Poverty, targeting and food security


back to issue

THERE has been a growing tendency in recent years to restrict – or attempt to restrict – social benefits to those who are ‘below the poverty line’. The idea is deceptively simple: focus your resources on the poor. In practice, however, this approach has proved quite problematic, and often counterproductive. This article looks at the problem from various angles and discusses an alternative in the context of the National Food Security Bill.

In the good old days, the poverty line was a relatively simple concept. By and large, it was just a statistical benchmark for making ‘poverty comparisons’ – for instance, to track poverty over time, or to compare poverty levels in different parts of the country. Many of these comparisons are not particularly sensitive to the choice of poverty line, within a reasonable range. There is even a situation going by the odd nomenclature of ‘first-order stochastic dominance’, which arises when the relevant comparison holds for any poverty line. For instance, Bihar is clearly poorer than Punjab, wherever one draws the line.


The choice of a poverty line, therefore, was not particularly controversial. One widely used initial benchmark was the level of per capita expenditure required to meet pre-specified calorie norms. One may or may not like this benchmark (the calorie ‘norms’ themselves are quite shaky), but at least it was fairly transparent. There was no claim that reaching the poverty line was a guarantee of being well nourished (since good nutrition requires much more than calorie adequacy), let alone healthy or well educated.

Then came the whole idea of ‘BPL targeting’, that is, of restricting the benefits of various programmes and schemes (in particular, the Public Distribution System) to households below the poverty line. This quietly transformed the poverty line from a statistical benchmark into a real life social division. The division was all the more artificial as the identification of BPL households was highly unreliable. Indeed, the Planning Commission uses one method to count the poor, and the Ministry of Rural Development uses a different method to identify them. This cannot work. And this is just the beginning of a series of conceptual flaws and implementation problems that plague the BPL Census. It is not surprising that, according to the National Sample Survey (among other sources), about half of all poor households in rural India did not have a BPL card in 2004-5. The ‘real-life’ poverty line not only divides people, it divides them in a cruel and destructive manner.

The Tendulkar Committee Report further complicated matters by claiming, for the first time, that the poverty line ensures ‘adequacy of actual private expenditure… on food, education and health’ [sic]. That Rs 32 per person per day (the Tendulkar poverty line for urban areas at today’s prices) is wholly insufficient for this purpose is self-evident from a common sense point of view. This is all the more glaring when one looks at the breakdown of the Rs 32 packet, which includes, for instance, just about one rupee per day for health expenditure. Yet the experts managed to rustle up technical arguments to substantiate their startling claim.


Consider, for instance, the argument they invoked to establish ‘food adequacy’ at the poverty line. Briefly, it consists of observing that food expenditure at the poverty line is typically higher than ‘normative food expenditure.’ The latter is defined as follows (hold your breath): ‘When estimated population from NSS [National Sample Survey] is ranked according to ascending size of food expenditure per capita, normative food expenditure per capita is defined by that level of food expenditure per capita that corresponds to cumulative share of population from NSS that equals the index of malnutrition derived from NFHS-III [third National Family Health Survey] for that state.’ The ‘index of malnutrition’, for its part, is an unweighted sum of the proportions of underweight children, adult men with low body-mass index (BMI), and adult women with low BMI.

If you didn’t understand that don’t worry – because it makes no sense. However, this imaginative argument (somewhat tangential in the report) serves the government’s purpose quite well, since it helps to justify BPL targeting. Indeed, the adequacy argument is the central plank of the Planning Commission’s recent affidavit to the Supreme Court, defending the use of official poverty lines for the purpose of determining food allocations under the Public Distribution System (PDS).


To be fair to the Planning Commission, it is not quite saying (at least not explicitly) that the PDS should be restricted to BPL households. It is just saying that the central government’s commitment ends there. But the writing is on the wall, not just for the PDS but also for other social programmes that are being quietly earmarked for BPL targeting, conversion to cash transfers, and ‘self-liquidation’ as the official poverty estimates go down.

The way forward is not to ‘fix the poverty numbers’ but to find a way out of this bankrupt approach of BPL targeting. That is the appeal of universal entitlement programmes such as school meals, the Integrated Child Development Services (ICDS), and the National Rural Employment Guarantee Act (NREGA). Many states are now moving away from BPL targeting in the PDS too: not only states like Tamil Nadu and Himachal Pradesh that have a long-standing commitment to universalism in social policy, but also, increasingly, other states (such as Andhra Pradesh and Chhattisgarh) where the PDS extends well beyond the BPL category. The PDS tends to function much better in those states, because everyone – or almost everyone – has a strong stake in it. This approach is expensive, but at least it tends to work.

Beyond this, there is a need for informed debate on the future of social support in India. Do we want a divisive, unreliable and exclusionary system of targeted transfers that self-liquidates over time, or to build a comprehensive social security system inspired by constitutional principles, fundamental rights, and ideals of solidarity and universalism? The whole experience of the last twenty years is that over-reliance on economic growth for social progress is a recipe for disappointment. Even Bangladesh and Nepal (starting way behind, and with much slower economic growth) have made far more progress than India during this period, in terms of a wide range of social indicators. Greater attention to social support and economic redistribution is long overdue. Poverty lines, for their part, are best sent back to their statistical kennel.


Let me now introduce three acquaintances. Meena, age 50, lives in a two-room kaccha hut with her disabled husband Chhote Lal who studied up to class 2. They own half an acre of unirrigated land and a goat. Meena is unable to take up any remunerated work as Chhote Lal needs constant care. Without any specific means of subsistence, they live on one meal a day.

Zafar, age 35, never went to school but he learnt to read and write in a night school. Aside from harvesting the odd sack of grain from his small patch of land, he earns a pittance as a weaver. The family is struggling to make ends meet and two of his five children work as child labourers.

Jeetu, age 45, lives on his own – his family deserted him as he suffers from HIV/AIDS. He has been left to his own devices, in a one-room brick shed on the outskirts of the village. He is too weak to work. Compassionate villagers give him rice from time to time, with some vegetables on festival days – everyone is waiting for him to die.

What do these people have in common? Answer: each of them belongs to a ‘zero score’ household – a household that will get a score of zero in the Socio-Economic and Caste Census (SECC), if the census reaches them at all.

The SECC is supposed to rank rural households on a scale of 0 to 7. A household’s score is simply the number of ‘deprivations’ it has from the following list of seven: (i) living in a single-room kaccha house; (ii) having no adult member aged between 16 and 59 years; (iii) being a female-headed household with no adult male member aged between 16 and 59 years; (iv) having a disabled member and no able-bodied member; (v) being a Scheduled Caste or Scheduled Tribe; (vi) having no literate adult above 25 years; and (vii) being a landless household deriving a major part of its income from manual casual labour. None of these criteria apply in the above examples.


After ranking households in this manner, a cut-off is supposed to be applied to identify ‘priority’ households – the main beneficiaries of the public distribution system under the proposed National Food Security Bill. For instance, if the cut-off is two, then priority households will consist of all households with a score of two or more. The cut-off is supposed to be specified so that the share of priority households in the population is around 46 per cent, the proportion of the rural population below the ‘Tendulkar poverty line’ (about Rs 25 per person per day in rural areas), with a small margin for targeting errors. That, at any rate, seems to be the game plan as of now.

Since Meena, Zafar and Jeetu have a score of zero, they are certain to be left out from the priority list, even before the census begins. The good news is that they are fictional characters. But it would be easy to find real-life examples of such situations, or of other stark cases of poor – even destitute – households being left out of the priority list because they have a zero score. In fact, even households with a score of one are almost bound to be left out, since the cut-off is unlikely to be less than two.


The odd nature of this scoring system can be appreciated in more general terms by considering Adivasi (tribal) households – the most disadvantaged section of the rural population. Since most Adivasi households possess a little bit of land, however unproductive, and a house (often a mud house) with at least two rooms, the first and last ‘deprivations’ in the list will not apply to them (note that even land possessed as a matter of traditional rights, without legal title, is to be counted as ‘owned’ by the SECC). Further, since a large majority are likely to have at least one able-bodied male adult aged between 16 and 59 years, the second, third and fourth criteria will not apply to them either. It follows that most Adivasi households will have a score of only one, unless they are ‘lucky’ enough to have no literate adult, in which case their score will shoot up to two. But even a score of two may not catapult them into the priority club. And if it does, Adivasi communities will be oddly split down the middle, between ‘score one’ and ‘score two’ families – a very divisive situation.

This state of affairs is all the more absurd as the distinction between ‘priority’ and ‘general’ households in the National Food Security Bill is wholly unnecessary and counter-productive. As it is, the bill calls for 25 per cent of rural households to be entirely excluded from the public distribution system. Now, if the proportion of excluded households is as high as 25 per cent (instead of 10 per cent as the National Advisory Council had proposed), it is absolutely pointless to split the rest into two groups. It would be much simpler and more sensible to give common minimum entitlements to all households that do not meet the exclusion criteria. This approach would reinforce, rather than undermine, the positive trend towards a more inclusive PDS in many states – Tamil Nadu, Andhra Pradesh, Chhattisgarh, Rajasthan, Orissa, among others. More on this later.


How would you feel about a law that entitles half of all Indian children to free school textbooks, but leaves it to the government to decide which half? It would be an odd sort of law, which seeks to create a justiciable entitlement for the child without any guarantee that he or she is eligible for it. Suppose further that the government decided to give the free textbooks to shorter children, on the grounds that underprivileged children are generally shorter – with different height cut-offs in different states, depending on their poverty rates. Perhaps a statistician would be able to relate to this method, but how would it fare in terms of common sense, feasibility, equity, and political appeal?

Something like this is about to happen with the National Food Security Bill. Under this bill, the population will be divided into three groups: priority, general and excluded. Each group is to have different entitlements under the public distribution system: major entitlements, token benefits, and nothing, respectively. There is no clarity, however, as to how these groups are to be identified – the bill leaves it to the central government to specify identification criteria, to be applied by state governments.

Until now, the main beneficiaries of the PDS were ‘below poverty line’ families. A census was conducted every few years to identify BPL families, based on some sort of scoring system. The cut-off scores were supposed to be set state-wise in such a manner that the proportion of families with a score below the cut-off (i.e. BPL families) matches the proportion of families below the poverty line according to the Planning Commission’s poverty estimates.

As discussed earlier, this approach is very unreliable and divisive. Many states, however, have moved away from BPL targeting in recent years, and extended the PDS well beyond the BPL category. Further, this move seems to have helped them to improve their PDS by creating a broader and stronger constituency for it.


The National Food Security Bill threatens to undermine this positive trend. It effectively reimposes BPL targeting under another name, that too based on rigid national criteria. ‘Priority groups’ are not fundamentally different from BPL households, and the Socio-Economic and Caste Census, which seems to be expected to identify priority households, is much the same as earlier BPL censuses. In fact, the term ‘Socio-Economic and Caste Census 2011’ appears to be treated as interchangeable with ‘BPL Census 2011’ on the website of the Ministry of Rural Development. This is a serious problem: the ‘hit or miss’ approach involved in BPL censuses is bad enough when the PDS is run as a scheme, but it is especially problematic for the purpose of legal entitlements. A legal right cannot leave any ambiguity as to who is entitled to it.

In a statement released on 3 October 2011, the government asserted that the SECC would actually be a departure from the earlier BPL methodology. Specifically, it promised to stop setting state-wise ‘caps’ on the BPL list based on official poverty estimates. This sounds good, except that it is far from clear how the identification process would work without poverty caps. An ‘expert committee’ is supposed to sort this out after the SECC results ‘are available and have been analyzed’, but the track record of expert committees on this subject is not particularly encouraging.

There is a simple way out of this mess: abolish the distinction between general and priority groups, and give all households a common minimum entitlement under the PDS unless they meet well-defined ‘exclusion criteria’. In other words, target the rich instead of trying to target the poor.

Indeed, the rationale of the distinction between general and priority households is far from clear. Neither the National Advisory Council, nor the Rangarajan Committee, nor any other expert group recommended that the proportion of excluded households should be as high as 25 per cent in rural areas (and 50 per cent in urban areas), as the bill prescribes. Insisting, after this exclusion exercise, on a further distinction between priority and general households is unnecessary, counterproductive, and impractical.


There is, of course, a case for giving special treatment to the poorest households (e.g. by giving them pulses and edible oil, in addition to foodgrains, under the PDS). But this purpose would be better served by retaining and consolidating the Antyodaya programme, which is working reasonably well and already covers about 25 million rural households.

This simplified framework would be relatively practical, transparent, equitable and politically appealing. Most people would be clear about their entitlements, making it much more likely that the bill will succeed. This approach would also resolve much of the alarming confusion that surrounds the Socio-Economic and Caste Census, the National Food Security Bill, poverty lines, and related matters.


* This article is adapted from three earlier essays: ‘The Poverty Trap’ (The Hindustan Times, 22 September 2011), ‘Kaun Banega Scorepati?’ (The Hindu, 28 November 2011) and ‘Just Target the Rich’ (Times of India, 4 February 2012).