I have never understood the fashionable Delhi habit of treating a Census as if it were merely a very large questionnaire. A Census is the operating system of the state. It tells governments whom they are planning for, where public services are failing and whether old political assumptions have survived contact with a changing country. India’s prolonged wait for a fresh count has therefore been a real governance failure, and the Union government deserves criticism for allowing the pandemic-era delay to stretch so long. But the decision to conduct Census 2027, with caste enumeration and a digital-first design, is more consequential than the grudging applause it has received. It takes an issue the Congress and its allies had reduced to a campaign incantation and turns it into an administrative project with dates, rules and consequences.
My view is unfashionably simple: counting caste is sensible; governing by caste slogans is not. India cannot design targeted welfare while pretending that social identity has vanished, but neither can it allow every census table to become an auction catalogue for political quotas. The government’s approach creates an opportunity to replace insinuation with evidence. Whether parties use that evidence responsibly is another question. I am not optimistic about all of them.
From slogan to state capacity
The official notification fixes the reference date for most of India at March 1, 2027, with October 1, 2026 for snow-bound areas. The government’s announcement of the Census schedule also confirms caste enumeration. That matters because a serious count imposes disciplines that a rally speech does not: categories must be defined, enumerators trained, answers recorded consistently, privacy protected and disputes resolved.
Congress leaders were entitled to press for a caste census. On that narrow point, I think they identified a legitimate gap in public data. What irritated me was the packaging of the demand as moral revelation, as though no policy question remained once the words “jitni abadi” had been spoken. Population share is information, not a complete philosophy of justice. Need, deprivation, geography, gender, disability and access to state services do not disappear when a caste column is added to a form.
The BJP government has now accepted the count, depriving the opposition of its favourite accusation that the Centre feared the numbers. This is not capitulation; it is political adaptation of the useful sort. A governing party should be capable of absorbing a valid demand without swallowing the ideology wrapped around it. Narendra Modi’s governments have often preferred beneficiary-based politics—housing, toilets, bank accounts, cooking gas and food support—to the Congress tradition of appointing itself hereditary interpreter of social coalitions. Better caste data can sharpen that delivery model if it identifies communities and districts repeatedly missed by schemes.
It can also expose lazy claims. Parties routinely speak of enormous blocs as if every family within them has identical circumstances. They do not. A credible dataset may reveal striking internal differences across states, occupations and urbanisation levels. That will make politics noisier, but it can make policy more precise. I would rather argue over measured inequalities than over numbers reverse-engineered for television.
The digital gamble is the real reform
The caste headline has obscured the institutional innovation. According to the Office of the Registrar General and Census Commissioner, this will be India’s first digital Census, with mobile applications for data collection and an option for self-enumeration. That is not cosmetic modernisation. Done well, it can shorten processing times, reduce layers of manual transcription and make field supervision more responsive.
Anyone who has read old statistical publications knows the absurd lag between social change and usable official data. By the time tables are cleaned, printed and debated, migration has redrawn towns and welfare rolls have acquired new errors. Digital collection cannot abolish delay, but it can reduce the administrative archaeology. It also allows validation rules to catch obvious inconsistencies at entry rather than years later.
There are reasons for caution. An app does not magically improve a poorly framed question. Connectivity varies, devices fail, languages require careful handling and self-enumeration may work best for precisely the households the state already finds easiest to see. The digital divide is not a left-wing invention. Enumerators must remain central, and offline capability cannot be a footnote. India’s most vulnerable residents must not become missing values because a product manager admired a clean dashboard.
Privacy is the larger test. Caste data is sensitive; household-level information can be misused socially and politically. The government must explain plainly who can access raw records, how devices are secured, when identifiers are separated and what penalties follow unauthorised disclosure. “Trust us” is not a data architecture. A pro-government position, if it means anything beyond cheering, should demand that a reform survive beyond the government that introduced it.
I would also publish methodology notes, category guidance and quality audits early. The count will generate disputes because caste names, sub-castes and local spellings do not fit into tidy national boxes. Secrecy would feed conspiracy. Transparent rules would not eliminate grievance, but they would force critics to engage with the actual exercise.
What the numbers cannot decide
The predictable next move will be to claim that every population ratio commands an equivalent share of every public institution. I reject that arithmetic. Reservations address historical exclusion and unequal opportunity; they are not a corporate dividend distributed solely by headcount. Constitutional limits, judicial doctrine, administrative effectiveness and the interests of the poorest within broad categories all remain relevant.
This is where Rahul Gandhi’s politics often appears to me less like social justice than spreadsheet populism. He is right that power structures deserve scrutiny. He is wrong when he implies that a national count produces self-executing answers about representation, budgets or private property. Democratic policy requires choices among competing goods. Pretending otherwise is a way to claim virtue without accepting responsibility for trade-offs.
The BJP will face its own temptations. It may prefer to celebrate the decision and postpone discussion of uncomfortable findings. That would squander the exercise. If the data shows that particular communities have been bypassed by schooling, credit or government recruitment, ministers should say so and redesign programmes. Credit for authorising the count creates an obligation to release useful tables promptly, not to curate only politically convenient ones.
Nor should every scheme become caste-exclusive. Universal infrastructure often does more for dignity than endlessly fragmented entitlements: a reliable road, functioning school, piped water and primary health centre serve people without demanding that they perform identity before a clerk. The best use of granular social data is frequently to locate universal services where deprivation is concentrated, while reserving narrowly targeted measures for demonstrable barriers.
I want Parliament to debate these principles before the tables arrive. Otherwise the first viral chart will set the terms. A mature debate would distinguish descriptive data from normative decisions, household deprivation from group averages and temporary remedies from permanent political property. That may sound dry. Government is dry. The alternative is letting hashtag economics allocate the republic.
A count worth doing properly
The delayed Census has costs that cannot be wished away. Planning based on projections is inferior to an actual count, particularly after a pandemic and a decade of rapid urban migration. The Centre should acknowledge that more candidly. Defending the present government does not require inventing perfection where it plainly does not exist.
Yet politics must judge the decision now before us, not only the delay behind it. India is preparing a mammoth, digitally enabled enumeration that includes a question opposition parties insisted was being suppressed. That is evidence of a state capable of correction. It also demonstrates the difference between campaigning and governing: the first finds a grievance; the second must design the form, train millions, secure the data and live with the result.
My hope is that Census 2027 will puncture two myths at once. The first is the Congress claim that only its ideological vocabulary can recognise social disadvantage. The second is the complacent belief, occasionally visible on the right, that development alone has already dissolved inherited barriers. India is too complicated for either comfort.
Count everyone. Publish rigorously. Protect individual records. Then argue fiercely about policy without pretending the table has made the decision for us. If the government manages that, it will have done more for evidence-based social justice than years of opposition slogans. If it does not, critics will be justified. For once, however, they will have numbers rather than nostalgia.
