ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

How to Identify the Poor? A Proposal

The Census of 2002 to identify the poor in rural areas of India was the third in a quinquennial series. However, it has been appropriately criticised. This paper elaborates on the criticisms, and proposes an alternative set of criteria and methodology for conducting the next (now overdue) census of the rural population to identify the poor


How to Identify the Poor? A Proposal

Santosh Mehrotra, Harsh Mander

The Census of 2002 to identify the poor in rural areas of India was the third in a quinquennial series. However, it has been appropriately criticised. This paper elaborates on the criticisms, and proposes an alternative set of criteria and methodology for conducting the next (now overdue) census of the rural population to identify the poor.

Thanks are due to Biraj Patnaik, Dipa Sinha, Jean Dreze for comments on earlier drafts.

Santosh Mehrotra ( is in the Development Policy Division, Planning Commission, and Harsh Mander ( is a Special Commissioner of the Supreme Court of India on the Right to Food.

Economic & Political Weekly

may 9, 2009 vol xliv no 19

hile the estimation of poverty at the national level and at the state level is done from time to time by the Planning Commission, periodic censuses to identify those households that subsist below the deemed poverty line have been conducted by the Ministry of Rural Development (MoRD) of the central government. The purpose of these censuses has been to identify the below the poverty line (BPL) households living in the rural areas who could be targeted under its programmes.1 This census has been carried out in each of the states of India three times (1992, 1997 and 2002) in the last 17 years, and each time the methodology used has been different, because each time it has been recognised that the methodology that was used earlier was seriously flawed. It is now time to conduct another such census, which is expected to be carried out in the latter half of 2009. This paper proposes a methodology that hopefully will address the many large-scale exclusion and inclusion errors that have resulted from the previous three censuses, causing widespread discontent and injustice.2

The BPL list is of enormous practical importance to both central and state government officials and the rural population at large. This is because the central government uses it to identify the recipients of a series of MoRD programmes. The department of rural development is implementing a self-employment programme, i e, Swarnajayanti Gram Swarozgar Yojana (SGSY), under which the rural poor are organised into self-help groups, assisted for capacity-building and provided financial assistance to set up economic activity through a mix of credit and subsidy. The objective is to provide economic assets to BPL families so that they get an incremental income on a sustainable basis. Similarly, Indira Awaas Yojana (IAY), is a rural housing programme under which financial assistance is provided to the BPL families for constructing houses. Indira Gandhi National Old Age Pension Scheme (IGNOaPS) has now been extended to all the BPL households having members of the age of 65 years and above. Further, the department of drinking water supply is implementing a total sanitation campaign programme under which the BPL rural households are provided with financial assistance for construction of toilets.

Most importantly, the BPL list is used to give ration cards to beneficiaries who access the targeted public distribution system (TPDS) – which entitles them to a 35 kg ration of wheat or rice, depending upon the state. Further, state governments also used this database for variously targeting several schemes and services to the poor. These include maternity benefits and insurance paid at the death of a breadwinner of BPL families. Some states even require BPL cards for accessing free medicines, hospitalisation or medical investigations in government hospitals. Despite repeated directions, including of courts, to not use BPL ration cards for this purpose, these are used also as proxy identity cards, as de facto proof of identity, residence and even (in cases of contested citizenship such as of alleged Bangladeshis) of citizenship.

Given the enormous importance of the list, the methodology should have been as close to foolproof as possible. If anything, it was quite the opposite – as we will see in this paper. Section 1 reviews the criteria that were used over the last three surveys, focusing, in particular, on the first two. Section 2 spells out the basis of the third survey (of 2002), which is the current basis for the BPL list, and critiques it thoroughly. Section 3 proposes an alternative set of criteria, and elaborates on the procedure for conducting the census. It also clarifies the definitions of the terms used in the criteria presented in the previous section, given that administrative staff who will administer the census as well as the programmes that the list impacts should be very clear of what is intended. The final section uses a similar conceptual frame that drives the choice of criteria for the rural survey to suggest criteria for an urban census. These urban criteria have actually already been used in the state of Delhi in 2008, and we hope to further elaborate on these criteria for non-metropolitan areas in a f uture paper.

1 The Origins of the BPL Census

The Integrated Rural Development Programme (IRDP) was launched in October 1980 in all the blocks of the country as an instrument of poverty alleviation in the rural areas. Under IRDP, the approach was to assist the poorest of the poor rural families first and the target group was the rural families having annual income below the cut-off line of Rs 4,800. But families with an annual income up to Rs 3,500 were assisted first. Registers were maintained at the village level for recording persons identified to be under the poverty line as defined by the Planning Commission in monetary terms was adopted. There was no systematic survey or census to prepare the BPL list. The resources under IRDP were limited, therefore, under the guidelines, priority was given to the poorest of the poor. However, it was noticed that a large number of beneficiaries who were either ineligible or were comparatively better-off got the assistance under the programme.

1.1 BPL Census 1992

In 1992, for the Eighth Plan, a detailed procedure was prescribed to identify BPL families in the rural areas. The need for a systematic survey of BPL families was felt because a significant part of the benefits of the IRDP had gone to either ineligible categories or to the non-poor, the most eligible families were left out, including those living in outlying hamlets of the villages, nomadic families and women-headed households. The 1992 Census used income as a criteria, the annual income cut-off was Rs 11,000 per household, below which all were poor. The BPL families were classified into income ranges of Rs 0-4,000, Rs 4,000-6,000, Rs 6,000-8,500 and Rs 8,500-11,000.

However, in the 1997 Census the income criteria was dropped, since it was felt that there is always an inherent bias among the population to underestimate income so as to be covered under the target group.


1.2 BPL Census 1997

In the 1997 Census, it was decided to use exclusion criteria – and all rural households that were not excluded were considered as BPL. Five exclusion criteria were used: (1) whether the household was operating more than two hectares of land; (2) whether it had a pucca house; (3) whether any resident member of the household has an annual income from salary/self-employment exceeding Rs 20,000 per annum; (4) whether the household owns the following consumer durables: TV, refrigerator, ceiling fan, motorcycle/ scooter, three-wheeler; and (5) whether the household owns the following farm implements: tractor, power tiller, combined thresher/ harvester. This exclusion criterion could be used for weeding out those families for the census which are prima facie not poor.

Once the poor were thus identified, further information was sought about all BPL families, but this information was not intended to be used for ranking the BPL. It was intended merely to acquire further information about the characteristics of the household. These criteria were housing status: whether the house was kutcha or pucca; whether it was owned or rented. Information was also obtained about whether the family was landless or was a small/marginal farmer, and whether the land was irrigated or not. Other information was about the ownership of livestock assets, formal training acquired; expenditure of the household on various food items; receipt of assistance from government programmes (e g, IRDP, Training of Rural Youth for Self-employment (TRYSEM), Development of Women and Children in Rural Areas (DWCRA), Employment Assurance Scheme (EAS), Jawahar Rozgar Yojana (JRY), etc); perception of household regarding sufficiency of food (e g, two square meals a day); indebtedness of the household; and whether any household had migrated.

It is remarkable that most of these same questions were used, and then given a score when the 2002 Census was undertaken. In other words, roughly the same questions were used in 2002 as in 1997 – as you can see in Box 1 (p 39) (which lists the questions of the 2002 Census).

2 BPL Census 2002 – Problems with 13 Criteria

Box 1 lists the questions in the household questionnaire used to identify the poor in 2002. The criteria for identifying BPL, the household should be of a directly verifiable and observable nature. Most of the 13 criteria, on the other hand, are not easily or clearly verifiable.

These criteria have been widely and bitterly attacked, by rural poor people and their organisations and by non-governmental organisations. We will deal very briefly with each question in the list of 13, in turn, and some of the obvious difficulties with each:

  • (i) Landholdings: The question on landholdings does distinguish between irrigated and unirrigated land, but it does not distinguish between land irrigated as a result of private investment (for example, a privately dug tube well) from a publicly irrigated piece of land. The question makes no provision for distinguishing land by its quality, which goes beyond the issue of irrigation.
  • (ii) Type of House: This question is reasonably clear cut. Census criteria used by the registrar general of India enable us to clearly distinguish between a kutcha, semi-pucca and pucca house. But there is a contradiction: for nearly two decades governments,
  • may 9, 2009 vol xliv no 19


    both central and state, have been providing BPL households with houses either under IAY or state government programmes. How these families will be accounted for – given that despite being poor the now have pucca houses – is a question that requires a serious thought.

    (iii) Availability of Clothing: This question investigates whether a person owns less than two pair of clothing at a minimum, increasing up to “more than 10 pair”. This question is flawed for several reasons. First, in colder climates more clothing is required than in warmer climates. Second, it is unclear from the question whether undergarments are counted among clothing or not. Answers to such question are bound to lead to very unclear outcomes.

  • (iv) Food Security: This question is comparable to the NSS question on “whether a family gets two square meals a day”. Even if all the members of a household eat two meals per day there may be serious problems with the quality of the meal, as reflected in the calorie intake and protein intake per meal. Moreover, there is no way the answer to the question will reveal how the food is being shared between the men in the household on the one hand, and the women and girls on the other. There are also complex problems of recall, and cultural shame in the male head of household publicly admitting that there is not enough food to feed his family.
  • (v) Sanitation: This question investigates whether the family defecates in the open or in a group latrine or in a private latrine. If the family has a private latrine the household would get four points, i e, the maximum number of points. By getting the maximum number of points even for one question, the family runs the risk of scoring itself out of the BPL category. In another words, this is a perverse question and encourages totally perverse behaviour. This kind of question would discourage families from actually investing in a private latrine.
  • (vi) Ownership of Consumer Durables: Possession of a TV requires a much higher purchasing power than possession of an electric fan or radio. Therefore, simply counting, as the BPL Census questionnaire does, whether the household has any such item or “all items” is not a meaningful exercise. However, the expert group can consider using possession of consumer durables (for example, a motorcycle, colour TV, fridge) as exclusion criterion, as a criterion for excluding a household from the BPL category.
  • (vii) Literacy Status of Highest Literate: The lowest scoring category under this criterion is “Illiterate”, while the highest scoring category “Postgraduate”. There is a slight risk that this question will not yield very meaningful answers for many reasons. First, even with a postgraduate degree, a person may still be unemployed. An unemployed person may still be thus a head of a poor household. Second, what means will the investigator use to interrogate and then confirm the educational status of the highest literate is also not clear. Once again this provision also perversely places at a disadvantage a poor family which has sacrificed money and the opportunity costs of child labour earnings to send a child to school.

    (viii) Status of Household Labour: As formulated in the questionnaire, this question does not address the problems that in most rural household women and children work along side the adult male on the family farm. In any case, it is unclear how the investigator will ascertain that only the adult male in the household is working, while the woman and children are not.

    Box 1: Indicators Used for Identification of BPL in Census 2002

    1 Landholdings (tick √ one and indicate score)

  • (i) Nil holdings
  • (ii) Less than one ha unirrigated or 0.5 ha irrigated
  • (iii) More than one and less than 2 ha unirrigated or 0.5 to 1 ha irrigated

    (iv) More than 2 ha and less than 5 ha unirrigated or 1.0-.2.5 irrigated

  • (v) More than 5 ha unirrigated or more than 2.5 ha irrigated 2 Type of house
  • (i) Houseless
  • (ii) Kutcha
  • (iii) Semi pucca
  • (iv) Pucca
  • (v) Urban type 3 Availability of clothing
  • (i) Less than two pairs
  • (ii) More than two pairs but less than four
  • (iii) More that four pairs but less than six
  • (iv) More than six pairs but less than ten
  • (v) More than ten pairs 4 Food security
  • (i) Less than one meal per day in major part of the year
  • (ii) Normal one meal but sometimes less
  • (iii) Normal one meal throughout the year
  • (iv) Two meals per day and occasional shortage
  • (v) Enough food 5 Sanitation
  • (i) Open defecation
  • (ii) Group latrines with irregular water supply
  • (iii) Group latrines with regular water supply

  • (iv) Group latrines with irregular water supply and sweeper
  • (v) Private latrine
  • 6 Ownership of consumer durables, viz, TV, electric fan, kitchen appliances, cooker, radio, etc.

  • (i) Nil
  • (ii) Any one item
  • (iii) Two items only

    (iv) Any three

  • (v) All items 7 Literacy status of highest literate
  • (i) Illiterate
  • (ii) Up to primary
  • (iii) Completed secondary/passed 10th
  • (iv) Graduate/professional
  • (v) Postgraduate/professional 8 Status of household labour
  • (i) Bonded labour
  • (ii) Women and child labour
  • (iii) Only adult female and no child labour
  • (iv) Adult males only
  • (v) Others 9 Means of livelihood
  • (i) Casual labour
  • (ii) Subsistence cultivation
  • (iii) Artisan
  • (iv) Salary
  • (v) Others 10 Status of children
  • (i) Not going to school and working
  • (ii) Going to school and working
  • (iii) Nil
  • (iv) Nil
  • (v) Going to school and not working 11 Type of Indebtedness
  • (i) Daily consumption purposes from normal sources
  • (ii) For production purposes from normal sources
  • (iii) For other purposes from normal sources
  • (iv) Borrowing from institutional agencies
  • (v) No indebtedness and process assets 12 Reason for migration
  • (i) Casual work
  • (ii) Seasonal employment
  • (iii) Other forms of livelihood
  • (iv) Non-migrant
  • (v) Other purpose 13 Preference for assistance
  • (i) Wage employment
  • (ii) Self-employment
  • (iii) Training and skill upgradation

  • (iv) Housing
  • (v) Loan/subsidy more than Rs One lakh or no assistance require
  • Economic & Political Weekly

    may 9, 2009 vol xliv no 19

  • (ix) Means of Livelihood: This question asks whether means of livelihood is casual labour (zero points), or subsistence cultivation
  • (1) or artisan (2), or salary (3), or others (4). This question simply fails to understand that the poor always have multiple sources of livelihood, and members of a household are likely to be engaged in different kinds of work, and even the same individual may engage in different kinds of work and derive income from those sources at different times of the year. In other word, it is impossible for household, while answering the question, to merely tick one of the five listed answers to the question.3
  • (x) Status of Children: This question investigates whether the children are “not going to school and working”, or “going to school and working”, or “going to school and not working”. As with the question on literacy, the answer to this question would be either perverse or not meaningful, if it encourages parents to force the child to not go to school and only work, simply because it would entitle them to zero points, and hence, increase the chances of the household being counted as BPL.
  • (xi) Type of Indebtedness: This question, of all questions among the 13 criteria, shows the least understanding of ground realities of the rural indebtedness. The poorest people lack assets, which can be used as collateral, and hence, are unable to borrow. But this question gives the highest score “4” to the household that is not indebted. In another words, the poorest household might end up getting four points and thus rule itself out of the BPL category.
  • (xii) Reason for Migration: This question gives the lowest score to the household which has a member migrating for casual work, and gives the score of 3 to non-migrants, and score of 4 to those migrating for reasons other than casual work. Quite apart from the fact that this question does not distinguish households from which only one member has migrated from and those from which more than one member might have migrated. It is entirely unclear why the reason for migration should merit different scores for the purposes of determining whether a family is BPL or not.4 Fittingly, the West Bengal government excluded this c riterion in its rural survey.

    (xiii) Preference of Assistance: This question is also among the most meaningless among the all criteria. It does not attempt to establish the current status of a household in respect of a particular tangible attribute of well-being, but rather interrogates

    (1) whether the household would like to receive assistance in the form of wage employment (zero) or self-employment, (2) or training, (3) or housing, and (4) or loan more than rupees one lakh. Given that such a question would yield answers that are less than meaningful when comparing across categories of a nswers, the West Bengal government eliminated this criterion from its rural survey.

    2.1 Other Issues Related to Census 2002

    For each question, the enumerator was supposed to tick one answer. For the first answer the score given was 0, for the second it was 1, for the third it was 2, for the fourth 3, while for the fifth answer the score was 4. The maximum score would be 54 for the 13 questions taken together. The cut-off for BPL category was determined by the numbers as given by the Planning Commission’s

    40 estimates for poor in the state concerned. In other words, there could only as many poor as the Planning Commission had estimated. Naturally, the cut-off for determining those who would be identified as poor were going to be different in each state. In most cases, the cut-off was between 16 and 25.

    One of the important features of the guidelines issued for BPL Census 2002 was to put the ceiling on the number of BPL households to be identified in conformity with the poverty estimates of Planning Commission. Originally, the states were asked to identify the number of BPL families in such a manner that it should be equal to the Poverty Estimates of 1999-2000 of Planning Commission. However, subsequently, the states were given the option of deciding the total number of BPL households equal to the Poverty Estimates of 1999-2000 or the adjusted share computed by the Planning Commission, whichever is higher. In addition, the states were also given the flexibility of another 10% to account for the transient poor.

    The fact that there is a cut-off for determining the total number of poor had several implications. First, there was, as Hirway (2003) notes, a “mad rush” to be enrolled as poor, and “the rich and powerful in a village frequently pressurise the talati and the sarpanch to include their names in BPL lists” (p 4805). Our approach, as outlined in the next section, is quite different:

  • (1) There is no cut-off that we are suggesting that corresponds to the Planning Commission cap for the number of poor in a state. In fact, we suggest that each scheme should have its own cut-off, depending upon the resources available, and the resources are a function of the allocation made not just by the central government (which is usually the bone of contention) but by the state government. If the state government decides that the allocation required by the needs of the population is greater than can be achieved from the central government’s allocation, then the state government should be willing to utilise its own resources for the programme.
  • (2) Our exercise is simply an exercise in giving scores to families, and thus producing a ranking of the population in the village. In fact, the score for each dimension in the criteria should not be made public by the enumerator (it was quite public in the 2002 Census). There were no weights in the 2002 criteria, but we have weights. More importantly, our weights are not awarded by the enumerator, but will be entered by the computer, since the code for the dimension will generate the score automatically in the computer programme, and the enumerator should have no information about the weights of the dimension, and thus the affected family too will not know their final score until the exercise is over. And the appeal to the gram panchayat is not going to be about inclusion in the BPL list or exclusion from it, but about the score that the family might have received – which will be well after the enumerator’s job is totally completed.
  • (3) The criteria are so directly verifiable or observable that almost no one can – poor or rich – can pretend to be what they are not.
  • Another criticism, pointed out by Jain (2004), was that in M adhya Pradesh the block panchayat officers announced that the first list of BPL families was final and did not entertain any grievances, and that the cross-checking did not function. However, this kind of problem should not arise with what we are proposing

    may 9, 2009 vol xliv no 19


    since our exercise is not intended to lead to a “BPL List”; we are only proposing that the census objective will be to generate a ranking of households according to certain directly observable criteria, and the cut-off will come later for each scheme/ programme, and the cut-off may well vary for different p rogrammes.

    Another criticism (Alkire and Seth 2008) of the 2002 criteria is that the BPL Census focused mainly on resources (land, house, clothing, food, bathroom, consumer goods, loans), rather than capabilities, i e, the things that households are able to do and be (nourished, educated, and in good health). We should state at the outset that our proposed criteria also do not capture capabilities. But we are not apologetic about that, since such capabilities are not easily measured in the field by ill-equipped officers, who are expected to do a census of the entire rural population. For instance, we earlier felt that there is a case for targeting households by body mass index (BMI), given that a third of Indian households, according to Third National Family Health Survey (NFHS 3 of 2005-06), are malnourished (i e, have a BMI of less than 18.5). However, we had to abandon the idea since we know that the sturdy weighing scales that will also measure height are very e xpensive, and are likely to malfunction after repeated use, e specially when the weighing scale must be carried around from village to village and from house to house in dusty conditions.

    Some state governments raised objections to putting a cap on the total number of BPL households to be identified. This problem has arisen primarily because this database of BPL Census is now being used by the state governments for other centrally sponsored and state sector schemes also. Therefore, a number of states had expressed their difficulties in restricting the number of BPL households in their state to the limit fixed by Planning Commission or and also to the number identified on the basis of BPL Census 1997. The number of BPL households identified by the state governments has not been in line with the official poverty estimates of Planning Commission.

    The finalisation of results of the BPL Census 2002 was delayed because of the Supreme Court order passed in the matter of PUCL vs Union of India 9 in Writ Petition NO 196/20010 directing the Union of India not to insist upon the state governments to delete any name from the existing BPL list.

    Given the variety of problems that have arisen with the 2002 Census, it has been felt that it is necessary to completely revise the criteria for identifying the poor.

    3 A Proposal for BPL Census 2009

    We propose the following procedure for the new census of BPL families, which we hope will overcome most of the problems encountered in earlier censuses, by methods which are indeed transparent, objective and verifiable. We will list our proposals separately for rural and urban areas.

    3.1 For BPL Rural Families

    The survey of BPL rural families should be undertaken simultaneously in all states and union territories (UTs) from July to December 2009, over a six-month period. The survey will be undertaken by government functionaries from a different block or mandal

    Economic & Political Weekly

    may 9, 2009 vol xliv no 19

    from the one in which the village is located. It will be supervised by the gram panchayat. The survey will be shared in a meeting of the gram sabha and its approval has to be obtained before it is fina lised. The first appeal against both inclusion and exclusion errors alleged in the list will lie with the gram sabha. The second appeal will lie with the chief executive officer of the district panchayat, and his/her decision which will state reasons in writing will be final and binding.

    The survey should be valid for a period of 10 years, i e, up to June 2019. However, each state government/UT will invite objections and undertake a summary revision of the list every two years, specifically to take account of (a) deaths; (b) new households;

    (c) in- and outmigration; and (d) changes in fortunes of the households.5 The revised lists prepared on the basis of this summary process will be read out and approved by the gram sabha.

    Any individual will also retain the right to apply for correction in the list at any time to the gram panchayat, and his/her application will have to be disposed of within a period of three months of the application. The first appeal against the application will be to the gram sabha, and the final appeal will be to the chief executive officer of the district panchayat.

    The basic unit of the survey will be a household as defined in the next section. Care must be taken to not exclude homeless households who may not have a roof, but who reside in the village; and destitute and isolated single individuals who should be treated as separate households for purposes of the survey.

    Care must also be taken not to exclude outmigrants. The survey must coincide with the kharif sowing and harvesting season, when seasonal migrants are expected to return home. But in regions when the sowing and harvesting season does not fall within the prescribed time period, a special dispensation should be sought from the MoRD, to reschedule the dates to include the sowing and harvesting season.

    The households which fulfil any one of the following conditions will not be surveyed for BPL status: (1) families which have two standard hectares of agricultural land or its equivalent plantation land; (2) families which have four-wheeled diesel and petrol vehicles; (3) families which have at least one running bore well; (4) any person in the family is drawing a salary of over Rs 10,000 per month in government/non-government/private organisations; and (5) income taxpayers.

    The survey will include all households in the village except those that are excluded by criteria listed in the preceding paragraph, and will not have any cut-off line. It will only rank families in terms of surveyed relative poverty levels, according to the following scale and weightages Each department which uses the list will have to take a decision about where it wishes to establish its cut-off.

    The final point about procedure is that this method must be piloted in a few states with different characteristics. The possible principle on which the pilot states could be chosen are geographical spread; at least one hilly state; at least one north-eastern state; high tribal density population; at least one rainfed area; at least one area in the well-irrigated belts of the country. Without piloting, we will not be able to discover the flaws in this c ompletely revised methodology, so that adjustments can be made to it before the actual census is launched in the latter half of 2009.

    3.2 Theoretical Framework

    There is a theoretical framework that led us to the choice of criteria. First, the exclusion criterion above shows that we propose auto matically to exclude those households that have at least one member working in the formal economy, either in the public or private sector. This is because they constitute barely 7% of the n ation’s population, and by definition they do not experience the kind of vulnerability to exogenous shocks that we are trying to capture here. In other words, we are concerned exclusively with those who toil and subsist without any social security in the informal or unorganised sector of the economy. However, 93% of the workforce is in the unorganised sector, so we must have a variety of filters to clearly identify the poor, through directly verifiable and observable characteristics (to make for administrative ease and convenience)

  • so they are not amenable to easy manipulation by the local elite.
  • The inclusion criteria use a multiple of principles for identifying the poor. There are essentially three inclusion criteria:
  • Workers in low income-yielding, undignified, unsafe or highly vulnerable (to exogenous shocks) occupational categories;
  • Households where the bread-earners’ working (and earning) capacity is significantly compromised by mostly constraints over which they have no control;
  • Households subjected to social exclusion on the basis of their ascribed status in historically disadvantaged groups, which we call for short affirmative action categories.
  • The occupational categories are in some ways self-explanatory since most of the members in this category listed relate to agriculture, which is the principal occupation in rural areas. Inclusion of the self-employed artisan may require some explanation; quite often they will have limited land, and in a rapidly modernising economy, they face competition to their craft from manufacturing industry that is gradually making their earnings highly vulnerable. In addition, there are those residing in forests or close to heavily-forested areas that are dependent mainly upon forest products for acquiring their food, and thus again quite vulnerable.

    It is the remaining two categories that perhaps require further explanation. They are both categories that involve social exclusion. Karl Marx (1862-1990: 603) wrote evocatively of the exclusion of destitute populations from what he described as “political economy”:

    Political economy does not recognise the unoccupied worker… The beggar, the unemployed, the starving [and] the destitute are figures which exist not for it, but only for the eyes of doctors, judges, gravediggers and beadles. Nebulous… figures which do not belong within the province of political economy.

    Incidentally, Marx was right about their exclusion, but not about their being “unoccupied workers”.

    On the contrary, we have found that the destitute (many of which are captured in the third category) are forced to labour in arduous, low-paid, undignified work in order even to stay barely alive as each new day dawns.

    In a perceptive paper, Barbara Harriss-White tries to unravel the features and sources of destitution. First, it involves the a bsence of any control over assets and the loss of access to income from one’s own labour. This loss of control may result from m ishaps, addictions, disasters, health emergencies, and collapse or withdrawal of family support.

    A plausible common sequence involves the progressive liquidation of small stock, livestock, consumer goods and eventually the failure to protect from sale the key productive assets… The right to the asset of one’s own labour may be forfeited. This right may be sold (bonded) to others. The concept of dependence may be transformed and the labour of non-labouring dependents sold or bonded. The most extreme tactics do not involve the sale of labour so much as the marketing of the body itself (as in the sale of blood or of organs or the renting of the body as in sex work) (Harriss-White 2003: 2, 4).

    The destitution and helplessness of highly marginalised groups do not arise frequently from low incomes or even from their own intrinsic and irrevocable biological infirmities (such as of age and disability), but from the fact that in many cases these infirmities are externally imposed by social arrangements themselves. There are some echoes of this idea in some of the recent literature on social exclusion. Whereas concepts such as poverty, vulnerability, deprivation and inequality do not impute causality, a social exclusion framework implies not only that a person or persons are being excluded but that someone or something “is doing the excluding”. The word “exclusion” suggests that there is a core and a periphery, and that “excluded” people are those who are actively blocked access to the core. The importance of these perspectives is that poverty is not perceived to be a mere attribute of certain categories of people. Instead, it is seen as something that is actively done to people. It is not what they are, but what they have been made. It is interesting that the ex-untouchables of India have discarded the appellation given to them by Gandhi – harijan, meaning children of god – which they regard as patronising. They prefer dalit – which means one who is crushed – because the term implies that they have been oppressed, and it has therefore, acquired a cultural context of assertion and anger (which is what we try and capture in the affirmative action categories like most backward castes, scheduled castes, primitive tribes). In this sense, the term “exclusion” is useful.

    So also is the word “social”. The most evolved definition of food security, so far, at the time of writing that we could locate in the literature appears in The State of Food Insecurity in the World 2001: “(Food security is) a situation that exists when all people, at all times, have physical, social and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life” (FAO 2001; emphasis added).

    The inclusion in the definition of social access is highly significant, because it acknowledges that people may be barred from access to food even if it is locally available and they have the economic means. These social barriers to food security may include gender, caste, race, disability or stigmatised ailments. Herein lies the foundations for the last two categories in our criteria for identifying the poor.

    3.3 The Criteria

    For the sake of simplicity and clarity, each household should be marked out of 100 (this method has worked well in BPL Census done in Kerala recently). It is easy even for illiterate poor people

    may 9, 2009 vol xliv no 19


    to understand that they obtain marks out of 100 (instead of the 54 maximum that a household could have received in the 2002 Census). It is possible for highly disadvantaged families to be awarded more than 100 points on this scale (such as for a landless aged single woman-headed household from a designated primitive tribal group (PTG).

    The ranking should be done on the following parameters:

    A Occupational categories:

  • (a) Destitute/dependent on alms: 40
  • (b) Forest gatherer: 40
  • (c) Landless worker: 35
  • (d) Tenant/sharecropper: 30
  • (e) Marginal farmer: 30
  • (f) Small farmer: 25
  • (g) Self-employed artisan: 30 Affirmative action categories:
  • (i) SC/ST: 30
  • (ii) MBC (Designated Most Backward Castes): 15
  • (iii) Muslims:6 15.

  • (iv) Designated Primitive Tribal Group: 50 B Social categories:
  • (i) Single women headed household: 40
  • (ii) Disabled bread-earner: 40
  • (iii) Bonded workers (bread-earner or dependent): 40

  • (iv) Old person headed household: 40
  • (v) Bread-earner with HIV AIDS, leprosy, mental illness: 40
  • (vi) Bread-earner with TB: 20
  • (vii) Disabled dependent: 20

    Many of these categories require a definition. Definition of a household: “Household” will mean a nuclear family comprising mother, father, and their (unmarried) children, and may include any person wholly or substantially dependent on the head of the family. Household will also mean a single-member family. Within households which may even share a kitchen and roof, the following will be treated as separate households: (a) a single woman and her children; (b) old individuals or couples in which one or both are beyond the age of 65 years; (c) every d isabled adult with spouse and children; (d) every adult with l eprosy, mental illness or HIV/AIDS with spouse and children; and (e) bonded labourers with spouse and children.

    A Occupational Categories: (1) Destitute/dependent on alms: A household in which no member is “employed” in regular, casual or seasonal employment and the household is completely dependent on alms for survival.

  • (2) Forest gatherer: A household in which no member is “employed” in regular, casual or seasonal employment and the household is completely dependent on foraging, gathering and hunting from the forest for survival.
  • (3) Landless worker: A landless worker household is one in which no member owns any cultivable land either himself/herself or through his/her parents.
  • (4) Tenant/sharecropper: A farmer who does not own but c ultivates land owned by other persons and leased from those other persons, provided that the total land that the farmer c ultivates (both as landowner and leaseholder should not exceed two hectares).
  • Economic & Political Weekly

    may 9, 2009 vol xliv no 19

  • (5) Small farmers and marginal farmers: For the definitions of small and marginal farmers it is suggested that the Planning Commission sets up a committee to decide the definitions for each state (or sub-regions of states) based on state-specific local conditions.
  • (6) Self-employed artisan and worker: This is a household which is primarily dependent on income from non-agricultural work which is not regular and is based on either daily wages or piece rate basis, and which is based entirely on family labour with no wage employment of other workers.
  • B Affirmative Action Categories: (1) SC/ST: All the groups that come under SC/ST in each state, as scheduled in the Constitution.
  • (2) MBC (designated most backward castes): In states where there is a defined list of MBCs, the castes under these lists should be treated as MBC for this purpose as well. (3) Muslims: All Muslim families. (4) Designated primitive tribal group: all designated PTG families based on the state lists.
  • C Social Categories: (1) Single women-headed household: A woman who is either a widow, deserted/separated, divorced or is above 35 years of age and is unmarried is a single woman. A single woman and her unmarried children must be considered as a separate household even if they are living in a larger unit sharing a common roof and kitchen.
  • (2) Disabled adult (bread-earner): All persons who are defined as disabled by the Persons with Disabilities (PWD) Act, 1995 should be considered as being “disabled”. The act defines a disabled person as one who is “suffering from 40% or more disabi lity”. This can be certified by the gram panchayat. If any adult member of the household, between 18 and 65 years of age is disabled, then the household should be given the 40 points under this category.
  • (3) Bonded workers (bread-earner or dependent): A household in which any one (or more) of the adult or minor members is working as a bonded worker – i e, is working in lieu of an advance (or loan) taken, is not free to work elsewhere and is receiving less than minimum wages.
  • (4) Old person-headed household: Any person above 65 years of age is an old person. An old person or a couple where one or both the members are above 65 years of age must be considered as a separate household even if they are living in a larger unit sharing a common roof and kitchen.
  • (5) Bread-earner with HIV/AIDS, leprosy, mental illness: A household in which any member has been diagnosed as being HIV +, has leprosy or is mentally ill.
  • (6) Bread-earner with TB: A household in which any member has been diagnosed with TB in the last five years.
  • (7) Disabled dependent: All persons who are defined as disabled by the PWD Act, 1995 should be considered as being “disabled”. The act defines a disabled person as one who is “suffering from 40% or more disability”. This can be certified by the gram p anchayat, based on culturally accepted criteria of determining the extent of disability. If any member of the household who is under 18 years of age or is above 65 years of age is disabled, then the household should be given the 20 points under this category.
  • Since such a massive survey is being undertaken, certain issues will be surveyed which will not be included in the scoring system, but which will give useful feedback about the status of poor families, and their access to government schemes and the services that individual families should be able to access. The list of questions that should be asked in the survey is given in Box 2.

    Box 2: List of Possible Questions
  • (1) Does the household have a job card under National Rural Employment Guarantee Act (NREGA)?
  • (2) Do members of the household migrate seasonally?
  • (3) Does the household have a BPL or Antyodaya Anna Yojana (AAY) ration card at present?
  • (4) Does the household have title to their house-site?
  • (5) Is their home (a) kutcha; ( b) pucca; or (c) they are houseless?
  • (6) Does the household have individual sanitation?
  • (7) Are all children below six years, if any, enrolled in Integrated Child Development Services (ICDS)?
  • (8) Are all children between six and 14 years, if any, enrolled in ICDS?
  • (9) Is there any person above the age of 65 who receives old age pension?
  • 4 Criteria for Identifying Urban Poor

    While there has been some attempt over the past two decades to identify the poor in rural areas, even though very inadequately, in the urban areas, no serious attention has been given to the tasks of identifying the poor correctly. As a result politicians and officials have had a field day in using the identification of poor as a means of political manipulation for decades. We have, therefore, made an effort here to suggest some criteria for identifying the poor for a metropolitan city like Delhi. In fact, the outline suggested here has been the basis for determining the questionnaire that has been used in late 2008 for a first round of survey.

    Again, the principles are the same as for rural areas. First, the criteria should be easily verifiable and directly observable, and hence,


    1 The estimation of poverty by the Planning Commission is done on the basis of a large sample s urvey of consumer expenditure carried out by the National Sample Survey Organisation (NSSO) after an interval of five years approximately. The latest poverty estimates released by the Planning Commission, based on the 61st Round of NSSO of 2004-05, are reported in the 11th Five-Year Plan document (Planning C ommission 2008).

    2 In fact, a many Standing Committees of Parliament have over the years repeatedly castigated the central government to use BPL list resulting from the flawed exercise.

    3 Jalan and Murgai (2007) point out that the ordering of the livelihood category is a serious problem: e g, they say that the assumption that an artisan is always better off than one employed in agriculture or a landowner may not be true. Similarly, Sundaram (2003) refers to a similar problem in that the question assumes an artisan is always better off than a household engaged in subsistence agriculture.

    4 Thus, Alkire and Seth (2008) point out that the more educated and more empowered also face migratory pressures and at the same time many rural poor people are left behind.

    5 The 2002 Census was criticised (Alkire and Seth 2008) that BPL surveys are conducted every five years, but household status can change well before. Hence, we are suggesting a method of updating the list earlier than five years.

    6 A question can be raised why among the minorities

    44 they should be simple and transparent. Only then can they avoid misuse or misinterpretation to suit special interests or used as a basis for building a network of patronage. Second, the social vulnerability criteria are the same, e g, child-headed households, single women and single-women headed households, disabled people and families with disabled people, old people without caregivers, and so on.

    Essentially, there should be three filters to identify the poor, the first being social vulnerability (as just noted). The second filter should be occupational categories – as in rural areas, except that the occupational categories would be different in urban areas. The occupational categories we have suggested for urban areas are: ragpickers, casual daily wage workers, rickshaw pullers, porters, construction workers, street vendors and hawkers, domestic help. The difficulty with this category is how to come up with an exhaustive list to capture the poorest people in the unorganised sector.

    The third filter is completely different and relies upon place of residence – which is absolutely critical as a clear, transparent means of identifying the poor in urban India. In Delhi, we suggested the use of three categories: (1) shelterless, (2) dwellers of unauthorised slums, and (3) dwellers of authorised slums and residents of resettlement colonies (in the latter case only direct allottees would be eligible, not later purchasers). The place of residence is an indicator of access to public services, as it may safely be assumed that the shelterless will have least access to public services, followed by residents of unauthorised slums (where government would have undertaken no upgradation of services), followed by authorised slums and resettlement colonies.

    It is important that as many people as possible participate to this debate of how government should identify poor rural and urban households, and how it should rank them, in ways that are transparent, objective and verifiable. The contribution of the state to the survival with dignity of millions of indigent households in India hinges critically on these decisions.

    of India, only the Muslims are singled out. First, there is ample evidence (see e g, Sachar Committee report) that Muslims have worse social indicators on a number of counts. Second, Muslims in rural areas are a smaller proportion of the total Muslim population than for other communities, and in urban areas the entire urban population (not just Muslims) have better access to services, and thus the rural Muslims are particularly worse off. Third, for other communities there is no particular evidence that the incidence of poverty is worse among them than for the general population. Fourth, STs are over-represented among Christians (another minority community), and they are already being captured in the ST category in our proposed criteria for the 2009 census. And finally, Muslims are the single largest minority community (accounting for 13% of the total population of the country), and not taking the fact into account that rural Muslims are disadvantaged would be a mistake.


    Addison, T David Hulme, Ravi Kanbur (nd): “Poverty Dynamics: Measurement and Understanding from an Interdisciplinary Perspective”, www.chronicpoverty. org/ resources/ working_papers.html.

    Alkire, S and S Seth (2008): “Multidimensional Poverty and BPL Easures in India: A Comparison of Methods”, Ophi Working Paper Series.

    FAO (Food and Agriculture Organisation of the United Nations) (2001): The State of Food Insecurity in the

    may 9, 2009

    World 2001 (Rome: FAO). docrep/ 003/ Y1500E/ y1500e06.htm#P0_2 (consulted 6 March 2009).

    Harriss-White, Barbara (2003): “Destitution in India”, Paper for the Conference on Inequality, Poverty and Human Well-being, Helsinki, 30-31 May. papers-pdf/Harriss-White%20260303.pdf (consulted 6 March 2009).

    Hirway, I (2003): “Identification of BPL Households for Poverty Alleviation Programmes”, Economic & Political Weekly, 38 (45): 4803-38.

    Jain, S K (2004): “Identification of the Poor: Flaws in Government Surveys”, Economic & Political Weekly 39 (46): 4981-84.

    Jalan, J and R Murgai (2007): “An Effective “Targeting Shortcut”? An Assessment of the Below-Poverty-Line Census Method”, mimeo (New Delhi: World Bank).

    Marx, Karl (1862-1990): Capital, Vol 1 (London: Lawrence and Wishart).

    Ministry of Rural Development (2002): Report of The Expert Group on Identification of Households Below Poverty Line (BPL Census 2002) (New Delhi: Government of India).

    Sundaram, K (2003): “On identification of Households Below Poverty Line in BPL Census 2002, Some Comments on Proposed Methodology”, Economic & Political Weekly, 38 (9): 4803-08.

    World Bank Human Development Unit (2008): Social Protection for a Changing India (South Asia: Washington DC).

    vol xliv no 19

    To read the full text Login

    Get instant access

    New 3 Month Subscription
    to Digital Archives at

    ₹826for India

    $50for overseas users


    (-) Hide

    EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

    Back to Top