CASTE, CENSUS, AND THE BIRTH OF ‘DATA STATE’
Anand Teltumbde
The BJP’s sudden embrace of the caste census marks one of the most striking political reversals in recent Indian history. For over a decade, the party had vehemently opposed caste enumeration, dismissing it as “vote-bank politics”, a “divisive agenda”, and an “urban naxal idea”. It refused to release the Socio-Economic and Caste Census of 2011 and described the data before the Supreme Court as “flawed”. The ideological rationale seemed clear: Hindutva sought to subsume caste differences within a larger, unified Hindu identity, while a caste census threatened to expose the internal fractures that this project sought to transcend.
Yet, within months of the 2024 Lok Sabha election-which delivered the BJP a diminished mandate-the same government announced that caste would be included in the forthcoming census. Many observers welcomed this as a victory for democratic pressure or as a long-overdue commitment to social justice. Both explanations are inadequate. Political parties rarely abandon entrenched ideological positions without compelling reasons. The real question is not why the BJP accepted the caste census but what political purpose it now expects the exercise to serve.
The answer lies in the shifting arithmetic of Indian politics. Since the Mandal era, the BJP had neutralised caste by subsuming it within Hindu consolidation. However, the 2024 election exposed the limits of that strategy. In key States, particularly Uttar Pradesh, caste re-emerged as the dominant language of political mobilisation.
Congress leader Rahul Gandhi’s persistent advocacy for a caste census successfully reframed enumeration from a statist exercise into a legitimate democratic demand concerning representation and resource distribution. Continued opposition became electorally costlier than acceptance. But electoral compulsion explains the “change” in position; it does not explain how the state intends to use the census. To understand that, we must examine the census as a technology of governance.
A census is not a neutral statistical mirror. It is a primary instrument through which the modern state renders society legible. To count is to classify; to classify is to define; and to define is to acquire the capacity to administer. The history of the census in India illustrates this with clarity.
History of counting caste
Colonial administrators did not invent caste, but they transformed fluid, localised social formations into standardised, all-India administrative categories. This classification had profound consequences: it simultaneously froze previously negotiable identities and politicised them, as communities began organising around official labels to secure state patronage. Enumeration, thus, became a mechanism not merely for recording reality but for reorganising it.
Independent India inherited this colonial apparatus but redirected it towards democratic redistribution. Issues such as reservation, enumeration of Scheduled Castes (SCs) and Scheduled Tribes, and the Mandal framework all depended on the state’s capacity to identify historically disadvantaged communities. Enumeration acquired an emancipatory purpose.
The Constitution institutionalised caste as an indispensable administrative category. The state could neither abolish caste nor ignore it; it was forced to govern through it. This system can be understood as “constitutional caste”-a framework where caste derives its significance not from scripture alone but from law, policy, quotas, and welfare entitlements.
For decades, this paradox remained politically manageable. The Mandal Commission transformed it by turning caste into the principal axis of democratic mobilisation. The BJP’s rise was a direct response to this landscape.
Hindutva did not deny caste; it sought to politically subordinate it to a larger religious identity. The Ayodhya movement encouraged Hindus to view religious allegiance as more significant than internal caste divisions. Consequently, the BJP’s long-standing opposition to a caste census reflected a strategic calculation: a comprehensive enumeration threatened to revive questions of internal inequality and privileged-caste dominance that Hindutva had sought to marginalise.
Transition to “data state”
The 2024 election altered that calculation. Yet, to view the BJP’s reversal merely as a concession to opposition pressure is to miss the deeper transformation occurring within the state itself. We are witnessing the transition from the constitutional state to the “data state”. The 21st century state no longer governs merely through law or bureaucracy; it increasingly governs through information. Aadhaar, digital welfare platforms, electoral databases, and artificial intelligence have created capacities unimaginable to colonial administrators or the framers of the Constitution.
In this new context, census data no longer merely describes society; it can be integrated with multiple databases to classify, target, and administer populations with unprecedented precision. Detailed caste data permits not only more effective welfare planning but also far more sophisticated political management. Smaller communities can be separately identified, targeted with specific benefits, encouraged to seek subcategorisation, or cultivated as distinct electoral constituencies. The census ceases to be a mirror of social reality and becomes an instrument for actively managing social differentiation itself.
This is the deeper significance of the BJP’s reversal. It does not necessarily signal an abandonment of Hindutva; rather, it may mark its adaptation to the age of the data state. If the earlier phase of Hindutva sought to govern despite caste by subsuming it within Hindu unity, the next phase may seek to govern through caste by measuring, classifying, and politically managing its internal diversity.
The objective is not to weaken reservation but to prevent broad Bahujan solidarities from re-emerging as an independent political force. The administrative precision of the data state allows the ruling party to fragment large constitutional categories-like Other Backward Classes (OBCs) or Dalits-into smaller, manageable vote banks, thereby undermining their collective political weight.
Consider the demands for subcategorisation within SCs and OBCs. These demands often arise from genuine grievances about the unequal distribution of reservation benefits. However, under the data state, the same information that enables equitable subcategorisation can also be used to reorganise political constituencies.
The instruments of social justice and the instruments of electoral strategy are identical; their purposes are not. Social justice requires recognising internal inequalities to distribute resources fairly. Electoral strategy exploits those same inequalities to fragment political solidarity. Data possesses no intrinsic politics; its democratic or anti-democratic character depends entirely on the purposes it serves.
This possibility is particularly relevant to the evolution of Hindutva. The BJP’s success has rested on subsuming caste divisions within a larger Hindu identity. The emerging strategy is more sophisticated: preserving Hindu unity at the ideological level while politically managing caste differences through differentiated welfare, representation, and electoral targeting. The state acquires the capacity to identify, reward, and cultivate specific sub-castes, effectively atomising the broader political coalitions that challenge its dominance.
There is a striking historical irony here. The colonial census transformed localised jatis into administrative communities. The constitutional state converted many of those communities into categories for democratic redistribution. The data state may now fragment those constitutional categories into increasingly granular electoral constituencies, or datafied castes. Enumeration continues to reshape caste, not by abolishing it but by repeatedly reorganising its political expression to suit the governing logic of the age.
This analysis should caution both supporters and opponents of the caste census. For supporters, it is crucial to recognise that enumeration is not an end in itself. If administrative recognition instead becomes a means for new forms of political fragmentation and managerial control, the emancipatory promise of constitutional policy stands fundamentally compromised. Conversely, for opponents, refusing to count caste does not make caste disappear; it only makes castes so much more indestructible, if destruction of caste is what they aim at.
The challenge, therefore, is not simply to conduct a census but to ensure that the knowledge it generates serves democratic equality rather than merely enhancing the state’s capacity to manage social divisions.
The census does not speak; governments speak through it. The information it produces will be deployed according to the political logic of the regime in power. If that logic is oriented towards deepening representation and addressing historical injustice, the census can be a progressive tool. If it is oriented towards fragmentation and electoral management, it will become an instrument of a new kind of authoritarian governance-one that governs not through force but through data. History cautions us that the latter is more likely.
Caste and the state
In conclusion, the history of the caste census is ultimately not a history of caste itself but of the changing relationship between caste and the state. The colonial state used enumeration to administer an empire. The postcolonial constitutional state used it too for a similar purpose in the name of social justice. The emerging data state threatens to use it to perfect the political management of social divisions.
To the extent the caste census enables subcategorisation, it will certainly serve the BJP’s ideological agenda of abolishing caste-based reservation in favour of economic conditions, as it has already seeded in the form of EWS (economically weaker section) reservation. (I have explained this politics in my book The Caste Con Census.) The debate over the caste census is, therefore, framed too narrowly. The real question is no longer whether caste should be counted. It is who counts, for what purpose, and in whose political interest. If the colonial state governed through enumeration and the constitutional state governed through recognition, the emerging data state may govern through information itself.
Whether that transition deepens democracy or merely refines the management of inequality will depend not on the census alone but on the capacity of democratic institutions and social movements to ensure that enumeration remains subordinate to B.R. Ambedkar’s larger vision: the annihilation of caste and the deepening of substantive democracy. The struggle, ultimately, is not over numbers; it is over the political uses to which those numbers are put.
Anand Teltumbde is a former CEO of Petronet India Ltd, a professor at IIT Kharagpur and Goa Institute of Management, a civil rights activist, and the author of over 30 books.