Polity10 September 20263 min read
Registrar General notifies 40-question Census 2027 schedule, drawing privacy objections
The Office of the Registrar General & Census Commissioner, India (ORGI) has notified a 40-question schedule for Population Enumeration, the second phase of Census 2027. The 2011 Census had 29 questions. New fields include Aadhaar, Voter ID, mobile number, nationality, parents' details and bank accounts.
Many of the questions are relevant to demographic and socio-economic planning. The objection is to the extent of personal information being collected.
Why the extra questions are contested
Critics say several questions closely resemble what is sought for the National Population Register (NPR): parents' birth details, declared nationality, Aadhaar, Voter ID, passport, driving licence and mobile number. They fear that collecting parental details and nationality could shift the Census from a statistical exercise towards population and citizenship profiling, potentially facilitating a National Register of Citizens (NRC).
The UN Fundamental Principles of Official Statistics call for keeping individual information confidential and using it strictly for statistical purposes. The Census Act, 1948 contains confidentiality safeguards, but concerns remain that Census data could be shared with the NPR or other administrative systems, which would undermine that principle.
Privacy worries are sharper because Census 2027 will be the country's first fully digital and self-enumerated demographic exercise. Critics say Aadhaar, Voter ID, passport numbers and the location of Covid-19 vaccination have little relevance for the aggregate statistics needed for socio-economic planning. Centralising such identifiers into linkable profiles raises the risk of data breaches, unauthorised access and de-anonymisation. They also say the shift from statistical gathering to individual-level tracking violates purpose limitation and could ultimately facilitate state surveillance.
The caste field
Entries for Scheduled Castes and Scheduled Tribes will be selected from a predefined list. Other castes, including OBCs, will be recorded in an open-text field based on the respondent's declaration.
Different spellings, regional names and sub-caste identities could produce a large number of duplicate or unclassified entries. A similar approach in the 2011 Socio-Economic and Caste Census (SECC) reportedly produced around 46 lakh caste names, making standardisation and meaningful analysis difficult. Without a strong standardisation and validation mechanism, the exercise may generate extensive data but limit its usability for policymaking, critics say.
Practical and methodological challenges
A 40-question form may increase respondent burden. The primary respondent may not have immediate details of every household member, leading to incomplete or inaccurate answers. The problem is greater in hostels and old-age homes, where administrators may not hold detailed parental or identity information on all residents.
The de facto enumeration approach may also undercount migrant workers, who may be away from their usual place of residence during the Census.
Suggested safeguards
The remedies put forward are:
- Replace the open-ended caste field with a standardised, consultative caste list in the form of a drop-down menu.
- Remove Aadhaar, passport and bank account numbers. If financial inclusion data is needed, a Yes/No question on bank account ownership would suffice.
- Conduct a thorough Post Enumeration Survey (PES) to identify undercounting and inaccurate responses.
- Set up legal and administrative firewalls between the Census and the NPR and NRC, so that Census data is not repurposed for citizenship profiling or administrative tracking.
- Put in place cybersecurity protocols aligned with the Digital Personal Data Protection (DPDP) Act, 2023, including end-to-end encryption, independent security audits and strong access controls, to prevent unauthorised access and cross-linking with other government databases.