Will UID fix India's social services? Brazil might say yes.
The AP's South Asia Bureau Chief Ravi Nessman earlier this week wrote that "if technology can fix [India's] Public Distribution System, anything is possible." He was referring partly to the huge problem of fake registrations for the country's food program, a system so leaky that the Asian Development Bank estimates that it only delivers about 10% of its benefits to the intended recipients.
It's important to point out that the chain from government-bought grain to the poor is an extremely complex one, and the graft that occurs goes far beyond fraud at the point of registration. However, is the fact that some recipients are now being registered (under Nandan Nilekani's UID scheme) using biometric data and are given microchip-embedded cards reason for optimism, that some of the leakages of the PDS might be reduced? Nessman seems cautiously optimistic, citing the fact that "once everyone has an ID number, they won't need ration cards. Their information, stored on secure servers, can be verified by a cell phone."
For sure, the use of biometric data is essential to prevent fraud: there has already been several prominently reported incidences of frauds in the registration scheme. With more than 170 million people already in the database, it's critical that points of sale have adequate means to verify people's identities rather than just relying on cards and UID numbers. Nilekani, chairman of the UID scheme, is undeterred, and says the registrations will vastly improve targeting of social services, and serve as "a platform for inclusive growth."
Are these grand hopes justified? Since India's UID will on completion be the world's largest identity scheme, it has no obvious peers. However, the success of Brazil's Bolsa Familia, the world's largest conditional cash transfer program and one of the world's best targeted social welfare schemes, suggests there is cause for optimism.
Bolsa Familia is based on Cadastro Unico, a so-called single registry database started in 2001 that provides beneficiary data for various government programs. The database doesn't provide the specific functions for each program, but does provide a common pool of potential beneficiaries that all programs can access, and has multi-layered verification mechanisms to reduce leakages. Unlike the registration centers used in India, Brazilians are registered for the database by government officials who visit households to conduct interviews. Despite the fact that the data set is vast, with 40 variables for households and 100 for individuals, and the data requirements are strict (most incomplete entries are rejected), the Cadastro had grown to reach more than 92% of the population as early as 2007.
Since the database is not tied to a specific welfare program, all Brazilians are eligible to - and do - register. Local officials have an incentive to register their citizens since more information allows them to build up a more accurate picture of their constituency. As has already been observed by India's Centre for Policy Studies, such data has become a starting point for local planning in Brazil, and that's a model India would do well to emulate.
What's striking is that Brazil has achieved a leakage rate of around 20% - admirable among developing nation social programs - without the use of biometric data (which some academics have criticized for being very insecure). This suggests that following a similar centralised database strategy, but with a stricter identification system (as UID in theory should be), India should be on the path to a much better-targeted PDS.
But Brazil's success with Cadastro Unico isn't down to the enthusiasm of individuals and planning-happy officials alone. Its accuracy is ensured by a complex system of checks and balances at the local and federal level and a sophisticated system of performance management for Caixa, the state-owned bank that operates the CU system and provides the technology for it. The Ministry for Social Development (MDS) is a critical player, since it oversees Caixa's contract and monitors CU's performance - awarding bonuses for targets being met and charging penalties for those that are missed. In its early years the expected miracle provided by Caixa's technology was slow to materialise and the program was beset by hardware and software problems, but the development of these oversight mechanisms by the MDS has helped Brazil to achieve maximum leverage from technology.
The creation of a UID database for India is a promising step, and its technologies and scale have the potential to far surpass Brazil's achievements with Bolsa Familia. But irrespective of its sophisticated technology, the registration scheme will not on its own deliver a revolution in the PDS (or any other social program). For any technology policy to succeed, the humans in the process must immediately see the benefit of participating. That means India should consider how it will incentivise UID's operators and local officials to ensure the database is accurate, and punish them if it is not. The scheme will undoubtedly run into problems, but it needs to be resilient and accountable. If the country can start cutting fraud at the point of entry, perhaps it can then move on to tackling the significant problems elsewhere in the system.