Subscribe
World Population Day usually turns public attention to the population size. Is India 1.4 billion, 1.45 billion or 1.5 billion? Has fertility fallen? Which states are growing faster? Which regions will gain or lose political weight?
These questions matter. But for an urbanising country, another question is equally important: where are people counted?
Population is not only a matter of births and deaths. In demography, the population of any place changes through four components: births, deaths, in-migration and out-migration. A city grows not only because children are born there, but because workers move into it. A village shrinks not only because people die, but because people leave. Migration is therefore not a side story in population statistics. It is one of the main engines of urbanisation.
This is why Census 2027 matters so much.
India’s last census was conducted in 2011, when the country’s urban population was 377.1 million, or 31.2 per cent of the total population. Rural India still accounted for 68.8 per cent. But 2011 is now an old baseline for a country that has seen rapid migration, peri-urban expansion, construction growth, climate stress, new industrial corridors and new informal settlements over 16 years. The World Bank estimates that by 2036, India’s towns and cities will house 600 million people, or 40 per cent of the population, and that urban areas already contribute almost 70 per cent to GDP. It also estimates that India will need $840 billion in urban infrastructure investment by 2036.
India’s urban dream — smart cities, Viksit Bharat 2047, metro networks, housing, climate-resilient infrastructure and municipal finance reform — rests on one basic statistical foundation: the census count of urban India.
Urbanisation is usually reported as a simple ratio:
Urbanisation rate = urban population / total population × 100
But the simplicity is deceptive. The numerator — urban population — is not a natural fact. It is produced through census classification.
In India, urban areas include statutory towns such as municipal corporations, municipalities, cantonment boards and notified town areas. They also include census towns: settlements that are not legally urban but meet three criteria — population of at least 5,000, density of at least 400 persons per sq km, and at least 75 per cent of male main workers engaged in non-agricultural work.
This means urban India is calculated through three connected questions: how many people are counted in a settlement, how dense that settlement is, and what kind of work its male main workers do.
Migration affects all three.
If migrant workers are counted at destination, they increase the city’s population, density and labour profile. If they are counted at origin, the city may remain smaller in data than it is in daily life. If workers are absent, proxy-counted, misreported or missed, the urbanisation rate, census town classification and city-level service demand can all be distorted.
This is not only a statistical issue. It is an urban economics issue.
The worker may be urban, but the statistic may remain rural
India’s cities run on migrant labour. Workers from Odisha, Bihar, Uttar Pradesh, Jharkhand, Chhattisgarh, West Bengal, Rajasthan and Madhya Pradesh build and service Bengaluru, Surat, Hyderabad, Delhi NCR, Mumbai, Pune, Chennai, Kochi and many other urban economies.
They work in construction, textiles, logistics, sanitation, domestic work, security, street vending, hotels, factories and platform services. Their labour is urban. But their documents, caste certificates, ration cards, voter identity, family obligations and social recognition often remain tied to the village.
This creates what may be called statistical rurality: people who are economically urban but counted as rural.
Consider one worker from rural Odisha living in Bengaluru. If he is counted in Bengaluru, he enters Karnataka’s urban population. If he is counted in his village, he remains part of Odisha’s rural population. The total population of India does not change. But the geography of population changes. Karnataka’s urban population is understated; Odisha’s rural population is overstated relative to lived economic activity.
That matters for city finance, water supply, sanitation, housing, heat-action planning, transport, ward population and labour-market estimates. A city may bear the cost of migrant residence while the migrant remains statistically attached to the village.
Census 2027 matters for five reasons.
First, it comes after a long delay. Census 2021 was postponed following the COVID-19 pandemic, and Census 2027 will now become the first full census after 2011. The official reference date for Population Census 2027 is March 1, 2027.
Second, it will be India’s first digital census, using mobile-based data collection, geo-referenced jurisdictions, near-real-time monitoring and an optional self-enumeration facility.
Third, it will include caste enumeration during the Population Enumeration phase. Earlier censuses systematically enumerated Scheduled Castes and Scheduled Tribes; Census 2027 will undertake broader caste enumeration.
Fourth, it will generate the new baseline for urban projections, infrastructure planning, welfare targeting, environmental planning, and future state and city population estimates.
Fifth, it is happening at a time of wider documentary anxiety. Public debates around welfare databases, nativity, citizenship, caste certification and Special Intensive Revision of electoral rolls have made people more sensitive to names, documents and official lists. Recent reports on electoral roll revisions and deletions show why citizens may not always separate one form of state enumeration from another, even though a census is legally and administratively different.
The Census Act protects individual census information. The government has stated that Section 15 treats personal information as strictly confidential; it cannot be made public under the Right to Information Act, used as evidence in court, or shared with any institution. But law and perception are not the same thing. Poor and mobile households may still fear being counted wrongly.
For migrant workers, the risk is not simply being missed. It is being counted incorrectly.
Will a locked village house be followed up? Will an absent household head be recorded? Will the caste name be spelled correctly? Will a worker counted in Bengaluru lose recognition in Odisha? Will a family at a brick kiln be counted at the worksite, at origin, or not at all? Will a Banjara or other nomadic household be recorded under the correct community name?
Such anxieties can produce what may be called enumeration-induced return migration: migrants returning home, or ensuring origin-based reporting, because they fear omission, caste misclassification or future loss of recognition.
The scale of such behaviour is unknown. It should not be exaggerated. But even a small behavioural response matters statistically if it is concentrated in migrant-heavy corridors, census towns, slums, construction camps and industrial clusters.
Caste enumeration raises the stakes further. Many short-term, seasonal and circular migrants come from Scheduled Castes, Scheduled Tribes, OBCs, Denotified Tribes and Nomadic communities. For them, caste is not merely a social label. It is linked to dignity, reservation, welfare and political voice. If caste-marginalised migrants are counted mainly at origin, cities may appear less caste-structured than their labour markets really are.
The city uses their labour but may not fully record their social location.
At the national level, the effect of migrant mislocation may appear modest. Suppose India’s population is about 1.45 billion. If 10 million functionally urban migrants are counted at rural origin instead of urban destination, the measured urban share falls by around 0.69 percentage points. If 20 million are counted at origin, the fall is around 1.38 percentage points. If 50 million are counted at origin, the fall is around 3.45 percentage points.
This is not a prediction. It is a sensitivity exercise.
But for urban economics, even 0.5 percentage point is not small. It means millions of people. At city scale, a few lakh uncounted migrant residents can change estimates of rental housing demand, bus ridership, sanitation burden, hospital capacity, school enrolment and heat exposure.
India may then build the infrastructure of 2047 on the population baseline of a mismeasured 2027.
This World Population Day, the question is not only how large India’s population is. It is whether India can locate its population correctly.
Census 2027 will not only tell us how many people live in India. It will tell us how urban India really is, how many workers sustain its cities, how settlements are transforming, and how climate and infrastructure risks are distributed.
A census should not force migrants to choose between urban visibility and rural belonging. A worker should not fear that being counted in Bengaluru will weaken his caste identity in Odisha. A family should not disappear because it lives between a village and a worksite. A nomadic community should not be statistically fragmented because it does not fit the settled household model.
India wants to build the cities of 2047. First, it must count the people who are already building them.
Madhusudan Nag is currently based at the Maulana Azad National Institute of Technology (MANIT), Bhopal. He has Ph.D in Economics from Centre for Development Studies, Kerala. He tweets at @MSnag_writes on X.
Shakuntala Ghadai is a PhD scholar at the Centre for Development Studies, Kerala. She tweets at @shakuntala_read on X.
Views expressed are the authors’ own and don’t necessarily reflect those of Down To Earth