Every CEO or MD of the top 50 companies on the National Stock Exchange of India (NIFTY 50) belongs to the upper castes. Every one of the nearly 12,000 manual scavengers are Dalits. Ninety-five per cent of the founders of India’s billion-dollar startups are from the upper castes; 95 per cent of the low-paid, bottom-of-the-rung gig workers in these startups are from the backward castes, Dalits and minorities.
These are facts. But are they mere coincidences? Is such stark disparity across castes an outcome of discrimination? Or is it because of a lack of opportunities and networks for certain castes? Or is it because of poverty conditions at birth and not identity?
The Leader of the Opposition Rahul Gandhi has long believed that these are not mere coincidences and answers can be obtained only through a detailed caste census, which he has termed an “X-ray” of Indian society. Prime Minister Narendra Modi and the RSS have long believed that while these may be India’s ugly truths, it is not worth pondering over and dismissed the idea of a caste census as “divisive” and “regressive”. Until a few days ago.
On April 30, the Modi government did a somersault and announced its intent to conduct a caste census by enumerating caste in the next Census exercise but didn’t specify timelines, process — the when and how of it. Politics aside, what is a caste census, how must it be done and what can it achieve?
The Telangana 2024 Social Educational Employment Economic Caste Survey (SEEEPC), colloquially referred to as “caste census”, was concluded in February 2025. This was its scale: About 3.5 crore (35 million) people across 245 sub-castes with 74 points of information on each. Collected by 1,03,889 enumerators visiting every household. Across 33 districts and 94,261 blocks in two months. The 74 fields of information range from demographic, social, behavioural, discriminatory, educational, occupational, economic, political to caste-related information of every individual or household.
Subsequently, the Telangana government appointed an expert group to study this extensive dataset and submit a report of its findings and interpretations.
The Telangana SEEEPC survey is perhaps the most detailed and rigorous such exercise conducted in India since the Mandal Commission’s field survey in 1980. To put this in context, the Mandal Commission report, which is rightfully hailed as the “gold standard” for caste census, surveyed 50 lakh people out of a population of 35 crore then, collecting 11 points of information on each, which was a remarkable effort for 1980. But the Telangana 2024 exercise is several hundred times larger in scope, size and scale than the Mandal Commission survey. While SEEEPC is called a “survey” for legal reasons, empirically it is a “census” that covered 98 per cent of the population of Telangana.
If the very purpose of embarking on a caste census is to gather insights into social disparities and craft policies based on data and evidence, then the quality and accuracy of the information collected in the exercise are key. The three broad measures used to study Telangana society as part of its SEEEPC exercise are the share of the population, share of participation and share of representation of each social group. Share of participation refers to the share of educational, occupational and economic attainment of each social group. Share of representation refers to the power of social networks of each social group in various facets of society such as politics, government, business, media, judiciary etc.
Contrary to popular belief, a caste census is not just an enumeration of the population by various caste groups. It is a detailed and scientific exercise to collect various information to glean a holistic picture of the living conditions of people. The design of the questionnaire is the most critical aspect of this exercise, but the social, geographical and cultural diversity of India make it a very complex and challenging task. For example, parameters such as inter-caste marriage, which is best understood only by representatives of each caste in each district, can reveal a lot about social mobility, as the Telangana exercise reveals. The questionnaire cannot be scientifically designed without involving citizen groups, sociologists, anthropologists, historians, empiricists and civil society organisations to understand local conditions in every district in the country. This is what Rahul Gandhi meant when he urged the Modi government to conduct a “people census” as was done in Telangana and not a “bureaucratic census”.
Politicians and media commentators often obsess only over the share of the population of each caste group. What is far more important is the level, trend and distance of disparity among various caste groups. If a greater percentage of children of illiterate parents of one caste group have become graduates than those of another caste group in the same category, it indicates greater educational, social and economic mobility, regardless of the share of population of each. Or if one group is doing better economically but is being discriminated against socially and culturally by not being allowed to visit places of worship or not allowed to be in common public places with other caste groups, it calls for greater attention. A well-conducted, rigorous, objective, transparent caste census exercise can reveal such rich, qualitative information about the status and conditions of various caste groups, beyond just their population.
There are those who strongly believe one’s economic conditions at birth are a more important determinant of eventual outcomes than one’s birth caste, such as PM Modi who has said that there is no caste but only poverty. A caste census is precisely the objective way to settle this debate. The Telangana exercise clearly shows that caste is a more significant variable in shaping one’s life outcomes than birth poverty.
There is also a temptation to conflate caste census with reservation. Reservation is a policy tool to help oppressed groups overcome the handicap of lack of social networks and adequate representation. Caste census is a scientific empirical exercise. The two are not the same. The Supreme Court in several judgments has ruled that the mere definition of a caste as backward is not enough to justify further caste-based policies. The SC has consistently held that if rigorous data, empirical evidence and analysis done by experts show extreme backwardness of certain caste groups, then such caste legislation is justifiable.
The Modi government’s sudden and surprising decision to conduct a caste census is a significant effort towards data and evidence-driven policy-making that must be welcomed. The harsh reality is that we are exercised by caste as a nation. The solution may lie in a caste census exercise.
Chakravarty is Convenor of the Government of Telangana’s Expert Group to study the SEEEPC Survey. Views are his own
