Home >Mint-lounge >Features >SECC is not something to measure poverty: Abhijit Sen
Abhijit Sen, professor at Jawaharlal Nehru University and former member of the Planning Commission. Photo: Pradeep Gaur/Mint
Abhijit Sen, professor at Jawaharlal Nehru University and former member of the Planning Commission. Photo: Pradeep Gaur/Mint

SECC is not something to measure poverty: Abhijit Sen

Former member of the Planning Commission and a key expert behind the Socio Economic and Caste Census on reframing the poverty debate

New Delhi: Ever since the Socio Economic and Caste Census (SECC) was formally released, it has been embroiled in controversy. First, political parties like the Rashtriya Janata Dal led by Lalu Prasad alleged that the government was being selective and had held back the data on caste. Second, bulk of the media interpreted it as a new measure of poverty, leading to claims that poverty had worsened substantially. To address this and other questions, Mint spoke with Abhijit Sen, professor at Jawaharlal Nehru University, former member of the Planning Commission and chairman of the expert group on SECC to recommend the methodology for determining the class of beneficiaries for different rural development programmes. Edited excerpts:

First off, isn’t the SECC data meant to identify the poor, as opposed to the task of the erstwhile Planning Commission which was entrusted with the responsibility of counting poor? A lot of the narrative appearing in the media seem to confuse the two.

The SECC was never expected to estimate the poor. Nor were any other previous BPL (below the poverty line) censuses. The origins of SECC lie in the original BPL census, which is essentially that while the Planning Commission estimates the poor, we identify the poor. Therefore, we devise some way where a number of people can be fitted in within the number of people which the Planning Commission has indicated.

This was till (N.C) Saxena (committee report on deriving the suitable methodology for BPL census) said that you do that only for a limited set of people. According to him some people should be automatically excluded, similarly some should always be included and the remainder should be ranked. He also added that why have only one cut-off—BPL and APL (above the poverty line). (Instead) why don’t you actually collect data, which can be used to select people for different schemes.

Within BPL you will be identifying people’s eligibility for different schemes using various parameters?

Basically, the idea was that rather than call it BPL or anything, this (census) is a database, which having excluded some people, for the remaining part it is an identification process geared to a particular process. Therefore, in some sense it also ensures better targeting of those schemes.

How does SECC differ from the BPL census?

The BPL census has varied from one census to another. The last one had some 13 criteria, most of which were questions—like has anyone passed higher secondary in your family or does your home have a toilet and so on. This could either be of short- run or long-run in nature—that is, true today not tomorrow. The purpose was to give marks. Not only that, once the marks were given, the original questions disappeared. These were then added up and there was a cut-off mark for each state such that the number of people below that are the guys who will be BPL.

In that sense SECC is scientific?

I don’t know whether it is more scientific, but it is basically saying let us build a database that relates to every individual. And this database should be available to those at the state or central level to design particular programmes. Basically every rural household has one page of data, which has been uploaded onto a central database.

This idea of an online national registry, is it unique to India?

The only one I know is Mexico. I don’t think anyone has done something like this.

You have previously referred to the binary BPL trap in the context of identifying beneficiaries of welfare programmes. Can you elaborate?

The binary BPL trap is the following. All that the data is doing is to ask whether you are BPL or not BPL. After that till the next census takes place you have a BPL card that is used for PDS (public distribution system), free services in hospitals, schools, pensions and so on. It is a card which is worth a hell of a lot if you have it.

Now, the point is that some people may be very deserving of a PDS, in a much larger number. The number deserving of scholarships may be much smaller and certainly those deserving of disability benefits even smaller. It would be very peculiar to say then that only a BPL person with a disability will get the benefit; because there might be a guy who is somewhat richer than the BPL person, but much more disabled. But if you follow the only BPL criteria then what you will do is to deny this person any benefits.

The data is not just a number. It is a full data set. So you can actually rank people for particular schemes. And of course how much you can give and how many people you can give to will depend on your fiscal resources. What this database can do is tell you that given an amount of money, who are the people this can be given to.

Is it like a deprivation matrix?

It is a set of deprivation indicators.

Viewed in this way it reframes the poverty debate, which at present is unidimensional as it is based only on the measure of consumption?

It does. But the main thing is that SECC is not, I repeat not, something to measure poverty. It is something to identify people who you will, with your limited resources, be in a position to give benefits to. In this it is multidimensional. Not because you want to look at the multidimensionality of a particular poverty measure or of a particular individual, but simply because different schemes address different kinds of problems.

Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.

Click here to read the Mint ePapermint is now on Telegram. Join mint channel in your Telegram and stay updated

My Reads Redeem a Gift Card Logout