The Union government’s decision to expedite the 2027 phase of the Census in the four poll-bound states, coupled with the introduction of several new questions in the census schedule, has prompted some former Census officials to express concerns over the quality of data to be collected and the potential linkages between the Census, the National Population Register (NPR) and eventually the National Register of Citizens (NRC).
There is precedent for using census data to prepare the NRC. Records show that this happened in 1951, during the first census of Independent India, when the NRC was prepared across the country by ‘copying’ data from the census schedules. However, for unknown reasons, a nationwide NRC was never published except in undivided Assam, which at that time also included Manipur and Tripura.
Talking to HinduFormer Deputy Registrar General of India K. Narayanan Unni stressed that any link between the census and population registration activities remains speculative. However, some of the information sought in the latest census questionnaire resembled data collected during previous NPR updates.
The NPR, first created in 2011 and originally proposed to be updated to accommodate the first phase of the now-cancelled 2021 census, was shelved following stiff opposition from political parties and civil society groups who raised concerns that the data could be used to create a nationwide NRC amid allegations that the process could lead to large numbers of people being left stateless.
Mr. Unni said, “NPR is created under the Citizenship Act and the data has to be published. However, Census data is confidential and can only be used for statistical purposes. But in 1951, Census data was used to prepare NRC in Assam…” He added that although Census data is confidential, it does not mean that the 1951 practice of using it to prepare NRC will not be repeated.
On August 14, 2026, the Registrar General and Census Commissioner of India (RG&CCI) notified a set of 40 questions to be asked in the census phase or second phase of the census by adding 13 new questions or data fields that were not part of the 2011 Census questionnaire.
Some additional questions, such as details of a person’s parents including their name, religion, place of birth, Aadhaar number, mobile number, voter ID and driver’s license details, mirror those asked in NPR’s 2019 rehearsal form.
NPR was updated in 2015 and already has a household database of 119 million residents.
1951 exercise
1951 Census Report prepared by R.B. Vagaiwalloy, Superintendent of Census Operations in Assam, Manipur and Tripura, shows that the NRC was intended primarily as a census and administrative tool rather than as a means of verifying citizenship.
“An important innovation of this census was the preparation of the NRC, in which all important census data were transcribed from the census forms. The register uses the same symbols and abbreviations and is a copy of the census forms,” the 1951 report stated.
The missing NRC records were copied from the census forms and the missing census records were recreated from the NRC records as part of a special reconciliation exercise carried out by Vagaiwalla. This comes despite the Registrar General’s countrywide direction to limit such checks “in the interest of saving time and money”.
According to Vagaiwalla, the register was to be maintained on an ongoing basis and updated by the village authorities, thereby ensuring continuity between censuses. Although Section 15 of the Census Act 1948 provides for confidentiality and states that “Census records [are] not verifiable and not admissible in evidence,” the NRC was prepared using the data, which set a precedent for using census data to prepare the register.
A former census official, speaking on condition of anonymity, said NPR and the census are legally separate activities governed by different laws. The official noted that the NRC update in Assam in 2019 was not based on NPR data and required separate data collection.
Underestimating problems
On September 5, the RG&CCI advanced the 2027 census phase in the poll-bound states of Uttar Pradesh, Punjab, Uttarakhand and Goa, while postponing activities in violence-hit Manipur. The census will also be the first in independent India to list castes beyond Scheduled Castes and Scheduled Tribes.
Mr. Unni said the decision to conduct the census in four states earlier than the rest of the country could lead to an undercount of migrant workers, especially from Uttar Pradesh.
“Census dates have traditionally been set around March 1 because migration was relatively low during this period. Advancing the process in states where Assembly elections are due may mean that a few migrant workers will return home to vote and will not be there where they usually reside. By the time they reach their home state, the census there may have already ended, creating the possibility that some migrants will not be counted in any place,” he said he.
Barring these four states, as well as Ladakh and the snow-covered areas of Himachal Pradesh, Uttarakhand and Jammu and Kashmir, the census in the rest of the country will take place in February 2027 with a reference date of March 1, 2027. Snow-covered areas are traditionally counted before the rest of the country.
India follows de facto a census system in which a person is counted where he or she is present during the census period. Mr Unni said even a small increase in undercounting could affect the quality of the data, as past censuses had already recorded some degree of undercounting.
On the government’s decision to postpone the census in Manipur, he said conducting the census without the cooperation of the public would jeopardize the quality of data. However, he argued that a delay could deprive the state of demographic data relevant to claims about migration and demographic change.
“Respondent fatigue”
Mr Unni questioned the usefulness of some of the new questions, especially those related to identity documents and vaccination history. “Many of these details are already available in administrative databases. It is unclear what statistical purpose some of these questions will serve,” he said.
He also warned that the lengthy questionnaire could increase “respondent fatigue” and create practical difficulties for enumerators, most of whom are public school teachers.
“If a household has to answer dozens of questions for each family member, this will increase the time spent per household and can lead to resistance, especially in urban areas where enumerators are often unknown to residents,” he said.
Another former census official raised questions over the inclusion of Aadhaar data in the 2027 census, noting that collection of Aadhaar data may require changes to existing legal provisions.