Experiments with Aadhaar-Bharat Bhatti, Jean Dreze and Reetika Khera
Technical glitches in the unique identification method make it unreliable in disbursing wages under the employment guarantee scheme
Within a few weeks of “Aadhaar-enabled” payments of Mahatma Gandhi National Rural Employment Guarantee Scheme wages being initiated in Jharkhand, earlier this year, glowing accounts of this experiment started appearing in the national media. Some of them also gave the impression, intentionally or otherwise, that this successful experiment covered most of Jharkhand. A fairly typical excerpt, which condenses five grand claims in a few lines, is as follows: “As the new system ensures payment of wages within a week, the demand for work under MGNREGS has gone up. Consequently, migration has been checked, families have been reunited and, no less important, some workers have a saving in the bank.”
Enthused by these upbeat reports, we tried to trace the evidence behind them, but quickly reached a dead end. The authorities in Ranchi referred us to the website of the Unique Identification Authority of India (UIDAI), but we did not find any evaluation of the experiment there or, for that matter, any details of it. There was no alternative, it seemed, than to check the facts for ourselves.
We headed for the Ratu Block in Ranchi District, the source of most of the reports. It was, at that time (early March), one of the five Blocks where the experiment had been launched. On arrival, we found that only three gram panchayats (GPs) were involved, out of 14 in Ratu Block. The showpiece appeared to be Tigra GP, but it turned out that even there, only one worksite had enjoyed the blessings of Aadhaar-enabled wage payments. In the three GPs together, the system had been implemented at five worksites, employing a total of about 50 workers. We managed to interview 42 of them with the help of a small team of student volunteers.
The main role of Aadhaar in the Jharkhand experiment is to facilitate the implementation of the “business correspondent” (BC) model. Under this model, accredited agents provide doorstep banking services to MGNREGS workers using a micro-ATM. They act as extension counters of the local bank (in this case, Bank of India), disbursing wages close to people’s homes. Biometric authentication is meant to prevent identity fraud, e.g. someone’s wages being withdrawn by someone else. Aadhaar is one possible foundation of biometric identification, though not the only one. In this approach, wages are paid through Aadhaar-enabled accounts that are supposed to be opened at the time of UID enrolment. Authentication requires internet connectivity, so that workers’ fingerprints and Aadhaar numbers can be matched with the UIDAI’s Central Identities Data Repository.
The BC model widens the reach of the banking system in rural areas. This, in turn, helps to bring more MGNREGS workers under the umbrella of the banking system, as opposed to post offices, where corruption (including identify fraud) is a serious problem. Doorstep banking facilities are also a significant convenience for workers in areas where bank offices are distant, overcrowded, or unfriendly.
A little farcical
Coming back to Ratu, some aspects of the experiment were a little farcical. For instance, on one occasion, workers from Tigra were asked to collect their wages 10 kilometres away, so that Aadhaar-enabled payments could be done in front of a visiting Minister. On a more positive note, the system seemed to work, at least under close supervision. Further, most of the workers had a positive view of it. They appreciated being able to collect their wages closer to their homes, without the hassles of queuing in overcrowded banks or of depending on corrupt middlemen to extract their wages from the post office. They did not fully understand the new technology, but nor were they afraid or suspicious of it.
Having said this, there were problems too. Dependence on fingerprint recognition, internet connectivity, and the goodwill of the BC created new vulnerabilities. Fingerprint recognition problems alone affected 12 out of 42 respondents. Some workers did not have a UID number, and some had a UID number but no Aadhaar-enabled account. None of them had received bank passbooks, making it difficult for them to withdraw their wages from the bank when the Aadhaar system failed.
Four respondents were yet to find a way of getting hold of their wages. Otherwise, the payment of wages was reasonably timely, but this had more to do with intensive supervision than with Aadhaar. It is important to understand that Aadhaar, on its own, is of limited help in reducing delays in MGNREGS wage payments. This is because the bulk of the delays occur before the banking system is involved — at the stage of submission of muster rolls, work measurement, preparation of payment advice, and so on. At every step, there is a lot of foot-dragging, and Aadhaar is not the answer.
(According to the MGNREGA Commissioner in Jharkhand, quoted in one of the articles mentioned earlier, “Against one month now, payments will reach workers’ accounts in one week.” This statement is typical of the delusional mindset of the Jharkhand administration. Not only are current delays much longer than one month, the claim that Aadhaar will reduce them to one week has no basis.)
What next? It is easy to envisage a certain way of extending this experiment that would turn it into a nightmare for MGNREGS workers. Three steps would be a potent recipe for chaos: depriving MGNREGS workers of bank passbooks, imposing the system even where there is no internet connectivity, and insisting on a single bank operating in each Block (the odd “one Block, one bank” rule). All this may seem far-fetched, but there are precedents of this sort of irresponsibility. Short of this, if the Aadhaar-based BC model is hastily extended without the system being ready (as happened earlier with the transition from cash to bank and post-office payments of MGNREGS wages), it could easily compound rather than alleviate other sources of delays in wage payments.
It is also possible to see a more constructive roll-out of the BC model across the country. In this constructive approach, the BC model would act as an additional facility for MGNREGS workers, supplementing ordinary bank procedures instead of becoming a compulsory alternative. This would enable labourers to bypass the BC in cases of fingerprint recognition problems, or when the BC is corrupt or unreliable. For this purpose, the first step is to issue bank passbooks to MGNREGS workers — this had not been done in Ratu.
The question remains whether Aadhaar adds value to other versions of the BC model. In the adjacent Block of Itki, the BC model is being implemented without Aadhaar, in partnership with FINO, a private company. Workers’ fingerprints are stored on a smart card, used for authentication and tamper-proof record-keeping. This obviates the need for internet connectivity, an important advantage of the Itki system in areas like rural Jharkhand.
Aadhaar, for its part, has two potential advantages. First, it facilitates multiple biometric applications based on single UID enrolment. Second, Aadhaar facilitates “inter-operability”, that is, linking of different UID-enabled databases. But the same features also have costs. For instance, dependence on a centralised enrolment system (as opposed to local biometrics) makes it much harder to correct or update the database, or to include workers who missed the initial enrolment drive. Similarly, inter-operability raises a host of privacy and civil liberties issues. A brief exploratory visit to Itki did not uncover any obvious reason to prefer the Aadhaar system to local biometrics.
It is also worth noting that the Jharkhand experiment is a very poor cousin of much earlier and larger efforts to implement the BC model in Andhra Pradesh. Unlike UIDAI, the government of Andhra Pradesh has conducted serious experiments with the BC model and learnt from them. Biometric micro-ATMs are now being installed at local post offices, an important idea for the whole country: micro-ATMs could give post offices a new lease of life as effective payment agencies.
In short, Aadhaar-enabled payments for MGNREGS workers raise many issues that are yet to be properly examined and debated. The Ratu project, for one, looked more like a public relations exercise than a serious experiment. Incidentally, we learnt in June 2012 that Aadhaar-enabled wage payments had been discontinued in Tigra, due to resilient fingerprint recognition problems. That, of course, was not reported in the national media.
Last but not the least, it is not clear why MGNREGS should be used as a testing ground for UID applications when other, more useful options are available. For instance, UID could be used quite easily to monitor office attendance of government employees.
The social benefits are likely to be large, and this is a more natural setting for early UID applications than the jungles of Jharkhand. Any takers?
The Hindu, 27 June, 2012, http://www.thehindu.com/opinion/lead/article3573730.ece?homepage=true