Identifying a billion Indians

Economist: Identifying a billion Indians

Reliable identity numbers could create many opportunities for business.

IN A small village north-west of Bangalore, peasants queue for identities. Each man fills in a form with his name and rough date of birth, or gets someone who can read to do it for him. He places his fingertips on one scanner and stares at another. A photograph of his face is snapped. These images are uploaded to a computer. Within a few weeks he will have an identity number.

The Indian government is trying to give all 1.2 billion Indians something like an American Social Security number, but more secure. Because each “universal identity number” (UID) will be tied to biometric markers, it will prove beyond reasonable doubt that anyone who has one is who he says he is. In a country where hundreds of millions of people lack documents, addresses or even surnames, this will be rather useful. It should also boost a wide range of businesses.

So far the process has gone smoothly. More than 1m people have been enrolled since October, and the pace is accelerating. It needs to: 1m is less than 0.1% of the population. The scheme presents difficulties both for the people in charge, many of whom were recruited from software firms, and for the private contractors who are doing much of the work. How do you ensure that the data are accurate? How do you build a robust database containing biometric information about more people than any other? How do you deal with peasants whose fingerprints are unreadable after years of manual work? (Adding moisture to their fingertips helps.)

When an individual is enrolled, his biometric data must be compared with everyone else’s to ensure there is no duplication. Sometimes the workers who show people how to place their fingers on the scanner accidentally scan their own fingerprints. As enrolments hit a peak of about 1m a day, the system will need to carry out a staggering 14 billion matches per second.

This mighty task has been awarded to private contractors in an unusual way. There are three vendors: Accenture and L-1 Identity Solutions of America, plus Morpho of France. The firm that does the fastest, most accurate job gets 50% of the work; the others get 30% or 20%. This allocation is frequently reassessed, so if the second-best firm starts doing better, it picks up some work from the leading firm. This keeps everyone sharp.

One database, many possibilities

The government’s aim is to improve services and reduce corruption. A shocking two-thirds of the subsidised grain that the government allocates to the poor is either stolen or adulterated. When middlemen say they have delivered so many bags of rice to so many thousands of peasants, there is no way to tell if they are lying. But if each peasant has to scan her irises every time she picks up her ration, it will be harder to scam the system. Similar controls could be used to curb voter fraud.

A reliable way of identifying people would also smooth financial transactions. …