Opinion S Y Quraishi writes: Census is about who we are. It cannot ignore caste and migration

Census 2027 will be India’s first digital census. It will also trigger women’s reservation — though the government’s 2029 promise is mathematically impossible given delimitation’s four-to-six-year track record — and enable the first Lok Sabha delimitation since 1976, frozen by the 84th Constitutional Amendment until “the first Census after 2026”. The most politically charged aspect remains unresolved: Will Census 2027 enumerate caste? The government has not said if caste will be included, relegated to a separate exercise, or avoided altogether.

Caste was last comprehensively counted in 1931 under the colonial administration. Independent India counted Scheduled Castes and Scheduled Tribes for reservation. The rationale was nation-building: The republic would not reinforce caste categories by officially enumerating them. The 2011 Socio-Economic and Caste Census (SECC) attempted to count all castes, but its findings remain unreleased — officially due to data quality concerns, unofficially due to political sensitivities. Over a decade later, that data sits unused, neither published nor discarded.

Comprehensive caste data would help reshape debates around OBC reservations, social justice policies, and resource allocation. It would give marginalised communities demographic evidence to support their claims. It would also intensify caste-based political mobilisation. But the alternative — ignorance — is not neutral. Without accurate caste data, policy operates on political assertions, not demographic reality.

Announcing a census without specifying whether it includes caste suggests indecision or deliberate ambiguity. Both are problematic. Census 2027 should enumerate caste comprehensively, with the same rigour applied to age, literacy, or occupation. Democracy requires information, even uncomfortable information.

Census 2027 must correct another blind spot — migration. Tens of millions work outside their home states. Construction workers build Mumbai but are counted in Bihar. Factory workers manufacture in Haryana but are enumerated in Uttar Pradesh. The Census will count them based on where they are found — but that doesn’t translate into political or civic belonging. The deeper problem is electoral. Migrants are registered in ancestral villages they left years ago, not cities where they live, work, and some even pay taxes. Many don’t vote because travelling home is expensive. Others vote in villages whose issues they no longer experience, while remaining voiceless in cities whose governance affects them directly.

This creates a distortion: Cities are governed based on census populations, including migrants, but those migrants cannot vote in municipal elections because they are registered as voters elsewhere. Urban governance becomes unaccountable to urban residents. Rural areas receive allocations for citizens who are no longer there.

Electoral laws do allow — even require — registration at places of work, not ancestral places where they are not “normally resident” for more than six months. This requires coordination between states and updated voter rolls — the alternative is millions contributing to India’s economy while not playing a role in democratic processes.

Census 2027 will be digital-first, conducted on tablets rather than paper. This offers advantages: Faster enumeration, reduced errors, real-time monitoring, and quicker results. But it also creates risks. The Census will collect personal information — names, addresses, biometric data, possibly caste, economic status, migration history. Once digitised, that data becomes linkable and searchable. The critical question is: Will Census data remain isolated for statistical purposes, or will it be linked with Aadhaar, National Population Register, or voter rolls to create comprehensive citizen profiles?

Citizens must know that being counted doesn’t mean being tracked. Strict legal constraints are essential: Census data cannot be shared with law enforcement or used for citizenship verification. Independent audits must verify data protection. Technology can make the Census faster and more accurate, but only legal safeguards make it trustworthy.

The Census must be seen as an instrument of fairness, not central control. Every state must have real-time access to its enumeration data during collection. Public dashboards should track progress district by district. Independent audits must verify accuracy before results are finalised. Data should not be weaponised or selectively released. The 2011 SECC experience — caste data unreleased for over a decade — must not be repeated. If Census 2027 counts something, it must publish it. Federal trust depends on procedural integrity and transparency.

Census 2027 is the foundation of everything that follows: Resource allocation, representation, and planning. The six-year delay has already imposed costs. We cannot afford more delays or half-measures.

Census 2027 must be comprehensive (including caste), accurate (counting people where they live), transparent (giving states data access), and protected (ensuring enumeration doesn’t become surveillance). A republic that stops counting eventually stops caring. When we count with fairness and foresight, we govern with justice. Census 2027 is not just about how many we are. It is about who we choose to be as a democracy.

Quraishi is former Chief Election Commissioner of India and the author of An Undocumented Wonder: The Making of the Great Indian Election