Govt. yet to decide how open-field caste data will be sorted, tabulated

1. At a Glance

• The 2027 Census second phase (population enumeration) will record caste through an open column. People not from SC/ST write in their caste name, with no pick-list. [1]
• The open field also allows a "no-caste" option, and respondents may decline to disclose caste. [1]
• As of the 26 Sept 2026 report, officials said no decision had been taken on how to sort, tabulate and rationalise the data. [1]
• Why it matters: this is the technical core of the caste-census debate (data quality, OBC sub-categorisation, reservation policy, federalism).

2. Why in the News

• The Union government notified the second-phase questionnaire, finalising the open-column method, "more than a month" before the 26 Sept 2026 report. [1]
• The Hindu reports that officials have not yet decided how the open-field data will be sorted, tabulated and rationalised. [1]
• Opposition parties, OBC associations and experts oppose the open-column method. They say it will repeat the 2011 SECC, which yielded over 46 lakh caste names. They add that successive governments shelved that data. [1]
• The second-phase questions cover caste, family particulars and COVID vaccinations. [1]
• The notification date is reported elsewhere as 14 Aug 2026, with 40 questions. I could not cite this from a whitelisted source, so treat it as unverified.

3. Background & Evolution

• 2011 SECC: ran with an open field to record caste, yielding over 46 lakh caste names. Successive governments did not act on it. [1]
• Post-2011 SECC caste data was not officially released. Secondary explainers cite data-quality concerns and the very large number of names. These are non-whitelisted sources, so I have not cited them.
• Bihar and Telangana caste surveys used the pick-list method, unlike the open-field method. [1]
• Phase 1 (houselisting and housing census) preceded Phase 2. [2]
• Static background, from general knowledge and uncited: the Census is conducted under the Census Act, 1948. Caste other than SC/ST was last enumerated in the 1931 Census. The last full caste count was 1931; the 1941 caste data was not tabulated.

4. Core Static Facts

Item	Fact
Method	Open column: free-text caste entry for non-SC/ST respondents [1]
Options	"No-caste" option; may decline to disclose [1]
Comparison	Bihar and Telangana used pick-list surveys [1]
Phase 2 content	Caste, family particulars, COVID vaccination fields [1]
SECC 2011 result	46 lakh+ caste names [1]
Status of tabulation	Undecided per officials [1]
Phase 1	Houselisting and housing census; the PIB document titled "Census 2027: India's First Digital Enumeration Exercise" (25 Apr 2026) refers to a digital census [2]

5. Multi-Dimensional Analysis

Administrative

• Free-text entry produces spelling variants, synonyms, sub-castes, gotras and surnames recorded as castes. Rationalisation needs a classification or master-list decision, which is still pending. [1]
• The government has not said who will do the coding: the Registrar General, an expert body, or states.

Social

• OBC groups fear an unusable dataset. The 2011 precedent is invoked as evidence. [1]
• The "no-caste" and non-disclosure options protect individual choice, but they may cause undercounting. [1]

Legal / Constitutional

• Uncited static knowledge: the Census Act, 1948 makes responses compulsory and protects individual records. Caste data also bears on Article 340 (backward classes commissions) and on Articles 15(4) and 16(4).

Ethical / Governance

• Opposition parties allege repeating a flawed method. Officials have not committed to a tabulation protocol, which is a transparency concern. [1]
• Past non-publication of SECC data is a precedent for accountability questions. [1]

Scientific / Technological

• The census is described as a digital enumeration. [2] Tools such as text normalisation and fuzzy matching could help clean the entries, but the article reports no such decision.

6. Recent Developments (last 12-18 months)

• 25 Apr 2026: PIB document on Census 2027 as India's first digital enumeration exercise. [2]
• Phase 1 (houselisting): under way in 2026. The reported completion date appears only in non-whitelisted sources, so it is unverified.
• Aug 2026 (approx.): second-phase questionnaire notified, with an open-column caste method. [1]
• 26 Sept 2026: officials say the sorting and tabulation method is undecided. Opposition and OBC bodies criticise the open field. [1]

7. Prelims Hooks

• Caste in Census 2027 phase 2 is recorded through an open column. [1]
• The open column applies to people not from SC/ST. [1]
• The method includes a "no-caste" option. [1]
• Respondents can decline to disclose their caste. [1]
• The 2011 SECC open field yielded over 46 lakh caste names. [1]
• Bihar and Telangana caste surveys used pick-list methods. [1]
• Phase 2 also includes fields on family particulars and COVID vaccinations. [1]
• The 2011 SECC caste data was shelved by successive governments. [1]
• Static, uncited: the Census is conducted under the Census Act, 1948.
• Static, uncited: the last caste-wise enumeration across all castes was in 1931.

8. The 2011 Data Hit a Sorting Wall — and That Wall Is Still There

• A committee was already built for exactly this job, and it never delivered
• After SECC 2011, the Union Cabinet approved an expert group under Arvind Panagariya, then Vice-Chairperson of NITI Aayog, only to classify the caste and tribe data [3].
• Its whole task was to turn about 46 lakh entries — caste names, sub-castes, synonyms, spellings, surnames, clan and gotra names — into usable groups [3].

• It was serviced by the Ministry of Social Justice and Empowerment and the tribal welfare side of government [3]. Its report has still not been made public.

• So the 2027 problem is not new, and the harder half is being repeated first

• Step 1 is collecting names. Step 2 is grouping them. 2011 finished step 1 and failed at step 2.

• Census 2027 has notified step 1 again, with the open column, while officials say step 2 is undecided [1].

• Why the raw list cannot be used as it is

• Of the ~47 lakh names recorded in 2011, many were the same community written differently — spelling changes, different languages, sub-caste names, even occupation names written in the caste box [7].
• Until someone decides which of those names are one community, you cannot say how many people belong to any caste. No count means no policy use.

9. What Ungrouped Names Cost: The Rohini Commission's Missing Numbers

• OBC reservation is already known to be unevenly shared, and fixing it needs caste-wise counts
• The Rohini Commission, set up in 2017 to study sub-categorisation of OBCs (splitting the OBC quota into smaller slabs so the weakest get a share), found that 97% of OBC-quota central jobs and seats went to about 25% of OBC castes [5].

• 983 of about 2,600 OBC communities — roughly 37% — had zero representation in those jobs and institutions [5].

• The Commission itself said it needed the data and asked for more time to get it [4]

• To split a quota fairly you need two numbers per caste: how many people it has, and how much of the benefit it already gets.

• A free-text list of lakhs of names gives neither until the names are merged into communities.

• This is the practical loss if sorting stays undecided

• The 27% OBC quota is still argued over using 1931 figures [6]. A 2027 count that cannot be grouped leaves that gap exactly where it is.
• Sub-categorisation, state OBC list revisions and any court defence of quota size all wait on the same missing table.

10. The Strongest Case For the Open Column — and Where It Actually Fails

• The open column has a real argument behind it, and it should be stated honestly
• A pick-list decides in advance which castes exist. If your community is not on the list, the enumerator must push you into the nearest box, or leave you out.
• Every state's list is different. Bihar and Telangana used pick-lists built for their own states [1], so their numbers cannot be added up into a national picture.
• Government has said caste is being counted inside the main Census, not as a separate survey, so that one uniform method applies across all states instead of many state surveys done differently [6].

• An open box also lets a person name their community in their own words. That protects small and unlisted groups from being erased at the moment of counting.

• Where the argument stops working

• Openness is a virtue at the collection stage only. The decision about who belongs where does not disappear — it just shifts to the coding table afterwards.
• In a pick-list, that decision is visible and can be challenged before counting. In the open-field method, it happens later, inside the office, with no published rule [1].
• So the honest reading is: the critics are wrong that the open box itself is the flaw, and right that 2011 proves the flaw appears immediately after it [3][7].

11. Who Decides What Counts as One Caste — and What Should Be Settled First

• Merging two names, or keeping them apart, changes a community's counted size
• If two names are treated as one caste, that group looks bigger. If they are split, both look smaller.
• Group size is what later feeds arguments about quota share and about which communities are the most backward [5]. So coding rules are a policy decision, not clerical work.

• In 2015 the government recognised this by sending it to a Cabinet-approved expert group rather than to clerks [3]. For 2027, officials have not said who will do it [1].

• The Registrar General should publish the classification rules before Phase 2 begins

• Say in advance how spelling variants, surnames and gotra names will be handled, and which body signs off — as the 2011 exercise needed a named expert group [3].

• A rule published before counting can be argued about in the open. A rule made after counting cannot be checked against anything.

• The Ministry of Social Justice and Empowerment should release the Panagariya group's classification method

• That group was created and serviced for this exact task [3]. Its method, even if its report is dated, is the only existing national attempt at grouping these names.

• Publishing it lets experts point out mistakes once, instead of the same mistakes being made twice.

• State OBC lists should be used as a coding aid, not as the collection form

• Bihar's and Telangana's pick-lists already name the communities each state recognises [1].
• Those lists can help match a written-in name to a known community afterwards, keeping the open box at the door and still producing groupable data.

12. Anchors for Answers

• Data: About 46 lakh caste, sub-caste, synonym, surname, clan and gotra names recorded in SECC 2011, all awaiting classification [3]; 4.7 million names including spelling and language variants and occupation entries [7]
• Data: 97% of OBC-quota central jobs and seats went to 25% of OBC castes; 983 of about 2,600 OBC communities had zero representation [5]
• Report/Committee: Expert Group under Arvind Panagariya (NITI Aayog), approved by Cabinet in 2015, to classify SECC 2011 caste and tribe data — report not made public [3]
• Report/Committee: Justice G. Rohini Commission on sub-categorisation of OBCs, constituted 2017, which sought caste-wise data [4][5]
• Law/Case: Census Act, 1948; Article 340 (backward classes commission); Articles 15(4) and 16(4)
• Comparison: Bihar and Telangana caste surveys used state-specific pick-lists, so their counts are not comparable nationally [1]
• Scheme: SECC 2011 — the same open-field method, whose caste data was never published [1][3]

13. Mains Relevance

• GS-I: Indian society, social empowerment, caste. GS-II: government policies and interventions; statutory bodies; federalism; welfare of vulnerable sections. GS-IV: transparency and data ethics.
• Question stems:
• Evaluate the open-column method for caste enumeration in the Census against pick-list surveys. How can data quality be ensured?
• Caste data without a classification protocol is of limited policy value. Discuss with reference to the SECC 2011.
• Discuss the implications of a caste census for reservation policy and Centre-state relations.

14. Related Topics to Study Next

• SECC 2011: the precedent, and why its caste data was not released.
• Bihar and Telangana caste surveys: the pick-list method and its litigation and policy fallout.
• Census Act, 1948: confidentiality and the powers of the Registrar General.
• Article 340 and backward classes commissions: constitutional basis for OBC identification.
• OBC sub-categorisation (Rohini Commission): the potential use of caste data.
• Reservation ceilings and the Indra Sawhney case: how data affects the 50% cap debate.
• Delimitation: a live 2026 national issue, listed in The Hindu's topics bar. [1]
• Digital Census and data protection: privacy in self-enumeration.

15. Common Errors / Trap Areas

• Confusing SECC 2011 (a survey by the Ministries of Rural Development and Housing and Urban Poverty Alleviation, outside the Census Act) with the Census 2027 exercise. The Act point is from general knowledge and uncited.
• Assuming the 46 lakh figure is the number of castes. It is the number of caste names recorded, including variants. [1]
• Assuming the open column applies to everyone. It applies to non-SC/ST respondents. [1]
• Assuming that tabulation rules have been notified. Officials say they are undecided. [1]
• Confusing the two phases: Phase 1 is houselisting and Phase 2 is population enumeration with caste. [1][2]