India is about to run the largest counting exercise on the planet, and it will look nothing like the one that came before it. Census 2027 is the country’s 16th national census, its first since 2011, the first ever to be conducted digitally, and the first to enumerate caste across all communities since 1931. Those are not small upgrades: they reset how India measures itself after a fifteen-year blind spot, and the data produced will shape welfare schemes, constituency decisions, resource allocation and a long-delayed political debate about who exactly lives in the country and how. Here is what the exercise involves, when it happens and why the caste column is the part everyone will be arguing about.
A census fifteen years in the making
India’s decennial census is the backbone of nearly every statistic the country runs on — population projections, welfare targeting, constituency data, urban planning, even vaccine distribution models. Yet the previous count dates to 2011, and the one scheduled for 2021 was postponed indefinitely by the COVID-19 pandemic, creating the longest gap between censuses since the practice began. That gap has forced governments to work with estimates and projections for years, and the further the data ages, the shakier every figure built on it becomes.
The government formally set things in motion in June 2025, when a Gazette notification under Section 3 of the Census Act, 1948 kicked off the process. In December 2025, the Union Cabinet approved a budget of ₹11,718 crore for the exercise, which is being described as the largest administrative and statistical operation in the world. Around 30 lakh field workers — enumerators and supervisors, mostly school teachers on enumeration duty — will fan out across the country, supported by roughly 1.3 lakh dedicated census functionaries, many carrying smartphones or handheld devices preloaded with the census application.
How the census unfolds: two phases, three reference dates
The census runs in two phases. The first, the House Listing and Housing Census, stretched from April 1 to September 30, 2026, with each state and Union Territory conducting its houselisting for a continuous 30-day window. Enumerators visited every structure in the country to record building characteristics, household size, water and electricity access, sanitation, assets and, for the first time, internet connectivity at home. The second phase, Population Enumeration, moves door to door across most of India in February 2027, with a reference date of 00:00 hours on March 1, 2027.
Weather makes the schedule more complicated than that. Ladakh and the snow-bound areas of Jammu and Kashmir, Himachal Pradesh and Uttarakhand conduct population enumeration far earlier — September 1 to 30, 2026 — with a reference date of October 1, 2026. A third reference date of January 5, 2027 applies to Goa, Uttar Pradesh, Punjab and the non-snow-bound parts of Uttarakhand, whose enumeration runs from December 1, 2026 to January 4, 2027. Provisional population totals are expected within roughly ten days of enumeration wrapping up, with detailed tables arriving in stages over the following months and years.
The caste question returns after 95 years
This is the headline change and the most consequential one. India last counted caste across all communities in 1931, under colonial rule. Every census since independence has recorded only Scheduled Caste and Scheduled Tribe status, while caste data for the rest of the population came from separate, partially released exercises — most recently the Socio-Economic and Caste Census of 2011, whose caste figures were never published in a usable national form.
The 2027 population enumeration schedule changes that. The notified form carries 40 questions, and where the 2011 version asked simply whether a respondent was SC or ST, the new question 10 reads “SC/ST/Caste” — meaning every household will be asked to state its caste. The political sensitivity is hard to overstate, and the debate is running on both sides. Supporters argue that accurate caste data is essential for measuring inequality and fine-tuning reservations and welfare schemes. Critics have raised process concerns: the Congress party’s communications chief, Jairam Ramesh, publicly questioned why enumeration is not based on a pre-prepared, state-wise caste list as done in recent caste surveys in Bihar and Telangana, arguing that the absence of such a list raises doubts about how responses will be recorded. There is also a technical knot: there is no single unified OBC list, since the Centre and each state maintain their own, and one caste can fall under a backward category in one state and the general category in another. How enumerators and the software handle these mismatches will matter enormously for the quality of the final data.
Going digital: how the first tech-first census works
The other transformation is methodological. Census 2027 is India’s first fully digital census, replacing the paper forms of 2011 with mobile applications, a central Census Management and Monitoring System, GPS-tagged enumeration blocks and real-time supervision dashboards. Paper schedules remain available as a fallback, but the bulk of collection is expected to happen on devices.
The most visible change for ordinary households is self-enumeration — the option to fill in your own data online before an enumerator ever knocks. Roughly 55,000 households used the feature on its first day of rollout, including senior constitutional functionaries testing the system themselves. The process works like this:
- Check the notified window for your state. Self-enumeration opens for a set period ahead of the house-to-house visits in each jurisdiction — 15 days before houselisting, and separate windows ahead of population enumeration.
- Log in to the official self-enumeration portal or app during that window using your registered mobile number.
- Fill in the household and member details through the guided form, answering the same questions the enumerators carry.
- Submit the form and receive a unique reference ID tied to your household.
- Show that ID to the enumerator when they visit, so the field worker can verify and link your self-submitted record instead of re-collecting everything from scratch.
Self-enumeration is entirely optional, and households that prefer the traditional visit lose nothing by skipping it. For the system, though, every self-enumerated record lightens the field workload and reduces transcription errors, which is precisely the point of building the digital layer.
What the new questionnaire actually asks
The demographic core of the census — name, age, marital status, education, occupation, religion, mother tongue, migration, fertility — carries over from previous rounds. What surrounds it has shifted to match how India lives in the 2020s. Among the notable additions:
- Digital footprint: questions on digital literacy, whether individuals use the internet or smartphones, and internet availability at home.
- Documentary identity: Aadhaar number, voter ID, passport, driving licence, mobile number and the number of bank accounts held.
- Caste for all: the expanded SC/ST/Caste question applying to every respondent, not just notified groups.
- Migration reasons refreshed: new categories including displacement due to climate events and natural disasters.
- Inclusive gender recording: explicit options to record transgender identity.
- Health and family details: place of COVID-19 vaccination and separate particulars for both parents and spouse.
Together these additions turn the census into something closer to a national identity and connectivity audit, layered on top of the traditional demographic count.
How 2027 compares with 2011
The scale of the overhaul becomes clear when the last census and the new one are placed side by side:
| Feature | Census 2011 | Census 2027 |
|---|---|---|
| Data collection | Paper forms | Mobile apps with paper fallback |
| Self-enumeration | Not available | Optional, via official portal and app |
| Caste enumeration | SC/ST only | All communities, first time since 1931 |
| Houselisting questions | 34 columns | 33 notified questions, internet access added |
| Population enumeration | 28 columns | 40 questions, including digital and identity data |
| Budget and gap | Standard decennial cycle | ₹11,718 crore approved after a 15-year gap |
The comparison highlights two things at once. The continuity is real — the structure, the two-phase design and most demographic questions are recognisably the same exercise India has run for a century and a half. But the additions, particularly the caste column and the identity-linked fields, make the 2027 dataset far more granular, and correspondingly more politically charged, than anything the country has produced before.
The risks the digital census has to manage
A first-ever digital exercise of this scale carries its own set of failure points. Connectivity is the obvious one: enumerators in remote and rural areas will need the app to work offline and sync when a signal returns, and the training of 30 lakh field workers on new devices in a compressed timeline is a genuine logistical feat. Data protection is the second concern — a form that asks for Aadhaar numbers, bank account counts and mobile numbers concentrates an enormous amount of personal information in one system, and questions about storage, access and safeguards are already being raised in Parliament and outside it. The third is data quality in the caste column itself: without a standardised master list, respondents and enumerators may record the same caste in different ways, which could complicate analysis unless the government publishes clear coding rules.
There is also the boundary-freezing rule to remember: with the notification issued, states may not alter the maps of districts, tehsils and police stations until the census completes, since enumeration blocks are built directly on administrative boundaries.
Why the results will echo for years
Census data is never just a headcount. The figures feed the formulas behind centrally sponsored welfare schemes, become the denominators for every per-capita statistic the government and economists quote, refresh the population projections used for everything from school planning to constituency preparation, and, for the first time in a century, give India a nationwide picture of its caste composition. That last dataset in particular will feed directly into the loudest policy debate of the coming years — how reservation and social justice policy should be calibrated once actual numbers replace the estimates everyone has been working from.
What will genuinely change is the resolution. A digital census with self-enumeration can publish provisional totals in days rather than months, update tables faster and slice data at a level paper forms never supported. A census that finally counts caste can close the gap between the India of 1931 and the India of 2027 in one column. The exercise that ends on March 1, 2027 will not answer every question people are asking of it — but for the first time in fifteen years, the country will at least be counting from facts rather than from memory.