Socio-Economic Caste Census: Has it ignored too many poor households?

N C Saxena, Distinguished Fellow, SKOCH Development Foundation

Two questions are integral to any discussion on poverty: How many poor, and who are the poor? Finding answers to these questions needs different approaches and methodology. An analyses by N C Saxena

Whereas the number of poor can be fixed simply through a sample survey, identifying them would need visiting each and every household. While Planning Commission was the official agency to estimate the percentage of population below poverty line (BPL) once in five years based on the National Sample Survey on Consumption Expenditure, a Census to identify the BPL households has been conducted by the states through the Ministry of Rural Development of the Central Government three times (1992, 1997 and 2002) in the last 25 years. These surveys hardly attracted media’s attention to the extent the debate on the number of poor and fixing the poverty line did. The primary purpose of conducting village-wise census of poor households by contacting each household is to identify the deprived households that could be assisted under various programmes of the Ministry. Such a survey is necessary if there are programmes and benefits exclusively targeted to the poor.

A large number of flagship programmes are universal, such as SSA, MDM, NRHM, ICDS, NREGA, etc. Then there are programmes which could be called ‘BPL Plus’, such as Rashtriya Swasthya Bima Yojana (RSBY), a cashless health insurance cover of the Union government, and construction of household toilets. Here the subsidy is available to all BPL families plus many others such as artisans, domestic workers (for RSBY), and small and marginal farmers, SCs, STs, and women headed households (for toilets). Similarly the rural housing programme gives weightage to other indicators besides being BPL, such as being houseless, or belonging to PTG groups, etc.

In the third category are programmes, such as Old Age Pension, PDS, etc. that are specifically directed to BPL. In addition, each state has a plethora of schemes that benefit the BPL. Hence it is vital to evolve a workable methodology to separate them from the non-poor and identify them truthfully.

There has been acrimonious discussion in India about ‘how many poor’, but not on the more important issue, ‘who are the poor’ and how they are being identified.

There has been acrimonious discussion in India about ‘how many poor’, but not on the more important issue, ‘who are the poor’ and how they are being identified. Many studies have shown that the practice of including those who should not be included, while excluding those who should be included is widespread. Errors of exclusion are those that misclassify the poor in the non-poor category, while errors of inclusion include the non-poor in the poor category. According to the XI Plan (volume 2, chapter 4), there are huge exclusion and inclusion errors in identifying the poor. More than half of the poor in 2004-05 either had no card or were given APL cards, and were thus excluded from the BPL benefits. These must be presumably the most poor tribal groups, women headed households, and people living in remote hamlets where administration does not reach. Thus the people most deserving of government help are deprived of such assistance. On the other hand, almost 60 per cent of the BPL or Antyodaya cards had been given to households belonging to the non-poor category. There is a great deal of clamour amongst all people to benefit from government largesse, and obviously the non-poor are able to get their names included in the list by using political pressure or through bribes. Identifying the poor correctly and the criterion for inclusion is thus central to poverty alleviation.

In view of the inadequacies pointed out by the state governments and others about the methodology followed in 2002 for identifying the poor, the Ministry of Rural Development constituted an Expert Group under my chairmanship in 2009 to recommend a more suitable methodology for conducting the next BPL census with simple, transparent and objectively measurable indicators for identification of BPL for providing assistance under various programmes in the rural areas. My report submitted in August 2009 suggested dividing all rural households into three categories:

  • families who need to be automatically excluded; such as those owning three or four wheeled motorised vehicles, or a mechanised farm equipment, or drawing a salary of over
    Rs 10,000 per month, or employed in government, or paying income tax;
  • the poorest such as homeless, destitute households, ‘Primitive’ Tribal Groups, households with disabled persons as bread-earners, bonded labour, to be automatically included; and
  • the rest of the households to be graded on pre-determined deprivations, and ranked accordingly.

The Ministry of Rural Development accepted the above categorisation with some changes. However the most important recommendation of conducting census and collecting information openly in a gram sabha meeting with video graphing so as to promote transparency and avoid backdoor influences was not accepted. Instead the Ministry decided to send a surveyor to each person’s house to fill up a detailed form. It accepted Home Ministry’s offer of combining caste census with compiling data on economic deprivations, and the scheme was named as Socio-Economic Caste Census (SECC), the results of which have just been declared. Figures for urban areas and on caste are yet to be made public.

The decennial Census methodology is based on a concept that neither the respondent nor the surveyor has any vested interest in giving false information. In the instant case where inclusion determines access to many privileges, the temptation for the respondent to give false information to the outsider surveyor who has no knowledge of that village would be very high. The surveyor too may use this opportunity to extract a bribe for recording answers that may place the respondent in the category of the poor. Although on paper surveyor’s information is to be checked by his supervisor and the Sarpanch but we all know how collusion between supervisors and the ground staff encourages distortion in preparing such lists of beneficiaries. Therefore the chances that the SECC would continue with large errors of inclusion and exclusion cannot be ruled out. There should have been greater transparency and closer involvement of gram sabhas and civil society in the identification process.

Two decisions taken by the last government are relevant to our discussion. First, on October 3, 2011 a joint statement was issued by Deputy Chairman, Planning Commission and Minister of Rural Development that ‘present state-wise poverty estimates using the Planning Commission methodology will not be used to impose any ceilings on the number of households to be included in different government programmes and schemes’. It meant that ministries and states are free to extend benefits of targeted programmes to any number they like. Secondly, and in keeping with the spirit of the above decision, the government passed a Food Security Act in 2013 extending benefits of subsidised food to 70 per cent of rural and 50 per cent of urban population, irrespective of the fact that the percentage of BPL households was only 26 and 14 per cent in these areas. This was keeping with the reality that although the poverty line in 2011-12 was fixed at Rs 27 and Rs 33 per capita per day for rural and urban areas respectively (I call it kutta-billi line, as only cats and dogs can survive on such a low income), 70 per cent of rural and half of urban people lived below Rs 50 and Rs 70 per capita per day respectively, which was barely enough to meet all basic needs. Thus the most important targeted programme of PDS was de-linked from the Planning Commission numbers on poverty.

In tune with the above decisions the SECC survey rightly decided not to give any finding on the number of poor in any village or district or state. It is a door-to-door census covering the entire country to study socio-economic status of each rural household and allows ranking of households based on predefined parameters so that state government departments can objectively prepare the list of beneficiary families and decide their number according to their budget and guidelines. It is expected that while doing so, they would first select all those who have to be compulsorily included and then take those who suffer from all the seven deprivations, and then those with six, and so on.

Coming now to the most important question: how reliable and authentic is the SECC data? One would be able to pass judgement on this only when the SECC lists are verified in the coming months by objective ground level tests by the civil society and professional organisations. However, a cursory look at the macro data does raise some doubts about the methodology followed and the results obtained.

According to the press release, out of the total 17.91 crore households, those who were excluded on the basis of their assets or profession were as high as 7.05 crore (39.39 per cent), whereas the poorest of the poor to be automatically included were only 16.50 lakh (0.92 per cent). This left 10.69 crore households who were considered for deprivation, but strangely out of these as many as 2 crore households did not report any deprivation! This does hint that the deprivations were very narrowly defined, leaving out 2 crore households who, despite not having sufficient income or assets to fit into the category of compulsorily excluded, would not be entitled to any benefit from the government.

It was expected that the list of deprivations on which households are to be measured would be so comprehensive as to include all the rural households, leaving apart those who have been compulsorily included or excluded. In other words the three categories in aggregate should have covered the entire rural population. Unfortunately this has not happened, and the number of households not reporting any deprivation is as large as 2 crore! It is, I would argue, because the deprivations have been very narrowly defined.

For instance, according to SECC the total number of female headed households was 12.83 per cent (2.23 crore) of the total, but out of these only 3.85 per cent (69 lakh) satisfied the deprivation criteria of not having any adult male member between 16-59 years, thus depriving almost three-fourth of female headed households from being considered for any government benefit. The narrow definition used in SECC meant that a widow who is supporting her 17 year son’s education through petty cultivation is not to be considered for any government benefit! My committee had suggested that all female headed households should be compulsorily included in the list of deprived households. This was unfortunately not accepted. At least the definition should have included all female headed households who do not have any working and earning male member in the family. The number of 69 lakh female headed households who are considered deprived by SECC needs to be compared with another set of two figures, both collected by SECC; 1.81 crore female headed households where monthly income of the highest earning household member is less than Rs 5000, or with 85 lakh female headed households who are without any mobile or landline.

“Uddi Gujjar, a widow in her late 30’s, lost her husband 10 years ago and lives with her two sons, Hanuman (16) and Prithviraj (15), in Tikel Purohitaan village of Jaipur district in Rajasthan. She has around 1.35 hectare of rain-fed farmland. She has been a beneficiary of BPL all these years and gets 25 kg per month of cereals at Rs 2 kg from the ration shop. But now she could be dropped out from the BPL list under the SECC. She says, “I live in a house made of mud, but it has a roof of cement and it is being considered as pukka. My 16- year-old son is being considered adult under the survey, but he is not allowed to take work under MGNREGA.” She questions the new criteria and asks the decision- makers to visit her house to assess whether she should be a beneficiary or not.”
Down to Earth, 10th July 2015

According to SECC itself, 13.35 crore households (74.5 per cent of the total) have monthly income of the highest earning household member as less than Rs 5000, whereas households with any one of the 7 deprivations is only 8.69 crore. Does it mean that almost one-third of those earning less than Rs 5,000 (Planning Commission’s poverty cut-off line was Rs 4080 per month per household in 2011-12) do not have any deprivation? It is also surprising that even in rural Kerala where only 9 per cent people are below the poverty line, households with less than Rs 5000 income is 70.5 per cent, almost the same as in UP (71.5 per cent). In any case it would have been better if the total household income was assessed, rather than just the income of the highest earning member of the household.

Kerala data on phones is also intriguing. SECC shows only 3.63 per cent households own a landline phone in the state, but the number of households owning both landline and mobile is shown as 28.33 per cent. How can this number be more than 3.63 is a mystery to me. Interestingly, the data shows that 87 per cent households in UP have a mobile, whereas only 61 per cent families in Kerala have it! Moreover household without any phone in Kerala is shown as only 7.17 per cent, which does not jell with other numbers given above.

The most disturbing data is about those poorest households who had to be automatically and compulsorily included. These are households without shelter; destitute, living on alms; manual scavenger families; primitive tribal groups; and legally released bonded labour. Their total number is shown as only 16.50 lakh, or much less than 1 per cent of all rural households in India. This gross under-estimation of the poorest of the poor does not tally with other government figures. For instance, the overall number of 16.50 lakhs is far less than the number of houseless alone as estimated by the Ministry of Rural Development. The ministry’s target in the next four years to provide IAY house for all rural homeless is 2 crore families! However, the two definitions are different leading to wide disparity in figures; but shouldn’t we have worked on common definitions so that the results of SECC can be operationalised in the field and IAY beneficiaries selected according to the SECC, and not be subject to pulls and pressures so common in IAY scheme. SECC is also not compatible with Census; as according to Census 2011, 4.3 per cent households (74 lakh) in rural India are homeless whereas another 39.4 per cent (6.6 crore) live in just one room. As against 6.6 crore, SECC identified only 2.37 crore households which included not only households with one room with kuccha walls and kuccha roof.

I will be very happy if I am proved wrong by detailed ground level analysis and verification. If the list is correctly prepared and to the satisfaction of rural people, it can certainly eliminate leakages and errors of inclusion and exclusion.

A survey done by the Ministry of Labour in 2014 shows that in about 5 per cent of the households, there are no workers aged 15 years and above. These 90 lakh households should have been automatically included in this category. This number is shown in the SECC as only 6.7 lakh, or 0.37 per cent as opposed to 5 per cent estimated by the Labour Ministry. Coming to the number of manual scavengers, according to a Lok Sabha report, 7.70 lakh manual scavengers and their dependents were identified by States/UTs during implementation of the National Scheme for Liberation and Rehabilitation of Manual Scavenging in India in 2007. SECC admits far less, only 1.8 lakh. Mihir Shah, who as a Member, Planning Commission was incharge of rural development, also thinks (The Indian Express, 10th July 2015) that the number of manual scavengers is greatly underreported, especially in certain states that remain in insistent denial about the existence of the problem, which in reality refuses to go away, given deeply entrenched social prejudices. While the 2011 Census of India lists 2,15,885 dry latrines in Assam, Andhra Pradesh, Tamil Nadu and Manipur, the SECC reports a mere 653 manual scavengers in these states!

The inference is inescapable; the poorest have not been able to convince our enumerators to recognise them.

Whereas the poor have gone largely unrecognised, there appears to be over-reporting on the number of rich households who were excluded from the survey if they fulfilled any of the 14 parameters of exclusion such as owned tractor or paid income tax etc. Their number is reported as 7.05 crore. If we add another 2 crore households who did not have any deprivation, those excluded from all government benefits would come to 9.05 crore or more than half of rural population. What is the justification then to give food subsidy to 70 per cent households when more than half are quite prosperous? Interestingly the BPL pilot survey had indicated that only 28 per cent households would be excluded, but the actual exclusion is more than half of all rural households.

Here again the details show many anomalies. The indicators in the ‘automatic exclusion’ too call for some discussion. Owning a motorised two-wheeler, for example, is an inaccurate measure of ‘richness’ considering it could have been given as dowry or bought on credit.

Then the number of rural households with regular salaried job in government is shown as 89.6 lakh, excluding contractual employees such as Anganwadi workers and para teachers. This needs to be compared with the total number of government servants in India, which as per the Labour Ministry is 176 lakh, including all categories. Do we then conclude that the families of more than half of them live in rural areas?

These are my initial reactions. Either the definitions of deprivations are deliberately kept narrow so as to exclude a large number of people, or enumerators have lived upto their past reputation of collusion with the rich. In the process a large number of poor would be denied access to targeted programmes. Government has not been able to eliminate poverty, but it has certainly succeeded in eliminating the poor from its radar.

I will be very happy if I am proved wrong by detailed ground level analysis and verification. If the list is correctly prepared and to the satisfaction of rural people, it can certainly eliminate leakages and errors of inclusion and exclusion, so common in welfare programmes. Let us hope that the process has remained immune to political or local interference and the rich and powerful have not tampered and corrupted the list.

Interestingly, the Ministry of Consumer Affairs, Food and Public Distribution has been pestering the states to revise their list of entitlement for subsidised food as per the SECC results (these were available for many states for the last six months or more) so that they can be entitled to enhanced quota under the Food Security Act. However states are dragging their feet and wish to continue with the old list prepared in 2002 and revised from time to time. Whether it is due to the pressure exerted by the vested interests of the old beneficiaries, or the states doubt the findings of the SECC, which they themselves collected is not known.

(The views and opinions expressed in this article are those of the author and do not necessarily reflect the views of INCLUSION. Comments are welcome at

comments powered by Disqus