The OYI Review · One Young India Press
From Enumeration to Entitlement Redesigning India's Digital Welfare Architecture for Domestic Workers
This white paper argues that India’s e-Shram portal has succeeded at counting informal workers but failed at getting them any benefits. Over 31 crore workers are registered, yet only about 3 in 100 of those eligible are enrolled in the pension scheme meant for them, and fieldwork with domestic workers in Mumbai found that registered workers had received nothing at all. The paper explains why registering doesn’t lead to benefits and proposes a fix: when a worker registers, the system should automatically check what she qualifies for, enrol her where possible, and tell her in writing what she now has.
Published 17 November 2026
Write a paper like this with Learnacy Essential. Publication depends on review and is not guaranteed. See how
On this page
From Enumeration to Entitlement
Redesigning India's Digital Welfare Architecture for Domestic Workers
Ayaan Wadhwa
ABSTRACT
By 14 July 2026 the e-Shram portal had registered 31,78,31,749 unorganised workers. Over the same stretch, enrolment in the Pradhan Mantri Shram Yogi Maan-dhan pension scheme stood at about 52.5 lakh. PM-SYM admits only workers aged 18 to 40, and 55.21% of registrants fall in that band, so the fair comparison is 52.5 lakh against roughly 17.5 crore age-eligible registrants: about three in every hundred. The database has grown by an order of magnitude in five years. The pension that registration was supposed to open a door to has not moved with it.
Fieldwork among 57 domestic workers in Mumbai in September 2025 found the same gap at street level. Twenty-two had heard of e-Shram and eleven had registered. All eleven reported that nothing had reached them since. Not a pension credit, not an insurance payout, not a rupee of any transfer. Registration had produced a twelve-digit number and nothing after it.
The failure is architectural rather than administrative. E-Shram was built to solve enumeration and has solved it extremely well; it was never built to convert a database entry into a claim on the state. Two bottlenecks stand between a domestic worker and a benefit, and Indian policy keeps treating them as one. The first is awareness, which is cheap and, on the evidence below, already solved. The second is entitlement, which is not. What follows proposes an automatic entitlement mapping layer: a check of the Universal Account Number against eligibility for already-mapped schemes at the moment of registration, default enrolment wherever the benefit costs the worker nothing, and a written statement handed to every worker setting out what she now holds. The labour codes came into force on 21 November 2025 and the Social Security (Central) Rules, 2026 followed on 8 May 2026, so this can be built by amending those rules rather than through fresh legislation.
I. Registration Without Entitlement
The portal launched on 26 August 2021 with three stated purposes: a National Database of Unorganised Workers, welfare delivered through that database, and formalisation of the informal workforce. Taking them in order shows where the architecture gives way.
On the first, the government deserves more credit than its critics usually give it. Before 2021 India had no national record of who its informal workers were. In the 2020 lockdown, relief could not reach people the state could not name. A database of 31.78 crore workers, seeded to Aadhaar and carrying occupation, skill and address fields, is a real feat of state capacity. What it cannot tell us is how much of the target population it holds. Official statistics put the number of domestic workers in India at 4.2 million while union and civil society estimates run past 50 million (Gudibande & Jacob, 2020). A denominator that uncertain makes a headline numerator hard to read.
The second pillar is where things give way. The e-Shram One-Stop Solution, launched on 21 October 2024, maps fifteen central schemes, Ayushman Bharat-PMJAY and PM-SYM among them. But aggregation is discovery, not delivery. It tells a worker which schemes she might qualify for, enrols her in none of them, and assumes she will complete a second unassisted journey for each. For a domestic worker who cannot read English, does not own the phone the OTP lands on, and cannot lose half a day's wages in a Common Service Centre queue, that second journey is where the process ends.
The Ministry's reply seems to cut the other way. Its Annexure II counts 8.45 crore registrants in PMSBY and 15.1 crore in Ayushman Bharat-PMJAY (Ministry of Labour and Employment, 2026b). Those are matches between databases, not enrolments the portal produced: the Ministry said in 2023 that about 40% of registrants were already on PMJAY rolls when the data was shared, and neither figure says whether a claim was ever paid (Ministry of Labour and Employment, 2023). They show the state can tell who holds which cover, not that registration got it to her.
The third pillar misleads more than it fails. A UAN is an identifier, not a status. It creates no employment relationship, imposes no obligation on any employer, and generates no claim a worker could enforce. She has been counted.
That distinction now carries legal weight. An expert committee constituted on the Supreme Court’s direction of 29 January 2025 in Ajay Malik reported in July 2025 that domestic workers need no separate legislation, on the reasoning that the labour codes already reach them (Bhagirath, 2026). Neetha had anticipated this, arguing that the Code on Social Security, 2020 dissolves domestic workers into the undifferentiated mass of unorganised workers and loses their specific protections on the way (Neetha, 2021). The eleven registered workers in this study are what that dissolution looks like from underneath. Several produced the card when asked. What none could name was anything the card had done.
So the database is not broken. Read alongside Scott's account of state legibility, e-Shram looks less like a welfare instrument that underperforms and more like an enumeration infrastructure that succeeded (Foucault, 1991; Scott, 1998). The error was in assuming that counting was the hard part.
II. Two Independent Bottlenecks
A. The awareness bottleneck
The survey, run across two mornings in September 2025 in one cluster of Mumbai's western suburbs with 57 domestic workers who each agreed to take part and whose names are changed here, splits the population cleanly. Twenty-two of the 57 workers had heard of e-Shram at all, and eleven had registered, a rate of 19.3% in India's financial capital four years after launch. Among the 46 unregistered, 35 gave one reason: they had never heard of the scheme. That is 76.1%. Another eight knew it existed but not how to enrol, and three had approached it and been turned back by the documentation. Not distrust. Not a considered decision to stay outside. Plain absence of information.
Male respondents, mostly drivers, registered at 31.3%; the women, who are the maids, cooks and cleaners, at 14.6%. A driver holds a continuous relationship with one household, which makes him easy to find, while the fragmented multi-employer schedule of feminised domestic work is what a registration drive is worst at reaching (Neetha, 2009). The system is least visible to the workers it was most loudly announced for.
And the fix is almost embarrassingly cheap. After a single ten-minute session, 52 of the 57 workers said they would now register, a shift across 91.2% of the sample. Very few problems in Indian social policy respond to ten minutes of anything.
B. The entitlement bottleneck
The eleven registered workers describe the second bottleneck, and credibly, because they had already cleared the first. Asked what benefits they had accessed, all eleven said none.
Sunita has cleaned four apartments a day in Mumbai for more than two decades. She produced her card with pride and said that it tells the government she exists but does not tell her employer to speak to her with respect. She still cannot use the main lift. Lakshmi, who is fifty and has chronic pain in her joints, made the same point in material terms: the government has her fingerprint, but when she is sick the card does not pay for her medicine and does not stop her salary being cut. Priya, twenty-four, registered during a local awareness drive and found that the portal gave her no contract, no grievance channel and nothing to put in front of an employer.
The Devereux and Sabates-Wheeler taxonomy places e-Shram lower than the government's own framing suggests (Devereux & Sabates-Wheeler, 2004). Their protective tier covers instruments relieving deprivation and the preventative tier those averting it, while the promotive and transformative tiers cover what builds capability and shifts the relations producing vulnerability. E-Shram aspires to the promotive tier through skill training and scheme access. In operation it does not reach even the protective one, because it transfers nothing.
Singh and Jackson's idea of infrastructural torque describes the friction produced when a life has to be bent into the shape an infrastructure expects (Singh & Jackson, 2017). Seema is what that looks like in one person. She is Mumbai-born, supports five members of a seven-person household on a combined income of roughly Rs. 25,000 a month, and she wants to register. She cannot. Her Aadhaar is not linked to an active mobile number, and until then the portal is shut to her. She is excluded neither by choice nor by ignorance but by a system demanding an unbroken chain of digital documentation she has no way to maintain (Khera, 2019).
C. Why collapsing the two produces a predictable error
The halves need different fixes, and merging them generates one foreseeable error. Awareness campaigns raise registration numbers, and registration numbers are what the state reports and what its intermediaries are paid on. So the published figure climbs, the press release is accurate, and the worker's position does not move. That is not bad intent. It is what happens when an output metric is mistaken for an outcome.
III. What a Solved Awareness Problem Looks Like
The survey that produced these numbers was not only a survey. Participants were told what e-Shram is, how registration works and what it promises, and were then asked to evaluate it. As a research instrument that is a limitation. As an accidental experiment in outreach it is the most useful thing in the dataset.
One student, a clipboard, a bilingual sheet, two mornings in September. Before the session, 38.6% of the 57 workers had heard of the scheme. After it, 52 of them intended to register. What they expected: 75.4% named welfare schemes, 52.6% help finding work, 42.1% skill training. And 78.9% said registration would matter most for their children, a quieter finding than it looks, since they were not expecting it to change anything for themselves.
Seema shows what the session did: she had heard the name beforehand and nothing else, and afterwards she wanted official proof from the government that she does this work.
That rate is encouraging on its face and uncomfortable underneath. A state four years into building and publicising a national welfare database had not managed to tell the domestic workers of Mumbai's western suburbs that it existed, and a school student with a printed sheet closed the gap in ten minutes.
Workers of India was built around that finding. It runs e-Shram seminars and registration drives across Mumbai, alongside Naukri Nama, a micro-contract service giving domestic workers a written record of terms, and a multilingual WhatsApp assistant in English, Hindi and Marathi. The assistant takes voice input and answers in voice. The binding constraint is not language, which the state has partly addressed, but literacy, which it has not. A worker who cannot read a Marathi form cannot read a Marathi chatbot either.
The point is not that one organisation fixed outreach in Mumbai. It is that this bottleneck yields to an intervention cheap enough that its absence at national scale is a choice rather than a constraint.
IV. Proposed Framework: Automatic Entitlement Mapping
The governing principle is a reversal of sequence. Registration should be the last thing a worker has to do, not the first of several errands she is left to finish alone. By the time her UAN is issued, the state already holds everything it needs to know what she is owed.
A. Eligibility inference at the point of registration
The fields e-Shram captures at enrolment decide eligibility outright for three of its mapped schemes: age, declared income, occupation code, Aadhaar number and bank seeding status between them resolve PMSBY, PMJJBY and PM-SYM. Ayushman Bharat-PMJAY and the Public Distribution System are different, since their eligibility comes from the Socio Economic Caste Census 2011 list, state expansions and ration rolls, so there the check can only match a registrant against the list rather than compute it from her fields. Running that check server-side at registration is a rules engine, not a research programme.
B. Default enrolment where the worker owes nothing
This category is narrow. PMSBY and PMJJBY carry yearly premiums of Rs. 20 and Rs. 436, so they are not free. What costs the worker nothing is Ayushman Bharat-PMJAY cover and a ration card, and for those, wherever the list match succeeds, enrolment should follow automatically from registration, with a real opt-out replacing the current opt-in. Defaults matter more here than in the pensions literature, because these workers have the least time to waste.
C. Assisted enrolment where a contribution is required
PM-SYM cannot be auto-enrolled. It needs a monthly contribution and a subscriber decision, and no interoperability layer should make either on somebody's behalf. What it can be is offered inside the registration sitting rather than as a separate errand months later, to every registrant aged 18 to 40 earning Rs. 15,000 a month or less. The two insurance premiums, Rs. 20 and Rs. 436 a year, belong in the same sitting, explained and debited only where the worker agrees. What is missing is the link that puts that offer in front of a worker while she is still seated with the enrolment agent and her documents out.
D. A written entitlement statement
Every UAN should generate a plain-language statement in the worker's registered language, printed at the registration point and sent by SMS, setting out what she now holds, what stays conditional, and what she must do to claim the rest. The card currently says nothing about entitlement. The only workers who find out what registration is worth are those who happen to meet somebody who tells them, which is exactly what happened to the 52 in this study.
E. What gets measured, and what gets paid for
The harder change is not technical. India measures this programme by registrations and pays its intermediaries per registration completed, and it has 31.78 crore registrations. That is not an accident. It is an incentive working correctly. An operator paid on enrolment has no reason to spend another twenty minutes walking a worker into a pension scheme, and faces no penalty when the card he issues never produces a rupee. Two changes follow. The headline figure reported to Parliament should become the benefit realisation rate, the share of registered workers who have received at least one traceable benefit, disaggregated by state and occupation. And intermediary payment should move to a fee per verified claim, so that the people paid to grow the database are paid instead to grow what it delivers.
The objection worth meeting head-on is that the One-Stop Solution already does this. It does not. It is a discovery layer, answering what a worker might be eligible for. Automatic entitlement mapping answers what she is now enrolled in. Only the second is a question a worker short on time and literacy can act on.
Cost is the easy part. This is software on rails that already exist, since the Direct Benefit Transfer architecture, Aadhaar-seeded accounts and the NDUW database are built and in daily use. The marginal build is a rules engine and an API layer between e-Shram and fifteen schemes already mapped to it. Set against what it cost to produce 31.78 crore registrations, that is close to a rounding error.
V. Implementation Roadmap
A. Phase one: single-state pilot, months 0 to 12
Maharashtra is the natural site. The Maharashtra Domestic Workers Welfare Board Act, 2008 constitutes district Domestic Labour Welfare Boards and makes every domestic worker aged eighteen to sixty eligible to register as a beneficiary. The Board runs a live registration portal, so the administrative surface exists. What it has never had is an automatic link between that surface and the central schemes a registered beneficiary already qualifies for. Run the pilot in Mumbai and Pune across three schemes only, chosen for how computable their eligibility conditions are rather than how important they sound, and evaluate on benefit realisation rather than enrolment.
B. Phase two: interoperability build-out, months 12 to 30
Extend nationally to two or three flagship schemes, starting with those whose eligibility is fully computable from existing fields and whose disbursal already runs on DBT. Publish the rules engine's logic. Where a worker is found ineligible, she and anyone assisting her should see which condition failed. That is the difference between a decision and a refusal.
C. Phase three: national scale-up with funded intermediaries, months 30 to 48
Shift intermediary payment onto a per-verified-claim basis and open the facilitation role to accredited civil society organisations alongside Common Service Centres. The awareness half is labour-intensive and locally specific, and organisations already embedded in worker communities do it better and more cheaply than any portal will.
The strongest practical argument for moving now is that the vehicle already exists. The labour codes came into force on 21 November 2025, and the Social Security (Central) Rules, 2026, notified on 8 May 2026, superseded the Unorganised Workers' Social Security Rules, 2009. Rules made by notification can be amended by notification: an entitlement mapping duty can be added to the chapter on unorganised workers under section 154 of the Code, with the draft published for objections and laid before Parliament as the 2026 Rules were, which needs neither new primary legislation nor the political capital that would cost.
VI. Risks, Limitations and Counterarguments
A. Privacy
Alston's warning about the digital welfare state is that digitisation arrives as a tool of surveillance and conditionality dressed as efficiency, and automatic entitlement mapping does mean more inference on more data about the people least able to contest it (Alston, 2019). India has a concrete precedent rather than a hypothetical one. About three crore ration cards were reported cancelled over Aadhaar mismatches, a figure put before the Supreme Court in 2021 and disputed by the Centre, but even the contested version describes exclusion by automation, in this population, within living memory (Khera, 2019). So the safeguards belong inside the rules rather than in assurances offered alongside them. Inference should be limited by purpose to the mapped schemes and barred from enforcement, verification or exclusion. The opt-out must be genuine and exercisable by voice. And no eligibility determination should ever become grounds for removing a benefit a worker already receives.
B. Fiscal cost
If realisation climbs from about 3% to even 20%, the subsidy bill rises steeply, and it would be dishonest to present that as a saving. It is the intended consequence. PM-SYM's central contribution fell from Rs. 324.23 crore in 2021-22 to Rs. 162.51 crore in 2023-24, and that fall was read in places as prudence. A scheme that is cheap because nobody claims it is not efficient. It is inoperative, and the low spend is the clearest measure of its failure.
C. Political resistance
Registration is invisible to employers and entitlement is not. Households employing domestic workers have no stake in a database and a considerable one in anything that makes obligations legible. The answer is that this proposal creates no employer liability at all. It routes state-funded benefits to workers the state has already counted, which is politically survivable in a way that employer contribution mandates are not. That is a deliberate narrowing, and it leaves the harder question of regulating the private household where it was.
D. Limits of the evidence
The survey is small, and the benefit findings are self-reported rather than checked against scheme records. Eleven registered respondents is a thin denominator for the sharpest claim. The intent figures come from an instrument that informed participants before asking them, so they measure intent after an intervention rather than latent demand. Still, a ratio of 17.5 crore age-eligible registrants to 52.5 lakh pension enrolments is not a sampling artefact. The fieldwork illustrates a pattern the aggregate data independently establishes.
VII. Conclusion
Under automatic entitlement mapping, Lakshmi's card would have done something. She is registered and she is fifty, which puts her past PM-SYM's entry age of 40, so the pension was never hers to join. But on the face of her own registration record she is inside the age band for PMSBY accident cover and PMJJBY life cover, and a list match would settle in seconds whether her household is on the Ayushman Bharat-PMJAY rolls. The check would have run the moment her UAN was issued. She would have left that registration point holding a sheet in Marathi telling her what she now had, and what the two premiums would buy. None of it requires persuading her employer of anything.
Seema is the harder case. She is not registered at all, because her Aadhaar is not linked to an active mobile number, and no interoperability layer reaches a worker who cannot get through the door. What she needs is a linkage camp and a ten-minute conversation. That is the cheap half, and the argument of this paper is that the cheap half is already solved wherever anybody bothers to solve it.
Nothing proposed here requires a new scheme, a new act of Parliament or a new portal. The database exists, the payment rails exist, and fifteen schemes are already mapped to the very registration record that fails to trigger any of them. What is missing is the connection between the counting and the covering, and the willingness to report the second number as seriously as the first.
India has built one of the largest worker databases in the world and it should say so. But a database counts people, and a welfare state is meant to reach them. Until reaching is what gets measured, the counting will keep improving while the lives do not.
REFERENCES
Alston, P. (2019). Report of the Special Rapporteur on extreme poverty and human rights (A/74/493). United Nations General Assembly. https://undocs.org/A/74/493
Bhagirath. (2026, January 18). Domestic workers do not need a separate law, says committee formed on Supreme Court directions. Down To Earth. https://www.downtoearth.org.in/governance/domestic-workers-do-not-need-a-separate-law-says-committee-formed-on-supreme-court-directions
Devereux, S., & Sabates-Wheeler, R. (2004). Transformative social protection (IDS Working Paper 232). Institute of Development Studies.
Foucault, M. (1991). Governmentality. In G. Burchell, C. Gordon, & P. Miller (Eds.), The Foucault effect: Studies in governmentality (pp. 87–104). University of Chicago Press.
Gudibande, R. R., & Jacob, A. (2020). Minimum wages for domestic workers: Impact evaluation of the Indian experience. World Development, 130, Article 104943. https://doi.org/10.1016/j.worlddev.2020.104943
Khera, R. (Ed.). (2019). Dissent on Aadhaar: Big data meets big brother. Orient BlackSwan.
Maharashtra Domestic Workers Welfare Board Act 2008, Maharashtra Act No. 1 of 2009 (India). https://www.indiacode.nic.in/handle/123456789/19597
Ministry of Labour and Employment. (2021). e-Shram portal: National database of unorganised workers. Government of India. https://eshram.gov.in
Ministry of Labour and Employment. (2023, March 13). e-SHRAM portal integrated with PM-SYM scheme (PIB Release ID 1906432). Press Information Bureau. https://pib.gov.in/PressReleasePage.aspx?PRID=1906432
Ministry of Labour and Employment. (2024, October 21). e-Shram One-Stop Solution. Government of India.
Ministry of Labour and Employment. (2025, November 21). Notification bringing the four labour codes into force. Government of India.
Ministry of Labour and Employment. (2026a, May 8). Social Security (Central) Rules, 2026, G.S.R. 344(E). Gazette of India.
Ministry of Labour and Employment. (2026b, July 20). Reply to Parliament: 31,78,31,749 workers registered on the e-Shram portal as on 14 July 2026 (PIB Release ID 2286757). Press Information Bureau. https://pib.gov.in/PressReleasePage.aspx?PRID=2286757
Ministry of Labour and Employment. (2026c, August 25). 5 years of e-Shram portal [Factsheet]. Press Information Bureau.
Ministry of Labour and Employment. (n.d.). Brief on Pradhan Mantri Shram Yogi Maan-dhan (PM-SYM). Government of India.
Neetha, N. (2009). Contours of domestic service: Characteristics, work relations and regulations. Indian Journal of Labour Economics, 52(3), 489–506.
Neetha, N. (2021). Misconstrued notions and misplaced interventions: An assessment of state policy on domestic work in India. The Indian Journal of Labour Economics, 64(3), 543–564. https://doi.org/10.1007/s41027-021-00334-w
Parliamentary Standing Committee on Labour. (2025). Report on the Pradhan Mantri Shram Yogi Maan-dhan Yojana. Lok Sabha Secretariat.
Scott, J. C. (1998). Seeing like a state: How certain schemes to improve the human condition have failed. Yale University Press.
Singh, R., & Jackson, S. J. (2017). From margins to seams: Imbrication, inclusion, and torque in the Aadhaar identification project. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (pp. 4776–4824). Association for Computing Machinery. https://doi.org/10.1145/3025453.3025910
Wadhwa, A. (2026). The illusion of inclusion: E-Shram, dignity, and Mumbai's domestic workers [Manuscript submitted for publication]. Convergence: A Journal for Young Researchers. Survey of 57 domestic workers, Mumbai, September 2025.
Cite this paper
Wadhwa, A. (2026). From Enumeration to Entitlement Redesigning India's Digital Welfare Architecture for Domestic Workers (Certificate No. OYI-2026-97DCML). The OYI Review. One Young India Press. https://www.oneyoungindia.com/white-papers/from-enumeration-to-entitlement-redesigning-india-s-digital-welfare-ar-5516f6
Wadhwa, Ayaan. “From Enumeration to Entitlement Redesigning India's Digital Welfare Architecture for Domestic Workers.” The OYI Review, certificate no. OYI-2026-97DCML, One Young India Press, 17 Nov. 2026, www.oneyoungindia.com/white-papers/from-enumeration-to-entitlement-redesigning-india-s-digital-welfare-ar-5516f6.
@article{wadhwa2026from,
author = {Wadhwa, Ayaan},
title = {From Enumeration to Entitlement Redesigning India's Digital Welfare Architecture for Domestic Workers},
journal = {The OYI Review},
publisher = {One Young India Press},
year = {2026},
url = {https://www.oneyoungindia.com/white-papers/from-enumeration-to-entitlement-redesigning-india-s-digital-welfare-ar-5516f6},
note = {Certificate ID: OYI-2026-97DCML}
} Write your own
Write a paper like this, and submit it for review.
In the Writing Studio, a student plans the paper with a structure guide, a rubric and citation tools, free with an account. Learnacy Essential adds feedback from an AI writing coach and submission to The OYI Review, plus 10 self-paced courses, each with a completion certificate, World 101’s member tools and the 3D labs. ₹1,499 for the year. One payment, including GST. No automatic renewal.
Learnacy Essential includes editorial review, and publication of one paper in the year if it is accepted. Publication is not guaranteed.How review works
Or start a draft free in the Writing Studio
Every programme registration also includes a year of Learnacy Essential. See programmes
