Choosing between the poor….Anil Padmanabhan
Disparities emerge between official poverty estimates and results of pilot study for BPL census
Disparities emerge between official poverty estimates and results of pilot study for BPL census
New Delhi: The United Progressive Alliance (UPA) is posed with the piquant problem of having to actually choose between the poor with some mismatch between official poverty estimates and the results of a socio-economic survey conducted by the rural development ministry.
The second, a pilot study, was conducted in 254 villages across all states in preparation of the below the poverty line (BPL) census, normally carried out every five years at the beginning of each so-called Plan period, to identify poor households in rural areas that are eligible to receive benefits under the governments various social sector programmes. This year the census is that much more significant as it will form the basis of deciding who will benefit from the proposed food security Act.
The pilot study has come up with estimates of poverty that match official claims only at the national level. The national estimate of poverty, as determined by the Tendulkar panel, puts the proportion of poor in rural areas at 42%. But the estimates differ vastly at the state level; while the pilot study estimates rural poverty in Punjab at 35%, the official estimate puts it at 20%; similarly, for Bihar the pilot reports nearly 70% poor while the Tendulkar panel pegs the proportion at 55.7%.
The discrepancies arise because the official estimates of poverty are based on consumption, while the BPL census uses socio-economic characteristics such as household assets since consumption is impossible to measure.
The two estimates do not differ in their identification of the super-rich (about 26% of the rural population), who need to be compulsorily excluded, or the poorest of the poor (about 32%), who need to be included. The problem arises with those in-between (42%).
The government is, therefore, proposing a compromise formula, wherein it will accept the official estimates as a ceiling and stack the poor in increasing order of poverty. The problem with this approach is that the formula which will be used to stack the households has an error rate as high as 50%, which means that for every person rightfully included in the poverty list, another would be excluded.
If the government goes ahead with this formulation, it could encounter a serious political problem as it will mean excluding some households from the list of entitlements of poverty benefits.
The best thing would be to universalize. The second best thing would be take the essence of the Saxena committee report, which identified three groups: those who can be included, excluded and the ones in between, and provide entitlements without huge differences. The worst thing would be to find scores with which to fit numbers from a BPL census that match consumption poverty…that is statistically impossible and an invitation for corruption, says Abhijit Sen, economist and member of the Planning Commission.
The Saxena panel was appointed by the rural development ministry to suggest a methodology for identifying BPL households. It distributed rural households into three categories: those that have to be compulsorily excluded; those that have to be compulsorily included; and those that fall in-between. In the last category, it recommended a scoring scale so as to restrict the benefits and prevent a fiscal overrun.
Stacking is impossible to achieve in social science. To minimize exclusion/inclusion errors, you need to deduplicate your data base, for which you need biometrics. Without that, using the proposed methodology will not work, a state government official, who did not want to be identified, said.