An imaginary friend

An imaginary friend

Claims that Aadhaar has increased the demand and allocations of NREGA work rest on several faulty assumptions.

aadhaar, aadhaar MGNREGA, aadhaar boosting mgnrega, aadhaar card, mahatma gandhi nrega, Indian School of Business hyderabad, Indian School of Business nrega report
The media reports on the claim that Aadhaar reduced wage payment delays is misleading. That the DIRI team went to the press without verifying their results is irresponsible and can potentially misdirect a crucial policy debate. (Source: PTI/File)

Written by Sakina Dhorajiwala, Anmol Somanchi, and Rajendran Narayanan

Two newspaper articles in Times of India and Financial Express claimed that Aadhaar linking boosts MGNREGA and cut wage payment delays. Both articles were based on a working paper titled “A Friend Indeed” by members of the Digital Identity Research Initiative (DIRI) at the Indian School of Business (ISB), Hyderabad.

The ISB study uses MGNREGA and drought data from 2011-2017. It deploys a statistical technique called “Difference in Differences” (DID) to causally claim that demand and allocation of MGNREGA work during periods of drought has increased after introduction of Aadhaar making MGNREGA “counter-cyclical”. The conclusions, however, rest on shaky grounds of incorrect assumptions and ignores important details that affect MGNREGA.

First, the study conflates different financial interventions introduced in MGNREGA such as Direct Benefit Transfer, Aadhaar seeding, Aadhaar payment bridge, and JAM as a solitary entity and refers to them as “Aadhaar linked payment (ALP)”. Each of these have different structures and were rolled out across years. Clubbing all of them as ALP obfuscates the complex reality. Further, the authors believe that prior to what they call “ALP”, wages were paid through cash payments. The study states ” … until 2014, a MGNREGA worker had to physically visit the designated government office, wait in line, submit his job card to collect the wages in cash from a government official.”. This is far from truth. On the contrary, the MGNREGA payments system has undergone many changes since the programme’s inception.  With the electronic Fund Management System (e-FMS) that started as early as 2011-12, and subsequently through National e-FMS, wages have been  credited directly to banks/postal accounts of workers by the Centre.


Second, the working paper completely ignores the Supreme Court (SC) case between Swaraj Abhiyan and the Union of India which started in 2015. The SC stressed upon the government to implement the provisions of MGNREGA, especially in drought affected areas, such as providing 50 additional days of work per household, reducing wage delays, and clearing pending wage payments. Taking cognizance of the SC orders, the Ministry of Rural Development (MoRD) issued strict guidelines to this effect. Further, central and state governments, have, at various times during 2014-2017 issued directives to enforce orders of allocating additional days of work in drought-hit areas. These judicial-administrative directives, which came in effect after Aadhaar was introduced, perhaps played an important role in the increase in MGNREGA work uptake in drought areas.

Third, a critical assumption underpinning the validity of the DID technique used in the study is the “parallel trends” assumption. This implies that, before the introduction of Aadhaar, the trends of MGNREGA outcomes be roughly parallel for drought and non-drought areas. While checking for this assumption, however, the authors seem to cherry pick their data. In the latest version of the working paper, they present the trends for only one year of the pre-ALP period despite their dataset spanning at least three such years (2011-12 to 2013-14). Further, the test for the assumption is presented only for household-level trends. The assumption does not seem to hold for individual-level outcomes seen over 2 years of the pre-ALP period. These observed discrepancies raise serious questions over the validity of the results.

Regardless of the glaring methodological concerns, the plausible mechanisms that the authors posit to explain their results hold no ground. On the demand side, they argue that timely payment of wages explains the increased demand. This is completely misleading since the study only uses data of Stage 1 delays which  corresponds to the time taken by the blocks to generate electronic pay orders called the Funds Transfer Order (FTO). The FTO is subsequently processed by the Centre and then wages are transferred to the workers’ accounts. The time taken by the Centre to release wages after receiving the FTO are the Stage 2 delays. Aadhaar has no role to play in reduction of Stage 1 delays. In our recent work on MGNREGA, we showed, by analysing 9 million transactions, that Stage 2 delays alone were, on an average, 50 days in 2016-17 — a fact acknowledged by MoRD. Without accounting for Stage 2 delays, the claim that Aadhaar has reduced delays in wage payments is just wrong.

On the supply side, it is posited that ALP enhances state capacity by allowing the government to monitor the program better. However, the entire digital trail to monitor MGNREGA, such as electronic muster rolls and payments to the workers’ accounts, existed even before ALP. Consequently, ALP neither has any role in identifying areas of distress nor does it generate any unique data for the government to better “monitor” the programme. ALP would help with identifying ‘fakes’ and ‘duplicates’, but an RTI revealed that they accounted for only about 1.4% of total households in 2016-17. Instead, the judicial-administrative interventions mentioned above appear to be a promising mechanism to explain both the demand and supply side increases.

The study doesn’t consider the issue of exclusions resulting from various technological failures of Aadhaar. Nonetheless, Ashwini Chhattre of ISB claimed in a media report that “The interesting part of MGNREGA is that it does not have the biometric authentication requirement because the job card is directly linked to the Aadhaar and the bank account, the wages are paid directly into the bank, so there is no need for biometric authentication”. This is untrue since payments in many places, are being done through Banking Correspondents using Aadhaar-based biometric authentication. Further, contrary to what the study claims, it is clear from ground reports that MGNREGA workers were not “encouraged” but rather coerced to get an Aadhaar by mandating it for receiving wages.

In addition to the shortcomings of the paper, as highlighted above, the misleading media reportage on the claim that Aadhaar reduced wage payment delays is deeply problematic. That the ISB team went to the press without verifying the veracity of their results is irresponsible and can potentially misdirect a crucial policy debate that has far-reaching consequences