When the Telangana government released the findings of its Socio-Economic, Educational, Employment, Political and Caste (SEEEPC) Survey in early 2026, it did something India’s policy establishment has long resisted, to put numbers to the structure of inequality that millions navigate every day. Covering more than 35.4 million individuals across 242 castes, roughly 97% of the state’s population, it is the most granular caste deprivation dataset independent India has ever produced.The survey’s Composite Backwardness Index (CBI), built on 42 multidimensional parameters spanning education, land ownership, housing, sanitation, occupation, debt, and social integration, shows that Scheduled Castes (SC) score 96 and Scheduled Tribes 95 on a scale where higher means more deprived, while upper-caste Other Castes (OCs) score just 31. The average SC or Scheduled Tribe (ST) household is, by the state’s own multidimensional measure, approximately three times as backward as the average upper-caste household.This data triggered the need for national conversation about caste census and reservations. Which communities will demand larger quotas? What happens to the Other Backward Classes (OBC) share? While, these are not unimportant questions, they are the wrong ones to begin with.A census that proves the problem is much biggerThe single most important finding in the SEEEPC Survey is not about quotas. It is about the nature of deprivation itself. The CBI was constructed precisely because single indicators like especially income miss how disadvantage accumulates and persists across domains and generations. What the index reveals, across all 242 castes, is a multidimensional hierarchy that operates in housing, sanitation, schooling quality, land access, labour market position, debt burden, and spatial location simultaneously. Poverty, for the most backward castes, is not just one which no single welfare instrument can address.Consider what the data shows about domains that reservations do not touch. Only 8% of SC households own agricultural land, compared to 31% of OC households. Among daily wage workers, 46% of SCs and 41% of STs survive in the precarious informal labour market, against just 11% of OCs. Private school enrolment stands at 10% for SCs and 8% for STs, compared to 30% for OCs. Social closure through endogamy, which reproduces both advantage and disadvantage across generations, remains near-universal, with inter-caste marriages at roughly 5 to 6%. None of these gaps, in land, in informal labour, in schooling quality, in spatial concentration, can solely be resolved by adjusting the percentage of seats reserved in a central university.The reservation system is not being indicted here. It has been effective, within its scope, at increasing representation in formal public institutions. The problem is that the public debate has allowed reservation to stand in for the entire architecture of caste redressal, only relying as a seat in a government college is an answer to a family living on daily agricultural wages in a rain-dependent district with no pucca home and no toilet. The Telangana survey questions and demands that we expand the frame and rethink the dimensions.The EWS quotaThe SEEEPC Survey also settles, at the level of state-wide empirical evidence, a question that the Supreme Court left open in its 2022 judgment upholding the 103rd Constitutional Amendment, whether the Economically Weaker Section (EWS) reservation has an empirical basis.It does not. The EWS quota was introduced in 2019 without any commissioned study or estimation of economically weaker sections in the upper-caste OC population. The Telangana data now shows that every individual from all 18 OC caste groups in the state scores below the state average in backwardness, meaning, by a multidimensional measure, there are no significantly backward households in the OC category to speak of. The category designed to receive 10% reservation has a CBI of 31, while the category receiving 27% OBC reservation has a CBI of 86, and the category receiving 15% SC reservation has a CBI of 96.The OC community constitutes 11.9% of Telangana’s population and receives reservation nearly proportional to its share. The BC and OBC population constitute 56.4% of the state and receives 27% reservation. No serious account of constitutional, utilitarian, or justice can explain this outcome except as political settlement. A caste census, to be meaningful, must provide the evidentiary basis on which this settlement is publicly renegotiated.Data does not reform itselfThe Mandal Commission submitted its report in 1980. The Sachar Committee reported in 2006. The Ranganath Misra Commission reported in 2007. Each produced careful data on deprivation, discrimination, and underrepresentation. Each was partially implemented, partially shelved, and litigated. The Bihar caste survey of 2023, which revealed that OBCs and EBCs together constitute nearly 63% of the state’s population, produced immediate political turbulence and court challenges before any policy action could follow from it. What we see is that when data emerges, political mobilisation absorbs it, and the institutional machinery required to translate data into redistribution is not built.The Telangana survey is methodologically stronger than any of its predecessors. The SEEEPC’s 42-parameter CBI is the first composite backwardness index of its kind in the country. The Independent Expert Working Group that analysed the data was constituted separately from the government, giving the findings a degree of analytical insulation. But, caste census, without a statutory framework compelling action on its findings, will become the most expensive commission report India has ever shelved.What the world already knows: Three lessons India is yet to learnIndia is not the first democracy to attempt the statistical mapping of structural disadvantage. Three cases from vastly different political contexts offer a sharper mirror. Brazil is the most useful parallel. Brazil carries a history of race-based stratification, and the myth of racial democracy functioned in Brazil much as the myth of a casteless modernity functions in Indian urban professional discourse. Beginning in the early 2000s, Brazil moved systematically from racial enumeration to legally mandated redistribution. The Quota Law of 2012, amended and strengthened in 2023, reserves 50% of federal university places for students who attended public schools, with a further racial sub-quota for Black, Brown, and Indigenous students calibrated to each state’s census demographics.Brazil also created verification commissions to prevent fraud, recognising that without institutional enforcement mechanisms, identity-based entitlements are captured by the least needy. The outcomes being, racial quotas in federal universities raised the share of Black students in ways that income-based quotas alone could not achieve, and quota beneficiaries show substantial long-term labour market gains with no evidence of peer harm. What Brazil demonstrates is not that quotas are sufficient, they are not, and Brazilian scholars readily acknowledge that private-sector transformation remains incomplete and that a National Racial Equity Index has still not been established but that statutory obligation tied to continuously updated demographic data produces outcomes that discretionary political will never will.Malaysia offers the cautionary inverse. Since 1971, Malaysia’s New Economic Policy (NEP) and its successor programmes have provided affirmative action for the Bumiputera majority, a category encompassing Malays and other indigenous groups, through university admissions, civil service hiring, corporate equity ownership targets, and public procurement preferences. Fifty years of race-conscious policy have reduced overall poverty and produced a Bumiputera middle class. But the deeper structure of the programme’s failure is precisely what India must reckon with, the NEP lacked systematic articulation of policy objectives, instruments, and outcomes, and its benefits were disproportionately captured by a politically connected Bumiputera elite while intra-ethnic inequalities widened.Bumiputera graduates, enrolled predominantly in lower-ranked public institutions, continue to rely heavily on the public sector for professional employment, a structural dependency that affirmative action entrenched rather than dissolved. The parallel for India is direct and uncomfortable: a reservation system that does not periodically audit whether its most deprived beneficiaries are receiving its benefits risks becoming, over time, an instrument of intra-community elite consolidation rather than deprivation redressal.The Telangana data showing enormous CBI variation within SC, ST, and BC categories where some sub-groups score near 116 while others cluster near 85 is precisely the kind of evidence that Malaysia never systematically gathered and paid a generational price for ignoring.South Africa’s Employment Equity Act of 1998 provides the third lesson, the bleakest. Enacted with genuine constitutional weight in the aftermath of apartheid, the Act mandated race-disaggregated reporting, equity planning, and representational targets for designated employers. It is, on paper, the closest existing model to what this article argues India needs. In practice, it has been substantially undermined by limited enforcement capacity in the Department of Employment and Labour, corporate compliance that is often superficial targeting visible diversity at entry levels while preserving racial concentration at leadership, and a legal complexity that smaller employers routinely exploit.The South African case demonstrates that a statutory framework is necessary but insufficient without three additional elements: genuine institutional independence from the executive, adequate enforcement resources, and a regular cycle of data-driven target revision. A law without a well-resourced enforcement body is, in practice, a recommendation.What connects Brazil’s partial success, Malaysia’s elite capture, and South Africa’s enforcement deficit is the quality of the institutional architecture between the data and the outcome. In each case, the political will to enumerate was greater than the political will to act. In each case, the absence of a mandatory, independently audited implementation cycle allowed dominant interests to absorb, delay, or dilute redistribution. India is currently positioned now Brazil was in 2002 and South Africa was in 1997, with new data and old institutions. The question is whether it learns from what followed in both those countries and repeats it.What a caste equity budget framework would actually requireThe policy conversation needs to move from enumeration to obligation. Three institutional changes follow directly from what the Telangana data shows.First, India requires a Caste Equity Budget Framework embedded in public expenditure planning. Every major welfare scheme, the Pradhan Mantri Awas Yojana, the National Health Mission, the PM Poshan scheme, Mahatma Gandhi National Rural Employment Guarantee Scheme (MGNREGS), should be required to publish caste-disaggregated data on actual beneficiaries, not just eligibility categories. When a housing scheme’s beneficiaries are 40% upper caste in a district where upper castes constitute 12% of the population, that is a governance failure. The SEEEPC’s finding that Muslim communities rank lowest in accessing government welfare schemes despite high deprivation levels is precisely the kind of outcome that caste-disaggregated expenditure auditing would surface and force accountability for.Second, the affirmative action framework must move from blunt categories to calibrated sub-categorisation. Within STs, the Nakkala community has a CBI of 112 while the Lambadi community has a CBI of 95. Within SCs, the Dakkal community scores 116 while others cluster closer to 85. A reservation system that treats all STs as interchangeable is not targeting. Periodic caste deprivation audits, based on census data, should trigger mandatory sub-categorisation reviews every decade, with independent assessment of whether the communities most in need are actually receiving the largest share of institutional benefits.Third, urban planning must incorporate caste deprivation mapping as a mandatory input. The SEEEPC data shows that upper castes disproportionately capture the benefits of city growth, while SC and ST households remain concentrated in urban informal settlements with amenities barely superior to rural slums. Urbanisation does not dissolve caste hierarchy but relocates it to new sites of exclusion. Master plans for tier-one and tier-two cities should be required to identify caste deprivation clusters and allocate infrastructure investment, schools, water, roads, health centres on the basis of that mapping, not purely on population density or land-use classifications that obscure who lives where.Counting caste is necessary, but it is not sufficient.The Telangana survey has demonstrated that such an exercise is administratively feasible, analytically rigorous, and indispensable. But data alone will not do anything. The same political economy that resisted the caste census for decades will resist acting on its findings. Dominant communities that feared the census revealed their advantage will find other instruments such as judicial challenge, administrative delay, definitional disputes to protect that advantage once the numbers are published.The measure of a caste census is not the quality of its methodology. It is what the state is legally compelled to do with the results. Until that obligation is created, until a statutory body with real independence is empowered to convert deprivation scores into time-bound, audited redistribution mandates, with published compliance scorecards for every ministry and state government, the caste census will only act as an X-ray but unacted upon. At present, what the country needs is the institutional architecture to make that evidence consequential.Boddu Srujana is an Assistant Professor with Department of Economics, Easwari School of Liberal Arts, SRM University-Andhra Pradesh. She posts on X at @SrujanaBoddu