Why India must get the Caste Census right

The Hindu

11-05-2025

The Narendra Modi government's decision to include caste enumeration in the next Census is one that is bold, transformative and commendable. Counting caste is not capitulation to identity politics. It is a mirror to the lived realities of millions. It marks a vital step towards evidence-based policymaking to build a more just and inclusive India. A nation that refuses to see itself cannot hope to heal itself.
Post-Independence, India attempted to abolish caste while simultaneously pursuing social justice — a textbook example of policy schizophrenia, as the two goals are fundamentally incompatible. The refusal to count caste in the Census was a corollary of the policy of caste blindness. But the Constitution explicitly mandates the pursuit of social justice through reservations in education, public employment, and electoral constituencies — measures that require precise, disaggregated caste data. Although the Constitution uses the term 'class', the Supreme Court of India has repeatedly ruled that caste is a valid, and often necessary, proxy for identifying backwardness and has insisted on detailed caste-wise data to uphold reservation policies.
In his 1955 essay, 'Thoughts on Linguistic States', Dr. B.R. Ambedkar denounced the omission of caste tables from the 1951 Census as an act of 'petty intelligence'. Visibility in data is the first step toward meaningful inclusion. Caste data collection across all major social groups is essential not only for administering reservations, but also for equity-driven planning, targeted policymaking, and tracking disparities over time. Not collecting it has rendered many of India's marginalised communities invisible in official statistics. Worse, a narrow elite of upper castes and dominant Other Backward Classes (OBCs) has entrenched its grip over wealth, opportunity and power behind the smokescreen of caste anonymity. In hindsight, this ranks among India's gravest policy failures.
A legal and administrative necessity
Since 1951, the Census has enumerated Scheduled Castes (SCs) and Scheduled Tribes (STs) but excluded OBCs, even though all three groups are constitutionally eligible for reservations in education and public employment. The usual justification, that OBCs lack reserved seats in Lok Sabha and State Legislative Assemblies (that SC/ST have) collapsed with the 73rd and 74th Amendments, which mandated OBC reservation (in addition to SC/ST reservation) in electoral constituencies of panchayats and municipalities. Implementing these provisions requires granular, area-wise OBC data. With the introduction of reservations in education and public employment for the Economically Weaker Sections (EWS) among upper castes (2019), a comprehensive enumeration of all castes has now become a legal imperative.
India's reservation policy currently operates in an evidence vacuum, leaving it vulnerable to arbitrary demands from powerful caste groups and politically expedient decisions by governments. With reliable caste data, the demands of the Marathas, Patidars, Jats, and others can be assessed transparently and on merit. The limited data we do have reveal deep inequities. According to submissions made by the Government of India to the Justice G. Rohini Commission, just 10 OBC castes cornered 25% of all public jobs and education seats reserved for OBCs, while a quarter of OBC castes secured 97% of the benefits. Shockingly, 38% of OBC castes received only 3% of the benefits, and another 37% got nothing at all.
Hence, caste enumeration is also an administrative imperative — to prevent the elite capture, enable rational sub-categorisation within social groups, and allow a more precise definition of the 'creamy layer'.
Collection of caste data must go beyond the decennial Census. All periodic government surveys should enumerate OBCs and upper castes alongside SCs and STs. The era of partial counting must end.
Learning from failure and success
In 2010, Parliament unanimously resolved to count caste in the 2011 Census. The 1931 Census had recorded 4,147 castes (excluding the then-called Depressed Classes). The Anthropological Survey of India has identified 6,325 castes. But the Socio-Economic and Caste Census (SECC) of 2011, conducted by the United Progressive Alliance-II government, was a debacle. It produced a ludicrous figure of 46 lakh castes and was never released.
What went wrong? First, the SECC-2011 was not conducted under the Census Act, 1948 and lacked legal authority. Second, it was conducted through the Union Ministries of Rural Development and Urban Development with no expertise for handling a complex socio-anthropological survey. Third, its open-ended questions about caste created confusion. Undertrained enumerators conflated castes, aliases, sub-castes, gotras, clan names, surnames and broader caste groups. The result was a chaotic, unusable data set. Was it sabotage or incompetence? Either way, a historic opportunity was squandered.
In contrast, in Bihar's caste survey, enumerators were given a vetted list of 214 castes specific to the State, with a 215th option for 'Other Castes'. The survey was well-planned, well-executed, and showed that a credible caste count is entirely feasible.
Blueprint for a successful Caste Census
To avoid repeating the SECC-2011 fiasco, here is what must be done.
First, legal backing. Amend the Census Act, 1948 to explicitly mandate caste enumeration and insulate the process from shifting political agendas.
Second, the right institution. Entrust the exercise solely to the Office of the Registrar General and Census Commissioner of India, and not Ministries that lack domain expertise.
Third, a standardised questionnaire. Use closed-option questions with dropdown menus covering sub-caste, caste (including aliases), broader caste group, and caste-linked surname (optional). Having 'caste' alone as an option can lead to errors since some caste names such as Rao, Naik, Singh or Bhandari span multiple communities. Assign unique digital codes to avoid duplication and semantic confusion (e.g., grouping 'Iyer' and 'Aiyar' under one code).
Fourth, State-specific caste lists. Develop draft lists in consultation with State governments, sociologists, and community leaders. Publish them online and invite public feedback before finalisation. Use a similar participatory approach for questionnaire design.
Fifth, enumerator training. Conduct region-specific training sessions with mock examples, clear dos and don'ts, and guidance on local caste nuances.
Sixth, digital tools. Equip enumerators with handheld devices that are preloaded with validated caste lists. Restrict data entry to predefined options to minimise human error.
Seventh, representative staffing. To ensure data integrity, deploy enumerators from diverse communities and in areas where they have no conflict of interest.
Eighth, independent oversight. Establish district-level committees to audit samples and monitor data integrity.
Ninth, pilot testing. Run trials in diverse States such as Tamil Nadu, Gujarat, Uttar Pradesh and Assam to refine methodology before nationwide rollout.
In every Census since 1951, the Government has successfully enumerated nearly 2,000 castes and tribes under the SC/ST categories. Counting the remaining 4,000-odd OBCs and upper castes (most of them State-specific) is not only doable but also long overdue. The delayed 2021 Census offers a rare chance to finally close this data gap. The time for denial and delay is over. The time to get the Caste Census right is now.
K. Ashok Vardhan Shetty is a former IAS officer of the Tamil Nadu cadre and a former Vice-Chancellor of the Indian Maritime University, Chennai