Rural deprivation7 min read . Updated: 07 Aug 2015, 12:28 AM IST
The problem with the SECC is the absence of cross-tabulations showing the intersections between the seven deprivation sets
The original intent of the Socio-Economic and Caste Census (SECC), whose findings for rural India were made public in June, was to collect information on economic and caste identifiers for access to subsidized food under the National Food Security Act of 2013, and to define a priority set with higher access and deeper subsidy margins. The need for the second disappeared with the law having done away with priority access. Since I was not approached by an SECC enumerator myself, I was quite pleasantly surprised to learn that the census had indeed been conducted, and the rural results tabulated and available. Although just released, the data refer to 2011.
There have been earlier full enumeration attempts to identify poor households, in 1992, 1997 and 2002. The first two identified the poor based on a prior template of criteria. The third actually ranked all households based on thirteen indicators. The SECC is, however, the first to have collected caste information, not yet released, beyond the usual umbrella categories of scheduled castes and tribes (SC/ST).
Voices not possible to ignore, such as that of a former Registrar General of the Census, say that the data released are not fully verified, and that there was no nation-wide uniformity in data entry procedures. Public notification of the data by panchayats for verification purposes was done patchily across the country, more in the south than in north India.
Till revisions are issued, we are compelled to use the data as reported on the SECC website. The context of the food security bill informs the manner of presentation, and the terminology used. So, out of a total of 179.1 million rural households, 70.5 million are reported as automatically excluded, based on any of 14 parameters (such as ownership of assets like refrigerators and landline phones, employment with the government, a member earning ₹ 10,000 or more per month). At the other end of the spectrum, 1.7 million are automatically included, based on five parameters, which include livelihood from manual scavenging, and former bonded labourers. The estimated number of these is far too small, lending credence to those who question the data.
It is the remaining universe of 106.9 million households, neither excluded nor included outright, which were assessed for incidence of the seven selected deprivation indicators. From among these, 20 million are reported to not have had any of the deprivation indicators. That makes for 86.9 million deprived by one of the seven criteria—nearly half the total number of rural households.
It is important not to look at this consolidated number, and address it through some blanket measure. Each type of deprivation is quite distinct and asks for an intervention specific to it.
Landless households deriving a major part of their income from manual labour (the intersection of landlessness, and livelihood from manual labour, each of which is larger taken independently), are reported at 53.7 million, 30% of rural households. The key ratio to focus on within this category, is the ratio of adult income earners to dependants. The range from zero (no able bodied adults, whether resident or absentee remitters), up to 0.5 (implying one earner supporting two dependants in addition to himself), marks the severely deprived, in need of urgent income support.
The earner dependant ratio is a robust marker to go with, even though there are large disparities between manual labourers in the wages commanded by them. From a study of rural labour markets I did many years ago, I found that the competition to get the best labourer for a day’s ploughing made for a wage ranking, which persisted across all days within a season. High-wage labourers were also at the top of rankings by days of employment. The top-ranked labourers were typically from the scheduled castes, and were calibrated to performance in a caste-blind way.
The attempt among employers to reserve the best for themselves by entering into longer-term contracts, for a few weeks or even a whole season, was often rebuffed by the labourers themselves, who preferred to be hired by the day on wages determined by competitive bidding.
The problem even in households with high earning manual labourers, then as now, was that if the major earners fell ill for even a short period, the household crashed. So the factoid yielded by the SECC, of 30% of rural households sourcing their income from manual labour and owning no land, marks the subset of rural households in urgent need of insurance against disability and ill health. And successful households faced, then as now, the problem of finding a secure depository for their savings, towards consumption smoothing or asset acquisition.
The problem with the SECC as currently presented is the absence of cross-tabulations showing the intersections between the seven deprivation sets, although this could easily be remedied going forward. Households without a literate adult over the age of 25, account for a quarter of total households—the second highest incidence among the seven indicators. This is a grim reminder of the poor reach of primary education until the 1999-2000 policy mandating primary schools within a one kilometre radius of all pupils. Cross-tabulations will most likely reveal that these households are fully contained within the set of landless manual labourers.
Aside from the earner dependant ratio, a robust signaller of deprivation among the seven is quality of housing. The SECC count of rural households living in a single room structure with kutcha walls and a kutcha roof is 23.7 million, 13.27%. This number would have been much larger if the indicator had been just quality of wall. In a research paper I published with Abhiroop Mukhopadhyay using an NSS housing survey for 2008-09, we found that 40% of all rural households lived in structures with kutcha walls, but only 17% also had a kutcha roof. (In this respect at least, the SECC figure is entirely consistent with NSS data, since it reports the further subset of these with just one room at 13.27%). More than half of kutcha-walled structures carried a pucca roof of tile or asbestos, and this was true even in the poorest quartile. But the sharp drop in the share of kutcha-walled structures going from the lowest to the second-lowest quartile, revealed the high preference for pucca walls even among poor households.
This clear preference is constrained by the illiquidity of the asset acquired. The transition to a pucca wall is irreversible in the face of downside income risk, unlike acquisition of an electric fan or cycle that can be sold or pawned. So a pucca wall, notwithstanding the preference for it, would be transitioned into only when the household feels sufficiently sure of not having to cover downside income risk, to that extent. A kutcha wall becomes thereby a signal of lack of confidence in the sustainability of current consumption levels.
The problem with taking wall quality as an entitlement indicator of course is that it alters the preference structure of households if the transition to a pucca wall will result in loss of entitlement. So on balance the set in need of income or subsidy support is best defined by the ratio of income earners in the household, summing together present and absentee, to dependants. These have to be added on to the core deprived (the included set, with a more plausible count). There is also a critical need for income support for individuals with different types of disability, regardless of the households in which they may be embedded.
That leads me to my final point, that the household classification says nothing about the deprivation status of individuals within it. The total number of households in the country aggregating across rural and urban, at 243.9 million, is only twice that in the US, which has one-quarter the population. Clearly, household size in India is twice that in the US. Household size varies inversely with prosperity. There are economies of scale to people living together and sharing the costs of cooking and shelter.
That does not mean the benefits accrue equally to all members. A household with an exclusion characteristic like ownership of a motorized fishing boat, or a member earning ₹ 10,000 per month, does not necessarily lift all members out of deprivation. This is where the National Family Health Survey (NFHS) becomes so important, yielding as it does the health and well-being status of adults and child members of households. Results from the fourth round of the NFHS conducted in 2014-15 should be released soon.
The best feature of the SECC was the call for verification at panchayat level of the deprivation classification, but it does not seem to have been heeded very widely. Meetings of the panchayat general body, the gram sabha, are normally held across the country on 15 August. It is not too late for the ministry of rural development to issue a country-wide call for verification on this day. The tabulations will need to be corrected accordingly.
Indira Rajaraman is an economist and is currently on the board of directors of the Reserve Bank of India.