The decision to include caste enumeration in Census 2027 has gained significance because the Census questionnaire, notified on 14 August 2026, includes Question 10: “Scheduled Caste (SC)/Scheduled Tribe (ST)/Caste”. Reports on the notified schedule indicate that, for people who are neither SC nor ST, the caste field is open-ended: the respondent states the caste or sub-caste and the enumerator records it. This format was also used in the pre-test conducted before the final questionnaire was notified.
This differs from asking every respondent to choose a caste from a predetermined list. The distinction matters because India has no single national list of all caste identities. There are constitutionally notified lists of Scheduled Castes and Scheduled Tribes, and Central and State lists of socially and educationally backward classes, each serving particular legal and administrative purposes.
Census 2027 is therefore asking a basic population question: what caste does a person report belonging to? That is not automatically the same as asking whether that caste or community is an Other Backward Class (OBC). Understanding the distinction requires looking at how caste and “backward class” operate in the constitutional and administrative framework, and what happens when millions of individual responses have to be converted into statistical categories.
What is an OBC?
“Other Backward Classes” (OBC) is an administrative and policy category derived from the Constitution’s provisions concerning backward classes; it is not the name of a caste.
The Constitution uses the expression “backward class” in Articles 15 and 16. Article 15(4) permits special provisions for the advancement of socially and educationally backward classes of citizens, while Article 16(4) permits reservation in appointments or posts in favour of a backward class of citizens that is not adequately represented in the services of the State.
Article 340 provides for the appointment of a commission to investigate the conditions of socially and educationally backward classes and recommend measures for their advancement. The Constitution thus refers to “classes of citizens” rather than simply castes.
How did the Constituent Assembly deal with the issue?
This terminology was debated when the Constitution was being framed. During the Constituent Assembly discussions on the provision that became Article 16, members questioned what “backward class” meant and how such a class would be identified. T.T. Krishnamachari questioned whether backwardness was to be understood in terms of economic circumstances, education or birth. B.R. Ambedkar explained that the identification of backward communities would be left to the local government rather than being exhaustively defined in the Constitution itself.
That history shows that the distinction between “caste” and “class” is not a recent semantic debate. It was present at the beginning of the constitutional framework itself.
If OBC means “class”, why are OBC lists largely made up of caste and community names?
Because the constitutional category of “backward class” developed within a society in which caste was an important determinant of social position.
The first Backward Classes Commission, headed by Kaka Kalelkar, was appointed in 1953 under Article 340. Its report considered several indicators of backwardness, including the social position of a caste or community, educational advancement and representation in government employment.
The Kalelkar Commission’s approach to caste was itself contested. The report and the material surrounding it reflected disagreement over whether caste should be treated as a principal criterion or whether broader social and economic factors should predominate.
The Mandal Commission, appointed in 1979 under Article 340, identified 3,743 castes and communities as socially and educationally backward. It also noted the difficulty created by the absence of comprehensive caste enumeration after 1931.
The Supreme Court later addressed the relationship between caste and backward class in Indra Sawhney v. Union of India. The Court accepted that caste could be a starting point for identifying a backward class because of the relationship between caste and social backwardness in India. But it did not equate caste with backward class, and recognised occupational and other social groups as possible bases of identification.
The distinction is therefore important: caste may be relevant to identifying social backwardness, but a caste is not automatically an OBC.
So does caste have anything to do with OBC status?
Yes—historically and administratively, very much so. The relationship is one of classification, not identity.
That is why many recognised OBC entries are named after castes or communities. The Central List maintained by the National Commission for Backward Classes is organised State-wise and contains caste, sub-caste and community names. Official examples include entries in which several names appear together—for instance, “Ahir, Yadav, Gwala” in the Delhi list. The Telangana material similarly contains entries with alternative or associated names.
It would therefore be wrong to say either that “OBCs are simply castes” or that OBC has nothing to do with caste.
A more accurate description is that OBC is a policy and administrative category of “backward classes”, with caste and community identity historically playing an important role in identifying those classes.
Why isn’t there one definitive national list of OBC castes?
The National Commission for Backward Classes maintains the Central List of OBCs on a State-wise basis. States also maintain their own lists for State purposes.
The constitutional position changed significantly with the 102nd and 105th Amendments. The 102nd Amendment created the constitutional framework for the National Commission for Backward Classes and Article 342A. After the Supreme Court’s 2021 Maratha reservation judgment interpreted the amended framework as affecting the States’ power to identify their own backward classes, the 105th Amendment restored and expressly clarified that power for State and Union Territory purposes.
The significance for Census 2027 is that OBC status is not necessarily one uniform, nationwide classification. A community may appear in one list and not another, or have different relevance for Central and State purposes.
Why can’t the existing OBC lists simply be used as the Census caste list?
There is a strong statistical argument for using a predetermined list. Standardised categories make aggregation easier, reduce the problem of different spellings and names, and give enumerators a defined vocabulary.
But an OBC list is not a list of every caste in India. It identifies communities recognised as backward classes for particular administrative purposes. Using it as the complete caste list would therefore apply a classification designed for one purpose to another.
There is another complication: the lists themselves contain multiple names, alternative names and descriptions of communities. The Delhi and Telangana examples demonstrate that an administrative entry does not always correspond neatly to one simple caste name or one social identity.
A comprehensive caste master list would therefore be a different exercise from an OBC eligibility list.
Why is SC/ST enumeration different?
The answer lies in the Constitution. Articles 341 and 342 establish specific mechanisms for identifying Scheduled Castes and Scheduled Tribes. The President specifies the communities for particular States and Union Territories, and Parliament can subsequently include or exclude communities from those lists.
This gives SCs and STs a constitutionally prescribed, State/UT-specific list-based architecture. OBCs developed through a different constitutional route: Articles 15(4), 16(4), 340 and, following later amendments, Article 342A.
What happens when people are allowed to state their caste themselves?
This is where the practical problem begins.
One person may give the name of a caste; another may use a regional name; a third may give a sub-caste; a fourth an alternative spelling; a fifth a community name; and another an occupational identity historically associated with that community.
The enumerator can record each answer. But the database may then contain many different strings referring to identities that authorities may subsequently decide are the same, related or distinct.
If the objective is to publish raw responses, this may not present a major conceptual difficulty. But if the objective is to produce reliable population counts for individual caste or community groups, the responses have to be standardised according to defined rules.
What did SECC 2011 teach us about this problem?
The 2011 Socio-Economic and Caste Census (SECC) provides the clearest recent evidence.
In July 2015, the Government reported that the caste component had produced 46,73,034 distinct caste names.
That number did not mean India had 46.7 lakh castes. The Ministry of Home Affairs explained that the returns included caste and sub-caste names, synonyms, surnames, clan and gotra names, phonetic variations, sections and sub-groups.
The Government therefore constituted an expert group headed by Arvind Panagariya to classify and categorise the caste returns. Subsequent Government statements continued to describe the data as requiring classification and categorisation before it could be used as standardised data.
The lesson is important: collecting caste information and producing a standardised caste database are not the same exercise.
What did Bihar do differently?
The Bihar caste-based survey provides an example of a more structured approach. It used a list of 215 caste categories and reported population under those categories. Instead of treating every respondent’s description as a separate statistical category, it used codes to group identities.
This made the final numbers easier to aggregate. But classification itself became an issue.
Many Lohars objected to being grouped under a single category labelled “Kamar (Lohar and Karmkar)”. The dispute was therefore not merely about spelling. It concerned whether identities treated by the survey as belonging to one statistical category should instead have been kept separate.
The lesson from Bihar is not that its classification was necessarily right or wrong. It is that a standardised caste count inevitably involves decisions about classification. A fixed system prevents the database from exploding into millions of names, but someone has to decide which names belong together. An open system gives respondents greater freedom to state their identity, but makes subsequent aggregation more demanding.
Does an open column mean the Census will produce millions of castes?
No. It could produce a very large number of distinct reported names. That is not the same as producing the same number of final caste categories.
The SECC 2011 figure of 46,73,034 referred to distinct caste-name returns, including synonyms, surnames, clan and gotra names, phonetic variations and sub-groups.
If Census 2027 produces a similarly large number of raw names, that would demonstrate the diversity of responses. It would not, by itself, establish how many distinct castes exist. That could be determined only after the responses had been classified according to defined rules.
The quality of that classification exercise may therefore matter as much as the wording of Question 10 itself.
If the respondent states a caste, when does it become an OBC?
Conceptually, there are three separate questions:
Self-identification: What caste or sub-caste does the person report?
Standardisation: If the responses are subsequently classified, which recognised caste or community category does that response represent?
Legal or administrative classification: Does that category correspond to a recognised backward class under the applicable OBC framework?
Only at the third stage does the question of OBC status arise.
Article 16(4) does not say that every caste is a backward class. It refers to a “backward class of citizens”. Article 340 similarly concerns socially and educationally backward classes. The Supreme Court in Indra Sawhney treated caste as a possible basis for identifying backwardness rather than as an automatic definition of backward class.
Thus, if 100,000 people report caste X, the Census can establish that 100,000 people reported caste X, subject to the classification methodology. Whether those people should subsequently be counted as OBCs requires the additional determination that caste X corresponds to a recognised backward class under the applicable list.
The two numbers may coincide, but they do not necessarily have to. Nor is membership of a listed OBC community identical, for every purpose, to eligibility for an OBC reservation benefit. In the reservation context, other rules, including exclusion of the creamy layer, also apply.
Why does this matter for reservations?
Because population size and backwardness are different questions.
A caste enumeration can provide evidence about the numerical size and geographical distribution of caste communities. That information could be relevant to discussions about representation, welfare or affirmative action.
But constitutional provisions governing backward classes do not say that population size alone determines backwardness. The Supreme Court’s jurisprudence, particularly Indra Sawhney, makes that distinction clear.
Even a perfectly classified caste enumeration would therefore not, by itself, settle every legal question concerning OBC reservation. It would provide population evidence; the legal conclusions drawn from that evidence would remain a separate matter.
What are the two sides really arguing about?
The disagreement can be stated more precisely than simply as “open column versus list”.
Those who favour a predetermined caste list are primarily concerned with standardisation. Their argument is that the Census must ultimately produce numbers that can be aggregated and compared, requiring categories to be defined clearly enough before or during enumeration.
Those who favour an open response are primarily concerned with self-identification. Their argument is that the State should not determine a person’s caste identity by forcing the respondent into a category prepared in advance.
Both concerns have a factual basis. SECC 2011 demonstrates the difficulty of standardising large volumes of self-reported caste information. The Central and State OBC lists demonstrate that administrative classifications themselves may contain multiple caste and community names. Bihar demonstrates how a structured coding system can make aggregation possible while also creating disputes over where identities should be grouped.
The unresolved question for Census 2027 is therefore not simply whether respondents should be allowed to name their caste. It is how the Census will reconcile faithful recording of self-identified caste with the subsequent need to produce reliable, standardised and legally meaningful categories.