ASSESSMENT OF OUTREACH AND BENEFITS OF NATIONAL RURAL EMPLOYMENT GUARANTEE SCHEME OF INDIA

Transcription

1 The Indian Journal of Labour Economics, Vol. 52, No. 2, 2009 ASSESSMENT OF OUTREACH AND BENEFITS OF NATIONAL RURAL EMPLOYMENT GUARANTEE SCHEME OF INDIA Abusaleh Shariff* NREGA, an ambitious mass employment guarantee scheme implemented since the last four years, intends to sustain income and consumption in Indian rural outback. A large programme, backed by budgetary allocation promises 100 days of manual work to households who register and apply. Failure to provide employment through the gram sabhas creates cash entitlements as a matter of law. This paper analyses official statistics and survey data from seven northern states. The future of NREGA is strongly linked to the cherished national goal to strengthen and broadbase decentralisation of local governance. But there are wide variations amongst the states not only in the level of decentralisation but also in the capacity to implement such a large scheme and lack of convergence amongst relevant government departments and functionaries. NREGA has the potential to address both sustenance of income and enhance the social welfare of households in rural areas. I. INTRODUCTION India s National Rural Employment Guarantee Act or Scheme (NREGA, also NREGS) is being implemented since the last four years, with the multiple objectives of sustaining income and consumption through wage works, creating durable assets and empowering rural communities through the backing of law. The rural mass public works programmes (PWPs) are programmatic efforts to generate non-farm employment so as to sustain incomes and consumption, especially during conditions of distress such as those caused by drought, natural disasters, market failure in distribution of food items, and so on (Sen, 1981; Dreze, 1990). The importance of direct policy and programme action for employment generation ensuring food security to the poor, replacing food subsidy strategies in the developing economics has been compulsively argued by von Braun, This paper reviews the functioning of NREGA over the years and also analyses data from formal sources, sample surveys and qualitative village studies in selected states in India. In spite of NREGA being an open-toall programme, its outreach and coverage are low. An econometric analysis of survey data suggests that occupational and gender identities do influence access to a wage employment * The author is with the International Food Policy Research Institute, New Delhi. a.shariff@cgiar.org. He would like to thank Anushree Sinha for data access, and P.M. Kulkarni and Anil B. Deolalikar for their excellent comments. He would also like to acknowledge research support provided by B.L. Joshi, Anjor Bhaskar, Devendra Kumar, Prabir Ghosh, Umme Sadiqha and Jaya Koti. The comments received at a conference organised by the Institute for Human Development (IHD) and Centre De Science Humaines, New Delhi, were enriching. He is also thankful to the referees of the Journal for their useful comments. The usual disclaimers apply.

2 244 The Indian Journal of Labour Economics programme, yet even the well-off within the local context tend to enrol and seek manual wage employment across selected states. Thus, the most deprived such as households reporting acute food scarcity and distress migration are not able to reap the benefits of NREGA. This paper is divided into five sections. Section II introduces the scheme while Section III evaluates the available official data up to Section IV analyses a set of data from 16 deprived districts of 7 states in the northern part of India. A synthesis of qualitative village studies is part of this section. Section V summarises the paper, lists issues that need further research and draws relevant conclusions. II. NREGA A STEP TOWARDS INCLUSIVE DEVELOPMENT 1. Guiding Principles Leading to the Enactment of the NREG Act While the NREG Act itself does not refer to or lay down any specific guiding principles, the debates around it refer to two experiences: (a) a fairly successful and people appreciated, the Maharashtra Employment Guarantee Scheme (MEGS), and (b) the harsh experience of the Bengal famine, which was exacerbated by both an inefficient administrative response as well as market failure in food distribution. A deeper understanding of the dynamics of the famine highlighted how the lack of employment and earning opportunities lead to impoverishment, debilitation and death. The classic exposition of the entitlement approach to the analysis of hunger and deprivation points to such a possibility during natural calamities as well (Sen, 1981; 1986). The entitlement approach focuses on the ability of a person to acquire food and other relevant commodities within the prevailing economic, social and legal arrangements. Public employment provisioning as a policy response to alleviate such deprivations has become an obvious policy choice in recent years. Policydriven investment in PWPs has been in vogue since a long time and has also strengthened the foundations for mixed economies through externality gains caused by the emergence of manufacturing and business activities. Note that this approach does not consider direct food subsidy to the vulnerable a prudent strategy which is reflected in India s long-time experiments with innumerable PWPs often tied with food supplies (more on this is discussed below). It is useful to recollect the limitations of the direct food subsidy policies that were documented by Pinstrup Andersen in Further, mass programmes yield better results when the communities are adequately empowered so as to seek participation in the programmes (Streeten, 2002; Weinberger, 2000; Krishna, 2001; 2006). 2. The Scheme The National Rural Employment Guarantee Scheme (NREGS), 1 initiated in 2005, has received priority policy attention in India s Eleventh Five Year Plan ( ), under the broader objective of Bharat Nirman 2 aiming for the resurgence of rural India. As mentioned above, NREGS is based on the positive experiences of reducing human distress over the past three decades, beginning with the early 1970s in rural Maharashtra (see Moore and Jadhav, 2006; Dev, 1995; Dev, et al., 2004). The civil society and activists working for the welfare of the poor provided the much-needed persistence, which is normally needed to

3 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 245 influence a policy. NREGS, India s public works programme is unique, being large in size 3, and intended to cover long periods, disburse huge funds and be dynamically responsive to climatic and rainfall conditions, and above all open to anyone intending to work at wage rates that are often lower than the prevailing causal wages in local areas. Since the programme is designed to self-target the needy, besides chronic poverty manifested, for example, in food inadequacy, it also intends to address idiosyncratic risks and shocks faced by other households (Schedule II of NREGA, 2005). Although India has always resorted to large universal programmes such as the Public Distribution System (PDS), the Integrated Child Development Services (ICDS) and a plethora of employment programmes, the NREGA is radical as it guarantees 100 days of wage work on demand, failing which the claimants can legally demand cash compensation. The Act promotes the seeking of paid employment on the volition of at least one member from any household in manual rural works. The scope for self-selection of those who indeed require imminent income support for livelihood is an innovation, which should result in a high degree of automatically targeting the most deprived thus reflecting its potential for poverty reduction. This scheme can attract the attention of the otherwise unemployed or under-employed workforce because of immediate income (cash) transfer opportunities, which is an inbuilt short-term relief objective. What Indian policy has recognised is that such Keynesian style expenditures on public works can be important normally, even in the absence of price or income shocks, so as to smoothen and sustain seasonal fluctuations in labour demand and, therefore, wage rates in rural areas wherein rainfall patterns and insufficient irrigation preclude year-round crop cultivation. 4 The Act also aims at the generation of productive assets, protecting the environment, empowering rural women, reducing rural urban migration and fostering social equity, among others. NREGA, if implemented with due earnest, has the potential of not only reducing vulnerability and relieving chronic income deprivation and improving rural livelihood security but also of the construction of durable assets and markets in rural areas. The Act envisages free entry and exit into the NREGA, but has made it mandatory to follow a two-step procedure. The first step is that all households intending to work on NREGA have to register and show an expression of interest with the Gram Sabha/Panchayat, followed by the second step of making a written application for actual work. The NREGS is still in its formative stage of implementation; therefore, a comprehensive assessment has not been attempted so far. While the local level functionaries and the panchayat need to work in tandem to successfully implement NREGA, often the over-riding influence and interference occurs from the local contractor, who is often a politician or has political patronage. While the local civil society institutions are generally non-existent, even if they do exist, they have been unable to penetrate the collusion between the bureaucracy and contractors. This micro-local level operative space is the one that needs a thorough study and one would find large variability across time and space with respect to these dynamic relationships. The programme s multiple objectives can be summarised as: generating employment and creating physical assets/infrastructure. The balancing act between the two is difficult since the former can potentially take precedence over the other. Balancing the objectives of

4 246 The Indian Journal of Labour Economics the programme necessitates large organisational skills, and the proximate involvement and commitment of the local level functionaries. In contemporary India, such public functionaries are not available and the execution of public works has always been dependent upon private contractors, who can engineer shifts in focus, for example, by making the employment programme more of a normal/regular construction project, and NREGA cannot be an exception to this trend unless adequate safeguards are enforced. 3. Budgetary Allocations The Budget of has allocated Rs. 39,000 crores or about 0.66 per cent of the GDP for the NREGA, but the entire allocated amount may not be spent, as has been the case during the previous three years. The actual allocations for (the first year of implementation of NREGA) was a meagre Rs. 12,000 crores, of which only Rs. 8,823 crores (US$ 2.2 billion) was actually spent in 200 selected districts. The coverage in terms of the number of districts was enhanced by an additional 130 in and in , all the 610 districts were covered, yet the total allocation of funds was only Rs. 16,000 crores. Although Rs. 20,000 crores was available for , it is still only about one half of the original amount estimated. This shows a tardy and troubled implementation of the programme; while NREGA is a Centrally sponsored scheme, administratively managed by the Ministry of Rural Development, its implementation is exclusively done by the respective state governments. The states are expected to systematically make a claim over the allocated resources by planning NREGA work activities and turning them into projects. An accumulation of such projects is used to appropriate funds from the scheme resources earmarked through a budgetary process. Often, the states are unable to execute the programme efficiently due to the shortage of administrative and implementing personnel at the grassroots and sometimes due to interparty political differentials, if the party in power in the concerned state is different from the one ruling the Central Government. The funding of NREGA has been carved out by amalgamating all past programmes and it still constitutes just about one-third of the annual budgetary allocation of the Department of Rural Development, Government of India... NREGA stipulates that the respective states have to contribute 25 per cent of the cost of wages and materials used during the programme implementation. The Maharashtra EGS, for about the last three decades, followed a different funding model entirely supported by the state legislative body, which continued to raise its (the state s) share of funds for NREGA by imposing (state) tax specifically for this purpose (Deolalikar, 1995). 5 Whether the current NREGA cost sharing between the Centre and the states is feasible or not depends upon a number of factors including the state level political will, public demand, and the fiscal prudence of the state governments. At the national level, NREGA allocations and expenditures hover between 2-3 per cent of the Central (revenue account) outlays for the past three years. Any further expansion of NREGA in terms of a deepening of the programme need not be a cause for concern with respect to its adverse impact on the national level fiscal deficits, even if no additional tax or levy in the name of NREGA is imposed. Due to the economic slowdown during the last two quarters of the

5 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 247 financial year and its expected continuation through , one expects a strain on the fiscal deficit and therefore on NREGA allocations as well. III. STATE LEVEL PERFORMANCE OF NREGA The national level Ministry and associated State Departments of Rural Development are responsible for implementing NREGS across the entire country. This Ministry updates the progress of implementation of NREGS on its website twice a year on the basis of data provided by the states. In the following section, the data extracted from this source is used to assess the progress of the NREGS for the year Accordingly, overall 33.7 million households were provided with 1.43 billion man-days of employment under NREGS and received close to Rs. 86 billion during (see Table 1). These absolute numbers suggest a vibrant and highly efficient programme implementation and match the stated policy and the targets. The following discussion intends to highlight the relatively better performing states in terms of implementation of NREGA and link it to levels of backwardness or incidences of poverty. Overall, Rajasthan, Madhya Pradesh and Chhattisgarh stand out as the better performing states, both in terms of having a better coverage and creating a relatively larger number of days of employment per household, thus facilitating better netting of wage income. While many other states are poor performers, the worst are Bihar, Uttar Pradesh, Orissa and Jharkhand, which harbour a large number of the poor but do not utilise the opportunities created by NREGA to extend benefits to the needy. The state level progress for the year is discussed below, by using four parameters, namely the percentage of households covered under NREGS, the number of NREGA employment days, the total state level NREGA expenditures and the average wage accruals per household. Note that this analysis is based on routine service statistics made available by the Ministry of Rural Development and not elicited directly from the beneficiaries through a survey. 1. Coverage and Number of Days of Employment Given the large variation in the size of the states, it is useful to discuss NREGA coverage in terms of the proportion of households enrolled in the scheme. This can be further matched with the state of household level poverty but such a matching is difficult to undertake without unit level data, yet the size of the coverage can be linked with the aggregate poverty rates officially estimated at the national level. 7 The states that claim to have covered more than 50 per cent of the households are Chhattisgarh and Madhya Pradesh, with both of them having poverty rates that are much higher than the national average; followed by Bihar and Jharkhand, with over 30 per cent coverage but very high levels of poverty (see Table 1). Other states achieving above 30 per cent coverage are West Bengal, Rajasthan, Assam and Andhra Pradesh, with moderate to low levels of rural poverty. On the other hand, states achieving meagre NREGA coverage are not only Punjab and Haryana, with very low levels of rural poverty; but also Gujarat, Kerala and Maharashtra, with less than 5 per cent coverage, and Karnataka with 8 per cent, but with moderate levels of poverty prevalence. Uttar Pradesh, with fairly high levels of poverty, has extended NREGS to only about 20 per

6 248 The Indian Journal of Labour Economics cent of its rural households. When it comes to providing the maximum number of days of employment, Rajasthan stands out with an average of 77 days, followed by 63 in Madhya Pradesh and 58 in Chhattisgarh. As per NREGA, households can claim a maximum of 100 days of employment per year. Thus, these three states can be considered as success stories of NREGS in India, in spite of their large uncovered targets in terms of coverage and man-days of work provisioning. All the other states doing poorly in coverage are also faring badly in terms of maximising the netting of the number of days of employment. States ordered by avg. no. of days of employment Table 1 State Level Performance of NREGA Based on Official Data, Amount distributed on NREGA (millions Rs) No. of HHs on NREGA work (millions) % Rural HHs participating in NREGA Avg. days of NREGA work/hh NREGA wage rate in Rs. Wage accrual in Rs./ HH Major states Rajasthan 10, Madhya Pradesh 16, Chhattisgarh Tamil Nadu Haryana Jharkhand Andhra Pradesh 16, Maharashtra Punjab Orissa Karnataka Assam Uttar Pradesh Kerala Gujarat West Bengal Bihar Other states Arunachal Pradesh Himachal Pradesh J&K Manipur Meghalaya Mizoram Nagaland Sikkim Tripura Uttarakhand Total Note: The NREGA wages are not straightforward. Often they are the piece rate wages for one day of work. Source: GoI, 2009 and GoI, 2007.

7 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS State Expenditures and Household Wage Accruals A look at the total expenditures suggests that Madhya Pradesh, Andhra Pradesh and Rajasthan have distributed Rs billion 8 as wage payments followed by Uttar Pradesh, Chhattisgarh, West Bengal and Bihar, with the utilised amounts ranging between Rs. 5 and 10 billion each. Notwithstanding a considerable variation in NREGS wage (see more on this below), 9 which ranges between the high levels of Rs. 136 in Haryana, Rs. 125 in Kerala, and Rs. 100 in Rajasthan and Uttar Pradesh each, and a low of just over Rs. 70 in many of the remaining states; the average accruals per household has been the highest at Rs in Rajasthan (with 31 per cent of the households being covered), Rs in Haryana (but with only 2.8 per cent of the households covered), Rs in Madhya Pradesh (with 54 per cent of the households covered) and Rs in Chhattisgarh (with 70 per cent of the households covered). The amount of wage accruals is a meagre Rs in Bihar, Rs in Gujarat, Rs in West Bengal, Rs in Maharashtra (all with low coverage), and Rs in Uttar Pradesh (covering 20 per cent of the households). Three states, namely, Madhya Pradesh, Rajasthan and Chhattisgarh, again stand out in terms of the annual size of wage accruals; followed by Andhra Pradesh and Jharkhand. The two relatively backward states that are not performing well are Bihar and Uttar Pradesh. 3. NREGA Wage Rates The ideal that NREGS should serve the deprived and the poor through a self-selection mechanism was based on the belief that the programme wages will provide adequate motivation to people who intend to work but do not have access to work in the local economy. Besides, the amounts of the daily wage should be just about the level of the reservation wage of deprived households. The idea was that NREGA wages should not be too high or too low, but such that the poor, when they intend to do so, can seek such employment that is available on demand. Labourers are entitled to the statutory minimum wage applicable to agricultural workers in the state, unless the Central Government over-rides this by notifying a different wage rate. If the Central Government notifies a wage rate, it is subject to a minimum of Rs. 60 per day for (Section 6 of the Act), and the wage amount during and has been higher at Rs. 70. One, however, finds wide variations across states with respect to the actual prevailing casual wages, which appear to be based on the demand and supply factors regulating labour in the micro-regions such as a district or block (see Table 2). Although each state is free to frame its own scheme, it has to follow certain basic features that are listed in Schedule I of the Act, which specifies the type of works to be undertaken, the wage rates and minimum facilities that are to be provided at the worksite. Further, the state scheme is expected to follow the operational guidelines issued by the Ministry of Rural Development at the national level in January For the year , it was found that many states fixed their own NREGA wages, which were considerably higher than the Rs. 70 fixed as the minimum wage. These states are Haryana, Kerala, Rajasthan and Uttarakhand, with wage rates of Rs. 100 or more. Of these, Kerala and Haryana also have high levels of regular daily causal wages

8 250 The Indian Journal of Labour Economics and, therefore, the need to fix higher NREGA wages. But in the case of Rajasthan and Uttarakhand, though there does not seem to be such a need, yet the NREGA wage is fixed at Rs. 100, which seems a good enough reason to explain why the NREGA performance is relatively better in these two states as compared with the former two. In two other states where the NREGS seems to be working well, Madhya Pradesh, and to some extent, Orissa, the NREGA wages of Rs. 85 and Rs. 70, respectively, are much higher than the prevailing daily causal wage rates in the respective states. This difference is much larger in Madhya Pradesh the NREGS wage is almost double of the prevailing state level average casual male wage rates. The case of Bihar, Andhra Pradesh, Maharashtra Table 2 NREGA Wage Rates as Compared to Casual Wages, State NREGA wage (Rs.) Wage rate (Rs.) for casual labor (2007) NREGA wages as % to casual wage Men Women Men Women Major states Rajasthan Madhya Pradesh Chhattisgarh 70 Tamil Nadu 85 Haryana Jharkhand 86 Andhra Pradesh Maharashtra Punjab Orissa Karnataka Assam Uttar Pradesh Kerala Gujarat West Bengal 75 Bihar Other states Himachal Pradesh Jammu & Kashmir Uttarakhand Arunachal Pradesh Manipur 66 Meghalaya Mizoram Nagaland 91 Sikkim 100 Tripura Source: Official NREGA Reports from NREGA wages are the average amounts paid often for a specified amount of work during the day fixed on the basis of the piece rate. The daily casual rates reported are also averages of a number of daily wage rates given for manual agricultural work, which often varies depending upon the nature of the work and the micro region within a district.

9 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 251 and Karnataka is different. The NREGA wages in these states are reasonable at around Rs. 70 to Rs. 80, and the respective normal wage rates are still lower. Therefore, if NREGS is implemented properly, these states will show a huge demand for NREGA employment, but this does not seem to be the case except in Andhra Pradesh. 4. NREGA Infrastructure Focus As mentioned elsewhere in this paper, Schedule I of the Act lists eight categories of tasks that are designed to be the focus of the Scheme. Practically all of them relate to water-based tasks such as water conservation and harvesting, drought proofing/afforestation, micro and minor irrigation works ; renovation of traditional water bodies also know as de-silting and flood control. Rural connectivity is a land-based activity and in addition, a residual that any other work that may be notified by the Central Government in consultation with the State Government. All such tasks are labour-intensive; therefore, there is a 60:40 ratio of labour to material inputs, which is strictly enforced so as to ensure larger wage transfers through NREGA works. In order to facilitate a reasonable assessment of the quality of assets created and associated benefits to the people, a programme needs to be functioning for a reasonable time span. Further, most of the states are unable to provide data on asset formulations during the last three phases of NREGS. It is argued that operational dilemmas due to programmatic interventions in water and land make cost-benefit analysis somewhat difficult. Further, an assessment also needs to be made in the natural resources framework, especially factoring shadow prices as well as alternative investment strategies in assessment. Under such circumstances, synergy with other rural development tasks and strategies to protect natural resources can be factored into the programme. Low employment intensity in works, leakages due to mismanagement by labour contractors and over-use of machinery are some of the problems leading to non-existent or low quality assets. The NREGA strict rule of a 60:40 ratio can also lead to assets that are normally of low durability and could lead to a situation wherein the same asset is created over and over again due to poor quality work. Although the strict ratio rule is understandable for creating works that generate large wage employment, durable asset creation can be achieved by creating synergies with other developmental programmes and their associated allocation. The need to establish durable asset formations could be within the framework of the watershed development programme, for example, or even the minor road construction activities through departmental synergies in works and projects. The associated information includes the average size of a work project that has worked out to be only 22 households (see Table 3), with the exception of Tamil Nadu, West Bengal and Bihar, wherein the number of households per project is somewhat large, but as we have seen earlier, overall they are poorly performing states. The average sizes of a project in the better performing states of Rajasthan and Chhattisgarh are reasonable at about 44 and 31, respectively, whereas the size is just about for the other relatively better performing states of Madhya Pradesh and Andhra Pradesh.

10 252 The Indian Journal of Labour Economics Another factor that carries a substantial weight in assessing the benefits through the asset is the question of who benefits from NREGA works, and whose land and properties are developed or protected. Whether the assets generated produce substantial public good or lead to discrete private gain is a substantial issue that needs to be documented as a matter of routine in NREGA. The Act nonetheless allows that the employment scheme can undertake land improvement activities on holdings owned by the SCs and STs, but it is difficult to assess if such a programme focus is actually put in place. There is also a proposal to do away with the limit of 100 days of employment per households in districts that are dominated by the SCs and STs, which can help these communities access a higher number of employment days. States Table 3 Share of Works * by Type of Asset/Activity, Rural connectivity Flood control Water conservation Drought proofing Minor irrigation Traditional water bodies Total work projects No. of HHs per work** Major states Rajasthan , Madhya Pradesh ,07, Chhattisgarh , Tamil Nadu , Haryana Jharkhand , Andhra Pradesh ,47, Maharashtra Punjab Orissa , Karnataka , Assam Uttar Pradesh , Kerala Gujarat , West Bengal , Bihar , Other states Himachal Pradesh , J&K Uttarakhand Manipur Meghalaya Mizoram Nagaland Sikkim Tripura , Total ,52, Note: *Work is defined here as a work project which is a specified bundle of work assignment. **The averages for and are obtained by dividing the number of households provided with jobs with the number of works. Source: Official NREGA Reports.

11 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 253 IV. EMPIRICAL ANALYSIS OF DETERMINANTS OF NREGA USERS In this section, a rare data set is analysed, which helps in understanding the various aspects of outreach of the scheme and also in assessing characteristics that favour its utilisation. In early 2008, the UNDP initiated a study to find out the perceptions of the under-privileged regarding government s safety net programmes in 16 selected most deprived (backward) districts in Bihar, Chhattisgarh, Jharkhand, Madhya Pradesh, Orissa, Rajasthan and Uttar Pradesh. The study surveyed 200 households selected from 5 different villages of the districts and all the households were so identified so as to facilitate their classification as marginal and deprived. Such an identification profile was created by using a combination of social (caste, education) and economic criteria. Thus, all the 200 selected households were eligible to be enrolled into the NREGS in the respective districts and also to receive 100 days of employment during a reference year. 1. Outreach, Enrolment and Number of Employment Days in Selected Districts3 While a highly focused and pan-indian assessment of NREGS is still due, a set of survey data from 7 states and 16 districts located in northern India are analysed and discussed below (NCAER, 2008). 10 Data available from published sources (Internet downloads) of the Government of India are also used wherever necessary. The survey was conducted in two districts of Rajasthan, Uttar Pradesh (one additional district), Madhya Pradesh, Bihar (one additional district), Jharkhand, Chhattisgarh and Orissa. All the selected districts 11 are part of Phase I of NREGS implemented during (except one district that belongs to Phase II). The sample households were selected from 5 most backward villages within a district. However, the number of households interviewed in each of the selected districts was fixed at 200. Since the survey was undertaken to find out the perceptions of the disadvantaged towards the government programme, only the poor and marginalised households were interviewed in each of the selected villages. Table 4 summarises the relative economic differentials amongst the selected deprived households, which are categorised on the basis of caste and religious identities. The Scheduled Castes (SCs) and the Scheduled Tribes (STs) are classified on the basis of their respective caste identities and are usually landless labourers and hunter-gatherers spread all over India. The Other Backward Classes (OBCs) are identified as the middle level castes in the hierarchy of social classification amongst the Hindus, and often they own and cultivate some land. The high caste Hindus are at the top of the list. The term minority is a classification which includes only the religious groups, mostly the Muslims and Christians in this data. The classification of deprived households is estimated from a number of asset ownership variables in the sample. For example, the better-offs in this sample are those who would own at least two types of animals out of cows, buffaloes, and sheep, and any two types of assets amongst cycles, electric fans, pressure cookers, sewing machines and television sets.

12 254 The Indian Journal of Labour Economics Social group SCs STs OBC Table 4 Economic Status* of the Deprived Sampled Households according to Social Groups High Caste Hindu Minority Total No. Row % Col. % No. Row % Col. % No. Row % Col. % No. Row % Col. % No. Row % Col. % No. Row % Col. % Most Average within deprived the group , , Better-offs from within the group All 1, , , Notes: (i)* Economic status is assessed by using a combination of animal and durable asset ownership. (ii) The sampling of households for this survey was done so as to net mostly the deprived population. As expected, one finds that the incidence of the most backward families is higher amongst the SCs, STs and religious minorities (mostly Muslims and Christians). As a corollary, the representation of high caste Hindus and OBCs is relatively higher in the average economic condition category. As far as the better-offs within the group are concerned, they constitute a small percentage with a marginally higher representation of OBCs and high castes. This profiling of the sample makes it clear that practically all the selected households deserve to benefit from the NREGS, but as will subsequently be seen, participation in and the use of NREGS is found in only in a small fraction of the households. Table 5 presents empirical evidence relating to NREGS participation and employment generation from the household survey mentioned above. A large variation is noticed both within and across states with respect to the proportion of households reporting NREGA work during the previous year and the average number of days worked. It should be noted that all sample households belong to the relatively deprived sections and are thus eligible to seek NREGA employment. There are some encouraging findings as far as the districts of Mandla (Madhya Pradesh), which records 93 per cent of the surveyed households reporting NREGA work and meeting 54 per cent of the target of total number of employment days, followed by Dungarpur (Rajasthan) with 87 per cent enrolment and 64 per cent of the requirement being met, and Kanker (Chhattisgarh), covering 76 per cent households and 30 per cent of the requirement being met. All these districts are extremely backward as reflected in the

13 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 255 State/District Table 5 NREGA Employment Status in Selected Districts in North India, Percentage of HHs reporting work Mean no. of work days of reporting HHs/ from records Total work days in selected villages percentage to the required employment % HHs reporting easy access to women/women who got work Rajasthan/Barmer ** ***61 Rajasthan/Dungarpur UP/Sitapur UP/Lalitpur UP/Azamgarh MP/Mandla MP/Tikamgarh Bihar/Gaya Bihar/Muzaffarpur Bihar/Purnia Jharkhand/Palamau Jharkhand/Dumka Chhattisgarh/Kanker Chhattisgarh/Bilaspur Orissa/Ganjam Orissa/Keonjhar Notes: (i)* A sample of 200 HHs from each district, belonging to deprived communities, was taken and the respondents were interviewed from five randomly selected backward area villages. (ii) ** District aggregates from official NREGA reports. (iii)*** Survey data indicating the percentage of those who said that it was easy to get work for women/percentage of women getting work extracted from NREGA records. small developmental index values assigned to them by the Planning Commission of India, and also host very high levels of poverty-stricken households. NREGA s prime objective is to reach people living in the backward areas of India. One, however, finds very low levels of NREGA coverage in villages in Bihar 12 per cent in Purnia, 14 per cent in Gaya and 17 per cent in Muzaffarpur, though they rank as moderate in terms of the index of backwardness, they also record a meagre amount of the requirements met. Similar low levels of coverage and much lower levels of the needs met are recorded in the selected districts of Orissa and Jharkhand. The broad trend observed is that a higher number of households covered have also reported a relatively high mean number of work-days, and those reporting lower coverage have reported a lower average number of days of work. It appears that those who sought work have received about 30 days of work, but if one considers all those eligible households, then the NREGS has so far created less than 20 per cent of the required jobs. For example, all 200 households in the selected sample are eligible to work on NREGS, which signifies a total of 20,000 days of employment per year. The official district level average number of days of employment per household is also presented in Table 5. It is instructive to note that the average number of employment days generated is far above the reported averages, and

14 256 The Indian Journal of Labour Economics often the difference is twice over; with the exception of Azamgadh and Purnia in which the reported average is higher. All the above data need to be understood from the perspective that official data are not authenticated for their accuracy; often the evidence is that such data are escalated to unreasonably high levels. Thus, on the whole, the empirical estimates show a meagre penetration, suggesting a marginal reach of the NREGS to the deprived households with some notable exceptions. At the outset, it appears that Rajasthan, Madhya Pradesh, Orissa and Chhattisgarh are doing well in reaching out to the deprived households in providing NREGS employment; though there is wide variation between the selected districts. On the other hand, Bihar, Uttar Pradesh and Jharkhand have reached only a small fraction of the deprived and eligible households. It is useful to recollect that all the selected districts have been implementing NREGS for over three years. There can be a number of reasons that differentiate the districts and states with respect to their NREGS performance, including institutional constraints (the implementing agency, panchayats versus state bureaucracy), ability to chart out a functional strategy for implementation, societal constraints and practical difficulties. Low performance can also occur due to the sheer lack of interest amongst local functionaries. Such differentials, however, can only be studied through a comprehensive study of the NREGS. Interestingly, the size of the most deprived, who constitute 55 per cent of all the households in the sample, is lower than that of the economically average households within the group, even in the better performing districts, not to speak of districts that are poor performers. The case of the mean number of NREGS employment is similar. The relatively better performing districts also host a relatively higher proportion of the better-offs and generate a relatively larger mean number of days for this group as compared with the most deprived within the sample households. 2. Determinants of NREGS Enrolment and Utilisation Since NREGA is a demand-based programme, it is expected that a number of household, community and governance factors will have a strong bearing on its successful implementation 12. The programme is anchored first on enrolment, and conditional upon whether after being enrolled, the workers maximise the number of days that employment is received. It may be noted that enrolment is identified differently from enlisting one s name in the muster roll, which is the primary listing. Enrolment, on the other hand, implies all those households who would seek NREGA employment during the reference period, as there is a statutory requirement to make an application to work. The household level data was used to undertake advanced econometric analysis to find out what characteristics encourage households to reach out for NREGS, and once they get enrolled, how they maximise the number of employment days and wage payments. The estimates for the days of work maximisation are undertaken conditional upon enrolment and in both situations, a common set of variables are identified. The household level identities, namely, occupation and caste/religious identity, educational level of the household head, housing quality and the amount of land owned are used. The expectation is that those engaged in low paid wage labour, the low caste SCs and STs, the illiterates and

15 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 257 those living in mud or kaccha homes and the landless are expected to use NREGS in large numbers and also to maximise the number of employment days. Given that panchayats are functional across all India, information on the participation and access to community and societal institutions, namely the panchayats, mahila mandals (women s groups), school committees, and so on; and perceived transparency in NREGS meetings are expected to favour better enrolment and utilisation. It is useful to assess how household level vulnerable aspects such as low economic status, women in need of work, households with food inadequacy and missing members impact NREGS utilisation. (i) Socio-economic Factors The first stage of enrolment shows relationships in the expected direction (see Table 6). The household occupation and levels of education show a considerable impact on NREGA enrolment. As expected, the casual labour households show highly significant positive and the salaried and self-employed show highly significant negative coefficients, as compared with the farming community. This suggests a fair access to the wage labour households that is compatible with the occupational hierarchy status in rural India. Similarly, as compared with the illiterate households, those having higher level education have significant and negative coefficients, suggesting that the former have better access to NREGS enrolment. As far as the caste and religious identity is concerned, the coefficients are not significant except that the minorities record positive enrolments at a less than 5 per cent significance level; while the forward caste Hindus record negative but not significant coefficients. The quality of housing, which reflects economic standing in the rural setting, shows no impact on enrolment. When these variables are evaluated in terms of the number of days of employment, one finds that the most dominant household level variable favouring the maximisation of employment days is 1-4 standard and 5-9 standard levels of education as compared with that of the illiterates. This is exactly the opposite of the impact during the first stage, and also counters the ordinarily expected direction of relationship that the illiterates who are expected to be poor get to work a higher number of days of work. The second effect relates to the STs, who are significantly able to maximise the employment opportunities available to them. Even after controlling for the land asset variable, which shows no significant association, ownership of a pukka (good quality) house shows independent positive impact at a less than 5 per cent level; that is, the labour force living in pukka homes maximises the netting of NREGS employment days. Thus, there are contradictory forces operating for the number of days of employment maximisation, which favors attributes pointing to human capital-based abilities such as education and relatively good (relatively better off) housing that lead to higher wage realisation. (ii) Community/Institutional Participation The panchayats are the grassroots level institutions involved in the implementation of the NREGS. The impact of this factor is captured in two ways. Firstly, we explore whether the

16 258 The Indian Journal of Labour Economics fact that women were elected or nominated members in these institutions facilitate enrolment and use of NREGS. The other aspects related to institutions are governance and transparency, which are measured by finding out whether the respondents attended panchayat meetings and whether these meetings were transparent. The former variable, that is being a member, did not show any impact on NREGS enrolment but has a dominant effect on maximisation of the number of days of employment (see more on this later). On the other hand, the households reporting higher participation in panchayat meetings and those having an opinion about transparency in NREGA meetings have a favourable impact on the choice of NREGA work. Thus, women s membership per se in village level institutions/organisations such as the panchayat, school committee, mahila mandals, etc. does not show a noticeable impact on enrolment, but has a strong bearing on maximising the number of days of employment. What does show a large impact on enrolment is the view that panchayat meetings are transparent. Surprisingly, both divergent views that meetings were transparent and that they were not-transparent show a large and significant positive effect on NREGS as opposed to those households that did not have a view at all or the fact that they were truly not interested in having a view on local panchayat meetings. Given this strong and important relationship, additional tests were performed using interactive terms of transparency variables with casual wage worker category but did not find any significant impact (estimates suppressed). This finding brings to the fore the facts that having a view on the panchayat functioning is important but the kind of view that one has really does not matter; and it provides a strong signal to programme managers that in order to enhance the reach and efficacy of NREGS and other similar programmes, people s participation is an absolute necessity, irrespective of whether such participants approve or disapprove of a programme. To reiterate, it is puzzling as to why the institutional membership of a woman would wield a large impact on the maximisation of employment? It appears that the inner story of who gets a higher number of employment days depends upon a crucial fact which shows little influence on whether one gets to work or not; but once one gets into the programme, one derives the maximum benefit. This crucial fact is that of women s memberships in local self-governance such as the panchayat, school committees, mahila mandals, and so on, and this impact stands out in the entire analysis. Thus household gets its number of employment days maximised when a woman from the household or even from the community is a formal member in the panchayat-linked village level institutions. This finding is highly significant to both the national goal of democratic decentralisation, on the one hand, and the favourable implementation of NREGA, on the other. Thus, formal participation in panchayati raj institutions (PRIs) enables households to eliminate informational asymmetries in such a way that they are able to plan and strategise a larger netting of NREGA work-days. One can see this as an adverse exploitation of the opportunity which should otherwise have been passed on to other deserving households. Added to this puzzle is another variable, which had a positive signal in enrolment, but when it comes to maximisation of the number of employment days, it provides conflicting results, for example, those who considered that NREGA meetings are not transparent seem to be able to gain a higher number of employment

17 ASSESSMENT OF OUTREACH AND BENEFITS OF NREGS 259 Table 6 Factors which Facilitate Enrolment and Use of NREGS A Multivariate Analysis Factors/Determinants Sign and strength of variable with NREGA enrolment Sign and strength of variable with no. of NREGA employment days (Social Group) Other Backward Classes (OBCs) (Excluded Category) Schedule Castes (SCs) ns ns Schedule Tribes (STs) ns ++ Forward Caste Hindus ns ns Minority ++ ns (Household Occupation) Farming (Excluded Category) Casual Labour +++ ns Salaried --- ns Self-employed --- ns (Education of the Head of the Household) Illiterate (Excluded Category) 1-4 Standard Standard - ++ Matriculation and Above -- ns (Women s Work Opportunity) Household Expressing No Work Opportunity (Excluded Category) Household Expressing Work Opportunity +++ ns (Housing Condition) Kachcha House Owners (Excluded Category) Pukka House Owners ns + (Post-survey Assessment of Economic Status) Deprived (Excluded Category) Well-off +++ ns (Women s Community Participation) Not a Member in any Institution (Excluded Category) Member in at least One Institution ns +++ (Attend Panchayat Meetings) Low Participation (Excluded Category) High Participation in Panchayats + ns (Transparent Enrolment) Unable to Assess (Excluded Category) Enrolment not Transparent Transparent Enrolment +++ ns (Social and Professional Network) Not a Member in any Institution (Excluded Category) Member in at least One Institution ns ns (Food Adequacy Categories) Household Reporting Average Food Access (Excluded Category) Household Reporting High ood Access ns +++ Household Reporting Fair Food Access ns +++ Household Reporting Inadequate Food Access ns ns Household Reporting Highly Inadequate Food Access ns ns (Migrant Status) Non-migrant Household (Excluded Category) Migrant Household (Owned Land for Cultivation) Land Owned in Acres - Land Owned Square Term Variable not used ns Observations 3200 Note: +++p-value<0.01 and positive, --- p-value<0.01 and negative, ++ (--) p-value <0.05, + (-) p- value <0.10, ns - not significant.