Counted Out: Nine Years of Disability Rights and the Census That Never Found Them

The Lift Line

A welfare scheme fails where you can see it, in unspent money. A rights law fails where you cannot, in a right nobody was in a position to claim.

Why This Editorial Matters for Your Exam

Disability is one of the few social sector topics where most candidates can name the law and almost none can explain the failure mechanism. The standard answer says the RPwD Act, 2016 is progressive, lists 21 disabilities and 4 per cent reservation, then complains about implementation. That is a description, not an analysis, and it earns description marks.

This editorial gives you a causal chain instead: undercount to underfunding, certification gate to unclaimed entitlement, recommendatory rule to unenforced duty. It also gives you the case, Rajive Raturi, that converts the complaint into law, and a live policy hook in Census 2027. That combination works for GS2 governance and social justice, for GS1 society questions on marginalised groups, and for the ethics paper.

GS Paper 2: Welfare schemes for vulnerable sections of the population by the Centre and States and the performance of these schemes; mechanisms, laws, institutions and bodies constituted for the protection and betterment of these vulnerable sections; issues relating to development and management of social sector or services relating to health, education and human resources.

Background and Context

The Rights of Persons with Disabilities Act, 2016 was passed by Parliament in December 2016, received assent on 27 December 2016, and came into force on 19 April 2017. It repealed the Persons with Disabilities (Equal Opportunities, Protection of Rights and Full Participation) Act, 1995.

A precise note on the anniversary framing, because the exam rewards precision. As of August 2026 the statute is nine years old measured from assent and a little over nine years measured from commencement. The decade completes on 27 December 2026 from enactment and on 19 April 2027 from commencement. The “ten years” in the headline is the tenth year in progress, not a completed decade.

The 2016 Act gives domestic effect to the United Nations Convention on the Rights of Persons with Disabilities, adopted by the United Nations General Assembly in 2006, which India signed on 30 March 2007 and ratified on 1 October 2007. The Convention’s core move, which the Act adopts, is to relocate disability from the individual body to the interaction between an impairment and a disabling environment. A staircase, on this view, disables a wheelchair user; the wheelchair does not.

The immediate occasion for the argument is the record after nine years, and the immediate opportunity is Census 2027, India’s first digital census, in which disability data is to be collected in the population enumeration phase and, per the Ministry of Social Justice and Empowerment, is to recognise all 21 categories.

The Analysis

1. Start with the discrepancy, because everything downstream is derived from it. Census 2011 recorded 2.68 crore persons with disabilities, 2.21 per cent of the population. The National Sample Survey 76th round, conducted July to December 2018, found prevalence of 2.2 per cent, with 2.3 per cent rural and 2.0 per cent urban. The World Health Organization estimates that roughly 16 per cent of the world’s population, about 1.3 billion people, live with a significant disability. Even allowing for India’s younger age structure, which genuinely lowers prevalence, and for definitional differences between a census question and a clinical instrument, the gap is not a rounding error. It is a factor of several.

2. An undercount is not a description error. It is a budgeting instrument. This is the point students miss. If reservation quanta, scheme allocations, assistive device targets and accessibility retrofitting budgets are all sized against the recorded population, then undercounting does not merely understate the problem; it produces a remedy scaled to the understatement. The people outside the count are not visible as unmet demand, because they were never counted as demand. The undercount and the underfunding are the same act performed twice.

3. Why the count is low is a design question, not a diligence question. A census question that asks whether a person “is disabled” invites a stigmatised binary that households, particularly for women and children, answer defensively. A question set that asks about difficulty in seeing, hearing, walking, remembering, self care and communicating measures function rather than identity and consistently returns higher and more usable numbers. This is why the design of the Census 2027 disability schedule matters more than the promise to include all 21 categories.

4. Certification is the gate between an entitlement and a person, and it is where the rights model quietly reverts to the welfare model. Section 2(r) sets the benchmark threshold at 40 per cent of a specified disability, and Section 57 provides for designated certifying authorities competent to assess and certify. In practice that means a medical board, in a district hospital, on specified days, with specialists for the relevant condition. Each of those conditions is a filter. For a person with limited mobility, in a village, without money for repeated travel, the entitlement is formally universal and practically rationed by the cost of proving eligibility. The Unique Disability Identity card was designed precisely to end repeated re-certification and to make the certificate portable across states; the one croreth card was presented in December 2023. It reduced repetition. It did not remove the gate.

5. Accessibility was drafted as a duty and administered as advice, until the Supreme Court intervened. Section 40 requires the Central Government to formulate accessibility standards. The Accessible India Campaign (Sugamya Bharat Abhiyan), launched on 3 December 2015 on the International Day of Persons with Disabilities, set targets across three verticals: the built environment, transport, and information and communication technology. Its deadlines were extended more than once. The legal reason for the drift became explicit in Rajive Raturi v. Union of India, decided on 8 November 2024 by a bench of Chief Justice D. Y. Chandrachud with Justices J. B. Pardiwala and Manoj Misra, which held that Rule 15 of the RPwD Rules, 2017 was ultra vires the Act to the extent that it framed accessibility guidelines in recommendatory language when Sections 40 and the associated enforcement provisions required them to be mandatory. The Court gave the Union three months to frame binding rules in consultation with stakeholders. That judgment is the cleanest available illustration of the editorial’s thesis: the right existed, the duty existed, and a subordinate rule quietly converted the duty into a suggestion.

6. Intersectionality is not an add-on; it is the reason aggregate improvement can coexist with unchanged exclusion. Disability interacts with sex, caste, poverty and location, and it interacts multiplicatively rather than additively. A disabled woman faces both the certification gate and the mobility and safety constraints that limit women’s independent travel to a district hospital. A disabled person in a rural block faces the gate at a longer distance and a higher cost. A disabled person from a marginalised social group faces it with less social capital to navigate the bureaucracy. Because outcome data is not routinely disaggregated along these axes, national averages can improve while the intersection stays exactly where it was, and the administrative system will register that as success.

7. What a rights statute actually needs from the state. A welfare scheme needs money and delivery. A rights statute needs four things the Indian state supplies unevenly: an accurate count of rights holders, a low friction way to establish eligibility, an enforcement mechanism with teeth, and disaggregated public reporting so that failure is visible. The 2016 Act supplied the entitlements and the institutions, including the Chief Commissioner and two Commissioners under Section 74 and State Commissioners under Section 79. It could not supply the census question, the medical board in the block, or the political decision to notify mandatory standards. That is why the same statute looks transformative in the law reports and uneven on the ground.

Data and Institutions Vault

Prelims-grade facts:

• Rights of Persons with Disabilities Act, 2016: assent 27 December 2016, in force 19 April 2017; repealed the Persons with Disabilities Act, 1995
• Specified disabilities raised from 7 to 21, listed in the Schedule; the Central Government may notify more
• Benchmark disability, Section 2(r): certified at not less than 40 per cent of a specified disability
• Section 32: 5 per cent reservation of seats in government and government aided higher education institutions, with a five year upper age relaxation
• Section 34: 4 per cent reservation in government establishments (raised from 3 per cent under the 1995 Act), allocated 1 per cent each to blindness and low vision; deaf and hard of hearing; locomotor disability including cerebral palsy, leprosy cured, dwarfism, acid attack victims and muscular dystrophy; and autism, intellectual disability, specific learning disability and mental illness together with multiple disabilities
• Section 40: Central Government to formulate accessibility standards
• Section 57: designation of certifying authorities
• Section 74: Chief Commissioner for Persons with Disabilities and two Commissioners at the Centre. Section 79: State Commissioners
• UNCRPD: adopted by the UN General Assembly in 2006; India signed 30 March 2007, ratified 1 October 2007
• Census 2011: 2.68 crore persons with disabilities, 2.21 per cent of the population
• NSS 76th round, July to December 2018, Report No. 583: disability prevalence 2.2 per cent (rural 2.3, urban 2.0)
• WHO: about 1.3 billion people, roughly 16 per cent of the world’s population, live with a significant disability
• Accessible India Campaign (Sugamya Bharat Abhiyan): launched 3 December 2015, three verticals (built environment, transport, information and communication technology); deadlines repeatedly extended
• Rajive Raturi v. Union of India, 8 November 2024: Rule 15, RPwD Rules 2017 held ultra vires to the extent it made accessibility standards recommendatory; three months given to frame mandatory rules
• Unique Disability Identity (UDID) card, issued through the Swavlamban Card portal: the one croreth card was presented in December 2023
• Nodal ministry: Department of Empowerment of Persons with Disabilities (DEPwD), Ministry of Social Justice and Empowerment
• Census 2027: India’s first digital census; disability data to be collected in the population enumeration phase, covering all 21 categories

Watch the trap: the Act came into force on 19 April 2017, not on the date of assent, so a “ten years of the RPwD Act” claim in 2026 is a tenth year in progress, not a completed decade. Second trap: 4 per cent is employment (Section 34), 5 per cent is higher education (Section 32); candidates reverse these constantly. Third trap: 21 is the number of specified disabilities in the Schedule, not the number of persons or categories of benchmark disability. Fourth trap: 2.21 per cent is Census 2011; do not attribute it to a survey, and do not present the WHO 16 per cent as an Indian figure, since it is a global estimate.

The Debate

Argument FOR reading the record as substantial failure. A right that cannot be claimed is a right in name only. The state has had nine years to fix the count, the certification chokepoint and the accessibility standards, and it took a Supreme Court judgment in 2024 to establish that standards written as mandatory in the Act could not be diluted into advice by a rule. The Accessible India Campaign’s repeated deadline extensions are not an accident of capacity; they reflect the absence of any consequence for missing them. Meanwhile the population against which every target is sized remains the 2011 figure, which almost nobody believes. On the evidence, the paradigm shifted in the statute and did not shift in the administration.

Argument AGAINST. Judge the Act against what preceded it, not against an ideal. The 1995 law recognised seven disabilities and framed the state as benefactor; the 2016 Act recognises 21, raises employment reservation to 4 per cent, creates a 5 per cent higher education reservation, imposes duties with statutory sanctions, and makes reasonable accommodation a legal concept rather than a courtesy. Rajive Raturi is not evidence of failure but of the framework functioning: a subordinate rule was struck down because the parent Act was mandatory, an outcome impossible under a welfare statute. Institutional change on this scale takes a generation, universal retrofitting competes for the same fiscal space as health and nutrition, and the certification threshold exists to prevent leakage of scarce reservations, not to obstruct claimants.

Balanced verdict. Both readings are compatible once the failure mode is stated correctly. The Act is genuinely transformative in what it makes claimable, and genuinely uneven in what gets claimed, and the second follows from the first. Rights statutes do not fail through bad intentions; they fail at the interfaces where a person meets the state, which here are the enumerator’s question, the medical board’s calendar, and the building that was signed off without an accessibility audit. The correct conclusion is therefore neither that the law needs rewriting nor that patience will suffice, but that the three interfaces need fixing in sequence, beginning with Census 2027, because every other reform is sized against the number the census produces.

How to Think About This

The transferable pattern here is: a right is only as real as the cheapest way to prove you hold it.

Whenever you evaluate a rights based statute, do not begin with the entitlements. Begin at the counter where the citizen must establish eligibility, and price that step in money, time and distance.

Who does the state believe exists? Every entitlement is sized against an official count. If the count is wrong, the entitlement is correctly administered and still inadequate, and no audit of administration will reveal it.

What must a person do before the right becomes available? Certification, documentation, registration and verification are not neutral formalities. They are a price, and the price falls hardest on precisely the population the statute was written for.

What happens if the duty bearer does nothing? If the answer is nothing, the duty is a suggestion regardless of the verb used in the statute. This is the whole content of Rajive Raturi.

The pattern generalises across the syllabus. It is the same analysis as the Right to Education and the documentation required for admission, the Forest Rights Act and the evidentiary burden on claimants, the Food Security Act and the ration card that does not exist, and the Maternity Benefit and cash transfer schemes that turn on a bank account and a linked identity. In every case, the binding constraint is not the entitlement in the statute but the proof demanded before it.

Diagram-in-Words

THE RIGHTS MODEL AS WRITTEN (RPwD Act, 2016)

  State = DUTY BEARER            Citizen = RIGHTS HOLDER
        |                                |
        |  21 specified disabilities     |  4% jobs (Sec 34)
        |  Sec 40 accessibility duty     |  5% higher education (Sec 32)
        |  Sec 74/79 grievance offices   |  reasonable accommodation
        +----------------> RIGHT <-------+


THE RIGHTS MODEL AS ADMINISTERED: three interfaces

  [1] COUNT            [2] CERTIFY              [3] ENFORCE
  Census 2011:         Sec 2(r): 40% benchmark  Sec 40 standards
  2.21% (2.68 cr)      Sec 57: certifying       -> Rule 15 (2017) made
  NSS 2018: 2.2%            authority              them RECOMMENDATORY
  WHO global: ~16%     medical board:           -> Accessible India
        |               district, specialist,      deadlines extended
        |               repeat visits, travel      repeatedly
        v                     |                        |
  budgets, quotas,            v                        v
  targets all SIZED      entitlement is           duty carries no
  on the LOW number      RATIONED BY THE          consequence for
        |                COST OF PROOF            non-performance
        |                     |                        |
        +---------------------+------------------------+
                              |
                              v
              RIGHT EXISTS, CLAIM NEVER MADE
              (failure leaves no administrative trace)
                              |
        INTERSECTION MULTIPLIER: woman + Dalit + rural
        raises the cost at EVERY interface, and is not
        disaggregated in reporting, so it is never seen


   REPAIR SEQUENCE (order matters)
   Census 2027 functional-difficulty questions, all 21 categories
        -> permanent UDID, block-level assessment camps
        -> notify MANDATORY standards per Rajive Raturi (8 Nov 2024)
        -> disaggregated outcome reporting by sex, social group, location

Takeaway Box

Lift line for an answer:

India did not fail to write a disability rights law. It failed to count the people the law was written for, and an uncounted person cannot claim a right.

Prelims hooks: RPwD Act 2016, assent 27 December 2016, in force 19 April 2017, repealed the 1995 Act; disabilities 7 to 21; benchmark 40 per cent under Section 2(r); 4 per cent employment (Section 34), 5 per cent higher education (Section 32); Section 40 accessibility standards; Section 57 certifying authorities; Section 74 Chief Commissioner and two Commissioners, Section 79 State Commissioners; UNCRPD signed 30 March 2007, ratified 1 October 2007; Census 2011: 2.21 per cent, 2.68 crore; NSS 76th round 2018: 2.2 per cent; WHO: about 16 per cent globally, 1.3 billion; Accessible India Campaign launched 3 December 2015; Rajive Raturi v. Union of India, 8 November 2024, struck down Rule 15, RPwD Rules 2017 as recommendatory; UDID one croreth card December 2023; Census 2027 to cover all 21 categories.

Ethics and interview angle: the certification requirement is defended as a safeguard against leakage of scarce reservations. But a safeguard calibrated to stop the undeserving inevitably also stops some of the deserving, and the two errors do not fall on the same people. Where the cost of a false rejection is borne by a poor disabled woman in a village and the cost of a false acceptance is borne diffusely by the exchequer, how should a public servant set the threshold? Is administrative caution a virtue when its burden is distributed this unequally?

PYQ linkage: UPSC has repeatedly examined the shift from welfare to rights based approaches in Indian social legislation, the performance of welfare mechanisms for vulnerable sections, and the gap between legal entitlement and actual access. It has also asked about the role of the census and administrative data in social policy. This editorial is the model case for all three, because it connects a statutory entitlement to the data instrument that determines whether it can be delivered.

Probable question: “A rights based law fails differently from a welfare scheme: it fails through entitlements that are never claimed.” Critically examine this proposition with reference to the implementation of the Rights of Persons with Disabilities Act, 2016.

Sources: Hindustan Times, Department of Empowerment of Persons with Disabilities, India Code, PIB

Source: Counted Out: Nine Years of Disability Rights and the Census That Never Found Them — Ujiyari.com | Free UPSC & State PCS Editorial Analysis