New Delhi: C.Chandramouli, registrar general and census commissoner of India, is on the threshold of one of the most challenging months of his career. As the head of an army of 2.7 million enumerators who will fan out for almost a month beginning 9 February, Chandramouli talked to Mint about the methods and controversies of the second phase of India’s 15th census exercise. Edited excerpts:

The National Population Register (NPR) seems to be designed to serve as a platform for a lot of other government programmes such as UID (unique identity program); how the rural health mission has fared. What’s the thought behind this?

The office of the registrar general wears several hats. One of the hats that I wear is the Census commissioner. The other hat is the Registrar General of birth index. The third hat is the National Registrar of Citizens. The NPR that you are referring to is done under the Citizenship Act. So this is a completely different exercise than the Census. We are only piggybacking on the first phase of the Census to get the details required for the NPR collected. We have collected and scanned those forms, now we in the process of digitising. After Census is over, from April we will start collecting the biometrics (photographs, 10 fingerprints, 2 iris prints) and this will feed into the UID. After getting the UID, we will be issuing the National Population register card to every usual resident. (Usual residents are defined as ones who are staying or intends to stay at that particular place for at least six months) That will be the basic identity document which will be a smart card and this will be used for various purposes of the government. UID is giving only the numbers. We are taking that number, putting that on a card and giving it to Indian residents. NPR will have residents as well as non-residents. The entire usual residents of this country. The next step of that will be citizens. This is an intermediate stage where we first create the population register and after due verification we will create the citizens register. So this is a different exercise than the Census.

Loading video...

How do you cover the population in temporary households?

There are three types of households in our Census. One is normal household who live in proper structures. The second type is the institutional households like jails, hostels, hospitals, military and paramilitary where people stay for a longer period of time. Then the third type is the houseless. The instruction that we have given is we are also working with NGOs in this sector who are working with street children, homeless that throughout the period of enumeration between 9-28 February, they would mark the areas where homeless are found. And on 28 night we would go about enumerating the houseless. These are nomadic people. There is no point calculating the same people again and again. To reduce that we do it on a single night.

It is difficult to count the NRIs (non resident indian) though they are citizens of this country. By definition, a non-resident is not a resident of this country. When the Census of the country where they are settled is done, they will be covered. However, if they are present in India during the counting, they will be covered.

What about foreigners who live here for an earning?

All foreigners except those of diplomatic status are covered. Diplomatic area is supposed to be the territory of that country. So people of diplomatic status are actually not part of the population of this country. But all other foreigners who are present here in that period, they will be counted as the population of this country. Because we don’t have a column of citizenship in our form. Where the issue comes up is the migration column. We ask whether this is your place of birth and if not what is your place of birth. Second is, how long have you been staying at the place where you are being enumerated. These two things will bring out the foreigners.

What about the illegal immigrants? Don’t they get legitimised by this process?

No. They are part of our population. I am not concerned about their citizenship. I am only concerned about the population of the country and population includes both citizens and non-citizens. The whole process is about aggregates, not about individuals. And the usual resident will be provided a residence card. That does not give them any right to claim citizenship.

The primary aim of the NPR is to provide a fool-proof identity. Perhaps the more challenging part will be to keep that going because this is not something that you have done before. How do you prepare to deal with this process?

This has been going on since 2003 and we have been getting ready for this. So capacity has been added over the years. As far as NPR is concerned, it is an ongoing process. This is the beginning that we have done with collecting the basic information. Once the current enumeration is done, any addition or deletion will be done through the route of applications. You have to apply, your antecedents have to be verified, proper enquiry has to be done with reference to your documents and other proofs. Then you are allowed to enter the NPR.

Has there been any upgradation in technology through which you carry out the Census?

You would be surprised to know that we are one of the most sophisticated censuses as far as technology goes. Right from the time 1961 Census, when computers were first introduced, we have been using some machine or the other: the punch card, the mainframes. We introduced an absolutely state-of-the-art technology in the last Census-the intelligent character recognition. You have heard of machine text being read, but hand-written text being read was something that nobody gave much confidence. Everybody used to think that this is not going to work. Everybody used to think that this is not going to work. Filled up forms in the field with no formal training with so many different languages, whether that could be recognised by these engines, that was the challenge. But we successively proved that it was effective over 95%. Now even the recognition is better with over 98% which means only 2% of the data will be left for manual recognition. So what happened is the time has been crashed from nine years up to 1991 to about four years in the last Census. We are attempting to bring it down to two years. We hope to bring out most of the data by 2013. The other thing is earlier the processing used to be done on a sample basis. In 1991 Census, it was a 40% sample. But since 2001, the data are based on 100% sample which means all the data which has been captured in the form, each and every form is scanned, data is extracted and it is on a 100% basis.

Now you talk of China, it is definitely larger than us. But they don’t have the complications of 18 languages in which instructions have to be given. And they have two ways of surveying: long form like us and the short form. While the short form is a simple six questions which is given to everybody, but the longer forms are done only on a sample basis. Whereas we have two long forms, one for house listing and the other for the main Census and all on the basis of 100% house to house basis.

Could it not be done with handheld devices?

We did see if that could work but the main problem is the language. Where will we get so many data entry operators in 16 languages. Typing is not easy and that too a common teacher can never be able to do that. But we have tried it in the NPR. In all the coastal areas we have captured data directly on laptops. That was successful because that was in English.

There is this issue of privacy with NPR data. While Census is confidential, NPR data cannot be fully confidential.

NPR is not confidential. Electoral roll is not private document, telephone directory is not private document. These are public documents. So what are the elements that we are collecting? Your father’s name, your mother’s name, your permanent and present address. They are all in the public domain in some place or the other. We are not collecting any information which is of private nature. We have taken 15 pieces of information like occupation and educational qualification. But what will be published is what is already there in the public space.

What is the revenue potential for the government from census data?

I am not looking at revenue. I am looking at disseminating the information. This is service. Usually what happens is our data is taken, massaged and sold. I am happy with it, i’m producing the basic numbers. We are looking to see the data is used. I don’t think there is any other source that can give you the granularity of data.

These figures count for something. They are the basic inputs for programmes, for resource allocation. Empirical decision making is done on evidence. You need hard numbers.

You had earlier said grafting the caste census on to main census could vitiate the integrity of the exercise. What did you mean by that?

That is the government stand. We do not take anything in the census which can generate some sort of a controversy. We do not take income, for example. These are contentious issues. It was felt in the government if there is a tendency to boost numbers (by different castes), it would vitiate the integrity of the census as a whole. So, it was felt, you first conduct the census, get the number right, and then the composition of the number can be arrived at. That was the idea. They took a conscious decision, it should be separated and they should be two independent exercises. From June to September (caste census will be conducted).

A single question on caste?

Religion and caste. All castes will be enumerated. Categorisation will not be done by us, it will be done by an expert group, which will go into which caste should be categorised as what because the same caste will be treated differently in different states. We are not experts in that.