7 min readAug 25, 2026 06:28 AM IST First published on: Aug 25, 2026 at 06:28 AM ISTI received a pleasant surprise last week. In an article in this newspaper (‘Count caste by all means. But abandon belief that counting settles anything’ IE, August 21), Surjit Bhalla, no dimagi Naxal by any stretch, endorsed the central claims I have been making in these pages on the caste census . Bhalla favours a caste census. And he does so for the right reasons: “Yes. Count caste — India allocates education and employment on caste.” I have been arguing for years that it is a scandal for a country to have legally enforced quotas for a social group — the OBCs — and not count its population. Bhalla also shares my assessment that the caste census is now dead, actually being killed by the government: “Yadav is right that the [caste] Census is now being killed by design.” And he agrees with my analysis of why the Modi government has done so: “His diagnosis of the politics [behind the technical choice of the wording of the caste question] is sort of persuasive.”I could not ask for more. I should have just posted his article on social media, while thanking and tagging him. But there is a technical twist in this tantalising tale. Bhalla wants me to realise that even in the absence of sabotage, the caste census would not have told us “the truth” and advises us to “abandon the belief that counting [castes] settles anything”. This does disservice to his own position and begs a question: Why should we trouble the already burdened Census enumerator with this additional question if its answer does not settle “anything”?AdvertisementMinus such verbal excesses, the main argument in Bhalla’s article is fairly straightforward. Counting castes by using a drop-down menu of categories like SC/ST/OBC is bound to produce unreliable data. If the Census allows every person to choose a caste category, which is linked to his or her entitlements, we are creating a perverse incentive to lie, to claim membership of a category to which the person does not belong. This could lead to a massive exaggeration in the numbers for OBCs. Hence his suggestion: “Ask for caste, sub-caste and surname, never for category.” He lays out a dazzling array of data to make an innocuous point: Human beings respond to incentives. Indeed, no proof was required for this.Sadly, Bhalla chooses a bad example to make an otherwise valid point. He presents the unusual rise in the estimated population of OBCs between the 55th (1999-2000) and 61st (2004-05) rounds of the National Sample Survey as evidence of incentivised manipulation in caste numbers. He correctly shows that such a rise could not have happened due to an increase in population. But he wrongly assumes that this must have been caused by respondents falsely reporting OBC status to claim entitlements. He does not mention that 1999 was the first time OBCs were counted in the NSS and were thus prone to being under-counted. He does not examine a competing and compelling hypothesis: OBC was still a recent category in 1999, and many more OBCs were unfamiliar with it then compared to five years later. Most importantly, while dissecting the data from 1999 and 2004, Bhalla does not care to share the NSS figures for 2009-10, 2014 and 2017-18, which report a fairly stable share for OBCs — between 42 and 43 per cent. If the respondents were over-reporting OBC status, why did this trend halt in 2004?Regardless, let’s stick to his general argument about perverse incentives that might lead to exaggeration of population numbers for the OBCs in the coming Census. That would be a valid apprehension if you do not know something basic about the Census of India, a surprising omission on Bhalla’s part. Section 15 of the Census Act, 1948 guards the privacy of Census responses and says that Census records are not open to public scrutiny and cannot be used as evidence in a civil or criminal court. The Supreme Court has reiterated that no person can use his or her self-declaration in Census responses to claim entitlements, reservation, or any statutory benefits. In sum: There is zero individual incentive to misreport your category in the Census.AdvertisementIf not an individual one, is there a collective incentive for social groups to inflate their numbers in the Census to claim higher community benefits, say, a higher quota for OBCs? Hypothetically, yes. But why would the non-OBCs participate in such a conspiracy if there are no individual gains for misreporting? Besides, history belies such apprehensions: If entitlements incentivise lying in the Census, why has the population share of SCs and STs not leapt up beyond credible population trends in the last seven decennial censuses?Now that the perverse incentive theory is laid to rest, may I invite Bhalla to rethink whatever remains of our difference on this issue. Although he does not spell out his alternative, it seems Bhalla does not favour an open-ended question on caste that simply records any answer offered by the respondent. As a seasoned survey researcher himself, Bhalla would know that it would be an invitation to a data disaster — millions of responses that defy post-coding. I understand his opposition is to a stand-alone caste category question — Are you SC/ST/OBC/General — as this method is vulnerable to large-scale errors. But that is not the proposal from the proponents of the caste census.you may likeWhat we are suggesting is what survey researchers call a two-stage hierarchical question using cascading response categories. This is exactly how the question about SC and ST has always been asked in the census: First you are asked to choose a category, then presented with a list of caste names within that category to choose from. In case you chose the wrong category, the drop-down menu in the next question helps to rectify it there and then. So, our proposal is that the same should be extended this time to the OBCs and the General category. As for Bhalla’s fair insistence that professional sociologists should be involved in classifying castes and sub-castes, I may again point out that this job has already been done by various commissions in the case of OBCs. And by the 43-volume People of India series by the Anthropological Survey of India for the rest. Why reinvent the wheel, if the desire is not to sabotage the caste count?This is not meant to be a rebuttal. I really hope to persuade Surjit Bhalla, whose voice carries weight with the present dispensation, unlike mine.The writer is member, Swaraj India, and national convenor, Bharat Jodo Abhiyaan. Views are personal