a wellbeing perspective which takes account of the subjective aspect of people s experiences. For example, it can be sensitive to the way in which people appraise their situation not only in relation to past experiences, but also the experiences of those around them. Even when evaluated against the objectives it set itself NREGS performance has been mixed. This is partly due to the potentially contradictory nature of some its objectives, for example, human capital development and reducing migration. It may also be due to a lack of fit between means and ends, for example, using heavy manual labour as the means to increase rural workers skills and female participation in the workforce. Some of NREGS objectives might have benefitted from further consideration what does it means, for example, to foster social equity in a country with a historical tolerance of ingrained inequalities? 19

For NREGS to continue to be successful in Andhra Pradesh then greater attention needs to be paid to its impact on social equity and social cohesion and specifically the ways in which some of the administrative practices described in this paper might undermine this. NREGS has considerable support, within India, and outside it. However, it also poses a number of questions to policymakers: is it possible to have a scheme that is internationally celebrated with large numbers of participants that still has relatively limited impact? How can the scheme manage disparities in benefits for different communities and types of participant, and historic imbalances of power and information? While there have been improvements in NREGS governance over time, are the existing accountability measures sufficient? Finally, bearing in mind the sensitivity of child labour in India, how can children s presence within the scheme be acknowledged and managed? Acknowledgements The authors would like to thank Dr Kumari, Abhijeet Singh, Dr Anuradha, V.L. Krishna, D. Udaya, Dr Hymavathi, Young Lives participants and other respondents. Young Lives is core-funded by the UK Department for International Development (DFID). Sub-studies are funded by the Bernard van Leer Foundation, the Inter-American Development Bank (in Peru), the International Development Research Centre (in Ethiopia) and the Oak Foundation. The views expressed are those of the author(s). They are not necessarily those of, or endorsed by, Young Lives, the University of Oxford, DFID or other funders. 20

REFERENCES 1. Adato, M. (2000). The impact of PROGRESA on community social relationships. International Food Policy Research Institute working paper. 2. Adato, M. (2007) Combining Survey and Ethnographic Methods to Evaluate Conditional Cash Transfer Programs. Q-Squared Working Paper No. 40. 3. Aiyar, Y and S. Samji. (2009). Transparency and Accountability in NREGA. A Case Study of Andhra Pradesh. AI Working Paper No. 1, February 2009 4. Ambasta, P., Vijay Shankar and Mihir Shah (2008) Two Years of NREGA: The Road Ahead, Economic and Political Weekly, 23 February. 5. Deneulin, S. And J.A. McGregor. (2009). The Capability Approach and the Politics of a Social Conception of Wellbeing. Working Paper. Bath: Wellbeing in Developing Countries/University of Bath, (WeD Working Paper 09/43). 6. Drèze, J. (2006) National employment guarantee inaction. The Hindu. 12th September. 7. Drèze, J. and A.K. Sen. (1989). Hunger and Public Action. Oxford: Oxford University Press. 8. Gaiha, R. (2005). Does the Employment Guarantee Scheme Benefit the Rural Poor in India? Some Recent Evidence. Asian Survey, 45 (6): 949-969 9. Gopal, K. (2009). NREGA Social Audit: Myths and Reality. January 17, 2009 Economic & Political Weekly. 10. Hirschman, A.O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press. 11. Holmes, R., N. Sadana and S. Rath. (2011). An opportunity for change? Gender analysis of the Mahatma Gandhi National Rural Employment Guarantee Act. ODI Briefing Paper #53. 12. Imai, K. (2007). Targeting versus universalism: An evaluation of indirect effects of the Employment Guarantee Scheme in India. Journal of Policy Modelling 29: 99-113 21

13. Jha, R., S. Bhattacharyya, R. Gaiha and S. Shylashri (2009) "Capture" of Antipoverty Programs: An Analysis of the National Rural Employment Guarantee Program in India, Journal of Asian Economics 20(4): 456-64. 14. Johnson, D. and A. Tannirkulam (2009) NREGA in Andhra Pradesh: Seven Lessons from the Data, CMF Focus Note 4, Delhi: Centre for Micro Finance. 15. Kareemulla, K., K. Sriniva Reddy, C.A. Rama Rao, K. Shalander and B. Venkateswarlu. (2009) Soil and water conservation works through national rural employment guarantee scheme (NREGS) in Andhra Pradesh An analysis of livelihoods impact. Agricultural Economics Review 22: 443 450. 16. Gough, I., J. A. McGregor and L. Camfield. (2007). Introduction: conceiving well-being in development contexts. In: Gough, I., McGregor, J.A. (eds.) 2007. Well-being in developing countries: new approaches and research strategies, Cambridge: Cambridge University Press. 17. McGregor, J. A. (2010). Community responsive social protection: monitoring shifting vulnerabilities for women and children in Zambia. Background paper to the European Development Report. 18. Mehotra, S. (2008). NREGA Two Years On: Where Do We Go from Here? Economic & Political Weekly August 2, 2008 19. Moore, M and V. Jadhav. (2006). 'The politics and bureaucratics of rural public works: Maharashtra's employment guaranteed scheme', Journal of Development Studies 42 (8): 1271-1300. 20. Poorest Areas civil society Program (PACS) (2008) Two years of the NREGS: A critical review-i and II. Accessed on http://www.empowerpoor.com/backgrounder.asp?report=763 21. Porter, C. (2010). Social Protection and Children: A Synthesis of Evidence from the Young Lives Longitudinal Research Programme in Ethiopia, India and Peru. Young Lives working paper." 22. Reddy, D.N., Rukmini Tankha, C. Upendranadh and Alakh N. Sharma. (2010). National Rural Employment Guarantee as Social Protection. IDS Bulletin 41(4): 63-76. 22

23. Sainath P. (2007). Employment Guarantee No place for single women. India Together; 28 May 2007. 24. Scandizzo, P.L, R. Gaiha and K. Imai. (2009). 'Does the employment guarantee scheme stabilize household incomes in rural India?' Applied Economics Letters 16 (12): 1221-1225. 25. Sjoblom, D. and J. Farrington. (2008). The Indian National Rural Employment Guarantee Act: Will it reduce poverty and boost the economy? Overseas Development Institute Project Briefing #7. 26. Shankar, S, R. Gaiha, and R. Jha. (2010). Information and Corruption: The National Rural Employment Guarantee Scheme in India. ASARC Working Paper 2010/02 27. Streuli, N. (2010). Children s experiences of a conditional cash transfer in Peru and its implications on their social worlds. Young Lives working paper (in press). 28. Sudarshan, R.M., Rina Bhattacharya and Grace Fernandez. (2010). Women s Participation in the NREGA: Some Observations from Fieldwork in Himachal Pradesh, Kerala and Rajasthan. IDS Bulletin 41(4): 77-83. 29. Uppal, V. (2009) Is the NREGS a Safety Net for Children?, Young Lives Student Paper, Oxford: Young Lives. 30. White, S. C., (2010). Analysing wellbeing: a framework for development practice. Development in Practice 20(2): 158-172. 23