Delete the errors to save the census

Have the census enumerators recently knocked on your door with swanky tablet computers in hand? If they have, it's because they have begun to go door-to-door in most States to complete the final phase of the Socio Economic and Caste Census (SECC). This mammoth exercise is being coordinated by more than 7,00,000 enumerators, data-entry operators, supervisors, trainers and government officials.

For the first time since 1931, this decennial exercise will also ask families for their traditional “caste” name. Also as a means-testing exercise, the SECC will single-handedly determine which of India's 250 million families will be eligible for a slew of anti-poverty subsidies — from old age pensions and health insurance to subsidies for housing and electricity. Most significantly, it may be the basis to decide which families receive low-cost foodgrains.

While there has been much debate on the criteria for evaluation used in this exercise, the on-the-ground methodology too needs scrutiny.

Last week, with a billowing dust storm to accompany us in the parched desert landscape of western India, a team of university students and I were witness to this historic exercise in a few villages.

The negatives

Basic information on each household from the National Population Register (NPR) is already pre-loaded on the tablet. Each family is first asked to verify this and then answer a number of additional questions. As we walked door-to-door with the enumerators, we found that each household interview usually does not last for more than 20 minutes. Some are wrapped up in 10.

Though the tablets designed by Bharat Electronics Ltd are sturdy and of a relatively low cost, the main problem with using them is that there is no paper trail for families to verify data. And often there is a slip.

For example, in one case we noticed that the operator had wrongly entered data that the household had two instead of one kachha room. This minor error could bar them from being declared as a poor family.

The operators are also inadequately trained. While the enumerators are usually government employees from secondary or high schools, the operators have been sourced from private companies sub-contracted on short-term contracts. In some States this has resulted in manpower shortages and delays.

The most important area of concern is the definition of a household. The Ministry of Rural Development has clarified that “if any female member of a household [for example, widowed, separated, second wives, single women, etc] decides to declare herself as a separate household, she should be recorded as a separate household.”

All the enumerators and respondents we met were completely unaware of this provision. As a result a deserving, aged, visually handicapped, impoverished widow who lives with her sons (see box) could be pegged as being “above the poverty line” and lose her eligibility to earn old-age pension. Many enumerators also did not know the simple tech-fix of how to split data on households on their tablets.

Land is another question where anomalies could potentially exclude millions of deserving families. The trouble is that most enumerators do not even enquire whether the land owned by a household is irrigated and the number of crops sown each year — both key exclusion criteria.

On a positive note, since the census and poverty identification surveys have been combined for the first time, there is a greater chance of enumerators visiting many more households, unlike in the previous BPL survey exercises.

But this time around, too many villagers are not even aware of this exercise or its significance. Often they are not even at home when the enumerators arrive. So far, the data has also not been published at the gram sabha level for villagers to verify or apply for corrections.

In Peru

In fact, a similar exercise in Peru's Juntos programme is reported to have sown the seeds of discord as even small differences in assets can make a world of a difference to a household's official poverty status. Validation of eligible households through local assemblies too has proved to be divisive, pitching neighbours against each other.

While this unique Indian census is being conducted more systematically and professionally than previous BPL surveys, a number of important hurdles still do crop up on the ground. So before it is rolled out in more States in the summer, this would be the perfect time to refine it, clarify its implications and iron out loose-ends.

(Swati Narayan is an independent social policy specialist. E-mail: swatinarayan@gmail.com)