MGNREGA in Kashmir: An Analysis of Labour Market Outcomes and Livelihood Security

Is MGNREGA Self-selective and Demand-driven?

The Act provides for self-targeting or self-selection, and whoever is willing to work should formally demand work and should be provided work within 15 days of raising the demand. If not, then he/she should be provided with unemployment allowances (MoRD, 2012). There is also a specified line of implementers to ensure the smooth functioning of the system from the bottom to the top. However, the ground reality is totally different. One, there is a widespread lack of awareness among the people about the real essence of MGNREGA, its various components and functioning system.

Two, there is a systemic constraint embodied in the institutional evolution within which MGNREGA is operating currently. As mentioned earlier, this type of PWP is entirely new to the regional labour market, and the region is also politically fragile. Therefore, providing the stipulated institutional set-up became a challenging task at the beginning. On the bureaucratic front, appointing officials at the village, block, district and state levels was a time-consuming process. Therefore, until the mandatory appointments could be made, the responsibility was given to the existing bureaucratic institutions (BDO, VLW etc). On the political front, establishing specified political systems at all these levels became more challenging, the reasons being that the architecture of MGNREGA as a central strategy is rooted in the 73rd Constitutional Amendment (73rd CA), which was not extended to J&K, whose panchayat institutions were governed by the J&K Panchayati Raj Act, 1989. Notwithstanding when the Jammu and Kashmir Rural Employment Guarantee Scheme (JKREGS) was formulated no such institutions were in place (from about 1970s), due to the ongoing conflict. This forced the state to opt for an alternative and relatively weak institutional strategy. At the gram panchayat (GP) level, a three-member committee—deh majlis—was to be nominated for planning and execution. At the block level, the Block Development Council was to be constituted with members from the deh majlis and line departments. At the district level, the district development board was to be constituted (Alam et al., 2010). However, even after seven years (by the time of the survey in 2013–2014), this alternate arrangement too did not completely exist on the ground. On the bureaucratic front, a few program officers and gram rozgar sevaks were recently appointed, but the control continued to lie in the hands of BDOs and VLWs, for whom MGNREGA is an additional rather than primary responsibility. On the political front, deh majlises had remained either confined to paper or were captured by the local elites.

These initial institutional dilemmas had given rise to a new colluded system in which local bureaucrats and middlemen worked together to actualise their political and economic interests out of the MGNREGA implementation. The middlemen were local informal and/or formal leaders, contractors/elites who had replaced the position of mate (supervisor) in MGNREGA and implemented MGNREGA projects like any other commission-based government construction contract. Having been in existence since the beginning of the policy, this system got embedded into the structure and continued even after the elected panchayats in 2011, despite the appointment of some staff of MGNREGA after 2012 and further dissolution of panchayats in 2015. Within such a derailed implementation system, the self-selective and demand-driven principles of MGNREGA saw limited scope to function as provisioned.

The willingness of a household to work in MGNREGA was translated in two ways. Either it is an under-reported demand (Himanshu et al., 2015), that is, people will inform some representative or official that they want to work, or a latent demand (Dreze & Khera, 2009), meaning they will not ask but will wait, assuming that someone would bring a project and invite them to work. Another aspect of this was that owing to the lack of awareness about the policy being a guaranteed entitlement, some people who were willing to work in MGNREGA felt shameful of asking a formal/informal representative for work, as they fear being looked down upon. The outcomes from the last 12 months show that this willingness to work (latent and under-reported demand) is not getting fully realised, indicating further that the policy is not demand-driven. Out of the households that wanted to work in the last 12 months, only 44 per cent got work (Table 6). Further, there appears to be a little exclusion bias towards the poor, as out of the total poor households who were willing to work under MGNREGA, only 48 per cent received work, compared to 52 per cent non-poor households (Table 6). Further, out of those poor who were willing to work or work more (in the case of non-completion of 100 days), as well as requested for it, no one reported getting it, in contrast to the 5.5 per cent non-poor households, that too from the higher income quintile (MPCI > ₹700 ≤ ₹3,533) (Table 6).

Why is Poor-household Participation Rate Low?

There are three broad reasons for the low participation rate and the possibilities of an exclusion bias towards the poor. Two indirect contributing factors include supply-side constraint and the local government’s unpredictability in implementation. The other direct contributing factor is the control of the middlemen on the local distribution system.

One, at the macro level, similar to observations in Himanshu et al. (2015), there is a supply-side constraint. As shown in Table 6, there is no lack of latent or real demand on the part of the households, but there is a supply shortage, affecting the poor households. Consistent with Dutta et al. (2014), higher work supply increases participation and results in a better inclusion of the poor and other disadvantaged groups. The MGNREGA participation rate in HPAs is 92 per cent compared with 51 per cent in LPAs. Consequently, the participation gap between the poor and the non-poor is only 6 per cent in HPAs compared with 15 per cent in LPAs (Table 5). Furthermore, while the latent demand is almost equal, rationing is relatively higher in LPAs where only 26 per cent compared to 67 per cent in HPAs have got work in the last 12 months (Table 6). All of this indicates that better work supply not only leads to widening of access to benefits but, more importantly, also creates a space for the inclusion of a larger number of poor households in the policy.

OPEN IN VIEWER

Table 6.

Work Demanded and Received (%)

		Wanted to Work in MGNREGA in Last 12 Months	Participated in the Last 12 Months	Number of Days Worked in the Last 12 Months			Requested Work or Requested More Work, If Not Completed 100 Days	Got More Work After Requesting
				<100	100	>100
Overall		94	44.5	40.5	1	3	32	1
Area	HPA	96	67	89	1.5	6.25	52	4.8
	LPA	92	26	95	5	0	26	0
Economic groups	Poor	95	48	89.1	5.0	5.0	32.5	0
	Non-poor	93	47	92.3	0	7.6	31.7	5.2
MPCI (₹)	≤300	85.2	40.6	84.6	7.6	7.6	22.2	0
	>300–≤500	100	46.1	91.6	4.1	4.1	37.7	0
	>500–≤700	92.7	54.5	100	0	0	18.2	0
	>700–≤3,533	94	45.5	91.2	0	8.6	33	5.5
	>700–≤3,533	94	45.5	91.2	0	8.6	33.0	5.5

Source: The author.

Two, at the meso level (block/district), there is a continued lack of predictability within the governments about the operationalisation of MGNREGA work. Whatever work is approved and/or supplied gets constrained by local governments that start or stop a project at their will without considering the actual need and demand of the people on the ground (Himanshu et al., 2015; Imbert & Papp, 2013). While this disproportionality affects everyone aspiring to work in MGNREGA, it affects the poor even more. There is neither any time-bound response to their demand nor can they go to the block office every time asking about the commencement date of the work. They mainly remain at the mercy of the local village headmen (middlemen), to whom the priority of engagement are not the poor but their family and other clientele. Poor and job scarce are given work only if some space is left, after the engagement of family and other clientele. The option of the poor raising their voice against these middlemen remains limited due to the existence of a colluded system, as well as lack of transparency and accountability in the system.

This leads us to the third, micro-level distributional constraint, and in fact a crucial one, that is, middlemen’s control over the local distribution process. As shown earlier, there is huge under-reported or latent demand. However, in practice, it matters very little whether a household demands work or not, because it is entirely up to the middleman whether he calls a household for work or not. This is partly the reason that some of the households that reported demanding work during the initial years are not doing so now. Hence, instead of voluntary exclusion, ⁹ which the policy wanted to foster by inserting the provision of self-selection, this nature of implementation is leading to involuntary exclusion through middlemen’s rationing, thereby undermining both the self-selective and demand-driven provisions of the policy.

When projects come up, these middlemen capture them and use their own family members (in some cases their older school-going children as well), extended families and affiliated labourers, or distribute work as political patronage. There are two scenarios which come into play in the local labour engagement process. One, if there is ample time to complete a project, the middlemen would not engage any other labourers beyond the core circle. The key incentive for middlemen in this is that the major portion of the project money comes home without any hassle, through different bank accounts of family members. Documents are counterfeited conveniently to match the online administrative procedures, stating that these were the only people who demanded work. This capturing sometimes leads to these households working for more than the specified cap of 100 days, while others equally or more deserving of wage employment remain without a single day of work. Table 6 shows the outliers of the 1 per cent that completed 100 workdays and the 3 per cent that exceeded the limit of 100 workdays during the last FY. Exceeding the 100-workdays limit not only is a violation of policy norms but also shows the extreme perpetuation of middlemen’s control on the work distribution. Two, if there is a shorter time frame within which a project is to be completed, then the labourers beyond this circle are also engaged. Here too, the preference is not given to those who have the most job scarcity, but to those who are or can be politically significant for the middleman.

What Motivates the Non-poor for MGNREGA?

Because the participation of the non-poor is predominant, so far it has remained unclear what motivates them for MGNREGA when its wage rate is lower than the external market wage rate and is meant to protect those with variable income and those vulnerable to livelihood risks and shocks (Johnson, 2008). Even if they are affiliated to middlemen politically or otherwise, they have little incentive to work for such a low wage rate. One argument, as mentioned earlier, could be that the policy is self-selective and the non-poor might work to overcome their slack-season employment crisis (Dutta et al., 2014). This is true to a certain extent. However, the non-existence of a demand-driven system in the first place, higher participation of the non-poor than the poor and the sense of exclusion of the poor do not justify this argument. Also, winter is the peak slack season in the region when several non-poor, including a majority of the poor, are workless; however, no infrastructure or MGNREGA-type work takes places during this season in the rural areas.

The second argument could be that a poor individual might be caught in a non-poor household (Dutta et al., 2014). This is true in the case of women participation from middlemen’s households, as they are equally unemployed as any other woman of their community. However, the participation of three to six family members of a middleman and no participation from an equal or more deserving neighbouring household cannot justify the huge presence of the non-poor in MGNREGA. The third possible argument could be that the poor are self-excluding themselves due to delay in payments, as reported in some parts of India, allowing the non-poor to participate. This is nullified by the fact that there is a huge latent and under-reported demand from the poor, as reflected earlier. Moreover, it is found that those who participated in MGNREGA and needed urgent payments, middlemen provided the cash, against the people returning the amount after receiving it in their account. Many participants have even received payments through bank transfer, in some cases with little delay, but in no case this was given as the reason for lack of willingness to work in MGNREGA. A widespread feeling among the participants was that their pending money was with the government.

Besides the middlemen controlling the policy and trying to get maximum benefits to their households, the only other factor that could explain this unusual participation is a faulty wage payment system. The last 12 months’ wage structure at the local level—the statutory MGNREGA wage, the actual wage paid under MGNREGA, and the external market casual labour wage rate—shows that rarely any household was aware of or paid according to the statutory wage (Figure 4). ¹⁰ Instead, the majority were paid either equal or close to the external market rate, which was between ₹200 and ₹350 (Figure 4). This has made it an attraction for non-job-scarce/non-poor households (HHs). ¹¹ This wage distortion is actually embedded in the ideological misinterpretation of MGNREGA by the regional administrations, most of which, at various levels of interaction, gave an impression of MGNREGA being a developmental policy rather than an SP policy. Hence, project accomplishment (asset creation), instead of delivery of entitlement (providing demand-based jobs when there is job scarcity), remains the priority. Due to consistent delays owing to poor administrative systems and conflict-related strikes, not many projects are sanctioned/completed during the beginning and middle of an FY. Hence, a lot of pressure gets built for meeting the physical targets (asset and workday generation) towards the end of every FY. This creates opportunities for the local colluded system of middlemen and bureaucrats to fulfil their political–economic interests. They do so by adopting an “Exit Theory” (Niehaus & Sukhtankar, 2013) wherein they bargain MGNREGA labour on the external market wage rate to complete projects quickly. In the backdrop of lack of proper documentation, transparency and an accountability system, middlemen bring in non-job-scarce/non-poor households, besides their families and political clients. To adjust records with the official wage rate on Management Information System (MIS), they counterfeit job cards and muster rolls by creating ghost workers and doubling the actual number of workdays. All of them get impunity within the poor transparency and accountability systems (social audits and complaint redress), which have grown under the shadow of 70 years of armed conflict and have become institutionalised.

OPEN IN VIEWER

Figure 4.

Source: The author.

This equalisation or bringing of MGNREGA daily wage rate (₹131) close to the external market daily wage rate (₹250–₹350) has in turn kept several poor/job-scarce households out of policy benefit, mostly at the cost of crowding out from the external labour market. Crowding out from the external labour market is explained by forgone employment reported by 45.5 per cent (Table 8), predominantly from non-poor households. Forgone work opportunity in case of MGNREGA means loss of work opportunity/income from other sources—a labourer leaving equally remunerative work in the private labour market and working under MGNREGA. The calculation is usually done by asking participants what they would have done in the absence of MGNREGA. If they say that they would have worked, then the calculation is performed by asking how many days and at what wage (Dutta et al., 2014). Sometimes, people leave hazardous, exploitative or physically challenging work and come to work in MGNREGA, which cannot be counted as forgone employment. But in the case of this study, the respondents did not report leaving any such kind of work. Instead, similar works in agriculture and/or construction in their own community or adjacent communities were reported. This diversion from equally remunerative works in construction and/or agricultural and allied labour sectors has eliminated the basic SP essence of MGNREGA. It being a safety net (last-resort job) when no other non-hazardous job is available to a rural labourer—and made it like any other government contract scheme. ¹²

This distortion, however, comes with huge political–economic incentives for the middlemen and bureaucrats involved. For a middleman, a great portion of project money comes home, with many family members working on sites, and/or he attains strong political clout in the community. Bureaucrats who collude in the process get fixed, project-based commissions from middlemen. Middlemen have widely reported and confirmed that block-level bureaucrats and support staff, on average, take ₹2,000 as commission for transferring the authority of a project. Every panchayat, on average, consists of seven wards, and every ward, on average, gets one project in an FY. Simply put, the average commission earning from one panchayat by just making authorities of MGNREGA projects is about ₹14,000, and this does not include other commissions paid in preparing work completion reports and final billing. This gives a glimpse of the level of personal motivation or incentives bureaucratic staff have in keeping the policy process derailed; if they actually start implementing the policy as per the guidelines, they would incur a huge loss.

Why are Some Areas Low-performing?

Why labour market outcomes are poorer in some areas that otherwise fall within the same socio-economic and geographical ecosystem also needs some explanation. If the policy was demand-driven, the answer could have been found in people’s demand differences in HPAs and LPAs. An alternate data source, willingness to work under MGNREGA, also does not show much difference—96 per cent in HPAs and 92 per cent in LPAs (Table 6). This indicates that lower willingness does not lead to less work and poor performance in LPAs. The difference in average monthly external market work availability also shows similar results (Table 9). This indicates that the need for public work is equal in both HPAs and LPAs. Considering these similarities, logically, given equal chances, the work supply, targeting and labour market outcomes should have been similar in the two areas. Since there is a lack of similarity, the local political economy of the areas with MGNREGA implementation was examined, which showed some significant qualitative differences.

OPEN IN VIEWER

Table 9.

Average Monthly Workdays by an HH Member in the Private/External Labour Market

		1–5 Days	6–10 Days	11–15 Days	16–20 Days	21–30 Days (Including Contract Labour)	No One in the Labour Market
	Overall	3	27	38.5	20	7.5	4
Area	HPA	0.5	13.5	19.5	11	3	2.5
	LPA	2.5	13.5	19	9	4.5	1.5
Social groups	ST	2.7	29.7	33.8	27	6.8	0
	General	2.5	26.1	42.7	15.1	8.4	5.9
	SC	14.3	14.3	28.6	28.6	0	0
Economic groups	Poor	4.3	58.7	19.6	4.3	2.2	10.9
	Non-poor	2.6	17.5	41.6	24.7	11.6	1.9
MPCI (₹)	11–300	9.1	36.4	9.1	0	9.1	36.4
	301–500	2.9	65.7	22.9	5.7	0	2.9
	501–600	2	30	45	10	8	5
	601–27,772	2.2	15.7	41	26	12.6	1.5

Source: The author.

Interaction with the households, local leaders (formal and informal) and bureaucratic officials revealed that during the pre-panchayat period (PRPP), one factor played a role in the cross-sectional difference. There were fewer middlemen (political leaders, elites or contractors) in LPAs willing to take up MGNREGA projects. Politically, while leaders had strong affiliations, political opportunism (direct political gain) from MGNREGA implementation was not much, because panchayats had been dysfunctional over the last 35 years due to the ongoing conflict and their hope of revival was marginal. Hence, it was mainly the economic benefit that was driving local informal leaders’ interest in MGNREGA. However, even though some LPA panchayats had politically strong informal leaders, during the initial years, most of them found MGNREGA projects to be less incentive-based and having a complex documentation system, and therefore gave less attention to it.

With the election of halqa (gram) panchayats in mid-2011, some significant changes took place in the local implementation process, such as increased awareness about the existence of MGNREGA, increase in number of local leaders in the policy process and efforts to distribute work as per the number of wards in each panchayat. However, even after these changes, as shown earlier, the targeting and labour market outcomes of MGNREGA continued to be poor in LPAs compared to HPAs, the reason being that while all the elected panchayat leaders (some of whom were already acting as middlemen and some who became middlemen after getting elected) continued to prioritise their economic interests in MGREGA implementation, there were differences in the local political ecosystem. Consistent with the findings of Jha et al. (2009), there was a dominance of what can be called ‘Political Contestation’ in LPAs contrary to ‘Political Cohesion’ in HPAs. ¹⁴ In LPAs, the ruling panchayat party had a strong bias against the opposition party members and their wards. They continually excluded them from the planning of meetings and manipulated plans, with the main agenda being delivering MGNREGA and other welfare schemes as political patronage to party-affiliated wards and clients. This continuing political battle (rivalry/contestation) undermined MGNREGA implementation, through poor planning, capturing of projects, delays, recurring complaints and spillovers. Moreover, it also worked as a negative reinforcement for bureaucratic staff, resulting in their lack of interest in these panchayats, with the overall impact being poor performance—limited work availability, poor targeting and lower workday generation. In contrast, though political differences were also present in HPAs, use of political interest in social policy distribution through excluding the opposition party was rarely evident. The main reasons being either that the entire panchayat belonged to one political party or there was a very small opposition that did not report any such politically motivated exclusions. This political cohesion also worked as positive reinforcement for the bureaucratic staff, which was evident from their interactions, as they appeared to be more informed, interested and involved. The overall impact was the relatively better performance in job card provision, awareness generation, opening of bank accounts, distribution of projects beyond political relationships and timely work sanction and completion, thereby resulting in greater work supply, better targeting of the poor and higher workday generation.