On caste census & NPAs, same reluctance to make privilege publicly visible

Why in the News

A reported miscommunication between the office of the Registrar General of India (RGI), which conducts the Census, and the Ministry of Social Justice, which maintains the lists of Scheduled Castes and Other Backward Classes, has established that the Ministry had in fact offered to supply those lists to the RGI. That matters because the stated ground for putting an “open-ended” caste question into the Census, in place of drop-down lists with an “Others” option for unlisted names, was that caste lists were not available. The sequence runs further back. The Union government reversed its position on counting caste before the 2024 general election, then postponed the Census well beyond its due cycle, with the pending delimitation exercise the operative reason. The same state that refuses to publish a caste tabulation also refuses to name the borrowers whose large loans public sector banks have written off. What is contested is therefore not the feasibility of either count, but whether the state will make privilege publicly visible at all.

What is the open ended caste question?

1. The design at issue: An open-ended question records whatever caste name a respondent states, as free text, with no list offered on the schedule.
2. The alternative it displaced: A drop-down list offers the notified caste names for that State, with an “Others” field capturing any name the list does not carry.
3. Why the two diverge at tabulation: Free text returns have to be matched to notified caste names after enumeration, and a name that matches no entry cannot be counted against any category.

What does the reported miscommunication establish?

1. The stated excuse does not hold: The ground given for the open-ended question was the unavailability of caste lists, and the Ministry of Social Justice had offered the Scheduled Caste and Other Backward Class lists to the RGI.
2. The delay had a separate driver: The Census was postponed beyond reasonable limits, with the impending delimitation exercise the reason.
3. The obstruction is not new, only more open: Governments have avoided, prevented or diluted a caste count across administrations, the Congress in 2011 and the Bharatiya Janata Party now.
4. Opposition has been continuous since 2001: Every proposal to count caste since the 2001 Census has attracted immediate and insistent opposition.

Who opposes a caste count, and on what interest?

1. Anti-reservation organisations: Youth for Equality, which opposes reservation, was the primary petitioner against the Bihar caste survey, and the Akhila Karnataka Brahmana Mahasabha petitioned against a similar survey in Karnataka.
2. Dominant caste beneficiaries of reservation: The Akhila Bharata Veerashaiva Lingayat Mahasabha and the Rajya Vokkaliga Sangha petitioned the Karnataka High Court to stay the 2025 caste survey, on the apprehension that a count would reduce their existing share.
3. The legal arguments were recycled: Petitioners in Bihar and Karnataka invoked grounds already rejected by the Supreme Court or superseded by constitutional amendments, including the 105th Amendment.
4. The elite response is about status, not quota share: For those in the so-called General Category who occupy the highest positions, being asked their caste is treated as implicating them in something connected to reservation, which is why some technology proprietors, film actors and other prominent individuals dissociated themselves from these surveys.
5. A partial softening after 2019: Reservation for Economically Weaker Sections, introduced in 2019 and in effect available to the upper castes, softened attitudes to a degree, and beneficiaries of reservation have long attracted resentment from the General Category.

Why is a caste count conceded in principle and blocked in design?

1. The electoral bind: An elected government must accommodate the rich and the powerful, and must also seek votes from the mass of people who are neither, so it concedes the principle of a caste count and then designs the count so that it does not produce one.
2. The debate assumes one kind of benefit: Most participants treat a caste census as being about reservation, so opposition to reservation converts directly into opposition to counting caste.
3. The spectrum of positions is therefore truncated: A frame built only around reservation leaves out the array of benefits the state confers on those who count but are never counted.

What does the refusal to name written off defaulters show?

1. Disclosure was judicially authorised: Reserve Bank of India v. Jayantilal N. Mistry (2015) held that information on wilful defaulters owing large sums to public sector banks is liable to be released under the Right to Information Act, 2005.
2. Banks and the regulator have not complied on write-offs: Both the banks and the Reserve Bank of India (RBI) have refused to provide details on loans written off.
3. The 2024 Directions stop short of write-offs: The RBI’s Wilful Defaulters and Large Defaulters Directions, issued in 2024, require individual banks to list on their websites the instances of default where suits have been filed.
4. Only aggregates are released: No information beyond aggregate amounts is provided on loans written off. Right to Information applicants have asked for those details repeatedly.
5. The most recent refusal: Bank of Baroda declined to name the defaulters on loans totalling Rs 35,715 crore, written off between 2020-21 and 2025-26, on which average recovery was 28 per cent.

What links a caste tabulation to a defaulters’ list?

1. The same state behaves differently on these two datasets: An administration that collects personal data extensively is reluctant to collect or release data in both these instances.
2. The common root is visibility of privilege: The reluctance stems from an unwillingness to make privilege publicly visible, so the case against the caste census rests on an aversion to counting the privileged and to putting inequality on record.
3. Neither dataset would reveal anything new: Neither a caste tabulation nor a defaulters’ list would disclose anything about the privileged that is not already known in a general way.
4. What is being defended is a claim, not a secret: An official and public count or list is resisted because it breaches an implicit right of the privileged to control their own social visibility.

Challenges to counting caste in the Census

1. Stated caste names do not map to a fixed list: A respondent’s own term can be a synonym, a sub caste or a spelling variant that matches no entry in any notified list. Eg. The Socio Economic and Caste Census of 2011 threw up roughly 46 lakh distinct caste and sub caste returns.
   The Fix: Publish the enumeration schedule with State specific drop-down lists mapped to a standard code, retaining a free text field for returns outside the list.
2. There is no single national list to count against: Scheduled Caste and Other Backward Class lists are notified State by State, so one caste can be listed in one State and absent in the next. Eg. The Jat community sits in the central list of Other Backward Classes for some States and not for others.
   The Fix: Publish a concordance mapping every State list entry to a central code before enumeration begins, so a return is classifiable at the point of entry.
3. Enumeration records a declaration, not an entitlement: The count captures what a household states, with no check against a caste certificate. Eg. Bihar’s caste survey of 2022-23 recorded caste on the respondent’s own declaration.
   The Fix: Record the stated caste name and the existence of a certificate as separate fields, so the two are tabulated apart rather than conflated.
4. The count’s timetable carries a seat allocation stake: The Census schedule determines when readjustment of constituencies can begin, which gives the timing an interest independent of enumeration. Eg. The freeze on readjustment of Lok Sabha seats under the Constitution (Eighty-fourth Amendment) Act, 2001 runs until the first Census taken after 2026.
   The Fix: Separate publication of the caste tabulation from the readjustment exercise, so the count’s release does not wait on a seat allocation decision.
5. Collecting a return does not commit the state to publishing it: Enumeration and publication are distinct decisions, and the second can be withheld indefinitely. Eg. The caste data of the 2011 Socio Economic and Caste Census was referred to an expert group and never released.
   The Fix: Release the caste tables on the same notified schedule as the Census’s other tables, so publication is not a separate discretionary step.

Conclusion

Two disclosure questions now sit with the executive at the same time. One is whether the Census schedule will carry notified caste lists or free text, which decides whether the enumeration produces a usable tabulation at all. The other is whether the regulator will extend its default disclosure requirement from suits filed to loans written off, which is where the larger sums sit. What to watch is the final form of the Census caste question and any amendment to the RBI’s Directions covering write-offs, since both are administrative decisions that need no legislation and neither has been taken.

What is transparency and accountability?

1. Transparency: Public officials and institutions have a duty to act visibly and to provide clear, accessible information about their decisions and actions.
2. Accountability: Public authorities are obliged to explain their actions, justify them, and take responsibility for them.
3. Why the pair exists: Information held by the state is the precondition for a citizen questioning its use, so disclosure converts a grievance into a claim the state must answer.
4. The open government standard: The Organisation for Economic Co-operation and Development (OECD) defines open government as transparency in government actions, accessibility of government services and information, and responsiveness of government to new ideas, demands and needs.

Laws and Rules Governing Transparency and Accountability

1. Right to Information Act, 2005: Entitles any citizen to seek information from a public authority without stating a reason for the request.
2. Section 4(1)(b): Requires a public authority to publish specified categories of information on its own motion, so that fewer requests need to be filed.
3. Section 7: Sets 30 days for a reply, and 48 hours where the information concerns the life or liberty of a person.
4. Section 8: Lists the exemptions, and Section 8(2) permits disclosure where the public interest outweighs the protected harm.
5. Section 20: Allows a penalty of Rs 250 a day, to a ceiling of Rs 25,000, on an officer who wrongfully refuses information.
6. Right to Information (Amendment) Act, 2019: Removed the fixed five year tenure of Information Commissioners, leaving the term to be prescribed by the Centre, and ended the parity of their salaries with those of Election Commissioners.
7. Whistle Blowers Protection Act, 2014: Provides a mechanism to receive disclosures of corruption or misuse of power and to protect the person making them.

Challenges in Transparency and Accountability

1. Information Commissions run behind their own caseload: Appeals accumulate faster than commissions dispose of them, so a delayed disclosure loses its use. Eg. Over four lakh appeals were pending across 29 Information Commissions as of 2024.
   The Fix: Fix a statutory disposal limit for appeals, as the Second Administrative Reforms Commission recommended, and report disposal against it.
2. Commissions sit without heads: A commission lacking a Chief Information Commissioner cannot constitute benches, so its docket stops moving. Eg. Nine Information Commissions were without a chief in late 2025.
   The Fix: Begin the appointment process a fixed period before a vacancy arises, with the shortlist published.
3. The penalty provision is rarely used: The power to fine an officer for wrongful refusal is exercised in a small fraction of the cases that attract it, so refusal carries no cost. Eg. Penalties are imposed in about 4 per cent of the cases where they are warranted.
   The Fix: Require a commission to record written reasons whenever it declines to impose a penalty after finding wrongful refusal.
4. Exemptions are read wide and the public interest override narrow: The exemption grounds are invoked routinely and the override that answers them almost never is. Eg. Section 8(2)’s public interest override is invoked in under 1 per cent of cases.
   The Fix: Harmonise the Digital Personal Data Protection Act, 2023 with the disclosure regime so the public interest override, not the personal information exemption, settles a request naming individuals.
5. Requesters carry personal risk: Seeking records on local contracts, land and licences exposes the applicant to retaliation. Eg. Over 100 Right to Information users have been killed since 2005.
   The Fix: Notify the rules under the Whistle Blowers Protection Act, 2014 and extend its machinery to information applicants, so a threatened applicant has a statutory route.

Back2Basics: The Constitution (One Hundred and Fifth Amendment) Act, 2021

1. What it did: Restored the power of States and Union Territories to prepare and maintain their own list of socially and educationally backward classes.
2. Why it was needed: An earlier reading of the Constitution (One Hundred and Second Amendment) Act, 2018 had left the power to notify backward classes with the Centre alone.
3. Provisions touched: It amended Article 342A and clarified Article 366(26c), so a State list and the central list operate separately.

Matching Previous Year Question

“[2020, GS2, 10 marks] “Recent amendments to the Right to Information Act will have profound impact on the autonomy and independence of the Information Commission”. Discuss.”