Ex-officials flag concerns over Census data quality, potential NPR, NRC link

Why in the News

Two administrative decisions on Census 2027 have drawn objections from former Census officials. The Registrar General and Census Commissioner of India (RG&CCI) notified 40 questions for the Population Enumeration phase on 14 August 2026, several of which mirror the rehearsal form of the National Population Register (NPR) used in 2019. The RG&CCI then advanced the Population Enumeration phase in the poll bound States of Uttar Pradesh, Punjab, Uttarakhand and Goa on 5 September, while deferring it in violence hit Manipur. The objection is that the two decisions together degrade data quality and revive a link between the Census, the NPR and a National Register of Citizens (NRC). Section 15 of the Census Act, 1948 bars that use, and the 1951 Census is the precedent showing it happened anyway.

What is the National Population Register?

1. A register of residents: The NPR records persons residing in a local area, and it is generated under the Citizenship Act, with its data intended to be published.
2. Its current scale: It was first created in 2011, updated in 2015, and already holds a family wise database of 119 crore residents.
3. Its stalled status: It was to be updated alongside the first phase of the now scrapped Census 2021 and has been held back after opposition from political parties and civil society groups.

What has changed in the Census 2027 design?

1. The question set has grown: 40 questions were notified on 14 August 2026 for the Population Enumeration phase.
2. The new fields are identity fields: They include particulars of a person’s parents including their names, religion, place of birth, Aadhaar, mobile phone number, voter identity and driving licence details.
3. The resemblance is to a different exercise: Those additions mirror the questions asked in the 2019 rehearsal form of the NPR rather than anything in the previous Census schedule.
4. The sequence has been reordered: Population Enumeration has been advanced in four poll bound States and deferred in Manipur, breaking the simultaneity a national headcount normally runs on.

Why do former officials say the data quality will suffer?

1. Migrants fall between two counts: Counting four States ahead of the rest of the country could lead to under enumeration of migrant workers, particularly those from Uttar Pradesh, since a person counted nowhere at the time of enumeration is not counted at all.
2. Some questions serve no statistical purpose: Details relating to identity documents and vaccination history are already held in administrative databases, and it is not clear what statistical purpose their collection in a Census would serve.
3. The link remains speculative rather than stated: A former Deputy Registrar General of India has stressed that any link between the Census and population registration activities remains speculative, while noting that the information now sought resembles NPR data.

What separates the two exercises in law?

1. The confidentiality bar: Section 15 of the Census Act, 1948 provides that records of the Census are not open to inspection nor admissible in evidence, and Census data can be used only for statistical purposes.
2. Different statutes, different outputs: The NPR is generated under citizenship law and its data is meant to be published, which is the opposite disclosure rule from the one governing the Census.
3. The 2019 practice went the other way: The update of the NRC in Assam in 2019 was not based on NPR data and required a separate data collection exercise of its own.

What does the 1951 precedent actually show?

1. The register was copied from the Census: The NRC was prepared across the country in 1951 by transcribing important Census data from the Census slips, using the same symbols and abbreviations.
2. The two records were reconciled against each other: Missing NRC entries were copied from Census slips and missing Census slips were recreated from NRC entries under a special reconciliation exercise.
3. It was designed as an administrative tool: The 1951 Census report for Assam, Manipur and Tripura records the register as a Census and administrative instrument rather than a citizenship verification exercise, intended to be maintained permanently and updated by village officials.
4. It was never completed nationally: A countrywide NRC was never published for unspecified reasons, except in undivided Assam, which then also included Manipur and Tripura.

Challenges to Census 2027

1. A staggered count breaks comparability: Population figures collected at different reference moments cannot be compared across States without adjustment, and every downstream allocation rests on that comparison. Eg. Advancing enumeration in Uttar Pradesh, Punjab, Uttarakhand and Goa places four States on a different reference point from the rest of the country.
   The Fix: Publish the reference date used for each State alongside the provisional totals, and state the adjustment method before the figures enter any allocation formula.
2. Identity questions depress response quality: Asking for document numbers in a count that promises confidentiality invites non response and misreporting from precisely the groups most likely to be undercounted. Eg. The 2019 NPR rehearsal form, which the new questions resemble, drew organised opposition on exactly this ground.
   The Fix: Make the identity document fields optional and record the non response rate for each, so the degradation is measured rather than hidden inside the totals.
3. Confidentiality is asserted but not enforceable by the respondent: Section 15 bars inspection and evidentiary use of Census records, and offers the person counted no remedy where the bar is crossed. Eg. The 1951 register was built by copying Census slips despite the same statutory bar being in force.
   The Fix: Provide a statutory penalty for onward transfer of individual Census records to any other register, enforceable on complaint rather than only at the department’s discretion.
4. Enumeration timing collides with the electoral calendar: Running a household level count in poll bound States places enumerators and electoral officials in the same households in the same season, which invites conflation of two separate exercises. Eg. Uttarakhand is simultaneously running a Special Intensive Revision of electoral rolls with a final roll due in October.
   The Fix: Separate the enumeration window from the notified election period by a fixed statutory interval, on the model of the silence period in electoral law.

Conclusion

The objection is not that a register is being built, but that the instrument being used promises confidentiality and the instrument it resembles promises publication. That precedent shows the two can be merged in practice with the statutory bar formally intact, and nothing in the present design forecloses that. The concrete markers to watch are whether the identity document fields survive to the final schedule, and whether the four advanced States are enumerated against a stated reference date the rest of the country can be compared to.

Matching Previous Year Question

“[2009] Consider the following statements : 1. Between Census 1951 and Census 2001, the density of the population of India has increased more than three times. 2. Between Census 1951 and Census 2001, the annual growth rate (exponential) of the population of India has doubled. Which of the statements given abova is/are correct ? (a) 1 only (b) 2 only (c) Both 1 and 2 (d) Neither 1 nor 2 ANSWER: (d)”