Reservation and the caste census: India's affirmative-action ledger reopens

What the system is

India reserves a share of public-sector jobs and government-aided educational seats for specified categories. At the central level the shares are 15% for Scheduled Castes, 7.5% for Scheduled Tribes and 27% for Other Backward Classes, with a further 10% for economically weaker sections added in 2019 (Wikipedia, “Reservation in India”, https://en.wikipedia.org/wiki/Reservation_in_India). The constitutional basis sits in Articles 15(4) and 16(4), which permit special provisions for socially and educationally backward classes and for SCs and STs, and in Articles 341 and 342, which let the President specify which communities are listed (same source). The Ministry of Social Justice and Empowerment is the nodal ministry for the SC and OBC welfare architecture, while job-reservation rosters are administered by the Department of Personnel and Training and the census is run by the Registrar General under the Ministry of Home Affairs.

The 27% OBC quota traces to the Second Backward Classes Commission, known as the Mandal Commission, constituted in 1979, whose 1980 report recommended reservation for Other Backward Classes; the recommendation was implemented for central services in 1990 (Wikipedia, “Reservation in India”). The Ministry itself was reconstituted from the former Ministry of Welfare in May 1998 (Wikipedia, “Ministry of Social Justice and Empowerment”, https://en.wikipedia.org/wiki/Ministry_of_Social_Justice_and_Empowerment).

The 50% ceiling and how it was breached

In Indra Sawhney v. Union of India (1992), a nine-judge bench of the Supreme Court upheld the 27% OBC quota, introduced the “creamy layer” exclusion that bars better-off members of OBC groups from the benefit, and held that total reservation should ordinarily not exceed 50% save in extraordinary circumstances (Wikipedia, “Indra Sawhney and Others v. Union of India”, https://en.wikipedia.org/wiki/Indra_Sawhney_and_Others_v._Union_of_India). That 50% figure has since anchored every subsequent dispute.

The ceiling was formally pierced by the Constitution (103rd Amendment) Act, 2019, which added a 10% reservation for economically weaker sections outside the SC, ST and OBC categories. In Janhit Abhiyan v. Union of India, decided on 7 November 2022, a five-judge Constitution Bench upheld the amendment by a 3:2 majority, with the majority reasoning that an economic-criterion quota was a permissible classification and the dissent objecting to the exclusion of SC, ST and OBC groups from the EWS category (Wikipedia, “Janhit Abhiyan v. Union of India”, https://en.wikipedia.org/wiki/Janhit_Abhiyan_v._Union_of_India). Because the EWS quota sits outside the 50% pool, central reservation now totals roughly 60%, which is why the ceiling itself is contested rather than settled.

The missing number: caste data

The central fault line is measurement. India’s decennial census has not enumerated caste beyond Scheduled Castes and Scheduled Tribes since 1931 (Wikipedia, “Caste census”, https://en.wikipedia.org/wiki/Caste_census). The Socio-Economic and Caste Census of 2011 collected caste data but its caste tables were never released for use. Without a current count of how many people belong to each OBC or other caste group, the size of the population that reservation and welfare are meant to serve is estimated rather than measured, which is the empirical hole every side of the debate points at.

On 30 April 2025 the Cabinet Committee on Political Affairs decided to enumerate caste for all groups in the forthcoming census, the first such enumeration since 1931 (Wikipedia, “Caste census”). The Office of the Registrar General and Census Commissioner notified the census in the Gazette under the Census of India Act, 1948, on 16 June 2025, setting a reference date of 00:00 hours on 1 March 2027, with 1 October 2026 for the Union Territory of Ladakh and snow-bound areas of Jammu and Kashmir, Himachal Pradesh and Uttarakhand (Wikipedia, “2027 census of India”, https://en.wikipedia.org/wiki/2027_census_of_India). The census runs in two phases: house-listing from April to September 2026, and population enumeration in February 2027, with caste recorded in the second phase (same source).

The states moved first

Ahead of the national exercise, two state governments ran their own caste surveys and turned the results toward reservation policy. Bihar conducted a caste-based survey between 2022 and 2023 and released its findings on 2 October 2023 (Wikipedia, “2022 Bihar Caste-Based Survey”, https://en.wikipedia.org/wiki/2022_Bihar_Caste-Based_Survey). On the strength of that data the Bihar legislature raised reservation for Backward Classes, Extremely Backward Classes, SCs and STs from 50% to 65% in November 2023, which combined with the 10% EWS quota pushed the state’s total to 75% (same source). On 20 June 2024 the Patna High Court struck down the increase, and the matter was carried to the Supreme Court (Supreme Court Observer, https://www.scobserver.in/journal/what-is-the-bihar-governments-65-percent-reservation-quota-challenge-in-the-supreme-court/). The case puts before the Supreme Court the question of whether survey-backed data can justify crossing the Indra Sawhney ceiling.

Telangana released the results of its 2024 caste survey on 3 February 2025, reporting that Backward Classes made up about 56% of the surveyed population, and the state moved toward a 42% reservation for Backward Classes in local-body elections (Wikipedia, “2024 Telangana Social Educational Employment Economic Caste Survey”, https://en.wikipedia.org/wiki/2024_Telangana_Social_Educational_Employment_Economic_Caste_Survey). Together the Bihar and Telangana exercises supplied the template, and the political pressure, for a national caste count.

The range of positions

Supporters of a caste census, including several opposition parties and some governing- coalition allies, argue that precise caste-wise data is a precondition for rationally targeting welfare and reservation, and some further argue the 50% ceiling should be revisited in light of the numbers (Oxford Human Rights Hub, https://ohrh.law.ox.ac.uk/how-a-caste-census-could-transform-indias-reservation-policies/). The Union government’s decision to enumerate caste in the 2027 census reflects a shift on its part toward collecting the data (Wikipedia, “Caste census”). Critics and cautionary voices contend that a caste count may harden caste identities, that enumeration is administratively fraught, and that individual reservation and scholarship benefits continue to depend on separately verified caste certificates rather than the census tally itself (Wikipedia, “2027 census of India”). What is not in dispute is that the 2027 census has been designed to record caste for all groups for the first time since 1931, and that the 50% ceiling now sits under active litigation in the Bihar case.

The welfare footprint behind the debate is large and growing: the Ministry of Social Justice and Empowerment reported its highest-ever departmental expenditure, about Rs 11,810 crore, in FY 2025-26, on scholarships and schemes for SC, OBC and other groups (Press Information Bureau, https://www.pib.gov.in/PressReleasePage.aspx?PRID=2249066&reg=3&lang=2).

Who owns this topic (and why we’re here)

Public understanding of reservation and the caste census is dominated by exam-preparation explainers and encyclopedia pages: Vajiram & Ravi, PWOnlyIAS, Testbook and Drishti IAS serve the civil-service aspirant, and Wikipedia serves the general reader. They explain the concepts well but flatten a live, contested institutional story into a static syllabus entry. Legal trackers such as the Supreme Court Observer and PRS Legislative Research cover the litigation and the statutes cleanly but in isolation from the welfare machinery. IndiaStand’s structural advantage is to hold the institution and the policy in one place: we track the Ministry of Social Justice and Empowerment as a seat of power, tie the reservation percentages to the judgments that fixed them, tie the caste count to the Registrar General and the Ministry of Home Affairs that runs it, and attribute every number to a primary or reference source rather than an aggregated claim, keeping the state of play current as the census phases and the Bihar litigation move.

Maintained topic brief. Analysis by IndiaStand — it characterises the state of play and the range of positions actually held, attributes each claim, and makes no forecast and no recommendation.