SSRN Id2800699

Accepted for publication in World Development.
Recent Social Security Initiatives in India
Jean Dreze and Reetika Khera
(Abstract)
There has been a major expansion of social security programmes in India during the last fifteen
years or so, along with wider recognition of economic and social rights. This paper discusses
five programmes that can be seen as partial foundations of a possible social security system
for India: school meals, child care services, employment guarantee, food subsidies and social
security pensions. The record of these programmes varies a great deal between Indian states,
but there is growing evidence that they make an important contribution to human well-being,
and also that the achievements of the leading states are gradually spreading to other states as
well. Much scope remains for extending these efforts: despite the recent expansion, India’s
social security system is still very limited in international perspective. The paper also discusses
some general issues of social policy in India, such as the arguments for universalization versus
targeting and the value of a rights approach to social security.
Electronic copy available at: https://ssrn.com/abstract=2800699

Recent Social Security Initiatives in India
Until recently, India’s social security system was largely confined to the formal sector of the
economy. Social benefits such as old-age pensions, health insurance and maternity benefits did
not extend much beyond the public sector. The unorganised workforce – more than 90 per cent
of all workers – was largely left to its own devices (National Commission for Enterprises in
the Unorganised Sector, 2006).
During the last fifteen years, however, there have been important initiatives to put in place
more inclusive forms of social security. Today, India has five major programmes that can be
seen as partial foundations of a possible social security system: school meals; the Integrated
Child Development Services; the National Rural Employment Guarantee Act; the public
distribution system; and social security pensions for widows, the elderly and disabled persons.
Some of them, like the public distribution system, existed earlier, but in a very limited and
ineffective form. As discussed further on, there is an important complementarity between these
different interventions, both in terms of the needs they address and in terms of the form of
support involved. Other elements, such as maternity entitlements, are in the process of being
put in place.1 As things stand, however, these five programmes account for the bulk of public
expenditure on social protection. Our aim in this review is to give the reader a sense of how
these programmes were initiated, what they have achieved so far, and the debates that have
surrounded them. Given the broad sweep of the paper, it will not be possible to go into the
details of all these issues, but we shall try to provide interested readers with some guidance to
further readings.
Before examining each of these foundational programmes, we shall discuss some overarching
features of India’s recent social security initiatives. Two issues are of particular interest. First,

the reference period has been associated with a growing emphasis on legal entitlements as
opposed to discretionary benefits. In fact, except for social security pensions, the benefits of
all these programmes are now protected by national legislations or Supreme Court orders.
Second, this period has seen important changes in the methods used to define and identify
households eligible to various social benefits. Specifically, there has been a move away from
targeting “below poverty line” (BPL) households, in favour of alternative methods. These
issues are discussed in the next two sections.
This article draws not only on the academic literature but also on the insights we have gained
from monitoring these social security initiatives in a number of Indian states. In particular, we
have learnt a great deal from a series of field surveys conducted with student volunteers from
2002 onwards. These include, for instance, a 2011 survey of the public distribution system in
nine major Indian states, and a 2013 survey of the five programmes mentioned earlier in ten
states. We shall use the findings of these surveys along with those of national surveys such as
the National Sample Survey, the India Human Development Survey and the National Family
Health Survey.
Eight Indian states figured in many of these field surveys: Bihar, Chhattisgarh, Himachal
Pradesh, Jharkhand, Odisha, Rajasthan, Tamil Nadu and Uttar Pradesh. We shall use this
reference group of eight states (the ‘reference states’ for short) from time to time to give the
reader a sense of the diversity of social policies across the country, without giving cumbersome
details for all Indian states. The reference states account for nearly half of India’s population,
and include states with widely varying degrees of involvement in the field of social policy.
However, this group is not meant to be representative of India as a whole, and indeed it is
somewhat biased towards poorer and less well-governed states.

The eight reference states include two (Tamil Nadu and Himachal Pradesh) that are known,
from earlier literature, to have a relatively good record of efficient and equitable provision of
public services, and better social indicators than most other Indian states. They also include
some states (Bihar, Jharkhand and Uttar Pradesh) that are known to have the opposite records
– poor governance, massive corruption, extreme inequality and dismal social indicators.2 When
presenting comparative data for the reference states, we shall place these two groups at the top
and bottom of the list, respectively, to make the contrasts clearer. In between are three
intermediate states (Chhattisgarh, Odisha and Rajasthan) that also used to have a very poor
record of social development, but have significantly improved in recent years. As discussed
further on, there is some evidence of a growing ability of the lagging states to emulate the
achievements of the pioneer states in the field of social policy. This development has an
important bearing on the scope for further progress in the future.
1. The Rights Approach to Social Security
The coverage of the five social security programmes mentioned earlier has markedly expanded
in the last 15 years or so. The trend is illustrated in Table 1, based on the first two rounds of
the India Human Development Survey, a useful source of comparable data for 2004-5 and
2011-2 respectively.3 This expansion has been associated with a gradual recasting of Indian
social security programmes in a rights framework. More precisely, this applies to ‘centrally-
sponsored’ programmes, initiated by the central government. At the state level, some Indian
states have adopted important social security legislations (e.g. the Chhattisgarh Food Security
Act 2012), others have continued to implement social security schemes at the discretion of the
government.

Except for social security pensions, other benefits discussed in this paper are now legal
entitlements under the National Food Security Act 2013 (NFSA) and the National Rural
Employment Guarantee Act 2005 (NREGA). Some (notably midday meals and child
development services) initially became legal rights under Supreme Court orders, and were later
included in the NFSA.4 Aside from these two social security legislations, other important laws
aimed at guaranteeing economic and social rights have been enacted in India during the last
fifteen years, notably the Right to Information Act 2005 and the Right to Education Act 2009.
The basic argument for the rights approach is that it gives some power to the unprivileged
(Drèze, 2004; Joshi, 2010). Impressive as it is in many respects, Indian democracy often gives
little weight to the needs and demands of disadvantaged groups.5 Social security legislations
help them to secure some basic entitlements, e.g. those mentioned in the Indian Constitution
under the Directive Principles of State Policy. The rights approach also helps to impart a
continuity and durability that is often sorely lacking in ad hoc ‘schemes’.
[Table 1 around here]
Table 1
Expansion of Social Security Programmes, 2004-12
2004-5 2011-2
Midday meals: Proportion of children aged 6-14 years who are
receiving a midday meal at school (%)
All children 37 50
Children enrolled in government schools 62 81
Integrated Child Development Services: Proportion of pregnant

women and young children who receive any benefits from ICDS
Pregnant womena 20 53
Young children 27 57

Employment guarantee: Proportion of rural households that worked

on NREGA or similar local public works in the past 12 months (%) 0.5b 29
Public Distribution System: Proportion of households that bought

rice or wheat from the PDS in the past 30 days (%) 27 52
Social security pensions: Proportion of widows and elderly persons

(age 60+) receiving a social security pension (%)
Widows 7 22
Elderly women 7 19
Elderly men 6 22
a
Proportion of women who received some benefit from ICDS during pregnancy, among those
who gave birth during the preceding five years. The next row indicates whether the
corresponding child received any benefits from ICDS.
b
Under ‘Sampoorna Grameen Rozgar Yojana’ and Food-for-Work programmes.
Source: India Human Development Survey (IHDS-1 and IHDS-2), special tabulations.
Giving legal rights to the underprivileged empowers them in at least three ways (Drèze, 2004).
First, it enables them, in principle at least, to defend their rights in court if need be. Second,
legal rights place clear responsibilities on the government, and help to hold the government
accountable to these responsibilities through democratic means (for instance, elections),
without necessarily going to court. Third, legal rights can influence public perceptions and
social norms. These can make a difference even in the absence of organised action. To
illustrate, India’s Right to Information Act has not only created a powerful right of access to
public records for all citizens but also promoted a culture of transparency in public
administration. Government documents are now routinely placed in the public domain even
without anyone invoking the Act to access them.
This is not to say that the rights approach is unproblematic. There are obvious limits to what it
can achieve in the absence of larger administrative, legal and political reforms (Aiyar and
Walton, 2014). Legal entitlements, for instance, lose some of their cutting edge when the legal

system is largely dysfunctional: Indian courts are clogged with millions of pending cases. 6 The
fact that a law is difficult to modify helps to protect social programmes from being diluted or
withdrawn, but it also imparts some rigidity in social policy. Further, national laws often lead
to excessive centralization. We shall return to some of these issues.
2. Targeting, Universalization and Self-Selection
The rights approach has implications for one of the most difficult aspects of many social
security programmes: the identification of eligible persons or households. It may help to
discuss the problem with initial reference to a specific scheme, say social security pensions for
widows. Like many other welfare schemes in India, this one used to be restricted to poor
households: only ‘below poverty line’ (BPL) widows were eligible. If a widow from a well-off
household secured a pension, this would have counted as an ‘inclusion error’. If a poor widow
failed to get a pension, that might have been called an ‘exclusion error’. But exclusion errors
did not matter much, because it was understood that only some widows could be helped (even
within the BPL category), due to financial constraints.
In the rights approach, on the other hand, exclusion errors are a major concern. If all poor
widows are deemed to have a right to social support, then it is important to ensure that as few
as possible are left behind. This can be seen as an argument for relaxing the eligibility criteria,
and perhaps even for ‘universalization’ of widow pensions. Indeed, the Indian experience with
identifying BPL households has been very sobering. Typically, BPL households are sought to
be identified through proxy criteria such as housing characteristics or land ownership. This
process is fraught with conceptual as well as practical problems. At least three independent
national surveys (the National Sample Survey, the third National Family Health Survey and
the first India Human Development Survey) suggest that nearly half of all poor households in

rural India did not have a BPL card around 2005.7 In this situation, one way of reducing
exclusion errors is to cast the net wider than the restrictive and unreliable BPL category, or
even to extend pensions to all widows.
Further, one can take the view that pensions should be regarded, in any case, as a right of all
widows – not just poor widows. Indeed, Indian widows often lead very deprived lives even in
relatively well-off households (Chen, 1998, 2000). Having some secure income of their own
can help them to avoid extreme hardship and enhance their status within the household. This
would reinforce the case for universalization.
A further argument for universalization is that targeting (whether based on the BPL category
or other criteria) is divisive. When everyone has a stake in the system, it has a much greater
chance to work. As Richard Titmuss pointed out long ago, ‘services for the poor will always
be poor services’.8
Universalization, of course, tends to be expensive, since it involves a major expansion of
coverage compared with targeted benefits. In India, about one third of all rural households
belonged to the BPL category in 2005 (Drèze and Khera, 2010, Table 1), implying that, keeping
per-capita benefits constant, universalization would be about three times as expensive as BPL
targeting. This is one reason why the principle did not make much headway in India for a long
time (except in a few states, notably Tamil Nadu). Over time, however, universalization
became more affordable and acceptable in various contexts.
All the social security initiatives discussed in this paper were sought to be restricted to BPL
households at some stage. In every case, BPL targeting proved problematic or undesirable and
was abandoned in due course. In some cases, ‘self-selection’ came to the rescue. For instance,
NREGA is based on the principle of self-selection: every rural adult is entitled to apply for
employment. As discussed below, the self-selection approach seems to work in this case, in the

sense that most of those who are willing to do manual labour for a subsistence wage belong to
poor households. Similarly, there is a strong element of self-selection in the midday meal
programme, in so far as children attending government schools tend to come from
underprivileged households. As far as social security pensions are concerned, the trend today
is towards universalization.
It is mainly the Public Distribution System (PDS) that has triggered much debate on targeting
versus universalization. Targeting led to massive exclusion errors, but universalization of the
PDS is an expensive affair. In recent years, many states have moved towards a more inclusive
PDS, without necessarily going all the way to universalization as Tamil Nadu did (with some
success) a long time ago. Under the National Food Security Act, some states are adopting a
new approach known as the ‘exclusion approach’.9 This approach is based on the idea that
instead of targeting poor households, it is safer to exclude well-off households (using simple
and transparent exclusion criteria), and include everyone else. This reduces the risk of
exclusion errors, but is not as expensive as universalization. In at least some of the states that
have followed this route, the transition from BPL targeting to the exclusion approach has been
associated with a major improvement in the PDS.10 However, excluding well-off households
is not a simple operation by any means. It also tends to create a relatively powerful constituency
of households that have no stake in the PDS and may even try to sabotage it.
In India as elsewhere, the debate on targeting versus universalization is far from settled. The
merits of each approach need to be seen in context – the case for, say, universal pensions is not
the same as the case for a universal PDS. The context includes the state’s capability to collect
and handle household data. There have been significant developments in this regard, such as
the completion of the Socio-Economic and Caste Census in 2011 (in a more systematic and
professional mode than the earlier BPL Censuses) as well as state-specific attempts to create

reliable databases for social security programmes. Detailed studies of these experiments,
however, are still awaited.
Finally, it is important to mention that universalization does not mean ‘uniformity’ of
entitlements. The basic principle is that everyone should be included in the system, but this
does not preclude special provisions for the most vulnerable groups. India’s PDS, for instance,
includes special food entitlements for the poorest of the poor, known as ‘Antyodaya’
households. This does not conflict with the universalization principle.
3. School Meals
Turning to specific programmes, we begin with school meals because it is the simplest and
perhaps most successful of the five initiatives being discussed. The idea is not new: in some
Indian states, notably Tamil Nadu and Gujarat, school meals have been in place for a long time.
They were supposed to be extended across the country under the National Programme of
Nutritional Support for Primary Education (NPNSPE), initiated by the central government in
1995. For several years, however, only dry rations (monthly quotas of wheat or rice) were
distributed to school children under the NPNSPE. On 28 November 2001, in response to a
public-interest petition on the right to food, the Supreme Court of India directed all state
governments to start providing cooked midday meals in primary schools.11
The implementation of Supreme Court orders on midday meals was far from instantaneous.
Initially, many states pleaded financial penury. In some states, the scheme also faced opposition
from teachers and upper-caste parents. However, the progress of the scheme was closely
watched by the Supreme Court, the ‘commissioners’ appointed by the Court for this purpose,
India’s ‘right to food campaign’, and the media. Within three years or so, school meals were in
10

place across the country. Initially, they were very frugal (sometimes just rice and turmeric), but
the menu improved steadily over time.
The scheme received a further boost in 2004, when the National Common Minimum
Programme of the first United Progressive Alliance (UPA) government expressed a strong
commitment to a ‘national cooked nutritious mid-day meal scheme’. In pursuance of that
commitment, there was a substantial increase in financial allocations for midday meals in the
2005-6 Union Budget (see Figure 1 below). This helped the state governments to continue
improving the quality of midday meals. In 2008-9, central assistance for midday meals was
extended to the upper-primary stage (Classes 6 to 8). Today, India’s midday meal scheme
covers more than 100 million children at a cost of Rs 97 billion per year (a little less than 0.1
per cent of GDP).12
Aside from pulses and vegetables, many states have started providing eggs with the school
meal at least once a week – up to five times a week in Tamil Nadu. This has further enhanced
the nutrition value of midday meals. Eggs are not only an excellent source of protein for
growing children but also a compact source of other nutrients such as vitamin A, fat and
calcium – in fact, virtually all essential nutrients except vitamin C (Applegate, 2000; American
Council on Science and Health, 2002). The compactness aspect is significant as young children
have small stomachs and need nutrient-dense food. Eggs also have the advantage of a longer
shelf life than comparable alternatives such as milk or bananas. Further, they cannot be diluted
or adulterated, and the quantity is easy to monitor, reducing the risk of corruption. Some states,
however, are resisting the provision of eggs in school meals, partly for fear of antagonising
upper-caste vegetarian lobbies.
School meals have at least four types of potential social benefits. First, they help to reduce
classroom hunger, and, if the meal is nutritious, contribute to better child nutrition. Second,
11

school meals often enhance school participation, and possibly also pupil achievements, in so
far as better-fed children are better learners. Third, school meals can contribute to social equity.
For instance, they help to counter caste prejudices, by teaching children to sit together and
share a common meal. Class inequalities, too, are attenuated when midday meals facilitate
school participation among unprivileged children. Finally, school meals have significant
gender-related benefits. For instance, they create a valuable source of remunerative
employment for women: about 2.4 million women in India work as cooks or helpers under the
midday meal programme.13 Further, school meals liberate women (especially those employed
outside the house) from the burden of having to feed children during the day.
Some of these benefits are relatively well established, including the positive effects of midday
meals on school attendance, pupil achievements, food security and child nutrition (Drèze and
Kingdon, 2001; Drèze and Goyal, 2003; Khera, 2006, 2013; Drèze and Khera, 2009a; Afridi,
2010, 2011; Afridi, et al., 2013; Singh et al., 2014; Jayaraman and Simroth, 2015; Chakraborty
and Jayaraman, 2016). For instance, one early study (Afridi, 2010) based on random
assignment of dietary recalls between school and non-school days in Madhya Pradesh found
that the school meal reduced the daily calorie deficit among school children by 30 per cent and
the protein deficit by 100 per cent. Similarly, Jayaraman and Simroth (2015), using ‘difference
in differences’ methods involving comparisons between areas within and without school meals
as well as between private and government schools, found significant effects of school meals
on school enrolment, especially in Grade 1. Using a similar method, Chakraborty and
Jayaraman (2016) find that ‘midday meals have a dramatic positive effect on learning
achievement’. All three studies suggest that India’s school-meal programme is highly cost-
effective (in terms of enhancing nutrition, school participation and learning achievements,
respectively).
12

The benefits of school meals in terms of better socialization of children are a little more
speculative. Since restrictions on food sharing (or ‘inter-dining’) play an important part in the
enforcement of the caste system, midday meals probably help to break down caste barriers. On
the other hand, the midday meal itself is sometimes a site of caste discrimination, for instance
when upper-caste parents object to the meal being prepared by a Dalit (scheduled-caste) woman
(Thorat and Lee, 2005; Ramachandran and Naorem, 2013). The value of midday meals as a
means of bringing about greater social equity depends on how these prejudices are handled by
school authorities and the local community.
India’s midday meal scheme still has some way to go in achieving satisfactory quality
standards, particularly in the worst-governed states. On 16 July 2013, twenty-three school
children died in Bihar after ingesting cooking oil that had been carried in a bottle formerly used
for storing insecticide (Khera, 2013). Milder incidents of food poisoning have also occurred in
other states from time to time. Each of these incidents is a setback for the scheme, but also a
wake-up call that often leads to stronger safeguards and tighter supervision. India’s midday
meal scheme is unlikely to be withdrawn – it is now a permanent legal right of school children
under the National Food Security Act. Judging from the experience so far, the scheme is likely
to make a growing contribution to the nutrition, education and wellbeing of Indian children.
4. The Integrated Child Development Services
India’s Integrated Child Development Services (ICDS) provides nutrition, health and pre-
school education services for children under the age of six years. These services are delivered
through child care centres (or anganwadis) staffed by trained local women known as
anganwadi workers and helpers.14 Children in the age group of 3-6 years are expected to go to
the anganwadi every day for food, games, health checkups and so on. For younger children,
13

monthly or weekly ‘take-home rations’ (THR) are provided along with some health services
such as immunization. ICDS also provides some nutrition and health services to pregnant and
lactating women as well as adolescent girls.
Though the ICDS had an earlier start (in 1975) than the midday meal scheme, it has been slower
to take root. The Supreme Court order of 2001 on midday meals, mentioned earlier, also called
for nutritious food to be provided to all children below the age of six years under ICDS. In
December 2006, the Court declared that all children under six (as well as all pregnant women,
lactating mothers and adolescent girls) were entitled to all ICDS services as a matter of right.
These orders led to a massive expansion in the reach of ICDS: there are nearly 1.4 million
anganwadis in India today, serving more than 90 million children (the cost is around 0.1 per
cent of India’s GDP). However, progress in the quality of services has been slower.
The Focus on Children Under Six (FOCUS) report presents a detailed picture of ICDS in 2004,
based on a survey of about 200 randomly-selected anganwadis in six states.15 As with other
programmes, striking inter-state variations emerged. Three of the six sample states (Himachal
Pradesh, Maharashtra and Tamil Nadu) were classified as ‘active’, and the other three
(Chhattisgarh, Rajasthan and Uttar Pradesh) as ‘dormant’ as far as ICDS is concerned. The
active states, by and large, had lively and widely appreciated anganwadis where children were
receiving most of the prescribed services. In the dormant states, however, anganwadis were
seen mainly as feeding centres. The feeding programmes themselves left much to be desired,
especially in states like Uttar Pradesh where low-quality ‘ready-to-eat’ products (supplied by
corrupt contractors) were being served instead of hot cooked food. Even in the dormant states,
however, a majority of mothers appreciated the potential value of the programme.
A resurvey conducted in the same districts ten years later found significant quality
improvements (Centre for Equity Studies, 2016). For instance, two of the formerly dormant
14

states (Chhattisgarh and Rajasthan) had a much more active ICDS programme by 2014.
Overall, taking all six states together, the improvements (e.g. in quantity and quality of food
supplements, maintenance of growth charts, regularity of child attendance) are modest, as
Table 2 illustrates. Nevertheless, it is interesting that quality improvements took place during
a phase of rapid quantitative expansion. As quantitative expansion becomes less urgent, there
is likely to be greater scope for qualitative improvements in the near future.
Table 2
Small Leap Forward: India’s Anganwadis in 2004 and 2014
2004 2014
Proportion (%) of sample AWCs with their own building 33 87
Perceptions of sample mothers
Proportion (%) who stated that their child attends the AWC ‘regularly’ 50 80
Proportion (%) who said that the following services were provided at the
AWC:
Supplementary nutrition 76 79
Immunization 43 81
Home visits 29 46
Referral service 28 24
Growth monitoring 63 70
Pre-school education 45 53
Proportion (%) who:
Were dissatisfied with the quality of food provided at the AWC 24 12

Felt that the quantity of food was adequate 59 79
Reported that PSE activities are taking place at the AWC 45 53
Felt that PSE activities benefit their child 54 82
Felt that ICDS is important for their child’s welfare 48 84
Perceptions of survey teams

Proportion (%) of sample AWCs whose overall functioning was rated as 34 23
‘poor’ or ‘very poor’
Proportion (%) of sample villages where the motivation of mothers to 33 50
send their children to the AWC appears to be ‘high’ or ‘very high’
AWC = anganwadi centre PSE = pre-school education
15

Source: Centre for Equity Studies (2016), based on the FOCUS survey and re-survey. The
FOCUS survey, conducted in 2004, involved unannounced visits to about 200 randomly-
selected anganwadis of six states (Chhattisgarh, Himachal Pradesh, Maharashtra, Rajasthan,
Tamil Nadu and Uttar Pradesh) and interviews with a random sample of about 500 women with
at least one child below six years.
Further evidence of slow but steady progress in the performance of ICDS has recently emerged
from the Rapid Survey on Children 2013-14. Table 3 presents a sample of ICDS-related
indicators for the reference states based on that survey (and also, where available, the
corresponding figures for 2005-6). Here again, striking contrasts between different states are
evident. To illustrate, a useful indicator of the functionality of anganwadis is whether growth
charts are being maintained for children under three. This is now the norm in the leading states
– not only Tamil Nadu and Himachal Pradesh, but also Chhattisgarh and Odisha, where the
standards of functionality of ICDS now look quite similar to those of Tamil Nadu and Himachal
Pradesh. On the other hand, growth charts are still rarely maintained in Bihar, Jharkhand and
Uttar Pradesh.
Looking at change over time, over a period of eight years (2005-6 to 2013-14), the
improvements look relatively modest once again, yet significant considering that they took
place at a time of rapid quantitative expansion of ICDS. More importantly perhaps, there have
been major improvements in some of the states where ICDS used to be virtually non-functional,
such as Bihar. In fact, among all major states, Bihar achieved the largest improvements in child
immunization and coverage of antenatal care during this period, starting of course from a very
low base (Drèze and Khera, 2015b). To put it another way, there are tentative signs of some
sort of ‘convergence’ over time towards the standards of the leading states. One can read a
similar pattern in the context of midday meals: fifteen years ago, midday meals were confined
to a few better-governed states such as Tamil Nadu and Gujarat, but today they are in place
across the country.
16

Among other recent experiences of interest, in terms of emulation of the leading states, is that
of Odisha – one of India’s poorest states, not known for exemplary governance. In a recent
survey of ICDS in four districts of Odisha, discussed in Khera (2015), we found that
anganwadis opened regularly and provided most of the prescribed services: supplementary
nutrition is part of the daily routine, with an improved menu (including eggs twice or thrice a
week), and pre-school education is taking root (at more than three-quarters of anganwadis,
children were able to recite a poem when asked). Health services such as growth monitoring,
immunization and ante-natal care were also provided regularly in the sample anganwadis, with
the help of the local Auxiliary Nurse Midwife (ANM) and Accredited Social Health Activist
(ASHA), India’s frontline health workers. Odisha has also developed a relatively successful
model of decentralized THR production through women’s self-help groups. THR in Odisha
also includes eggs for children below the age of three as well as pregnant and lactating women.
This is an important innovation: children below three are the most critical age group, yet ICDS
has tended to be more focused, so far, on children in the age group of three to six years
(Gragnolati et al., 2006).
17

Table 3 ICDS In The Reference States, 2013-14

Infrastructure indicators (% of Proportion of Children Pregnant women who had Pregnant women
anganwadis with specified facility) AWCs where aged 12-23 ante-natal checkupsa (%) who receive
growth charts of months who supplementary
children below 3 are fully nutrition from the
Functional Toilet Electricity years was immunized At least one Three or anganwadi (%)
baby prepared (%) (%) more
weighing
scale
Tamil Nadu 86 58 83 78 76 (81) 98 (99) 92 (97) 46 (50)
Himachal P. 87 71 46 65 80 (74) 91 (90) 58 (63) 66 (34)
Chhattisgarh 88 35 23 64 67 (49) 96 (89) 80 (55) 65 (64)
Odisha 89 21 25 67 62 (52) 92 (87) 75 (61) 61 (45)
Rajasthan 58 32 15 35 61 (27) 82 (76) 51 (41) 34 (17)
Bihar 57 20 14 17 60 (33) 85 (34) 33 (17) 22 (1)
Jharkhand 58 27 23 20 65 (34) 81 (61) 47 (36) 47 (35)
Uttar Pradesh 50 (46)b 20 12 47 (23) 62 (67) 39 (26) 24 (10)
Reference 63 38 28 32 59 (38) 79 (69) 52 (41) 34 (21)
statesc
INDIA 72 43 32 44 65 (44) 85 (77) 63 (51) 41 (21)
a
Base: Women who had a live birth during the 35 months preceding the survey. b In Uttar Pradesh, many anganwadis are run from private homes.
c
Population-weighted average of state-specific figures.
AWC = Anganwadi centre; ANM = Auxiliary Nurse Midwife; AWW = Anganwadi worker; ASHA = Accredited Social Health Activist
Sources: Rapid Survey on Children 2013-14, state factsheets (available at http://wcd.nic.in/acts/rapid-survey-children-rsoc-2013-14); figures in brackets are the
corresponding figures for 2005-6 (from the third National Family Health Survey), where available.
18

Turning from processes to outcomes, evidence has recently emerged of the impact of ICDS on
child nutrition and related indicators. The expansion of ICDS during the last ten years has been
associated with accelerated improvement in anthropometric indicators of child nutrition as well
as in other outcomes relevant to ICDS, such as child vaccination and institutional deliveries
(Drèze, 2015; Sinha, 2015; Narayan, 2016). Several econometric studies based on household
survey data have also found evidence of positive effects of ICDS on child nutrition (Kandpal,
2011; Hazarika and Viren, 2013; Jain, 2015; Mittal, 2015; Nandi and Laxminarayan, 2016;
Nandi et al., 2016). For instance, Kandpal (2011), using propensity score matching methods to
control for endogenous programme placement, finds that ICDS increases children’s average
height-for-age z-score by 6 per cent. Using similar methods, but looking specifically at daily
feeding of children below the age of two years, Monica Jain (2015) also finds that nutrition
supplements have a substantial effect on height, especially among girls. Both studies point out,
based on tentative cost-benefit calculations, that ICDS has substantial economic returns. Note
also that both studies are based on data from the third National Family Health Survey, collected
in 2005-6 – at an early stage of recent efforts to expand and improve ICDS. If the programme
already had substantial effects on child nutrition and high economic returns at that time, it is
likely to be all the more valuable today.
5. The National Rural Employment Guarantee Act16
The use of public works as a tool of social security, particularly in the context of drought relief,
has a long history in India and elsewhere (Drèze, 1990; Subbarao et al., 2013). Compared with
other approaches, public works have two useful features: self-selection (people decide for
themselves whether they need social support) and the creation (or at least possible creation) of
productive assets. After a prolonged drought in the early 1970s, when relief works reached an
19

unprecedented scale, the state of Maharashtra in India enacted a pioneering Employment
Guarantee Act in 1977. Maharashtra’s ‘employment guarantee scheme’ employed nearly half
a million persons on an average day in the late 1970s and 1980s, but declined sharply from the
early 1990s onwards (Mahendra Dev and Ranade, 2001).
In August 2005, a National Rural Employment Guarantee Act (NREGA) inspired by
Maharashtra’s employment guarantee scheme was passed unanimously by the Indian
Parliament.17 The Act came into force on 2 February 2006 in 200 of India’s poorest districts,
and was extended to the whole country on 1 April 2008. Under the Act, any adult residing in
rural areas who demands work has to be employed on local public works within 15 days.
Failing that, an unemployment allowance is due. Other important entitlements under the Act
include payment of wages within 15 days, basic worksite facilities, and minimum wages. The
minimum wage guarantee, however, was withdrawn in 2009, when the central government
started exercising its right to fix NREGA wages in different states instead of endorsing the
states’ own minimum wages. The practice, since then, is to revise NREGA wages upward every
year based on state-specific price indexes (keeping real wages constant).
The NREGA includes a number of provisions aimed at ensuring that the local administration
is accountable to rural workers. That is the purpose, for instance, of the unemployment
allowance. Similarly, if wages are not paid within 15 days, workers are entitled to
compensation. A penalty clause also makes NREGA functionaries liable to a fine if they fail
to do their duty under the Act. Central and state governments, however, have tended to resist
these accountability provisions, and short of going to court (which is cumbersome and
expensive), NREGA workers have no obvious way of putting pressure on the government to
enforce them. As a result, lack of administrative accountability has been a serious problem for
NREGA since its inception (Ambasta et al., 2008; Khera, 2011a; Aggarwal, 2016).
20

A couple of examples may help. One of the main challenges in making good use of NREGA
funds is corruption. Early on, ‘muster rolls’ (work attendance sheets) used to be routinely
inflated by corrupt functionaries or middlemen as a means of siphoning off NREGA funds.
Later on, safeguards were introduced, such as mandatory ‘social audits’ of NREGA works.
These safeguards, though not always enforced as intended, helped to curb corruption, but their
effectiveness has been severely compromised by the reluctance of state governments to take
action against corrupt officials.18
Another example relates to the lack of timeliness in wage payments. Delays in wage payments
have plagued NREGA ever since the central government made it compulsory for wages to be
paid through banks or post offices, in mid-2008. Bank payments greatly helped to contain
corruption, but they also led to much foot-dragging: with the corruption tap turned off or at
least tightened, many functionaries lost interest in the scheme. Since there are many steps in
the payment process (filing of attendance data, work measurement, preparation of payment
advice, and so on), serial foot-dragging can cause long delays. Bank processes themselves often
take time and add to the chain of delays. Here again, the resilience of the problem reflects a
lack of accountability at every level: while NREGA workers wait week after week (or even
month after month) for their wages, no-one is held responsible.
In spite of these and other hurdles, the NREGA has led to some major achievements. According
to official data, the programme generated more than two billion person-days of work every
year from 2008-9 to 2013-4, with a peak of 2.8 billion (54 days each for 53 million rural
households) in 2009-10 – see Table 4.19 Independent data from the National Sample Survey
suggest lower figures, but the gap between the two sources has narrowed over time. By 2011-
2, about 68 to 78 per cent of official NREGA employment was reflected in National Sample
Survey data, and about 90 per cent of it was reflected in the second Indian Human Development
Survey.20 About half of all NREGA workers are women (much in contrast with low rates of
21

female participation in the Indian workforce as a whole), and more than half belong to
scheduled castes or scheduled tribes. The effectiveness of NREGA ‘self-targeting’ is
reasonably well established.21
Table 4: NREGA Factsheet
2009-10 2014-15
Employment generation (person-days)
Total (million) 2836 1662
Per household employed under NREGA 54 40
Per rural householda 17 10
Work participation
Number of persons employed (million) n/a 62
Proportion of rural households employed (%) 31 25
Share of marginalized groups in NREGA employment

(%)
Women 48 55
Scheduled tribes 21 17
Scheduled castes 30 22
Expenditure on NREGA
Total expenditure (Rs billion) 379 360
Total expenditure as a proportion of GDP (%) 0.6 0.3
Share of wages in total expenditure (%) 67 67
a
Total person-days divided by number of rural households as per Census of India 2011.
Sources: Government of India (2012), Table A, p.4, for 2009-10. Official NREGA website
(nrega.nic.in), ‘At a Glance’ section, for 2014-5. The exchange rate for Indian rupees (Rs) was
around 46 per US dollar in 2009-10 and 65 per US dollar in 2014-5.
22

Table 5, based on National Sample Survey data, presents a sample of NREGA indicators for
the reference states. As earlier studies have noted (Dutta et al., 2012), NREGA employment
and expenditure are not particularly well correlated with poverty across states – the self-
targeting process operates mainly within states. The main reason is that some of the poorer
states, like Bihar, have low levels of NREGA employment due to political inertia and weak
administrative capabilities. Similarly, the participation of women in NREGA, though high in
most states, is quite low (even below the mandatory minimum of 33 per cent) in some of the
poorest states, such as Bihar and Uttar Pradesh. In 2009-10, NREGA wage rates were lower
than market wages in some states and higher in others, with little difference between the two
at the all-India level.
[Table 5 around here.]
As one might expect, the expansion of public employment under NREGA led to some
acceleration in the growth of agricultural wages. After stagnating for the first half of the 2000s,
agricultural wages grew at around 3 to 4 per cent per year in real terms during the second half
(Drèze and Sen, 2013, Table 7.1). The break was particularly sharp for women workers. In
itself, this pattern does not establish that NREGA was the driving force behind the acceleration
of wage growth, but several econometric studies confirm that the programme had a substantial
impact on rural wages (Azam, 2011; Berg et al., 2012; Imbert and Papp, 2015b; Muralidharan
et al., 2016). By raising wages in the private sector, NREGA also generates important second-
round income increases for rural workers (Imbert and Papp, 2015b). One recent study, based
on a large-scale randomized controlled trial (RCT) in Andhra Pradesh, even suggests that the
second-round effects account for the bulk of the impact of NREGA on the earnings of low-
income households (Muralidharan et al., 2016). In some areas, upward pressure on private-
sector wages has generated some opposition to NREGA from employers, including large
farmers and the corporate sector.
23

Large-scale employment generation, effective self-targeting and second-round wage increases
have produced substantial benefits for the rural poor. These benefits are best-documented for
Andhra Pradesh (and to some extent Tamil Nadu), where the state government has actively
promoted NREGA and extensive survey data are available.22 For other states, the evidence is
patchy, but consistent with the hypothesis that NREGA has a substantial impact on rural
poverty. Klonner and Oldiges (2014), for instance, estimate (based on a regression
discontinuity model) that NREGA reduces lean-season poverty by half for scheduled-caste and
scheduled-tribe households. However, in states with low levels of NREGA employment
generation and high levels of corruption, such as Bihar, the impact of the programme on rural
poverty is very limited as things stand (Dutta et al., 2014).
On the productive value of NREGA assets, there is a major information gap. Media reports,
focusing as they tend to do on negative stories, often create an impression that most NREGA
assets are useless. This impression is not borne out by recent research. 23 The most
comprehensive study on this, a detailed survey of 4,100 NREGA assets scattered over 100
villages of Maharashtra (Ranaware et al., 2015) found that 87% of the sample works were
functional and 75% contributed directly or indirectly to better agriculture. An overwhelming
majority (90%) of the users of these NREGA works considered them “very useful” or
“somewhat useful”. Similarly, an analysis of 100 randomly-selected NREGA wells in
Jharkhand (Bhaskar et al., 2016) estimates that the wells have a financial rate of return of 6 per
cent or so in real terms, based on their impact on agricultural productivity alone. It is, however,
difficult to generalize from this limited evidence.
24

Table 5: NREGA in the Reference States, 2009-10
Person-days of Per-capita Proportion of rural households working on Share of women Average wage rate
employment expenditure on NREGAa (%) in NREGA (Rs/day)
NREGA (Rs) employment (%)
Per Per rural All Scheduled Scheduled NREGA Casual

employed household households castes tribes labour
household
Tamil Nadu 58 25 555 34 52 (29) 83 72 111
Himachal P. 64 22 936 33 41 39 46 110 140
Chhattisgarh 44 24 723 48 44 52 49 82 69
Odisha 38 7 281 22 22 32 36 106 76
Rajasthan 69 47 1,133 62 65 82 67 87 126
Bihar 32 7 214 10 19 (9) 30 98 79
Jharkhand 43 18 586 19 27 20 34 98 101
Uttar Pradesh 58 14 389 16 33 (14) 22 100 94
INDIA 54 17 464 25 34 42 48 90 93
a
Figures in brackets may be unreliable due to small sample (less than 100 households). The overall proportion of rural households employed on NREGA in
2009-10 according to NSS data (25% – last row of this table) is a little lower than the corresponding figure from official NREGA statistics (31% - see Table 4).
Sources: Person-days of employment: Government of India (2012), Table 6.4, and Government of India (2009). Other figures: Dutta et al. (2012) Tables 1, 2,
6 and 7 (all based on National Sample Survey data, except for ‘per-capita expenditure on NREGA’, calculated from official data available at http:\ \nrega.nic.in).
25

The value of NREGA goes much beyond these relatively standard economic benefits. For
instance, the programme has helped to revive the institutions of local democracy in rural India,
such as Gram Panchayats (village councils) and Gram Sabhas (village assemblies). These
institutions, indeed, play a central role in the implementation of NREGA at every step, from
planning of works to social audits. Similarly, NREGA makes important contributions to
women’s empowerment and gender equality, bearing in mind that a vast majority of women in
rural India have very limited opportunities for remunerated employment (Srivastava and
Srivastava, 2010; Drèze and Sen, 2013). One survey found that only 30 per cent of female
NREGA workers had earned any cash income other than NREGA wages during the preceding
three months. Most collected their own wages (today, they would also have their own bank
account), and kept them.24 Further, NREGA has important liberating effects for rural workers
in general: it makes them less dependent on local landlords for survival, enhances their
bargaining power vis-à-vis private employers, helps them to resist oppression from the upper
castes, and also gives them an opportunity to organise around common interests. As Indrajit
Roy (2014) observes, NREGA helps workers ‘to live and work in the villages without adhering
to its habits [including] caste-based discrimination and the servility associated with it’.25
While NREGA appears to have substantial benefits whenever employment is actually
provided, the basic principle of “work on demand” is far from being fulfilled. Indeed, there is
a large unmet demand for NREGA work (Dutta et al., 2012). Other NREGA entitlements are
also routinely violated, from basic worksite facilities to payment within 15 days (Khera,
2011a). In areas where workers were relatively organised to start with, or where NREGA was
used as an opportunity to organise rural workers, the programme has tended to do better.26 The
Act was partly intended to facilitate the formation of workers’ organisations in rural areas, but
this has happened to a very limited extent so far.
26

The future of NREGA is somewhat uncertain. Ten years on, there is no reason to give up the
basic idea that people without an adequate means of livelihood can be helpfully employed on
local public works. The experience so far also shows that effective implementation of NREGA
has wide-ranging benefits, well beyond poverty alleviation. However, realising this potential
(especially in the poorer states) poses huge organisational and political challenges. Much
remains to be done to achieve the initial objectives of the Act.
6. The Public Distribution System
India’s public distribution system (PDS), initiated in the 1940s, provides monthly rations of
rice or wheat at subsidized prices to eligible households through a network of government-
regulated outlets. In some states, PDS outlets also supply other subsidized food commodities
such as pulses, oil, sugar and salt. More recently, some states have started experimenting with
maize and millets. The PDS also provides subsidized kerosene, across the country, but our
main focus in this section is on food and especially on foodgrains.
As mentioned earlier, the question whether the PDS should be universal (at least in rural areas)
or targeted has been a subject of lively debate. Until the nineties, though universal in principle,
the PDS actually had a patchy coverage and a major urban bias (Howes and Jha, 1992; Jha and
Ramaswami, 1992). By the late nineties, as arguments for explicit targeting gained ground, the
central government decided to restrict the PDS to poor households, to be identified by state
governments. Under the ‘targeted PDS’ (TPDS), there were two categories of ration cards, for
‘below poverty line’ (BPL) and ‘above poverty line’ (APL) households respectively. A third
category, ‘Antyodaya’ (poorest of the poor), was created in 2000. In 2001, APL issue prices
were increased, effectively driving out APL households from the system in most states, though
they continued to receive subsidized kerosene. Later on, as market prices increased while PDS
27

issue prices did not, APL distribution picked up again. Some states, notably Tamil Nadu,
rejected the TPDS from the beginning in favour of a universal system.
The APL-BPL system proved problematic from the beginning. Stringent caps on the number
of BPL cards set by the central government, combined with a very unreliable identification
process, led to massive inclusion and exclusion errors (Ram et al., 2009; Drèze and Khera,
2010). The caps themselves were somewhat arbitrary, as they were tied to statistical poverty
estimates that had little to do with food security (Drèze 2010b). In recent years, many states
have moved towards a more inclusive (even near-universal) PDS coverage using their own
resources (see Tables 1 and 7). This move was consolidated by the enactment of the National
Food Security Act 2013, of which the PDS is one component, with an enhanced coverage of
75% in rural areas and 50% in urban areas.
Another important debate relates to corruption in the PDS. The difference between market
prices and PDS issue prices creates a strong incentive for intermediaries (especially the
‘dealers’ who manage PDS outlets) to embezzle foodgrains from the system and sell them on
the open market. ‘Leakages’ can be estimated by matching National Sample Survey (NSS) data
on household purchases from the PDS with official data on PDS offtake (Khera, 2011b).
Estimates based on successive NSS surveys from 1999-2000 to 2011-12 suggest that all-India
leakages were as high as 50 per cent or so at the beginning of that period, declining to 40 per
cent or so by the end of it. Estimates based on the India Human Development Surveys suggest
a somewhat faster rate of decline in PDS leakages, from 49 per cent in 2004-5 to 32 per cent
in 2011-2 (Drèze and Khera, 2015a).
NSS-based leakage estimates for the reference states are presented in Table 6. In 2004-5, the
leakages were massive in all these states except Tamil Nadu, and to a lesser extent Himachal
Pradesh. By 2011-12, the leakages were lower, but more importantly, the decline in leakages
28

was driven by huge improvements in a few states that are known to have initiated PDS reforms
during that period. Here again, we see evidence of an ability of some of the lagging states to
emulate the achievements of the pioneer states.
Chhattisgarh is an important and well-researched example (Drèze and Khera, 2010; Puri, 2012;
Garg, 2013; Krishnamurthy et al., 2014; Joshi et al., 2016). In 2004-5, about half of the PDS
grain in Chhattisgarh was diverted by corrupt intermediaries (Table 6). Around that time,
however, the political leadership started taking interest in fixing the PDS.27 Chhattisgarh
emulated two important initiatives from Tamil Nadu’s reasonably successful PDS: an
expansion in the coverage (though Chhattisgarh stopped short of universalization) and a
reduction in issue prices. Both have the effect of increasing pressure on the system - from
below. Other voice-enhancing measures included simplifying entitlements (e.g., 35 kg per
month per card at a token price of Rs. 3/kg). This simplification enabled greater awareness of
entitlements, further enhanced by other innovative methods (e.g., by marking cardholders’
homes with a symbol indicating their entitlements).
The state government also took measures to streamline the supply chain. When the reforms
were initiated, PDS outlets were under the control of a corrupt nexus of private dealers and
politicians. The monopoly of private dealers was broken by handing over PDS outlets to
community-based institutions, such as Gram Panchayats, cooperative societies and women’s
self-help groups. To improve their viability, official commissions were enhanced. Transport
agencies were separated from distribution agencies, so that the latter keep an eye on the former.
Computerization of records also helped, especially by tightening accounting practices. Finally,
credible grievance redressal mechanisms with effective disciplinary action were put in place.
By 2011-12, estimated leakages had come down to less than 10 per cent. In three successive
surveys of the PDS in Chhattisgarh (in 2011, 2013 and 2016 respectively), we found that
cardholders were receiving more than 95 per cent of their foodgrain entitlements.28
29

The Chhattisgarh experience provided a roadmap for PDS reform in other states. Over time,
Bihar, Jharkhand, Madhya Pradesh and Odisha borrowed from the Chhattisgarh experience
with significant results. In Odisha, between 2004-5 and 2011-12, leakages came down from 75
per cent to 25 per cent or so. In other states, where only part of the package of reforms has been
implemented, the results are less dramatic: e.g., Jharkhand implemented the voice-enhancing
measures, but not much else. In the same period, leakages have halved in Jharkhand (from 85
per cent to 44 per cent – see Table 6).
Around the same time as the 68th Round of the NSS, on which the latest leakage estimates are
based, we conducted a survey of the PDS (hereafter PDS Survey 2011) in nine states, including
the eight reference states.29 Selected findings are presented in Table 7, along with those of a
2013 survey called Public Evaluation of Entitlement Programmes (PEEP), which had a wider
focus but included PDS data. The figures suggest that, by then, BPL and Antyodaya households
were receiving the bulk of their PDS entitlements, not only in Tamil Nadu and Himachal
Pradesh but also in other states, though significant shortfalls remained in Bihar, Jharkhand and
Uttar Pradesh. In the APL quota, large leakages continued, a pattern that can also be inferred
from NSS data (Drèze and Khera, 2015a).
The National Food Security Act 2013 gave further momentum to PDS reforms in many states.
The Act includes three sets of entitlements: maternity entitlements in the form of cash transfers,
nutritious food for children (under the midday meal and ICDS programmes), and subsidized
grain through the PDS. The PDS-related provisions in the NFSA create transparent entitlements
(5 kg per person per month) across the country, with broad coverage (75% in rural areas) and
low issue prices (Rs. 3/kg for rice and Rs. 2/kg for wheat).30 They also incorporate many of the
reforms that proved effective in Chhattisgarh. The implementation of the NFSA (including the
phasing out of the leaky APL quota) is an important opportunity to consolidate and extend the
PDS improvements achieved in specific states over the past decade. Early evidence suggests
30

that some states at least are making relatively good use of this opportunity. Madhya Pradesh
and West Bengal, for instance, have recently joined the league of active PDS reformers, and
continued progress is also observed in many of India’s poorest states (Drèze et al., 2016).
[Tables 6 and 7 around here.]
31

Table 6: Estimates of PDS Leakages and Poverty Impact
Estimated PDS leakages (%) Estimated impact of the

PDS on rural poverty, 2009-
10a (percentage reduction in
2004-5 2011-2 the poverty-gap index)
Tamil Nadu 7 12 61
Himachal Pradesh 27 27 35
Chhattisgarh 52 9 39
Odisha 76 25 23
Rajasthan 94 61 12
Bihar 91 24 4
Jharkhand 85 44 13
Uttar Pradesh 58 58 11
INDIA 54 42 18
a
Using the all-India ‘Tendulkar poverty line’ for rural areas (Rs 673/month in 2009–10).
Source: Drèze and Khera (2013, 2015a). The impact of the PDS on rural poverty was evaluated
from National Sample Survey (NSS) data, by comparing the poverty-gap index with and
without adding the implicit value of food subsidies to NSS estimates of ‘monthly per-capita
expenditure’. Leakages were estimated by matching NSS data on household purchases from
the PDS with official data on PDS offtake.
32

Table 7: The PDS in the Reference States

PDS coveragea Proportion of Purchase-entitlement Proportion of Proportion of
respondents who did ratio in last 3 months, respondents who feel respondents who would
not get any PDS grain BPL householdsb (%) that the quality of PDS prefer cash (%)
in the last 3 months (%) grain is ‘poor’ (%)
2011 2013 2011 2013 2011 2013 2011 2013

Tamil Nadu Universal 3 1 92 96 17 47 11 18
Himachal Pradesh Universal 0 1 92-100 100 4 17 9 12
Chhattisgarh Near-universal 2 0 95 98 14 0 2 9
Odisha Expanded 0 0 97-100 98 14 15 6 1
Rajasthan Expanded 1 1 86-100 99 19 27 15 12
Bihar Targeted 27 4 45 69 31 13 54 6
Jharkhand Targeted 3 2 71 75 19 14 22 23
Uttar Pradesh Targeted 2 10 77-88 86 11 5 34 29
EIGHT STATESc - 7 5 76-81 87 17 16 28 18
a
‘Targeted’: coverage is restricted to BPL households. ‘Expanded’: coverage has been extended beyond BPL households by state government,
but is not universal.
b
Average foodgrain purchases of BPL households, as a proportion of entitlements. In some states, BPL entitlements in 2011 were a little ambiguous
due to temporary extra allocations under Supreme Court orders; for these states, a range is presented instead of a point estimate.
c
Sources: PDS Survey 2011 and PEEP Survey 2013.31 Base: households with a BPL or Antyodaya card, except for purchase-entitlement ratio (BPL
households only). For further details, see Khera (2011c) and Drèze and Khera with the PEEP team (2014).
33

Evidence of the impact of the PDS on nutrition is limited as things stand. Kaul (2013) studies
the effect of the PDS on calorie consumption and finds that a one per cent increase in the value
of the implicit subsidy increases calorie consumption by 0.14%. Krishnamurty et al. (2015)
find that the PDS in Chhattisgarh has helped households to diversify their food consumption
and increase their intake of protein-rich food items. Aside from this, there is little evidence of
significant nutrition gains from the PDS as things stand (Kochar, 2005; Tarozzi, 2005; Kaushal
and Muchomba, 2013; Rahman, 2014; Narayanan and Gerber, 2016). This may reflect
methodological difficulties in establishing these effects as well as the dismal functioning of the
PDS until reforms were initiated in the mid-2000s; available studies focus largely on the pre-
reform period.
The PDS, however, does seem to have a substantial impact on rural poverty. In earlier work
based on NSS data (Drèze and Khera, 2013), we found that implicit transfers through the PDS
reduced the all-India rural ‘poverty gap index’ by one fifth or so in 2009-10 (see Table 6). The
corresponding reduction was very small in states that had a poorly functioning PDS at that
time, such as Bihar (less than 5%), but quite large in states with a well-functioning PDS,
notably Tamil Nadu (61%) and Chhattisgarh (39%).32 Quite likely, recent PDS reforms in the
laggings states have consolidated the programme’s impact on rural poverty.
7. Social Security Pensions
Non-contributory pensions for widows, the elderly and disabled persons are provided under
three sister schemes sponsored by India’s central government: the Indira Gandhi National
Widow Pension Scheme (IGNWPS), Indira Gandhi National Old Age Pension Scheme
(IGNOAPS) and Indira Gandhi National Disability Pension Scheme (IGNDPS). All three are
part of the National Social Assistance Programme (NSAP) initiated in 1995.33 Many states use
34

their own resources to expand these programmes beyond the central guidelines (either in terms
of coverage or in terms of pension amounts), or to supplement them with pension schemes of
their own. The financial allocation for NSAP in the last Union Budget, for 2016-7, is around
0.06 per cent of India’s GDP.
As with the PDS before the National Food Security Act, central support for social security
pensions extends – to this day – to BPL households only. This helps to contain the costs, but it
is problematic for two reasons. First, as discussed earlier, the process of identifying BPL
households is extremely unreliable. Second, the aim of social security pensions is not to support
poor households but to help vulnerable persons. Even among households above the poverty
line, widows and the elderly are often exposed to severe deprivation or insecurity. This applies
especially to widows, who have a very low status in Indian society (Chen and Drèze 1995;
Chen, 1998, 2000). Further, the poverty line in India is really a kind of destitution line: it falls
far short of the minimum required for dignified living in any plausible sense of the term (Drèze
and Sen, 2013). Thus, even if the process of identifying BPL households were perfectly reliable
and widows received fair treatment within the household, there would still be an argument for
extending social security pensions beyond BPL households.
Even for BPL households, the central contribution to social security pensions is very modest.
For instance, old-age pensions have stagnated at 200 rupees per month since 2006 even as
inflation reduced their real value by half or so (today, 200 rupees roughly corresponds to 10
kilograms of rice, or one day’s earnings at the minimum wage). To address this problem of
restricted coverage and low pension amounts, many states have expanded pension benefits
using their own resources (see Table 8). Even the enhanced amounts, however, leave
pensioners well below the poverty line unless they have other sources of earnings.
[Table 8 around here.]
35

Table 8: Social Security Pensions in the Reference States

Proportion of Approximate Old-age Proportion of Reported delays in pension payments (%) Main mode of
confirmed pension pension pensioners who had payment
No delay Less than a More than a
beneficiaries coverageb amount to pay a bribe to get
monthc month or no
in the official (%) (Rs/month) their pension
pattern
lista (%) sanctioned (%)
Tamil Nadu 97 26 1,000 51 10 98 2 Postal money orders
Himachal Pradesh 100 47 400 8 0 1 99 Postal money orders
Chhattisgarh 100 34 300 11 2 96 4 Post office accounts
Odisha 100 79 300 19 99 99 1 Cash through GP
Rajasthan 92 31 500-750 15 1 3 97 Post office accounts
Bihar 97 72 300-500 37 0 0 100 Cash through GP
Jharkhand 93 51 400 38 3 50 50 Bank accounts
Uttar Pradesh 100 34 300 50 0 0 100 Bank accounts
EIGHT STATESd 98 44 458 38 9 28 72
GP = Gram Panchayat (village council)
a
Percentage of persons in the 64 sample villages’ official lists of pensioners (about 3,253 names) who confirmed that they were getting a pension.
b
Proportion of sample households with at least one member aged above 60 where someone is getting a pension (widow or old-age).
c
Including ‘no delay’.
d
Source: PEEP Survey 2013; the survey included a special module on pensions, which covered 735 pensioners in the reference states (the initial
target was 12 per village, in 64 villages) selected at random from the pension lists. The figures in the first column are based on a full verification
of the pension lists.
36

India’s pension schemes have faced significant implementation problems. (Nayak et al., 2002;
Bhattacharya et al., 2015). The pensioner selection process tends to be slow and cumbersome,
forcing applicants to wait anxiously for years, if they are able to apply at all. In some states,
payments are irregular and unpredictable, or made once or twice a year instead of every month.
Further, given the limited reach of India’s banking system, collecting pensions is an ordeal for
many pensioners, especially if they are not in good health. Post offices (also used for pension
payments) are closer to homes, but the convenience comes at a price: in some states, corrupt
post-office employees often expect an inducement.
Delays in pension payments deserve special mention, because they tend to defeat the purpose
of providing some security to the pensioners. In November 2001, the Supreme Court directed
the state governments to ensure that pensions are paid by the 7th of each month without fail.
Fifteen years later, this goal is still elusive, as Table 8 illustrates. In three states (Tamil Nadu,
Chhattisgarh and Odisha), pensions are almost always paid within a month of being due, but in
other states, long delays are the norm. The only state where pensions are paid on time every
month is Odisha, where payments are made in cash on the 15th of each month at the local Gram
Panchayat office and the process is strictly monitored. This effective payment system is one
aspect of Odisha’s special commitment to social security pensions in recent years (Chopra and
Pudussery, 2014). Like PDS reforms in Chhattisgarh, Odisha’s recent experience with social
security pensions is an interesting example of how social programmes can be improved, even
in states that are not particularly well governed, when they become a political priority.
Delayed payments aside, a series of recent studies indicate that India’s pension schemes are
doing well in some important respects (Dutta et al., 2010; Gupta, 2013; Kaushal, 2013; Chopra
and Pudussery, 2014; Drèze and Khera, 2014; Marulasiddappa et al., 2014; Bhattacharya et al.,
2015). Modest as they are, the pensions tend to be highly valued by the recipients. Most of
them live in penury, and the monthly pension is a chance to enjoy small comforts – relieving
37

their pain with some medicine, getting their sandals repaired or simply avoiding hunger. The
administrative costs are very low, compared with other social security initiatives reviewed in
this paper. Further, recent surveys found little evidence of serious corruption in pension
schemes. To illustrate, in a detailed verification of randomly-selected pension lists in ten states,
we found – in 2013 - that 98 per cent of the 3,789 persons on the lists were alive and receiving
their pension (see Table 8, first column, for the reference states).34 There are leakages here and
there, for instance when post-office employees take a cut to disburse pensions, but nothing like
the scams that plague many other forms of government expenditure in India.
In the light of these findings, there is a case for consolidating the recent trend towards more
inclusive and effective pension schemes for widows, the elderly and disabled persons.
However, the central government seems to be taking a different view. The NSAP budget, after
rising steadily between 2010-11 and 2014-5, remained much the same (in money terms) for
three years in a row. The central government, it seems, lost interest in NSAP to focus on a new
pension scheme (the Pradhan Mantri Atal Pension Yojana), initiated in 2015 in a very different
mode – contributory and largely self-financed. Under this scheme, there is a minimum
contribution period of 20 years, when monthly contributions are to be paid through auto-debit
from a bank account. The modalities are not particularly advantageous, even for someone with
a secure monthly income.35 For those who live on the margin of subsistence, like so many
beneficiaries of the non-contributory schemes discussed earlier, the package is even less
attractive. So far, there have been few takers: about 2.4 million, compared with 32 million for
NSAP according to official data.36
As things stand, central and state governments are pulling in different directions. State
governments are consolidating non-contributory pensions under NSAP, at the cost of bearing
an ever-larger share of the payments. The central government is promoting its own contributory
38

pension scheme, without much success so far. A clear and effective national strategy for social
security pensions is yet to emerge.
8. What About the Costs?
Brief reference has already been made to the cost of some of the social programmes discussed
in this paper. A fuller view of the costs, as a ratio of GDP, is presented in Figure 1 for midday
meals, ICDS, NREGA and social security pensions. This figure must be read in light of the fact
that India’s GDP was growing at about 7.5 per cent per year in real terms over this period.
Also, the figure focuses on central government expenditure. This accounts for the bulk of total
public expenditure on these four programmes, because the contribution of state governments
is only 10 per cent or so for NREGA, the costliest item by far. But state governments do make
a significant additional contribution to public expenditure on specific programmes, e.g. 25 per
cent of the total (recently raised to 40 per cent) in the case of school meals.
[Figure 1 around here.]
39

1.00
0.90
Central government expenditure (as % of GDP)
0.80
Four schemes combined
0.70
0.60
0.50
0.40
NREGA
0.30
0.20
MDM ICDS
0.10
Social security pensions
0.00
2004-5 2005-6 2006-7 2007-8 2008-9 2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 2016-7
Source: Budget at a Glance (various years) for social spending and Finance Ministry's annual Economic Survey for
GDP figures (the 2016-17 GDP figure is from the Reserve Bank of India's Handbook of Statistics on Indian Economy).
Figure 1: Public expenditure on selected schemes (% of GDP)

40

School meals, ICDS and social security pensions cost relatively little as a proportion of GDP
– somewhere around 0.1 per cent in each case. NREGA expenditure rose steadily in the initial
years of the programme, until 2009-10. Since then, it has remained more or less constant in
money terms, implying a sustained decline in real terms, although NREGA expenditure picked
up again in the last Union Budget. The combined central government expenditure on these four
programmes was around 0.5 per cent of GDP in 2016-7.
The National Food Security Act is not included in Figure 1, because the Act is difficult to cost.
India’s ‘food subsidy’ is large – about 1 per cent of GDP. This figure is often quoted as the
cost of the NFSA, but this is misleading because the food subsidy is partly farmer subsidy,
partly consumer subsidy, and partly a price stabilization scheme (Deuss and Narayanan, 2017).
Even if just half of the food subsidy is counted as the cost of NFSA, however, it would still
cost 0.5 per cent of GDP – as much as the other schemes combined.
India does not have much else in place by way of social security programmes. In international
perspective, it spends relatively little under this heading. According to the Asian Development
Bank (2013), India spent 1.7% of GDP on ‘social protection’ in 2009, compared with an
average of 3.4% in the lower-middle income countries of Asia (where India belongs), 5.4% in
China and 19.2% in Japan.37 Even among low-income countries of Asia, the corresponding
figure is 2.6% - much higher than in India. Clearly, there is some scope for further social
security initiatives in the near future.
9. The Cash Dilemma
There is an influential view, particularly among economists, that some or all of the programmes
discussed here should be replaced with cash transfers. The most radical version of this view is
41

that they should all be replaced with a universal basic income (UBI).38 This would be hard to
justify, however, considering that some of these programmes (e.g. school meals) have well-
documented benefits that go much beyond what equivalent cash transfers might be able to
deliver. More plausible proposals involve retaining some of these programmes and replacing
others with cash transfers, universal or targeted.
UBI proposals for India are somewhat futuristic, since their estimated costs range between 3.5
and 10 per cent of India’s GDP (Bardhan, 2016; Joshi, 2016). Creating this sort of fiscal space
for cash transfers will take many years, if it is desirable at all. Meanwhile, in practical terms,
the argument for cash is relevant mainly to the PDS. The National Food Security Act does not
preclude food subsidies being replaced with a ‘food security allowance’ in cash. Many
economists feel that this would reduce costs, since the PDS remains quite expensive (in terms
of transaction costs as well as leakages) even in its reformed state. Narayanan (2011) presents
an insightful summary of the debate, and stresses the need to assess these arguments in context.
We explored poor people’s views on the respective merits of cash and in-kind transfers in the
PDS Survey 2011 as well as in the PEEP Survey 2013.39 The proportion of respondents who
expressed a preference for cash over food was only 28 per cent in 2011, and even lower – 18
per cent – in 2013 (Table 7, last column). In both surveys, one clear pattern emerged: where
the PDS delivered, people preferred food. Among the main reasons for preferring food were
lack of trust that cash transfers would keep up with the price level, fear that cash might be
misused, inadequate banking facilities, and concerns that food prices might increase if the PDS
is dismantled. Replacing the PDS with cash transfers is likely to be unpopular, unless the
transfers are much larger than the implicit value of food subsidies, but then the move to cash
transfers may not lead to significant savings for the government.
Pilot cash-transfer schemes have been launched during the last few years in Chandigarh, Dadra
and Nagar Haveli, Delhi, and Puducherry – small urbanised areas thought to be particularly
42

well placed for the transition. According to media reports and one early survey, these pilot
schemes are chaotic and unpopular.40 Interim findings of an evaluation conducted by J-PAL
(2016) for the Indian government are also sobering. In phase 2 of the evaluation (mid-2016),
nearly one-fifth of the entitled respondents did not receive any cash. In Puducherry, the
situation got worse between the two phases — in phase 1, up to 25 per cent received no cash,
in phase 2, this rose to 37 per cent. Like our own surveys, the J-PAL evaluation finds that a
majority of respondents prefer food over cash. Even if the pilots eventually succeed, there are
good reasons to be cautious in extending them to other areas – especially in the poorer states –
where the conditions of success (including a good banking infrastructure) may not be met.
Meanwhile, little attention is given to the possibility of expanding cash transfer schemes that
have already proved their effectiveness. Social security pensions, discussed earlier, is one
example. Another important example is maternity entitlements. Two Indian states, Odisha and
Tamil Nadu, have implemented maternity entitlements with good effect (Balasubramanian and
Sundari Ravindran, 2012; Khera, 2015). The NFSA requires maternity entitlements to be
extended across the country, but this is yet to happen.
The debate on cash versus kind is far from settled, but what is now widely agreed at least is
that it is possible, in India, to put in place an effective system of income support – in cash or
kind – for poor households. That itself is a significant development.
10. Concluding Remarks
Most of the rich countries today have a comprehensive social security system that encompasses
education, health, shelter, pensions and unemployment benefits. This applies not only to
western Europe but also to other OECD countries (e.g. Australia, Canada, Iceland, Israel,
Japan, New Zealand), with a few notable exceptions such as the United States. This historical
43

transition was associated with a rejection of the ‘poor law’ mindset, whereby social support
was restricted to a minority of poor people and confined to the bare minimum. This
comprehensive approach played an important role in ensuring the acceptability of higher
taxation as a means of financing these provisions. In contrast with the poor-law approach,
where a minority of rich people were expected to support a minority of poor, modern welfare
states build on principles of reciprocity and solidarity among large sections of the population
(Horton and Gregory, 2010; Judt, 2010).
Many developing countries, especially those with a relatively democratic political system, are
now heading in a similar direction. For instance, Brazil’s democratic constitution explicitly
recognises universal rights to health care and elementary education, and spells out the
implications of these rights in some detail. In India, however, a comprehensive vision for social
policy is yet to enter public debates in a significant manner (though there are traces of it in the
Constitution, under the Directive Principles of State Policy). One reason for this is the severe
resource constraint. India’s per-capita GDP is still very low, the income tax base is small (about
3% of the population, compared with 7-8% in Brazil and China) and the overall tax-GDP ratio
is also low in international perspective. However, the resource barrier has eased significantly
in the recent past due to high growth rates: public revenue at constant prices is more than four
times as large today as it was twenty-five years ago (Drèze and Sen, 2013). If this trend
continues, comprehensive social security will soon look much more affordable than it does
today.
While resource constraints are easing over time, political support for active social policies is
still far from assured. India is a very divided society, where privileged groups are not always
convinced that they have a responsibility towards the underprivileged. The country’s deep-
rooted caste hierarchy often reinforces the social distance between the privileged and the rest
(Drèze and Sen, 2013). And yet, as discussed in this paper, there have been significant steps in
44

recent years (and earlier, in specific Indian states) towards a more extensive social security
system. There is no guarantee that this trend will continue, but the historical experience of other
democratic countries suggests that further expansion of social security provisions in the near
future is at least a strong possibility.
Two important lessons emerge from the empirical evidence reviewed in this paper. One is that
the standards of implementation of social security programmes in India can and do improve
over time. This is at variance with the narrative of cynicism that pervades the mainstream media
on this subject (midday meals, for instance, rarely make headlines except when there are
incidents of food poisoning). It is, of course, part of the media’s job to expose corruption and
waste where they occur. Opposition parties and civic organisations also have good reason to
demand much higher standards of delivery. However, it is also important to note the progress
that has been made, whether in terms of reforming the PDS, or making anganwadis come to
life, or improving the quality of midday meals. Indeed, cynicism tends to obscure the scope for
further progress in this field and stifle political support for social security initiatives.
The second lesson, which relates to the first, is that states with poor standards of governance
are capable, over time, of emulating the achievements of the leading states. Not so long ago,
Kerala stood out among Indian states as a case of relatively effective delivery of essential public
services. Today, Tamil Nadu and Himachal Pradesh are more or less in the same league. Many
other states also have significant achievements to their credit: Chhattisgarh has turned the PDS
around, Odisha has lively anganwadis, and even Bihar has made a leap forward in some aspects
of child development. In absolute terms, the quality of public services in India’s worst-
governed states are still abysmal in many cases, but recent experience suggests that this
situation is far from immutable.
Following on this, we end by noting an important limitation of the rights approach to social
security. As mentioned earlier, the rights framework has made an important contribution to the
45

emergence of a semblance of social security system in India. The conversion of many social
benefits into legal entitlements has helped to enhance functionality and accountability in many
social programmes, including some (like the PDS) that were often considered beyond repair.
Having said this, the rights approach has a tendency to create some rigidity and centralization
in social policies. Laws and court orders tend to be difficult to modify. That is a strength in so
far as it protects social benefits from being withdrawn, but it is also a weakness: modifications
are often required or desirable with the passage of time, or simply because the initial legislation
turned out to be defective in some way. National laws also tend to lead to centralization, by
giving sweeping powers to the central government. Sometimes these powers can be well used,
e.g. to nudge reluctant states towards more active social policies. In some cases, however, they
have led to counter-productive restraints on the freedom of state governments to design social
programmes in tune with their own circumstances. The National Rural Employment Guarantee
Act, for instance, was intended as an enabling legislation (a foundation for state-specific
‘employment guarantee schemes’), but in recent years it has become more and more
centralized, with the union government remote-controlling and micro-managing ever more
numerous aspects of the programme. Of course, states can also enact their own laws, as
Chhattisgarh has already done with the Chhattisgarh Food Security Act. But there is a potential
tussle here between the centre and the states, which is yet to be resolved.41
In short, India is still far from having found its feet as far as social security programmes are
concerned. However, the initiatives, experiences and debates of the last fifteen years or so
(along with the sustained growth of public revenue) have opened up vast possibilities in this
field for the relatively near future. Making good use of these possibilities is one of the major
challenges for Indian democracy today.
46

References
Afridi, Farzana (2010), ‘Child Welfare Programs and Child Nutrition: Evidence from a
Mandated School Meal Program in India’, Journal of Development Economics, 92.
Afridi, Farzana (2011), ‘The Impact of School Meals on Student Participation in Rural
India’, Journal of Development Studies, 47.
Afridi, F., Barooah, B. and Somanathan, R. (2013), ‘School Meals and Classroom Effort:
Evidence from India’, Working Paper, International Growth Centre, London School of
Economics.
Afridi, F., Mukhopadhyay, A., and Sahoo, S. (2013), ‘Female Labor Force Participation and
Child Education in India: Evidence from the National Rural Employment Guarantee
Scheme’, Discussion Paper 13-03, Economics and Planning Unit, Indian Statistical
Institute, New Delhi; forthcoming in IZA Journal of Labor and Development.
Aggarwal, Ankita (2016), ‘The MGNREGA Crisis: Insights from Jharkhand’, Economic and
Political Weekly, 28 May.
Aggarwal, A., Gupta, A., and Kumar, A. (2012), ‘Evaluation of NREGA Wells in Jharkhand’,
Economic and Political Weekly, 1 September.
47

Aiyar, Y., and Walton, M. (2014), ‘Rights, Accountability and Citizenship: Examining India’s
Emerging Welfare State’, Working Paper, Centre for Policy Research, New Delhi
(available at www.accountabilityindia.in).
Ambasta, P., Vijay Shankar, P.S., and Shah, M. (2008), ‘Two Years of NREGA: The Road
Ahead’, Economic and Political Weekly, 23 February.
American Council on Science and Health (2002), The Role of Eggs in the Diet: Update (New
York: ACSH).
Applegate, Elizabeth (2000), “Introduction: Nutritional and Functional Roles of Eggs in the
Diet”, Journal of the American College of Nutrition, 19.
Asian Development Bank (2013), The Social Protection Index: Assessing Results for Asia and
the Pacific (Manila: Asian Development Bank).
Azam, Mehtabul (2011), ‘The Impact of Indian Job Guarantee Scheme on Labor Market
Outcomes: Evidence from a Natural Experiment’, Discussion Paper 6548, Institute for
the Study of Labour, Bonn.
Balasubramanian, P., and Sundari Ravindran, T.K. (2012), ‘Pro-Poor Maternity Benefit
Schemes and Rural Women’, Economic and Political Weekly, 23 June.
Bardhan, Pranab (2016), ‘Basic Income in a Poor Country’. Available online
(http://www.ideasforindia.in/article.aspx?article_id=1694).
48

Basu, Alaka M. (2012), ‘More Benefit than Cost: For Women, the NREGA Would Bring
Important Social Gains’, The Telegraph, 27 March.
Berg, E., Bhattacharya, S., Durgam, R., and Ramachandra, M. (2012), ‘Can Rural Public
Works Affect Agricultural Wages? Evidence from India’, Working Paper 2012-05,
Centre for the Study of African Economics, University of Oxford.
Bhaskar, A., Gupta, S. and Yadav, P. (2016), Well Worth the Effort, Value of MGNREGA
Wells in Jharkhand, Economic and Political Weekly, 7 May.
Bhattacharya, S., Jos, M. M., Mehta, S.K., and Murgai, R. (2015), ‘From Policy to Practice:
How Should Social Pensions Be Scaled Up?’, Economic and Political Weekly, 4 April;
reprinted in Drèze (2016).
Bhuwania, Anuj (2016), Courting the People: Public Interest Litigation in Post Emergency
India (Cambridge: Cambridge University Press).
Carswell, G., and De Neve, G. (2010), ‘Women at the Crossroads: Implementation of
Employment Guarantee Scheme in Rural Tamil Nadu’, Economic and Political Weekly,
28 December.
Carswell, G., and De Neve, G. (2014), ‘MGNREGA in Tamil nadu: A Story of Success and
Transformation?’, Journal of Agrarian Change, 14.
49

Centre for Equity Studies (2016), Progress of Children Under Six: Revisiting ICDS in the
FOCUS Districts (New Delhi: Centre for Equity Studies).
Chakraborty, T., and Jayaraman, R. (2016), ‘School Feeding and Learning Achievement:
Evidence from India’s Midday Meal Program’, Discussion Paper 10086, Institute for
the Study of Labor, Bonn.
Chen, Marty (ed.)(1998), Widows in India: Social Neglect and Public Action (New Delhi:
Sage).
Chen, Marty (2000), Perpetual Mourning: Widowhood in Rural India (New Delhi: Oxford
University Press).
Chen, M., and Drèze, J. (1995), ‘Recent Research on Widows in India’, Economic and Political
Weekly, 30 September.
Chopra, S., and Pudussery, J. (2014), ‘Social Security Pensions in India: An Assessment’,
Economic and Political Weekly, 10 May; reprinted in Drèze (2016).
Citizen’s Initiative for the Rights of Children Under Six (2006), Focus on Children Under Six
(New Delhi: Secretariat of the Right to Food Campaign).
Corbridge, S., Harris, J., and Jeffrey, C. (2012), India Today: Economy, Politics and Society
(Cambridge: Polity Press).
50

Dasgupta, Aparajita (2013), ‘Can the Major Public Works Policy Buffer Negative Shocks in
Early Childhood? Evidence from Andhra Pradesh, India’, Working Paper 112, Young
Lives Project, Oxford University.
Deininger, K., and Yanyan Liu (2013), ‘Welfare and Poverty Impacts of India’s National Rural
Employment Guarantee Scheme: Evidence from Andhra Pradesh’, Discussion Paper
01289, International Food Policy Research Institute, Washington, DC.
Desai, S. B., Dubey, A., Joshi, B. L., Sen, M., Shariff, A. and Vanneman, R. (2010), Human
Development in India: Challenges for a Society in Transition (New Delhi: Oxford
University Press).
Desai, S., Vashishtha, P., and Joshi, O. (2015), Mahatma Gandhi National Rural Employment
Guarantee Act: A Catalyst for Rural Transformation (New Delhi: National Council of
Applied Economic Research).
Deuss, A., and Narayanan, S. (2017), ‘Food Subsidies in India: Revisiting the Estimates’,
Indira Gandhi Institute of Development Research, mimeo.
Dhaktode, N.B. (2016), ‘Democratic Accountability and Empowerment of Dalits in India: The
Case of Institutionalizing Social Audit in Telengana and Andhra Pradesh’, M.Phil.
dissertation, Tata Institute of Social Sciences, Mumbai.
Drèze, Jean (1990), ‘Famine Prevention in India’, in Drèze, J., and Sen, A.K. (eds.)(1990), The
Political Economy of Hunger (Oxford: Oxford University Press).
51

Drèze, Jean (2004) ‘Democracy and the Right to Food’, Economic and Political Weekly, 24
April.
Drèze, Jean (2006), ‘Universalization with Quality: ICDS in a Rights Perspective’, Economic
and Political Weekly, 26 August.
Drèze, Jean (2010a), ‘Employment Guarantee and the Right to Work’, in Jayal, N.G., and
Mehta, P.B. (eds.) (2010), The Oxford Companion to Politics in India (New Delhi:
Oxford University Press).
Drèze, Jean (2010b), ‘Poverty Estimates vs Food Entitlements’, The Hindu, 24 February.
Drèze, Jean (2014), ‘Learning from NREGA’, The Hindu, 23 August.
Drèze, Jean (2015), ‘Small Leap Forward in Child Health’, The Hindu, 16 September.
Drèze, Jean (2016), Social Policy: Essays from Economic and Political Weekly (New Delhi:
Orient BlackSwan).
Drèze, J.P., and Goyal, A. (2003), ‘The Future of Mid-day Meals’, Economic and Political
Weekly, 1 November.
Drèze, J.P., Gupta, P., Khera, R., and Pimenta, I. (2016), ‘Food Security Act: How are India’s
Poorest States Faring?’, available at www.ideasforindia.in
52

Drèze, J.P., and Khera, R. (2009a), ‘Mid-Day Meals in Primary Schools’, in Kumar, A., and
Singh, A.P. (eds.) (2009), Elementary Education in India: Issues and Challenges (New
Delhi: Uppal).
Drèze, J.P., and Khera, R. (2009b), ‘The Battle for Employment Guarantee’, Frontline, 3
January; reprinted in Khera (2011a).
Drèze, J.P., and Khera, R. (2010), ‘The BPL Census and a Possible Alternative’, Economic and
Political Weekly, 27 February.
Drèze, J.P., and Khera, R. (2012), ‘Regional Patterns of Human and Child Deprivation in
India’, Economic and Political Weekly, 29 September.
Drèze, J.P., and Khera, R. (2013), ‘Rural Poverty and the Public Distribution System’,
Economic and Political Weekly, 16 November.
Drèze, J.P., and Khera, R., with the PEEP team (2014) ‘A PEEP at another India’, survey
report, Indian Institute of Technology Delhi; partly published in Outlook, 24 March.
Drèze, J.P., and Khera, R. (2015a), ‘Understanding Leakages in the Public Distribution
System’, Economic and Political Weekly, 14 February.
Drèze, J.P., and Khera, R. (2015b), ‘Child Development: Kerala Tops, Gujarat Flops, Bihar
Hops’, available at www.ndtv.com.
53

Drèze, J.P., Khera, R., and Pudussery, J. (2015), ‘Food Security: Bihar on the Move’, Economic
Drèze, J.P., and Kingdon, G.G. (2001), ‘School Participation in Rural India’, Review of
Development Economics, 5.
Drèze, J.P., and Sen, A.K. (2013), An Uncertain Glory: India and Its Contradictions (New
Delhi: Penguin).
Dutta, P., Howes, S., and Murgai, R. (2010), ‘Small but Effective: India’s Targeted
Unconditional Cash Transfers’, Economic and Political Weekly, 25 December.
Dutta, P., Murgai, R., Ravallion, M., and van de Walle, D. (2012), ‘Does India’s Employment
Guarantee Scheme Guarantee Employment?’, Economic and Political Weekly, 21
April; reprinted in Drèze (2016).
Dutta, P., Murgai, R., Ravallion, M., and van de Walle, D. (2014), Right to Work? Assessing
India’s Employment Guarantee Scheme in Bihar (Washington, DC: World Bank).
Garg, Samir (2013), “Twin Strategies for Food Security and Productive Inclusion: PDS
Reforms in Chhattisgarh, India”, Poverty in Focus, No 25, International Policy Centre
for Inclusive Growth, Brasilia.
54

Government of India (various years), Union Budget: Budget at a Glance (New Delhi: Ministry
of Finance). Available online (http://indiabudget.nic.in/glance.asp).
Government of India (2009), The National Rural Employment Guarantee Act 2005 (NREGA),
Annual Report April 2008 - March 2009 (New Delhi: Ministry of Rural Development).
Available online (http://nrega.nic.in/NREGA_eng_2008-09.pdf).
Government of India (2012), MGNREGA Sameeksha: An Anthology of Research Studies on
the Mahatma Gandhi National Rural Employment Guarantee Act, 2005 (New Delhi:
Orient Blackswan).
Government of India (2017), Economic Survey 2016-17 (New Delhi: Finance Ministry).
Gragnolati, M., Bredenkamp, C., Das Gupta, M., Lee, Y., and Shekar, M. (2006), ‘ICDS and
Persistent Undernutrition’, Economic and Political Weekly, 41 (12), 25 March.
Gupta, Aashish (2013), ‘Old-age Pension Scheme in Jharkhand and Chhattisgarh’, Economic
Harriss, J., Jeyaranjan, J., and Nagaraj, K. (2010), ‘Land, Labour and Caste Politics in Rural
Tamil Nadu in the 20th Century: Iruvelpattu (1916-2008)’, Economic and Political
Weekly, 45(31): 47-61, 31 July.
Hazarika, G., and Viren, V. (2013), ‘The Effects of Early Child Developmental Attendance on
Future School Enrollment in Rural North India’, Economics of Education Review, 34.
55

Himanshu, and Sen, A. (2013), ‘In-kind Food Transfers-I: Impact on Poverty’, Economic and
Political Weekly, 16 November.
Horton, T., and Gregory, J. (2010), ‘Why Solidarity Matters: The Political Strategy of Welfare
Design’, The Political Quarterly, 81.
Howes, S., and Jha, S. (1992), ‘Urban Bias in the Indian Public Distribution System’, Economic
and Political Weekly, 27(9):1022-1030, 9 May.
Imbert, C., and Papp, J. (2011), ‘Estimating Leakages in India’s Employment Guarantee’, in
Khera (2011a).
Imbert, C., and Papp, J. (2015a), ‘Estimating Leakages in India’s Employment Guarantee: An
Update’, Working Paper, Research Institute for Compassionate Economics; available
at riceinstitute.org
Imbert, C., and Papp, J. (2015b), ‘Labor Market Effects of Social Programs: Evidence from
India’s Employment Guarantee’, American Economic Journal: Applied Economics,
7(2): 233-263.
Jain, Monica (2015), ‘India’s Struggle Against Malnutrition: Is the ICDS Program the
Answer?’, World Development, 67.
56

Jayaraman, R., and Simroth, D. (2015), ‘The Impact of School Lunches on Primary School
Enrollment: Evidence from India’s Midday Meal Scheme’, Scandinavian Journal of
Economics, 117.
Jenkins, R., and Manor, J. (2016), Politics and the Right to Work: India’s Mahatma Gandhi
National Rural Employment Guarantee Act (New Delhi: Orient Blackswan).
Jeyaranjan, J. (2011), ‘Women and Pro-Poor Policies in Rural Tamil Nadu: An Examination
of Practices and Responses’, Economic and Political Weekly, 22 October.
Jha, S., and Ramaswami, B. (1992), ‘Urban Bias in the Indian Public Distribution System’,
Economic and Political Weekly, 9 May.
Joshi, Anuradha (2010), ‘Do Rights Work? Law, Activism, and the Employment Guarantee
Scheme’, World Development, 38.
Joshi, A., Patnaik, B., and Sinha, D. (2016), ‘Credibility and Portability? Lessons from CORE
PDS Reforms in Chhattisgarh’, Economic and Political Weekly, 10 September.
Joshi, Vijay (2016), India’s Long Road: The Search for Prosperity (Oxford: Oxford University
Press).
J-PAL (2016), ‘Process Monitoring of Direct Benefit Transfer in the PDS’, second quarterly
report submitted to the Food Department, Government of India.
57

Judt, Tony (2010), Ill Fares the Land (New York: Penguin).
Kandpal, Eeshani (2011), ‘Beyond Average Treatment Effects: Distribution of Child Nutrition
Outcomes and Program Placement in India’s ICDS’, World Development, 39(8):1410-
1421.
Kalra, Aarushi (2016), ‘Cash Transfers: Lost in Transactions’, The Tribune, 27 January.
http://www.tribuneindia.com/news/comment/cash-transfers-lost-in-
transactions/215949.html
Kaul, Tara (2013), ‘Household Responses to Food Subsidies: Evidence from India’, Working
Paper, University of Maryland, College Park.
Kaushal, Neeraj (2013), ‘How Public Pension affects Elderly Labor Supply and Well-being:
Evidence from India’, World Development, 56.
Kaushal, N., and Muchomba, F. (2013), ‘How Consumer Price Subsidies Affect Nutrition’,
Working paper 19404, National Bureau of Economic Research, Cambridge, MA.
Khera, Reetika (2006), ‘Mid-Day Meals in Primary Schools: Achievements and Challenges’,
Economic and Political Weekly, 18 November.
Khera, Reetika (2008), ‘Empowerment Guarantee Act’, Economic and Political Weekly, 30
August.
58

Khera, Reetika (ed.) (2011a), The Battle for Employment Guarantee (New Delhi: Oxford
University Press).
Khera, Reetika (2011b), ‘Trends in Diversion of Grain from the Public Distribution System’,
Economic and Political Weekly, 21 May.
Khera, Reetika (2011c), ‘Revival of the Public Distribution System: Evidence and
Explanations’, Economic and Political Weekly, 5 November; reprinted in Drèze (2016).
Khera, Reetika (2013), ‘Mid-Day Meals: Looking Ahead’, Economic and Political Weekly, 10
August; reprinted in Drèze (2016).
Khera, Reetika (2014), ‘Cash vs. In-kind Transfers: Indian Data Meets Theory’, Food
Policy, 46.
Khera, Reetika (2015), ‘Children’s Development: Baby Steps in Odisha’, Economic and
Political Weekly, 3 October.
Khera, R., and Muthiah, K. (2010), ‘Slow but Steady Success’, The Hindu, 25 April.
Khera, R., and Nayak, N. (2009), ‘Women Workers and their Perceptions of National Rural
Employment Guarantee Act’, Economic and Political Weekly, 24 October; reprinted in
Drèze (2016).
59

Klonner, S., and Oldiges, C. (2014), ‘Safety Net for India’s Poor or Waste of Public Funds?
Poverty and Welfare in the Wake of the World’s Largest Job Guarantee Program’,
Discussion Paper 564, Department of Economics, University of Heidelberg.
Kochar, Anjani (2005), ‘Can Targeted Food Programmes Improve Nutrition? An Empirical
Analysis of India’s Public Distribution System’, Economic Development and Cultural
Change, 54.
Kohli, Atul (2012), Poverty Amid Plenty in the New India (Princeton: Princeton University
Press).
Krishnamurthy, P., Pathania, V. S., and Tandon, S. (2014), ‘Public Distribution System
Reforms and Consumption in Chhattisgarh: A Comparative Empirical Analysis’,
Economic and Political Weekly, 22 February.
Krishnamurthy, P., Pathania, V. S., and Tandon, S. (2015), ‘Food price subsidies and Nutrition:
Evidence from State Reforms to India’s Public Distribution System’, UC Berkeley
Public Law Research Paper No. 2345675, UC Berkeley.
Mahendra Dev, S. and Ranade, A. (2001), ‘Employment Guarantee Scheme and Employment
Security’, in Mahendra Dev, S., Antony, P., Gayathri, V. and Mamgain, R. P. (eds.)
(2001), Social and Economic Security in India (New Delhi: Institute of Human
Development).
60

Marulasiddappa, M., Raonka, P., and Sabhikhi, I. (2014), ‘Social Security Pensions for
Widows and the Elderly’, Indian Journal of Human Development, 8.
Mittal, Nitya (2015), ‘Essays in the Economics of Health in India’, PhD thesis, Delhi School
of Economics.
Mookherjee, Dilip (2014), ‘MNREGA: Populist Leaky Bucket or Successful Anti-poverty
Programme?’, available at ideasforindia.in
Muralidharan, K., Niehaus, P., and Sukhtantar, S. (2016), ‘General Equilibrium Effects of
(Improving) Public Employment Programs: Experimental Evidence from India’,
Department of Economics, University of California, San Diego, processed.
Nandi, A., Ashok, A., Kinra, S., Behrman, J.R., and Laxminarayan, R. (2016), ‘Early
Childhood Nutrition is Positively Associated with Adolescent Educational Outcomes:
Evidence from the Andhra Pradesh Child and Parents Study (APCAPS)’, Journal of
Nutrition, 146, 1-8.
Nandi, A., and Laxminarayan, R. (2016), ‘Can Early Childhood Nutrition Improve Adult
Educational Attainment in India? Evidence from the ICDS Program’, SSRN Working
Paper (available at papers.ssrn.com).
Narayan, Swati (2016), ‘Towards Equality in Healthcare: Trends Over Two Decades’,
Economic and Political Weekly, 19 March.
61

Narayanan, Sudha (2008), ‘Employment Guarantee, Women’s Work and Childcare’,
Economic and Political Weekly, 1 March.
Narayanan, Sudha (2011), ‘A Case for Reframing the Cash Transfer Debate in India’,
Economic and Political Weekly, 21 May; reprinted in Drèze (2016).
Narayanan, S., and Das, U. (2014), ‘Women Participation and Rationing in the Employment
Guarantee Scheme’, Economic and Political Weekly, 15 November.
Narayanan, S., and Gerber, N. (2016), ‘Social Safety Nets for Food and Nutritional Security in
India’, ZEF Working Paper 146, Centre for Development Research, University of
Bonn.
National Commission for Enterprises in the Unorganised Sector (2006), Social Security for
Unorganised Workers, report to the Prime Minister; available online
(http://nceuis.nic.in/nceus_earlier_reports.htm).
Nayak, R., Saxena, N.C., and Farrington, J. (2002), ‘Reaching the Poor’, Working Paper 175,
Overseas Development Institute, London.
Nehra, Shikha (2016), ‘Ration to Cash, A Harsh Transition’, India Together,
http://indiatogether.org/ration-to-cash-a-harsh-transition-poverty
62

Pattenden, Jonathan (2015), ‘Class and Social Policy: The National Rural Employment
Guarantee Scheme in Karnataka, India’, Journal of Agrarian Change, online preview,
20 August.
Puri, Raghav (2012), ‘Reforming the Public Distribution System: Lessons from Chhattisgarh’,
Economic and Political Weekly, 4 February.
Rahman, Andaleeb (2014), ‘Revival of the Public Distribution System: Expansion and
Outreach’, Economic and Political Weekly, 17 May.
Ram, F., Mohanty, S.K., and Ram, U. (2009), ‘Understanding the Distribution of BPL Cards:
All India and Selected States’, Economic and Political Weekly, 14 February.
Ramachandran, V., and Naorem, T. (2013), ‘What it Means to be a Dalit or Tribal Child in our
Schools’, Economic and Political Weekly, 2 November; reprinted in Drèze (2016).
Ranaware, K., Das, U., Kulkarni, A., and Narayanan, S. (2015), ‘MGNREGA Works and Their
Impacts: A Study of Maharashtra’, Economic and Political Weekly, 28 March; reprinted
in Drèze (2016).
Ravi, S., and Engler, M. (2015), ‘Workfare as an Effective Way to Fight Poverty: The Case of
India’s NREGS’, World Development, 67.
Rodriguez, Valerian (ed.) (2002), The Essential Writings of B. R. Ambedkar (New Delhi:
Oxford University Press).
63

Roy, Indrajit (2014), ‘Reserve Labor, Unreserved Politics: Dignified Encroachments under
India’s National Rural Employment Guarantee Act’, Journal of Peasant Studies, 41.
Roy, Indrajit (2015), ‘Class Politics and Social Protection: The Implementation of India’s
MGNREGA’, Working Paper 46, Effective States and Inclusive Development
Research Centre, University of Manchester.
Singh, A., Park, A., and Dercon, S. (2014), ‘School Meals as a Safety Net: An Evaluation of
the Midday Meal Scheme in India’, Economic Development and Cultural Change, 62.
Singh, Prerna (2016), How Solidarity Works for Welfare: Subnationalism and Social
Development in India (Cambridge: Cambridge University Press).
Sinha, Dipa (2015), ‘Maternal and Child Health: Inching Ahead, Miles to Go’, Economic and
Political Weekly, 5 December.
Srivastava, N., and Srivastava, R. (2010), ‘Women, Work and Employment Outcomes in Rural
India’, Economic and Political Weekly, 10 July.
Subbarao, K., del Ninno, C., Andrews, C., and Rodríguez-Alas, C. (2013), Public Works as a
Safety Net (Washington, DC: World Bank).
Tarozzi, Alessandro (2005), ‘The Indian Public Distribution System as Provider of Food
Security: Evidence from Child Anthropometry in Andhra Pradesh’, European
Economic Review, 62.
64

Thorat, S., and Lee, J. (2005), ‘Caste Discrimination and Food Security Programmes’,
Economic and Political Weekly, 24 September; reprinted in Drèze (2016).
Tillin, L., Deshpande, R., and Kailash, K. K. (eds.) (2015), Politics of Welfare: Comparisons
Across Indian States (New Delhi: Oxford University Press).
Tillin, L., Saxena, A., and Sisodia, Y. S. (2015), ‘Comparing the Politics of Food Subsidies in
Chhattisgarh and Madhya Pradesh’, in Tillin et al. (2015).
Titmuss, Richard (1967), ‘Universal and Selective Social Services’, New Statesman, 15
September.
Yanyan Liu and Barrett, C. (2013), ‘Heterogeneous Pro-Poor Targeting in the National Rural
Employment Guarantee Scheme’, Economic and Political Weekly, 9 March; reprinted
in Drèze (2016).
65

1
Maternity benefits (cash grants of Rs 6,000 per child, or US$ 90 at current exchange rates)
are a legal entitlement of Indian women under the National Food Security Act 2013. This
provision of the Act, however, is yet to be implemented.
2
On regional patterns of social development in India, see Drèze and Khera (2012), Drèze and
Sen (2013), Singh (2016), and the literature cited there.
3
In the case of midday meals (mandatory under Supreme Court orders from 28 November
2001), the phase of major expansion was 2002-4. Table 1, where the baseline is 2004-5, does
not capture that initial expansion, but it did happen within the last 15 years. In the case of
pensions, the expansion visible in Table 1 accelerated sharply in many states during the last
five years.
4
Strictly speaking, only some components of the Integrated Child Development Services (e.g.
nutrition supplements) are covered under the NFSA. Supreme Court orders, however, state that
all children below the age of six years are entitled to all ICDS services.
5
This aspect of Indian democracy is discussed in Kohli (2012), Corbridge et al. (2012), Drèze
and Sen (2013), among others. Dr. B.R. Ambedkar, architect of the Indian Constitution, had
anticipated it in a series of speeches and writings (see e.g. Rodriguez 2002).
6
Litigation is also expensive for poor people, and legal aid facilities in India are scarce. On the
other hand, India’s legal system provides ample room for ‘public interest litigations’ (PILs),
and these have been used from time to time to put pressure on the government to implement
social legislations (Bhuwania, 2016).
7
See Drèze and Khera (2010), Table 1, and Desai et al. (2010), Table A.13.1a.
8
Titmuss (1967), quoted in Horton and Gregory (2010), p. 273. This argument for
universalization is an important basis of the welfare state in Europe (Judt, 2010).
66

9
This approach is made possible by the expanded coverage of the PDS under the Act, discussed
further on: 75% in rural areas and 50% in urban areas, adjusted state-wise so that the coverage
is higher in the poorer states (e.g. 85% in rural Bihar).
10
See e.g. Drèze, Khera and Pudussery (2015) on Bihar’s recent experience.
11
The order applies to government and government-aided private schools. This public interest
litigation is known as the ‘right to food case’ (People’s Union for Civil Liberties vs.
Government of India and Others, Writ Petition [Civil] 196 of 2001), and continues to this day.
The original demand was that the government’s gigantic food stocks should be used to protect
people from hunger.
12
Strictly speaking, this is the cost to the central government; state governments also contribute
(about 25 per cent of total costs, recently raised to 40 per cent).
13
March 2011 figures from the central government’s midday meal website (mdm.nic.in).
14
Some states have two anganwadi workers and one helper, but the norm is to have one each.
15
See Citizens’ Initiative for the Rights of Children Under Six (2006). On the status of ICDS
at that time, see also Drèze (2006) and the studies cited there.
16
The literature on NREGA is too vast for us to be able to do justice to it in this short paper.
On the experience of the initial years, see Khera (2011a), Government of India (2012), and the
literature cited there. There have been rapid developments during the last few years, and also a
flurry of research. Important contributions include Dutta et al. (2012, 2014), Afridi et al.
(2013), Yanyan Liu and Barrett (2013), Klonner and Oldiges (2014), Mookherjee (2014),
Bhaskar, Gupta and Yadav (2016), Imbert and Papp (2015b), Ranaware et al. (2015),
Muralidharan et al. (2016), among others.
17
For a brief history of the National Rural Employment Guarantee Act (officially known today
as Mahatma Gandhi National Rural Employment Guarantee Act), see Drèze (2010a). Full
67

details of the Act, Schedules and Guidelines are available on the official NREGA website
(www.nrega.nic.in).
18
One state where the government often did take action against corrupt officials is Andhra
Pradesh. It is perhaps not an accident that NREGA has done relatively well in that state (see
the literature cited in note 23).
19
See also Government of India (2012), p. 4, and the ‘public data portal’ on the official NREGA
website (www.nrega.nic.in).
20
For further discussion, see Drèze (2014). The broad consistency between official NREGA
data and the second Indian Human Development Survey can also be verified from data
presented in Desai et al. (2015). On the gap between official figures and National Sample
Survey data, see also Imbert and Papp (2011, 2015a).
21
See particularly Yanyan Liu and Barrett (2013); also Drèze and Khera (2009b).
22
See Deininger and Yanyan Liu (2013), Dasgupta (2013), Ravi and Engler (2015),
Muralidharan et al. (2016), among others. On Tamil Nadu, see e.g. Narayanan (2008), Harriss
et al. (2010), Khera and Muthiah (2010), Jeyaranjan (2011), Carswell and De Neve (2010,
2014).
23
See e.g. Aggarwal et al. (2012), Bhaskar, Gupta and Yadav (2016), Ranaware et al. (2015);
also earlier studies cited in Drèze and Sen (2013), Chapter 7.
24
See Khera and Nayak (2009). The gender dimension of employment guarantee is a prominent
theme of the literature on NREGA in Tamil Nadu, cited earlier, and also of earlier writings on
Maharashtra’s employment guarantee scheme; see also Basu (2012), Naryananan (2008),
Narayanan and Das (2014), among others.
25
Roy (2014), p. 534. On the empowerment aspects of NREGA, see also Dhaktode (2016) and
Jenkins and Manor (2016).

68

26
For an illustration, see Khera (2008) on the experience of NREGA in parts of Badwani
district, Madhya Pradesh. On related experiences elsewhere, see e.g. Pattenden (2015) and Roy
(2015).
27
The role of electoral politics as well as public mobilization in creating political will for PDS
reform in Chhattisgarh is discussed in Tillin, Saxena and Sisodia (2015).
28
On the first two surveys, see Table 6; on the third, see Drèze et al. (2016).
29
The PDS Survey took place in May-June 2011 in twenty districts – two of the poorest districts
in each of nine states (the eight “reference states” and Andhra Pradesh). In each district, we
selected two Blocks. In each Block, three villages were selected at random, and in each of the
120 sample villages, 12 households entitled to the PDS were selected at random. In six sample
states, where the PDS is targeted, sample households were selected from the BPL list. In the
remaining three, where the PDS is universal or quasi-universal, the voter list was used. In the
time available, the survey teams interviewed 1227 (out of a target of 1440) households. For
further details, see Khera (2011c).
30
The Antyodaya scheme, intended for the poorest of the poor, continues under the NFSA,
with foodgrain entitlements of 35 kg per household per month across the country.
31
The PEEP Survey took place in May-June 2013 in twenty districts – two of the poorest
districts in each of ten states (the eight “reference states” as well as Madhya Pradesh and
Maharashtra). It was geared to studying all the five social programmes discussed in this paper
– not just the PDS. In each district, we selected one Block where a local organisation was
available to facilitate the survey. In each Block, four villages were selected at random, and in
each of the 80 sample villages, 30 households were selected at random from the official list of
NREGA Job Cards. In the time available, the survey teams were able to locate and interview
1,926 of these 2,400 households. There were some methodological differences between the
69

PDS Survey 2011 and the PEEP Survey 2013, but they are unlikely to matter much for the
purpose of the broad comparisons being made here.
32
These figures are based on the ‘Tendulkar poverty line’. Using a similar methodology,
Himanshu and Sen (2013) find that the PDS reduces the poverty gap for rural and urban areas
combined by 20 per cent or so.
33
This programme also includes two other schemes that are in the process of being phased out:
the National Family Benefit Scheme and Annapurna. We are not concerned, in this section,
with formal-sector pension schemes such as those available to public-sector employees.
34
These were the pension lists of the PEEP survey’s 80 sample villages. For further details,
see Chopra and Pudussery (2014).
35
An illustration may help. Under this scheme, a person aged 40 years today would have to
pay Rs 291 per month for 20 years in order to be eligible for a pension of Rs 1,000 per month
from then on. Assuming a life expectancy of 70 years and an inflation rate of 5 per cent or so
(reasonable assumptions for India), this person is better off putting the monthly contributions
in a bank account. Indeed, even with an interest rate of just 5 per cent in money terms (zero in
real terms), saving Rs 291 for 20 years from age 40 would enable him or her to spend Rs 1,000
for the next 10 years and end at age 70 with a good Rs 3,000.
36
See www.pradhanmantriyojana.co.in/atal-pension-yojana-apy and nsap.nic.in/, respectively.
37
The figure for India pertains to the central government only. However, even after adding a
generous margin (say 25 per cent) for state-government spending, social security expenditure
in India would look low compared with other Asian countries.
38
This is one possible reading of the chapter on UBI in the Government of India’s (2017) latest
Economic Survey. The chapter presents a UBI proposal costing about 5 per cent of India’s
GDP and lists a series of schemes and subsidies, including the five social programmes
70

discussed in this paper, also costing about 5 per cent of India’s GDP. Phasing out these schemes
and subsidies is envisaged as one possible way of financing the UBI proposal.
39
See Khera (2014) for a more detailed discussion of the findings.
40
The Delhi pilot (called Annashree) was quietly discontinued and we are not aware of any
evaluation of it. According to a survey by the Centre for Equity Studies (New Delhi), the initial
Puducherry pilot was discontinued within eight weeks following public protests and other
hurdles, and the Chandigarh pilot also led to massive discontent due to erratic transfers and
large exclusion errors (Kalra, 2016; Nehra, 2016).
41
Under the Indian Constitution, social security is on the Concurrent List of items on which
both centre and states have the power to legislate, but with national laws superseding state laws.
71

SSRN Id2800699

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SSRN Id2800699

Uploaded by

Copyright:

Available Formats

Accepted for publication in World Development.

Recent Social Security Initiatives in India

Jean Dreze and Reetika Khera

targeting and the value of a rights approach to social security.

Electronic copy available at: https://ssrn.com/abstract=2800699

Recent Social Security Initiatives in India

the Unorganised Sector, 2006).

Electronic copy available at: https://ssrn.com/abstract=2800699

issues are discussed in the next two sections.

somewhat biased towards poorer and less well-governed states.

Electronic copy available at: https://ssrn.com/abstract=2800699

important bearing on the scope for further progress in the future.

1. The Rights Approach to Social Security

Electronic copy available at: https://ssrn.com/abstract=2800699

continuity and durability that is often sorely lacking in ad hoc ‘schemes’.

[Table 1 around here]

Integrated Child Development Services: Proportion of pregnant

Electronic copy available at: https://ssrn.com/abstract=2800699

Employment guarantee: Proportion of rural households that worked

Public Distribution System: Proportion of households that bought

Social security pensions: Proportion of widows and elderly persons

accountable to these responsibilities through democratic means (for instance, elections),

without anyone invoking the Act to access them.

Electronic copy available at: https://ssrn.com/abstract=2800699

to excessive centralization. We shall return to some of these issues.

2. Targeting, Universalization and Self-Selection

security programmes: the identification of eligible persons or households. It may help to

within the BPL category), due to financial constraints.

Electronic copy available at: https://ssrn.com/abstract=2800699

even to extend pensions to all widows.

would reinforce the case for universalization.

Universalization, of course, tends to be expensive, since it involves a major expansion of

became more affordable and acceptable in various contexts.

Electronic copy available at: https://ssrn.com/abstract=2800699

programme, in so far as children attending government schools tend to come from

Electronic copy available at: https://ssrn.com/abstract=2800699

however, are still awaited.

Finally, it is important to mention that universalization does not mean ‘uniformity’ of

households. This does not conflict with the universalization principle.

distributed to school children under the NPNSPE. On 28 November 2001, in response to a

governments to start providing cooked midday meals in primary schools.11

Electronic copy available at: https://ssrn.com/abstract=2800699

the menu improved steadily over time.

commitment to a ‘national cooked nutritious mid-day meal scheme’. In pursuance of that

per cent of GDP).12

upper-caste vegetarian lobbies.

Electronic copy available at: https://ssrn.com/abstract=2800699

gender-related benefits. For instance, they create a valuable source of remunerative

on school enrolment, especially in Grade 1. Using a similar method, Chakraborty and

Electronic copy available at: https://ssrn.com/abstract=2800699

school authorities and the local community.

standards, particularly in the worst-governed states. On 16 July 2013, twenty-three school

4. The Integrated Child Development Services

Electronic copy available at: https://ssrn.com/abstract=2800699

lactating women as well as adolescent girls.

however, a majority of mothers appreciated the potential value of the programme.

Electronic copy available at: https://ssrn.com/abstract=2800699

supplements, maintenance of growth charts, regularity of child attendance) are modest, as

is likely to be greater scope for qualitative improvements in the near future.

[Table 2 around here]

Proportion (%) who:

Were dissatisfied with the quality of food provided at the AWC 24 12