Who Is Responsible When Technology Fails the Marginalised?

 The digitisation of welfare delivery systems cannot be at the cost of the marginalised, who continue to bear the brunt of the government’s techno-fetishism.

 

This is part of a six-article series on questions surrounding data, privacy, artificial intelligence, among others. You can read the introduction here

 

"Link fail" is a term now commonly used across rural India. It can mean anything from the internet not functioning to a biometric authentication failure. In Jharkhand, it is so common that a post office had to erect a special sign saying "link fail" to notify customers that they could not make transactions at the postal payments bank.

The last decade has seen a thrust in the digitisation of the delivery of welfare programmes. This means that application, approval, implementation, and also monitoring occurs online. Proponents of digitisation argue that technology makes processes more efficient and is more averse to the biases that our administrative systems are plagued with. However, digitisation also creates different challenges: there are newer forms of corruption and different ways in which people's rights can be subverted. 

This article highlights some concerns that arise due to the digitisation of welfare services and the human costs of the process. The first section discusses specific cases of failure of welfare delivery from different parts of the country, the second section places these case studies within the national context of digital welfare delivery, and the last section discusses structural problems within our digital systems.

Narratives from the Ground[1]

In November 2019, I met Radha Devi in the Rajsamand district of Rajasthan. She is a labourer under the Mahatma Gandhi National Rural Employment Guarantee Act (MGNREGA). In FY 2019–20, Devi worked for approximately 60 days between April and November, and wages amounting to Rs 12,000 were due to her. After completing her first round of work and seeing her fellow workers being paid, Devi enquired with officials in her village about her wages. She was instructed to wait—she would receive her wages eventually. As per the provisions of MGNREGA, Devi is entitled to 100 days of unskilled manual labour work at a stipulated minimum wage rate on demand. Her wages are supposed to be transferred to her bank account digitally within 15 days of having completed one round of work. If the wages are delayed beyond 15 days, she is entitled to compensation for the delay. Despite these provisions, Devi’s wages were pending for over eight months. Whenever Radha Devi enquired with the bank, she was sent back with the same response: “Wages were not deposited in the bank account.” Even block officials were unable to provide any clarification.

Radha Devi’s situation is not unlike many other MGNREGA workers, all of whose wages are in abeyance. As per the MGNREGA website, her name appears among those whose payments are “rejected.” The reason given was “inactive Aadhaar.” The website neither provides any explanation of the issue nor does it provide steps for redressal.

Before we consider the denial of dues, let us first understand the process of a successful MGNREGA payment. Before Devi begins work under the MGNREGA scheme, she has to submit a demand application with her Aadhaar[2] details and her job card number. The local MGNREGA office then updates her Aadhaar to her job card in the Management Information System (MIS), which is a real-time system for recording MGNREGA related work and payments. Once she completes one round of work, a fund transfer order (FTO) is generated in her name by block officials. This FTO is then sent electronically to the Ministry of Rural Development (MoRD), which initiates the payment to Devi’s Aadhaar number. The Aadhaar number issued by the Unique Identity Authority of India (UIDAI) is her financial address. The payment order goes through the Public Fund Management System (PFMS) to the “sponsor bank,” where the corresponding government department holds an account. The sponsor bank then sends an electronic notification to the National Payments Corporation of India (NPCI) which operationalises the payment and settlement systems for such transfers. At the NPCI server, one locates which bank the Aadhaar is mapped to through an institutional identification number (IIN). Once identified, the payment order reaches the concerned bank and there it searches for the bank account most recently linked to the concerned Aadhaar number. This is known as the Aadhaar Payment Bridge System, or more popularly as the Aadhaar Payments System. If successful, money is deposited to the corresponding bank account. However, if even a single link in this chain breaks, the payment can get delayed, fail (rejected), or even go to the incorrect bank account (misdirected payments). In Radha Devi’s case, her payments were rejected. 

Deposits into an account can be rejected for various reasons: the beneficiary’s account number is incorrect, the beneficiary’s account is dormant due to inactivity, and also if the Aadhaar number to which the payment is directed is delinked from the bank in the database. All of the government's Direct Benefit Transfer (DBT) schemes follow this payment procedure. What makes this process worse is that when a problem arises, there is often no clear method to resolve it. Radha Devi spent eight months petitioning various agencies for her dues before she decided to quit working under MGNREGA.

The issue of failed transactions is not restricted to Rajasthan and MGNREGA. Chintamma Golla,[3] a widow living in the Kurnool region of Andhra Pradesh, is a beneficiary of the Pradhan Mantri Kisan Samman Nidhi (PM-KISAN) Yojana (a cash transfer of Rs 6,000 to farmers by the central government). She is also a beneficiary of the Andhra Pradesh government's YSR Rythu Bharosa scheme, the state's cash support scheme for farmers. Golla’s payments under the PM-KISAN scheme were rejected due to problems with Aadhaar and bank mapping. The website threw up the error message "Aadhaar number not seeded in NPCI." Payments due to her under the YSR Rythu Bharosa scheme also failed due to the same reason. 

Digital cash transfers can also get “misdirected.” This happens when the cash transfer meant for one person goes into someone else’s account (Narayanan et al 2017). The government either transfers money into the incorrect Aadhaar number or the beneficiary’s Aadhaar is linked to someone else’s bank account. The latter is much harder to detect and official statistics on this are even harder to compute. Matters such as these only come to light when there is an in-depth investigation into the payment trail.

Premani Kunwar, a widow from the Garhwa district of Jharkhand, used to live with her schoolgoing son Uttam Kumar, and both of them survived on the ration Premani received from the PDS and a meagre social security pension. Premani’s Aadhaar had been incorrectly linked to someone else’s bank account, meaning her payments were being diverted elsewhere. Whenever Premani visited the bank to inquire about her pension payments, she was sent back because there were no deposits. In November 2017, Premani was refused ration by the local ration dealer. Facing the double blow of being denied both pension as well as ration, Premani died of starvation in December 2017 (Scroll 2017). Her story, however, is not unique. She and many others in Jharkhand have been victims of starvation due to the failure of technology (Saha 2018).

One Step Forward, Two Steps Back

The benefits of technology depend upon how it is introduced and implemented. For Aadhaar-based payments, the government set stiff targets and migrated beneficiaries from older systems to the newer payment platform. This required urgent action on the part of beneficiaries, local administration, and other intermediaries. In such a situation, there are bound to be errors. When this happens, it is women like Radha Devi, Chintamma, and Premani who have the most to lose. The MGNREGA scheme alone has had Rs 1,375 crore worth of payments[4] rejected in the last five years without any resolution. At a time when there is an unprecedented decline in women's labour force participation and more than three-fourths of women are not part of the labour force (PTI 2019), safety nets like MGNREGA are particularly important.

The DBT programme began in 2013 and despite the number of cases of DBT failures that have come to light, little has been done to resolve these issues systematically. A letter[5] from the office of the Chief Controller of Accounts, Ministry of Labour and Employment, which was addressed to the Additional Controller General of Accounts, PFMS, Ministry of Finance, in July 2018, explicitly stated that DBT failure rates were unusually high for the years 2017–18 and 2018–19 and the failure rate was over 10% for some schemes. For example, the failure rate for payments under "Stipend to differently-abled candidates under the scheme of Vocational Rehabilitation Centres for Handicapped (VRCs)" was 32% in 2017–18. Many such failures are treated as cases of “collateral damage” and are often expressed as statistics. While these failure rates may seem insignificant to authorities, it is an alarming violation of rights of the marginalised and often a matter of life and death for the afflicted.

While social security coverage remains poor, the proliferation of digital technologies is creating barriers to access existing safety nets. A 2019 survey[6] of pregnant and nursing women showed that about one-third of Anganwadi workers reported that women face issues due to Aadhaar-related problems, and approximately 15% of them reported bank-related issues. According to the Global Nutrition Report 2018,[7] India has either made no progress or is worsening in nearly all nutrition indicators. More than 50% of women in India are anaemic and in states like Jharkhand, nearly half of all children are underweight. It is critical that women and other vulnerable sections receive sufficient social security support in order to alleviate their condition. It is equally critical that they are able to access support with limited effort. Maternity entitlements are meaningful only if the mother receives sufficient money on time and can access it with ease. Evidence to understand the extent of payment-related problems is limited because reports on payment failures and delays are not published in the public domain by the implementing authorities. One has to rely on data from field surveys and on right to information (RTI) requests. It is even more difficult to capture the problems of access to payments at banks or via business correspondents. Data on biometric failures and misdirected payments remains elusive.

While digital technology is not the only reason for the failure of welfare programmes, each hurdle faced by beneficiaries reduces their chances of accessing benefits and makes them more vulnerable. Even when welfare delivery works, the complex technology used increases the beneficiary’s dependence upon those who have the power to understand and access it. This forms a toxic combination when there is an abject lack of accountability measures and weak grievance redressal systems.

Lack of Accountability

The introduction of complex technologies has only replicated the problems prevalent in the pre-existing administrative structures with an added layer of difficulty. Local officials maintain that they have no online control over payments while beneficiaries seldom understand the technical terminology and process, thus making it more difficult for them to demand an accountable administration. 

In January 2018, James Herenj, an activist from Jharkhand, found that many MGNREGA payments workers were being rejected, the primary reason for which was listed online as "inactive Aadhaar." Herenj, along with the afflicted beneficiaries, attempted to investigate what this meant, but local officials and bankers had no answers. Herenj then filed an RTI to the UIDAI. The UIDAI forwarded this request to the MoRD stating that the use of Aadhaar is determined by the implementing agency (MoRD). The MoRD, in turn, argued that the “subject matter” concerned the state government of Jharkhand and forwarded the request to the Rural Development Department (RDD) in Ranchi. The RDD responded that this matter concerned the MIS cell of the rural development ministry in Delhi and concluded, with calculated ambiguity, that the request “can be forwarded to MoRD.” This rigmarole took nearly three months and left Herenj exactly where he started. In February 2019, a team of four people—of which I was a part—brought this matter to the attention of the NPCI in a meeting with some senior officials. Finally, in April 2019, more than a year after Herenj first began his inquiry, the NPCI issued a circular clarifying that “inactive Aadhaar”' meant that the beneficiary’s Aadhaar number was de-seeded from the NPCI mapper (NPCI 2019). 

When there is a shift from an old system to a newer technological solution, logic dictates that the older methods should be available as alternatives. Newer technology must be human-centred and its operators must be sensitised so that it does not exacerbate the existing power structures in society. Alternative methods to provide services and some degree of flexibility are necessary for local administrators. It would not only make local accountability more meaningful but also allow for context-specific solutions to some problems.

A Missed Opportunity for Transparency

Considering the amount of data collected by the state on its citizens, what is published is disproportionately less. Section 4 of the RTI Act, 2005, states that "It shall be a constant endeavour of every public authority to take steps in accordance with the requirements of clause (b) of sub‑section (1) to provide as much information suo motu to the public at regular intervals through various means of communications, including internet, so that the public have minimum resort to the use of this Act to obtain information."  

An example of proactive disclosure is Rajasthan’s Jan Soochna portal.[8] It is a single portal created with the explicit purpose of disclosing important information to beneficiaries across schemes. The reports and interface are specifically designed keeping in mind programme beneficiaries as users. However, it is accessible only to literate internet users. Further measures must be taken to ensure that people who are unable to read and who do not have access to the website can also make use of its resources.

For people at the grassroots in the current scenario, if their details are unavailable in the system or if there is any sort of error, it is as though they do not exist in the eyes of the state. The government has been so focused on the efficient delivery of welfare programmes that the "welfare" aspect seems to have been forgotten in the process.

 

 

 

 

The author would like to thank Jean Drèze for helpful comments and members of the LibTech team for insightful discussions.
Back to Top