John Stuart Mill characterized the poverty alleviation problem as how to give the greatest amount of needful help, with the smallest encouragement to undue reliance on it. India has a long history of direct and targeted interventions to fight poverty through workfare schemes, subsidized food, farm-input and credit subsidies. More recently, and following the logic of Mills’ dictum, the Indian government wanted to ensure that rural households achieved a minimum income level cost-effectively, but without encouraging them to become dependent on public support. About 37% of Indians are poor; most are rural dwellers and earn their livelihood from agriculture (Tendulkar Report, 2009).

In November 2005, the Indian government embarked on an ambitious workfare scheme, the National Rural Employment Guarantee Scheme (NREGS), which guarantees hundred days of employment in unskilled manual labour at a minimum wage to every rural household each year. Some of its features include a time-bound employment guarantee and wage payment within 15 days (otherwise the government is penalized), prohibition of the use of contractors and machinery (to enhance direct benefits of the programme to the participants), facilities to be provided on the worksite, a 60:40 wage and materials cost ratio, and a mandatory 33 per cent participation for women. The scheme devolves considerable powers in planning and allocating resources to the local village councils (panchayats) and through social audits allows the community to monitor the progress. It also lists permissible works in the scheme such a s drought proofing, desalination of tanks and flood control. To minimize corruption, the scheme separates the agencies – banks and post offices – that pay from the ones that implement. During its initial years of operation, the NREGS involved an outlay of Rs 13,000 crore (for 330 districts); today that figure is Rs 40,000 crore.

The Comptroller and Auditor General’s report in 2007 sparked off a public debate on the effectiveness of workfare schemes like the NREGS. The report highlighted failures in the distribution of job cards, assignment of work, irregularities in the selection, design and execution of projects, and in the creation of good quality assets. We assess these issues by examining the results of a representative household sample (500 households per state) and ethnographic interviews in Madhya Pradesh (MP), Rajasthan, Andhra Pradesh (AP) and Tamil Nadu (TN).¹ 

 

Has the scheme reached the target population of the rural poor and the landless who are more likely to exist below the minimum poverty line and are most in need of a scheme that supplements their income? The poor in India are preponderantly scheduled caste (SC) and scheduled tribe (ST), landless, illiterate, and from female-headed households. Almost half (47%) of the rural poor are SC/ST (2000 figures). To ensure that genuinely poor people eking out their subsistence enrol, NREGS has a self-selecting mechanism that makes the transfer dependent on a work requirement, allows only unskilled manual labour, and prohibits the use of machines and contractors.

Over 90 per cent of the beneficiaries in the four states thought that the registration process was simple, and most completed the registration within 14 days. Most of the beneficiaries in the two southern states and MP were assisted by the village chief in the registration process; in Rajasthan, 60% of the beneficiaries were assisted by these representatives, while over a third said that an influential person (not a politician or a caste leader) helped them register.

 

In all four states, the historically, socially and economically disadvantaged SC/ST groups have participated in large numbers in the scheme. The shares of the landless among NREGS participants ranged from a low of 26% in Rajasthan and MP (37%) to 47% in AP and 53% in Tamil Nadu, and this corresponds to their population in these states. The larger the proportion of landless households in the village population, the larger is their share among the participants. So to the extent that many landless are also the poorest, the targeting of NREGS on this segment is satisfactory.

We considered targeting using an explicit criterion of poverty by constructing four groups – acutely poor, moderately poor, moderately non-poor and affluent – based on their self-reported monthly per-capita consumption expenditure. MP had the highest percentage of poor (i.e. acutely and moderately poor) participants (81%) followed by Rajasthan (50%) and TN (40%); in contrast, only 28% of AP participants were poor. In the two southern states, the acutely poor’s probability of participating in the scheme was lower than those of the other three categories, while the moderately non-poor and affluent had the higher probabilities, suggesting that the better off benefited more from the scheme. In MP and Rajasthan, the acutely poor had the highest probability of participating in the scheme, showing better targeting on the part of these states. In Rajasthan, however, the affluent had the second highest probability, and this displays an element of capture by the well-off groups. As the incidence of poverty in MP and Rajasthan is substantially higher than in Andhra Pradesh, the lower targeting accuracy in the latter is not so worrisome. Yet another aspect of targeting is within-group participation. Among the poor, participation rates ranged from 71% in MP to 87% in AP. So, a large majority of the poor participated in the four states.

 

One of the oft-repeated assertions is that the poor experience exclusion from the scheme or even if they work, earn less than the non-poor. NREGS earnings depend on the number of days worked, as well as the amounts earned per day. On average, a participating household worked for about 42 days in Tamil Nadu and MP, as compared to 63 days in Rajasthan and 87 days in AP. This means that AP performed best in terms of providing NREGS employment to rural households, while MP and Tamil Nadu performed much below par, providing less than half the guaranteed 100 days. Ethnographic evidence suggests that in these two states, collusion between the farmer and sarpanch resulted in fewer operational worksites. In other interviews, the sarpanches said that though many had applied for job-cards, fewer people turned up to work at the worksites. Poor and non-poor groups participated for similar number of days in the scheme in the four states. This implies that the shortfall in the number of days worked pertains to a systemic issue such as paucity of projects or funds, rather than a bias against the poor on the part of the authorities.

 

Do the non-poor earn more than the poor for the same amount of work? The variation between the daily average NREGS wage earned by poor and non-poor groups was low in the two southern states, implying that there was no discrimination against the poor. In Rajasthan too, the variation was relatively low. However, in MP, the non-poor’s daily average wage rate was Rs 82 as compared to Rs 73 for a poor participant; the acutely poor participant earned Rs 70 as compared to Rs 85 earned by an affluent one. This is worrisome and implies that in MP, though the affluent work for fewer days than the acutely poor, they get paid more than the latter. In MP, Rajasthan and AP, the affluent earned higher daily wages than the acutely poor; this is also troubling and needs to be addressed by the scheme’s administrators.

In all four states, participants earned lower daily wages relative to the promised amount of Rs 100 per day. This could be due to the way work is measured – all four states have adopted some combination of time rate and piece rate systems, and use a collective measurement method (total output on a worksite is measured collectively once a fortnight). Some less poor and more educated workers in AP, who knew about the piece rate system, used machines (prohibited under the scheme) to accomplish the work faster and earned higher wages. Lower wages could also be due to corruption. Beneficiaries in the four states were more likely to monitor the performance of public officials on the wage component of the NREGS than the material component. AP, where 70% were paid weekly, and 20% fortnightly, had lower levels of leakage in the wages component than Rajasthan and Madhya Pradesh. MP beneficiaries, for example, were prey to irregular payments: only 11% got weekly wages, 23% fortnightly, 9% monthly, and over half the beneficiaries got wages when it was available.

 

In the four states, the poor depended more than the non-poor on the scheme as a supplementary source of income, with the share being the highest (26%) in Andhra Pradesh, followed by Rajasthan (16%), and about 10% in the other two states. Even those with moderate amounts of land tend to rely on this source of income in the four states. The contribution of NREGS earnings to the total income for acutely poor households was the highest in AP (21%) while in the other three states it was about 17%. Landless households in Andhra Pradesh said that a quarter of their household income came from NREGS. Andhra Pradesh was clearly the best performer in terms of how the scheme’s wages supplemented the income of poor rural households. However, this overlooks the fact that a large majority of the participants in AP are non-poor, which is worrisome. In Rajasthan, along with the landless, those who had more than five acres had the highest reliance on NREGS income. Overall, though the attractiveness of this scheme to the affluent was considerably greater in the two southern states, the NREGS has achieved its goal of supplementing rural incomes.

Critics of the scheme have argued that it would be better to institute a direct cash transfer in the place of NREGS. We computed the average transfers (the average amount the government could transfer directly as cash to these groups) for the poor and acutely poor proportional to their share of the population and compared them with the actual earnings in NREGS. The results were that 75% of poor participants received more than the uniform transfer in AP, MP and Rajasthan, and half the poor participants did so in Tamil Nadu. For the Rawlsian variant, 10% of MP’s acutely poor participants received more than the Rawlsian transfer but in other states, the NREGS earnings were less than the Rawlsian transfers.

This implies that the acutely poor participants may be better off with a direct cash transfer provided that targeting is accurate, there are no irregularities or hiccups in its implementation, and more importantly, no disincentive effects discouraging job seekers from investing in human capital. In a separate exercise, we found that NREGS has enabled the SC/ST groups and women to earn more than a corresponding cash transfer in villages with political reservations. Educated participants and educated sarpanches enhanced the targeting in AP, Tamil Nadu and MP.

 

Some, including the agriculture minister of India, have criticized the scheme for making it hard and expensive for farmers to find labour during peak harvest seasons but our results challenge such an assertion. In the four states, the participation by the acutely poor decreases and that of the affluent increases when the scheme’s wages are higher than agricultural wages. So, linking the scheme’s wage to the minimum agricultural wage may, in fact, be counterproductive and lead to capture by the non-poor. While the Rajasthan panel dataset shows that women agricultural labourers have benefited slightly (their wages increased by Rs 3 between 2007-9), subsidizing farmers’ wage payments to workers (a proposal by the agriculture minister) may reduce the rural poor’s bargaining power and keep their wages at low levels. Interviews with farmers and sarpanches in MP, for instance, revealed that farmers were utilizing NREGS money to pay labourers a low daily rate (Rs 30-40) and pocketing the rest. The reasons for lack of viability of farming lie not so much in higher wages as in obtaining other inputs (fertilizer, seeds, and water) and remunerative prices for products. More generally, wage subsidies may worsen stagnating agricultural productivity. In any case, in principle, workfare ought to be used for building local public goods.

 

Concerns raised by activists and researchers include fudged muster rolls, pilferage, lack of information about the scheme and low quality works. Given the self-targeting nature of NREGS, access to information about the components of the scheme is important in increasing the rural poor’s participation and monitoring of the scheme. Access to information is essential to increase the bargaining power of the poor.

Higher gram sabha attendance in AP increased the access of the villagers to information about the NREGS as compared to Rajasthan and MP. While respondents in the four states knew that safe drinking water and first aid had to be provided on the worksite, respondents from the South knew more than their cohorts in the North. For instance, in AP, almost all knew that children had to be provided with shade from sunlight (as compared to only a fifth of the participants in the North). But a vast majority of participants in the four states did not know that they would receive additional wages if work was given at a distance of more than five kilometres, or that they were entitled to unemployment allowance in case they were not provided with work.

Another proxy to assess if beneficiaries possess information on the scheme is to examine the complaints, while keeping in mind that power asymmetries may not allow the poor to complain. Late payment of wages and low wages were among the main complaints in the official social audit reports in AP, MP and Rajasthan and were corroborated by the survey. In AP, 22% said that the duration of work was short, and 8% said that they got lower wages. In Rajasthan, 26% said that the duration of work was short, but 40% complained of lower wages, while in MP, the figures were 60% and 34%, respectively. However, 97% in Rajasthan and 78% in MP had not complained to anybody about these issues.

It is important for the government in collaboration with civil society actors to carry out periodic information drives among potential beneficiaries. While these measures may not stop the non-poor from benefiting at the expense of the poor, they might introduce better monitoring on the part of the community.

 

One of the debates in the scheme pertains to the utilization of the funds. NREGS funds are primarily used to generate wage employment and an auxiliary objective is to strengthen natural resource management and thus reduce the effect of some factors (such as drought) that cause chronic poverty. In the initial three years, most NREGS works – water conservation, water harvesting, drought proofing, afforestation and plantation, and rural connectivity to provide all weather access – were undertaken on common land. A majority of the 50 work-site supervisors in MP and Rajasthan interviewed said that the scheme had benefited the communities. They cited projects ranging from irrigation channels, storm drains, to public wells but critics could rightly point out that the supervisors may not be reliable informants. Newspapers have reported many instances of collusion between supervisors, sarpanches and NREGS officials in inflating the cost and providing low quality materials. It is hard to check the quality of materials used in a bund that was washed away in a storm. Unfortunately, it was not possible for us to assess the quality of these projects due to their temporary nature; so it is likely that their benefits are exaggerated.

 

Therefore, say some, the large outlays pumped in by the central government ought to create productive assets or risk increasing inflation. In contravention of the spirit of workfare, several states, including AP, Rajasthan and MP have instituted an individual beneficiary scheme to provide irrigation facilities, horticultural plantation and land development on land owned by households of SC, ST, land reform beneficiaries, Indira Awas Yojana, Below Poverty Line, and small or marginal farmers.

Our 2010 survey of 56 individual beneficiary households from 50 villages in Rajasthan and MP revealed that these households tended to be educated (above primary school), predominantly ST and male, and on average owned about 3.3 acres of land. Most of them cultivated food crops. Most of them, who had heard of the scheme from the sarpanch, had applied for and constructed (with the labour of their families) irrigation and land development facilities. Those who constructed wells on their lands said that they now had regular access to drinking water. So, the scheme has produced productive assets for those who have qualified. However, a question raised about the scheme is whether it is politically connected individuals who have obtained these projects.

 

To the question whether someone at the political/official level helped them in getting access to the scheme, about 70% replied in the affirmative. A fifth of these beneficiaries had worked for a political party and 18% had contacts with political leaders. These responses suggest that those with links to politicians or bureaucrats found it easier to access the individual beneficiary scheme provided they met the minimum requirements. In our worksite survey, about 60% in Rajasthan, half the workers in AP, and a third in MP had not heard of the individual beneficiary scheme. TN did not have this scheme during our survey. Of those who had heard of it in AP, a fifth were ineligible, a quarter had applied unsuccessfully, and a quarter were waiting to hear the result. In the other states, a large majority of those who had heard of it did not know how to apply for the scheme.

This is worrisome for the poor especially if more funds from NREGS are diverted to this scheme. Some sarpanches in MP told us that they were planning to assign 80% of the NREGS funds to such schemes. The creation of such assets – wells or irrigation channels or roads – would either exclude the poorest who are landless and/or involve contractors who would skim off more money from the scheme.

 

In conclusion, SCs and STs have participated in large numbers in NREGS, as have the landless and the poor. NREGS earnings were a fifth of the total yearly income for acutely poor households, implying that despite working for half the promised 100 days in three states (except AP), NREGS wages significantly boosted their income. In AP, TN and Rajasthan, there seems to be no systematic discrimination – in terms of wages and days worked – against the poorest groups who enrol in the scheme. In MP, however, despite the fact that the acutely poor participated in larger numbers than the affluent, their average daily wage was significantly less than the amount earned by the affluent. Here, the concern is that the poor are being paid less than the non-poor for the same amount of work. Another worrisome factor is that large numbers of (relatively) affluent also participated in the southern states and Rajasthan, thus raising serious concerns about the self-selection mechanism. These design and implementation issues therefore require careful scrutiny if the aim is to substantially enhance the benefits to the poor within current budgetary constraints.

Despite these problems, the scheme has succeeded in helping the poor supplement their incomes. Above all, dilution of workfare on specious grounds of enhancing assets of the poor and increasing farm productivity could seriously undermine its merits. Workfare has the potential to expand livelihood opportunities for the poor, raise agricultural wages and make them better aware of their entitlements. The enabling conditions for realizing this potential are somewhat daunting. Political competition, networking, and representation of disadvantaged groups in local councils are some elements that policy makers and social activists can examine to create these conditions.

 

* This article is a distillation of our forthcoming book, Battling Corruption: Making Government Responsive to the Poor.

Footnote:

1. The AP and Rajasthan surveys were conducted in 2007-8, and the other two surveys took place in 2009-10.