As before, all figures are subject to the general qualifications applying to official data on the NREGA. In some cases at least, there are serious credibility issues. For instance, the website presents impressive figures on the number of households who have demanded employment. And, as it happens, these figures are usually much the same (often exactly the same) as the number of households provided employment. Yet, we know from numerous field reports that the work application process is simply not in place in most districts. The figures on employment demanded are not worth the bytes they are made of.
Similarly, the website gives district-wise data on the share of women in the NREGA workforce, but the source of this information is not clear. Presumably, it is meant to come from the muster rolls (there is no obvious alternative). But in some districts at least, this information is often missing from the muster rolls. Having said this, the official figures on womens participation in the NREGA look plausible. Perhaps the gaps have been filled on the basis of informed guesses by data entry operators, programme officers or other staff.
The fact that at least some of the official figures are unreliable (to put it mildly) affects the credibility of the entire data set, since the accuracy of different variables is hard to guess. In this note, we have tried to stay clear of toxic data, but some uncomfortable questions remain. For instance, while expenditure data are likely to be reasonably accurate, the corresponding employment figures may well be inflated, to the extent that there are leakages in NREGA expenditure. The statistics presented in this note have to be read in this light or, rather, lack of light.
As a starter, a summary fact sheet on the all-India situation is presented in Table 1. Obvious as it may be, the first point to note is the extraordinary scale of the NREGA. In 2007-08, the programme generated 144 crore person-days of employment, according to official data. In other words, about four million persons were employed on an average day. Over the year, 34 million households participated in the NREGA. Assuming an average household size of five, this is equivalent, say, to the population of Bangladesh. The scale of the programme is even larger today, with the NREGA budget shooting up to Rs.25,000 crore or so in 2008-09 (from around Rs.16,000 crore in 2007-08). Nothing comparable has been done before.
In Table 1, an attempt is also made to compare the official figures for 2006-07 and 2007-08. Note that the comparison can be made in two ways: with or without restricting the focus to the 200 phase 1 districts where the Act came into force on February 2, 2006. Both comparisons are possible from Table 1. The unrestricted comparison (first and last columns in Table 1) must be read bearing in mind that the number of districts covered by the NREGA increased from 200 to 330 on April 1, 2007.
This unrestricted comparison suggests that the extension of the NREGA to 330 districts was relatively smooth, in the sense that what was achieved in 2006-07 was achieved again in 2007-08 on a larger scale. Total expenditure increased more or less in proportion with the increased coverage, and most indicators remained much the same in both years. This applies, for instance, to employment per rural household (and related indicators of the scale of works), as well to the share of women in the workforce, expenditure per district, and the share of wages in total expenditure.
Two minor changes can be noticed. First, the share of Scheduled Tribes (S.T.s) in the NREGA workforce declined (from 36 per cent to 29 per cent) between 2006-07 and 2007-08, presumably owing to the fact that the initial 200 districts had particularly high shares of S.T. households in the rural population. Second, there was an increase of about 15 per cent in the average wage rate (wage cost per day in the table), in nominal terms from Rs.65 to Rs.75 a day. In real terms, this corresponds to an increase of about 8 per cent over one year. This is hardly hyper-inflation of wages, as some commentators (including the Minister for Rural Development) described this pattern in the context of the alleged need to freeze NREGA wages.
On closer inspection (that is, looking also at the restricted comparison, confined to the phase-I districts), this aggregate picture hides important differences between the phase-I and phase-II districts.
Specifically, employment levels in 2007-08 were much higher in the phase-I districts than in the phase-II districts (about 20 and 10 days per rural household, respectively). This is as expected, not only because the phase-I districts were able to build on the foundations they had laid the previous year but also because the need for employment is greater in the phase-I districts since they are poorer. Note that, consistent with this interpretation, employment levels in the first year of implementation (2006-07 in the phase-I districts and 2007-08 in the phase-II districts) was considerably higher in the phase-I districts than in the phase-II districts. With these qualifications, the dominant pattern in Table 1 is one of relative continuity between 2006-07 and 2007-08.
Table 2 presents a sample of State-specific indicators. In this table, the unrestricted comparison (between the phase-I districts in 2006-07 and the Phase I + Phase II districts in 2007-08) is made State-wise. Here again, we find that the basic indicators for 2007-08 and 2006-07 are much the same, in most cases. However, there are interesting exceptions.
For instance, looking at the scale of NREGA works, as indicated by the number of person-days of employment per rural household, we find that the leaders in this respect, in 2007-08, were Rajasthan, Madhya Pradesh and Chhattisgarh (in that order) as well as the north-eastern region. In fact, the first three States together accounted for more than half the total employment generation under the NREGA in 2007-08. All three were in a leading position in 2006-07, too, with similar levels of employment per rural household. On the other hand, Assam did very well in this respect in 2006-07, but slipped back the next year.
Another State where there was a major setback was Orissa, where employment generation per rural household apparently declined from 21 days in 2006-07 to eight days in 2007-08. Orissa was the focus of a wave of embezzlement reports in 2006-07, and there are at least two possible reasons for this employment crash the following year. One is that there was a substantial reduction in the fudging of employment figures, owing to greater vigilance (there is some independent evidence of this from recent social audits initiated by the National Institute of Rural Development). This could bring down the official employment figures even without any reduction in actual employment. The other possibility is that employment actually declined, perhaps because the corruption scandals had some sort of dampening effect. It is quite likely that both factors played a role.
Looking at the other end of the scale, NREGA employment levels in 2006-07 were extremely low in Bihar, Gujarat, Haryana, Maharashtra, Punjab and West Bengal; they remained abysmally low in these States in 2007-08, and even declined in several cases (for example, Gujarat and Punjab). On the other hand, NREGA employment per rural household doubled in Andhra Pradesh, Kerala and Tamil Nadu, three States where there were active efforts to expand the programme in 2007-08 (starting from a relatively low base). The absence of a similar take-off in, say, Bihar and West Bengal is a matter of concern, given the high levels of rural poverty in those States. (All these figures have to be read bearing in mind that the need for NREGA employment varies from State to State. This helps explain the low levels of employment in, say, Kerala, but not in Bihar or West Bengal.)
It is a little ironic that all the three NREGA leaders in terms of scale of works had Bharatiya Janata Party (BJP) governments in 2007-08, considering that the Act is perceived as the Congress partys baby. If anything, the odd contrast between the BJPs activism in this regard and the Congress resilient inability to nurture its own baby became even sharper with Assams slide. Of course, it is not that every BJP government has done well in this respect Gujarat is a prominent exception. But on the whole, the BJP seems to have adopted the Congress orphan with open arms. Whether this played a part in the BJPs recent return to power in two of these three States (Chhattisgarh and Madhya Pradesh) in spite of the anti-incumbency factor is hard to guess.
Turning to other indicators, the share of women in the NREGA workforce was much the same in both years (around 40 per cent), though there was a noticeable increase (from 40 to 44 per cent) in the phase-I districts a healthy indication of growing participation of women over time. The State-specific figures and rankings for 2006-07 and 2007-08 are fairly stable, with Tamil Nadu having the highest participation of women in both years by a long margin (though Kerala was catching up in the second year). One interesting development is a major increase in womens participation in Bihar and Himachal Pradesh. However, the States where women constituted less than the statutory minimum of 33 per cent of the workforce in 2006-07 remained so in 2007-08 as well, except for Haryana and Uttarakhand. Punjab crossed the mark in the wrong direction. The share of S.C.s and S.T.s in NREGA employment declined a little between 2006-07 and 2007-08, mainly on account of the decline in S.T. participation (noted earlier), with no remarkable changes in State-specific figures.
The labour component of NREGA works accounts for almost 70 per cent of the total expenditure, a significant achievement considering the relatively low labour-intensity of many earlier employment schemes (often owing, in turn, to vested interests in favour of the material component). With the Central government footing the entire wage bill but only 75 per cent of the material costs, State governments have strong incentives to promote labour-intensive techniques under the NREGA. Tamil Nadu went furthest in this direction, with virtually no material component, on the grounds that material expenditure attracts private contractors and breeds corruption. Orissa is the only State where the labour component is substantially below the statutory minimum of 60 per cent of the total expenditure. It is worth mentioning in passing that Orissa also has an odd practice of charging hefty royalties (sometimes as high as 30 per cent of the total expenditure on particular works) whenever digging is involved in NREGA works. Even digging of plain sand apparently attracts royalties. This anomaly was exposed in November 2007 by a delegation of the Central Employment Guarantee Council to Orissa, but to our knowledge the matter is yet to be resolved.
Finally, there have been substantial changes in wage rates (last two columns of Table 2) in several States, and also at the all-India level. The changes have been mainly upward, with four exceptions, notably Maharashtra. The largest upward jumps, in proportionate terms, have occurred in Uttar Pradesh (66 per cent), Orissa (45 per cent), Haryana (28 per cent) and Rajasthan (16 per cent). In Uttar Pradesh, there have been successive increases in minimum wages, from Rs.58 a day to Rs.100 a day within a couple of years. The final hike to Rs.100 came on the personal intervention of Chief Minister and BSP leader Mayawati. It was a shrewd political move, which created unprecedented interest in the programme among the rural poor and took the wind out of the sails of the Congress as far as taking credit for the NREGA in Uttar Pradesh was concerned. This was also an interesting U-turn, considering that Mayawati (and her predecessor, Mulayam Singh Yadav) initially ignored and even disparaged the NREGA as a hare-brained initiative of the Congress.
In Rajasthan, the increase in wages between 2006-07 and 2007-08 is likely to reflect the combination of an increase in minimum wages (from Rs.73 a day to Rs.84 a day) and higher productivity linked to improved worksite management. Indeed, the Rajasthan government has taken important initiatives in this field, including mass training of mates (worksite supervisors) and group measurement of NREGA work. There is some evidence that this led to a significant increase in NREGA earnings under the piece-rate system. Even then, Rajasthan continues to have the lowest average NREGA wages in the country.
It is worth pondering whether there is any causal link between low wages and high employment levels in Rajasthan. As mentioned in our earlier article, this is also a broad pattern across States: high levels of NREGA employment tend to be associated with relatively low wages (for example, in Chhattisgarh, Madhya Pradesh and Rajasthan), and vice versa (for example, in Bihar, Haryana, Kerala, Maharashtra, Punjab and West Bengal). The association may be incidental, but it is also possible that low wages make it easier, in various ways, to implement NREGA on a large scale. Note, however, that the north-eastern region seems to be able to generate large-scale employment at relatively high wages.
Moving to more disaggregated data, Maps 1 and 2 display district-specific employment levels (number of days of NREGA employment per rural household) for 2006-07 and 2007-08. From these maps, particularly Map 2, it is clear that the NREGA, despite being a national programme, is still heavily concentrated in a small number of districts, mainly in the three leading States mentioned earlier (Chhattisgarh, Madhya Pradesh and Rajasthan), where most of the NREGA districts have high levels of employment generation. Outside that region, very few districts have comparable levels of NREGA employment, except for pockets of the North-East. Except for Andhra Pradesh, employment levels are still very low throughout the southern and western regions, even in highly deprived areas such as northern Karnataka, eastern Maharashtra or Nandurbar district, also in Maharashtra.
More importantly, perhaps, the programme is yet to take off in most of the eastern region (Bihar, Jharkhand, Orissa and West Bengal), where the incidence of poverty and hunger is very high and the NREGA could make a big difference.
Such is the picture that emerges from official data. As mentioned earlier, the accuracy of the official figures is an open question, which calls for urgent scrutiny. In any case, independent large-scale surveys of the NREGA (analogous to, say, the National Family Health Survey) would be very useful. It is not clear why a government that now spends about Rs.25,000 crore a year on the NREGA is unable to spare a few crores for a comprehensive independent survey to find out where the money is going and what the programme is achieving.
Last but not least, it is worth noting that there are crucial gaps in the present Monitoring and Information System (MIS). For instance, in our experience, most programme officers at the block level do not have a list of active worksites. All they know is that various sums of money were disbursed to various gram panchayats (or other implementing agencies) on various dates.
Similarly, there is no systematic record of delays in wage payments. If labourers remain unpaid for weeks at a particular worksite, there is no trace of it in the MIS. This is one reason why long delays in wage payments persist in many States, causing immense hardship to NREGA workers and even inducing some of them to quit.
These and other gaping holes in the statistical system need urgent attention if the NREGA is to achieve its ambitious transparency norms.