Tuesday, December 6, 2011

India: How Much Fraud is Acceptable in NPR, UID

Home Minister expresses concern on UIDAI data collection process (Deccan Herald)

It's hard to read the above linked article as other than a continuation of the rivalry between the National Population Register (NPR) being assembled by Home Minister Chidambaram and the Unique ID (UID) Project being conducted by UIDAI and its chairman, Nandan Nilekani. The two organizations have overlapping mandates and heat generated between the two has been vented through the press before.

But a very interesting question is concealed within Home Minister Chidambaram's statements.

Home Minister Chidambaram says:
"The data collected by multiple registrars of the UIDAI does not meet the degree of assurance required under the NPR from the point of view of internal security," he said.

The Home Minister said the UIDAI process of enrolment is based on production of documents and, in the absence of documents, through an introducer based mechanism. It was due to the fact that document based systems are not feasible in rural areas especially among the poor, illiterate, landless and women.

"If the UIDAI process is to be introduced in NPR, it would lead to large scale exclusions. The possibility of inclusion of non-usual residents in the local register and the creation of false identity profiles is also real. This would defeat the purpose of creation of NPR. There are also a number of legal, technical and practical issues that makes it difficult to accept data collected by other registrars," he said in the letter.
So, how do you go from a situation where you have a billion people and no rigorous ID management to a system where everyone has a permanent, singular, legally-enforceable, government-backed identity?

Answering two other questions will help answer the first.
How much time and money can be spent?
What error rate (level of deception) is tolerable?

India has (hundreds of?) millions of people without ID resulting in a social welfare system that is rife with corruption and an entire underclass that lacks access to the tools that allow for social mobility and the enjoyment of basic rights. Some of the individuals lacking ID don't even know all the information that will be asked of them in obtaining an ID document.

What is your date of birth?
Where were you born?
What is your father's name?

I don't know.
When this is the real root of the problem (and Chidambaram acknowledges this with his mention of the poor, illiterate, landless and women) how much worry about the "inclusion of non-usual residents in the local register and the creation of false identity profiles" is appropriate?

Bringing full citizenship rights to poor, illiterate landless Indians brings with it a near-certainty that full citizenship rights will be conferred upon some number of poor, illiterate, landless non-Indians living in India. Is it worth it?

Creating an identity profile for people who lack the details required to create a rigorously complete identity profile means that the system will be open to some measure of deception. If a person has no ID and must be trusted to provide information to obtain an ID, they can either err or lie in the information they provide. How much error/deception is acceptable?

Even with unlimited resources and unlimited time, something both NPR and UID lack, imperfection is guaranteed.

The saving grace is that through database de-duplication, everyone only gets one ID. What was fluid solidifies.

So, yes, some sort of illegal resident amnesty and opportunity for "individual self-reinvention" is going to happen as a result of either the NPR or UID. This must be balanced against the known ills associated with the lack of a functional ID management infrastructure.

The existence of NPR and UID means that these questions have already been answered. Lots of money is available. Reasonable efforts to prevent wholesale fraud will be undertaken. Some abuse will happen, but it's worth it to both bring huge numbers of poor people into the system and to make a sharp break with a past where poor ID management made it too easy to operate invisibly for any purpose whatsoever.

There may be a real difference between the auditing processes used by NPR and UID. NPR's may be more rigorous and, in the view of Mr. Chidambaram, superior. Fortunately, the truth will come out. The expenditures of both efforts will be known. The number of credentials issued by each will be known. And at some point, it will be possible to compare the two systems to each other both in terms of their efficacy and their cost.