OBC Politics and the Politics of Backwardness

The Other Backward Classes are the only major constitutional category in India whose defining term the Constitution never defines. “Socially and educationally backward classes” appears in the text, is nowhere explained in it, and has been filled in for seventy-five years by commissions, executive orders and — decisively — by the Supreme Court. That absence is not an oversight. It is what has made backwardness the most contested political category in the republic, because the claim to it is worth something, and the state must decide who may make it.

The Category Without a Definition

The constitutional vocabulary

• The Constitution supplies the entitlement and withholds the criterion, and every subsequent controversy grows in that gap.
  • Article 15(4), inserted by the First Amendment (1951), allows the state to make special provision for the advancement of “any socially and educationally backward classes of citizens” and for the Scheduled Castes and Scheduled Tribes.
  • Article 16(4) permits reservation in appointments for any “backward class of citizens” that in the state’s opinion is not adequately represented in the services — a differently worded and separately litigated test.
  • Article 340 empowers the President to appoint a commission to investigate the conditions of socially and educationally backward classes and recommend steps to remove their difficulties. Both the Kalelkar and Mandal Commissions were constituted under it.
  • Article 342A, added by the 102nd Amendment (2018), and Article 366(26C) finally gave SEBCs a place in the constitutional text — but as a list-making power, not as a definition.
• Note the asymmetry that runs through the whole subject: Scheduled Castes and Scheduled Tribes are identified by presidential notification under Articles 341 and 342, a closed and published list. Backward classes are identified by a test that no instrument states.
• The Constituent Assembly deliberately left the term open. Ambedkar, defending Article 16(4), described “backward classes” as a compromise between the claim of equal opportunity and the claim of communities that had been shut out of the administration — and declined to fix its content, leaving it to be worked out by the state.

Backwardness has no constitutional definition. It has only a constitutional history, written case by case.

The varna inheritance and what it does not settle

• In the classical scheme the OBCs occupy the position of the Shudra varna — inside the ritual order, unlike the untouchable castes placed outside it, but below the twice-born.
  • This is the source of a real disability: exclusion from learning, from land ownership in many regions, and from ritual authority.
  • It is also the source of the category’s essential ambiguity. Shudra status implied subordination, not exclusion, and the castes that carried it ranged from landless artisans to substantial peasant proprietors.
• The consequence is that OBC status has never mapped cleanly onto deprivation. A caste can be ritually low and economically dominant in the same district — precisely the combination Srinivas described as the dominant caste, treated at length in the companion note on caste in Indian politics.
• Muslim, Christian and Sikh communities also appear on backward-class lists where their traditional occupation places them in an equivalent position, which is why the category is not, strictly, a Hindu one.

Galanter and the loose conception

• Marc Galanter, in Competing Equalities: Law and the Backward Classes in India, describes the OBC not as a group but as a loose and shifting conception — one that changes with the administrative purpose for which it is being drawn.
  • Unlike the SC and ST schedules, the backward-class list is compiled separately by the Union and by each state, on different criteria, for different purposes, and revised at different times.
  • The consequence is that a caste can be backward in one state and forward in the next, backward for state employment and forward for central employment, and backward today because it was backward in a survey conducted decades ago.
• Galanter’s larger argument is that India’s compensatory discrimination is administered through group membership while being justified by individual disadvantage, and that the two never quite align. The creamy layer, sub-categorisation and the caste-census demand are all attempts to close that gap.

An internally unequal category

• The OBC label covers peasants, artisans, service castes, herders, fishers, oil-pressers, weavers, potters and toddy-tappers — occupational groups with almost nothing in common except a shared position below the twice-born and above the untouchable.
• Within it sits a stratum of dominant OBCs: land-controlling, numerically concentrated, politically organised castes that command state politics in their regions.

Stratum	Representative castes	Political characteristics
Dominant OBCs	Yadavs (UP, Bihar), Kurmis, Jats, Gujjars, Reddys, Vokkaligas, Patidars, Marathas, Lingayats, Kammas	Landowning, numerically concentrated, supply chief ministers and party leaderships, capture the bulk of reserved posts
Intermediate OBCs	Koeris, Nais, Kumhars, Telis, Ahirs outside their core belts	Some land, weak organisation, electoral swing groups rather than power centres
Extremely Backward Classes	Mallahs, Nishads, Dhanuks, Bind, Kahars, Musahar-adjacent service castes	Landless or near-landless, low literacy, almost no share of the reserved quota, mobilised as coalition partners rather than principals

• The political corollary is that the OBC category has an internal upper class of its own, and the whole of the sub-categorisation debate follows from it. The Rohini Commission later put a number on the imbalance.
• Several of the dominant castes are, in Srinivas’s terms, dominant castes rather than backward ones — which is why their claim to backwardness has become one of the most fraught questions in Indian politics.

The contested arithmetic

• No one knows how many OBCs there are, and this single fact shapes the entire field.
  • The Mandal Commission estimated the OBC share at 52 per cent, extrapolating from the 1931 census — the last enumeration to record caste comprehensively — after deducting SCs, STs and the communities it classed as forward.
  • The National Sample Survey Organisation put the figure nearer 41 per cent in its 2006 rounds, and later NSS and health-survey estimates have hovered in the 40-44 per cent range.
  • The central OBC list, which contained 2,399 communities on Kalelkar’s count and 3,743 on Mandal’s, now runs past 5,000 entries as states have added castes — growth that reflects successful lobbying at least as much as discovered deprivation.
• The number matters because 27 per cent was derived from it. Mandal’s recommendation was not proportional to 52 per cent; it was the largest figure that could be added to the existing 22.5 per cent for SCs and STs without crossing the 50 per cent line the Supreme Court had drawn.
• Satish Deshpande argues that OBC politics is the heart of Indian politics — that the OBCs are the single largest pan-Indian political bloc, that no national alliance can be built without them, and that state politics in most of India simply is OBC politics.

“Intermediate groups like the OBC are in the thick of caste — what it does to them, and what they choose to do with it, will have a major impact on when and how the republican goal of transcending caste is reached.”

• Deshpande’s point is sharper than a statement of numbers. Because the OBCs are neither the beneficiaries of the ritual order nor its clearest victims, whether caste hardens or dissolves will be settled by what they choose to do with it.

Naming the Backward: The Commissions

The First Backward Classes Commission

• The Kalelkar Commission was appointed by presidential order on 29 January 1953 under Article 340, chaired by Kaka Kalelkar, and reported in 1955.
  • It used four broad tests: low position in the traditional caste hierarchy, lack of general educational advancement, inadequate representation in government service, and inadequate representation in trade, commerce and industry.
  • It identified 2,399 backward castes, of which 837 were classed as “most backward” — the first official recognition that the category contained a still worse-off core.
  • It recommended 70 per cent of seats in technical and professional institutions for backward-class students and 25 to 40 per cent reservation across classes of government service.
• The report was shelved, and the reason is the most quoted fact about it: the chairman disowned his own commission’s method in the covering letter transmitting it.

“I am definitely against Reservation in Government services for any community for the simple reason, that services are not meant for the servants, but they are meant for the service of society as a whole.”

• Kalelkar argued in that letter that caste-based identification would perpetuate the very distinctions the Constitution sought to erase, and urged that backwardness be assessed on other criteria. Two members recorded formal dissents against the caste basis.
  • The Union government seized on the disagreement. Its memorandum to Parliament held that the Commission had applied no objective test, and it declined to act on the report.
  • The states were left to draw their own lists, which is why state backward-class lists predate the central one by three decades and why the two still diverge.
• Nehru’s own scepticism reinforced the shelving. He held that caste-based quotas in services would entrench caste and depress standards, and preferred general educational and economic uplift — a position that lost decisively in 1990 but has never disappeared from the argument.

The Second Backward Classes Commission

• The Mandal Commission was constituted on 1 January 1979 by the Janata Party government under B.P. Mandal, a former Bihar chief minister, and submitted its report in December 1980.
• Its central methodological achievement was to convert backwardness into a score. Eleven indicators were grouped under three heads and weighted, giving a maximum of 22 points; a caste scoring above 11 was classified as backward.

Head	Indicators	Weight each	Head total
Social	Considered backward by others; dependence on manual labour; early marriage above the state average; women’s participation in work above the state average	3 points	12
Educational	Children aged 5-15 never attending school; school drop-outs in that age group; matriculates below the state average	2 points	6
Economic	Family assets below the state average; kutcha houses; distance from a drinking-water source; households taking consumption loans	1 point	4

• The weighting is the argument. Social indicators carried three times the weight of economic ones, which encoded the Commission’s position that backwardness in India is a social fact expressed economically, not an economic fact that happens to follow caste lines.
• Its findings and recommendations:
  • 3,743 castes and communities identified as OBC, constituting 52 per cent of the population.
  • 27 per cent reservation in central government services and public undertakings, and in central educational institutions — set, as noted, by the arithmetic of the 50 per cent ceiling rather than by proportionality.
  • Reservation to apply to promotions as well, with carry-forward of unfilled vacancies for three years.
  • Land reform, financial assistance for traditional occupations, separate vocational institutions and residential schools for backward-class children — the non-quota recommendations, almost entirely ignored.
• The Commission is criticised on its data as sharply as it is defended on its purpose.
  • Its caste proportions rest on the 1931 census, half a century stale when it reported.
  • Its own survey covered two villages and one urban block per district, a sample too thin to support caste-level conclusions.
  • It made no allowance for internal differentiation within the category it created — the omission that produced the creamy layer and the Rohini Commission.

The announcement of 7 August 1990

• V.P. Singh announced in Parliament on 7 August 1990 that the government would implement the Mandal recommendation on reservation in central services, a decade after the report had been filed and left unopened by four governments.
• The politics were transparent and are not diminished by being so.
  • The National Front government was fracturing; Singh was losing ground to Devi Lal within his own party and needed a constituency that was not his rival’s.
  • Reservation delivered one at a stroke: an announcement that cost the exchequer nothing, required no legislation, and created a national OBC constituency where none had existed.
  • It also split the Hindu vote along caste lines at the exact moment the Ram Janmabhoomi movement was attempting to consolidate it — which is why the Mandal announcement and the Rath Yatra are best read as a single sequence.
• Jaffrelot’s formulation of Mandal against Kamandal captures the decade: two competing solvents applied to the same electorate, one dissolving Hindus into castes, the other dissolving castes into Hindus.

The anti-Mandal agitation

• The announcement produced the largest student agitation since 1974, concentrated in north Indian cities and in urban upper-caste households.
  • Around 200 attempted self-immolations and over 60 deaths were recorded; the attempt by the Delhi student Rajiv Goswami in September 1990 became the agitation’s defining image.
  • Protest took the form of bandhs, road and rail blockades, and the closure of universities across the north for months.
• The agitation’s composition is analytically important. It was urban, educated and upper-caste, drawn from precisely the stratum whose expectation of secure government employment the quota threatened — and it was almost absent in the south, where OBC reservation had been in place for decades and had ceased to be news.
• Its arguments — merit, efficiency, the fear that caste would be perpetuated by the remedy — became the standing vocabulary of anti-reservation politics and reappear, largely unchanged, in every subsequent controversy.
• The government fell in November 1990 for a different reason, but the agitation had already made the announcement the defining act of the ministry. The quota itself survived: implementation was stayed by the Supreme Court, litigated for two years, and then upheld.

Central lists, state lists and the machinery of naming

• Two parallel systems of classification operate, and their divergence is a permanent source of litigation.
  • The central list governs Union employment and central educational institutions; the state lists govern state services and state institutions. A caste on one need not be on the other.
  • State backward-class commissions, most of them predating the central one, are statutory bodies that recommend inclusion and exclusion. Karnataka’s Havanur, Venkataswamy and Chinnappa Reddy commissions, Tamil Nadu’s Sattanathan commission and Bihar’s Mungeri Lal commission each produced classifications that shaped state politics for a generation.
  • The Mungeri Lal Commission‘s division of Bihar’s backward classes into Annexure I (extremely backward) and Annexure II (backward) is the origin of the EBC category that has since organised Bihar politics.
• The National Commission for Backward Classes was created as a statutory body in 1993 on the Supreme Court’s direction in Indra Sawhney, and given constitutional status by the 102nd Amendment under Article 338B.
• In Ram Singh v. Union of India (2015) the Court quashed the notification adding Jats to the central OBC list in nine states, holding that the NCBC’s contrary advice is binding on the government “for strong and compelling reasons” alone, and that the Cabinet had overridden it on stale material.

“New practices, methods and yardsticks have to be continuously evolved moving away from caste centric definition.”

• That sentence, and the judgment’s reference to transgender persons as an emerging backward class following NALSA, mark the Court’s attempt to loosen backwardness from caste altogether — an attempt that has so far had far more rhetorical than administrative effect.

The Jurisprudence of Backwardness

The case law is not a commentary on OBC politics; it is the mechanism by which the category acquired content. Every element that now defines the OBC — the 50 per cent ceiling, caste as a permissible marker, the creamy layer, the requirement of empirical data — was invented by a court, not by a legislature. This section follows that construction in sequence.

The first four decades

• State of Madras v. Champakam Dorairajan (1951) struck down the Madras Communal Government Order, which reserved medical and engineering seats by community, as a violation of Article 29(2) and the equality guarantee.
  • The Court held that the Directive Principles could not override a Fundamental Right, and that a classification by caste or community in admissions was impermissible however benevolent its object.
  • Parliament reversed it within months. The First Amendment (1951) inserted Article 15(4), and the constitutional basis of OBC reservation dates from that reversal — the first amendment to the Constitution was an amendment about backwardness.
• M.R. Balaji v. State of Mysore (1963) is the source of two propositions that have governed the field ever since.
  • The Mysore order reserving 68 per cent of professional seats was struck down as “a fraud on the Constitution“; the Court laid down that reservation under Article 15(4) should ordinarily not exceed 50 per cent.
  • It held that Article 15(4) speaks of classes, not castes, and that caste cannot be the sole test of backwardness — social and educational backwardness must be shown, with economic condition and occupation as relevant considerations.
  • It also rejected the Mysore government’s division of backward classes into “backward” and “more backward” — a holding that stood for four decades and was steadily eroded thereafter.
• T. Devadasan v. Union of India (1964) struck down the carry-forward rule as applied, because unfilled reserved vacancies accumulated until reservation in a single year reached 65 per cent.
  • The majority treated the 50 per cent line as a limit on the effect of the rule in any given year, not merely on the rule in the abstract.
  • Justice Subba Rao dissented, arguing that Article 16(4) was a substantive guarantee to backward classes and not an exception to be narrowly confined — the dissent that Indra Sawhney would substantially vindicate.
• R. Chitralekha v. State of Mysore (1964) upheld a classification of backward classes made on occupation and income without reference to caste at all, holding that caste is “only a relevant circumstance” and that excluding it does not vitiate a classification that otherwise satisfies the test.
• P. Rajendran v. State of Madras (1968) approached from the opposite direction and held that a caste is also a class of citizens; if a caste as a whole is socially and educationally backward, reservation for that caste is permissible under Article 15(4).
• Read together, Chitralekha and Rajendran leave the law in an unresolved posture: caste is neither necessary nor forbidden as a marker of backwardness. Every later judgment works within that space.
• K.C. Vasanth Kumar v. State of Karnataka (1985) produced five separate opinions from five judges and no ratio, which is precisely what makes it useful.
  • Justice Chinnappa Reddy held that in Indian conditions caste is the relevant unit of backwardness and that demanding a caste-free test amounts to denying the remedy.
  • Justice Desai pressed the opposite case for a means test applied to individuals across communities.
  • Several opinions urged a periodic review of backward-class lists and the exclusion of the advanced sections within them — the creamy layer in embryo, seven years before it was mandated.
  • The Court also suggested that lists be revisited every five years and that reservation be reviewed after a generation, a suggestion no government has honoured.

Case	Year	The proposition it added
Champakam Dorairajan	1951	Communal quotas invalid; prompted Article 15(4)
Balaji	1963	50 per cent ceiling; caste not the sole test; no “more backward” sub-division
Devadasan	1964	Carry-forward cannot breach 50 per cent in any year
Chitralekha	1964	Backwardness may be identified without caste
Rajendran	1968	A caste may itself be a class
Vasanth Kumar	1985	Anticipates the means test, periodic review and exclusion of the advanced
Indra Sawhney	1992	Caste as starting point; creamy layer; ceiling affirmed; no promotion quota

Indra Sawhney (1992)

• The challenge was to two office memoranda.
  • The V.P. Singh government’s OM of 13 August 1990 implemented the 27 per cent quota in central services.
  • The Narasimha Rao government’s OM of 25 September 1991 added an economic preference within the 27 per cent and a further 10 per cent for the economically backward among the forward castes.
• A nine-judge bench decided the case on 16 November 1992. The 6:3 majority upheld the 27 per cent quota; Justice B.P. Jeevan Reddy wrote the leading opinion.
• Caste as the starting point of class. The majority held that “class” in Article 16(4) means a social class, and that in Indian conditions a caste is normally identifiable as one.

“A caste is nothing but a social class — a socially homogeneous class. It is also an occupational grouping, with this difference that its membership is hereditary.”

• The holding is finely balanced and is regularly misstated. The Court did not say caste equals class; it said a caste may be taken as the unit whose backwardness is then tested, so that caste is the starting point of the enquiry and not its conclusion.
• The other holdings, each of which has generated its own body of law:
  • Creamy layer. The socially advanced members of a backward class must be excluded, because once a person has escaped backwardness the constitutional justification for the preference lapses. Governments were directed to frame criteria within four months. The Court expressly confined this to OBCs and said it had no application to SCs and STs.
  • The 50 per cent ceiling was affirmed as a rule flowing from Article 16(1), breachable only in “extraordinary situations” — the example given being remote populations far outside the national mainstream. The exception has been invoked constantly and allowed almost never.
  • No reservation in promotions. Article 16(4) covers appointment, not promotion; existing promotion quotas were given five years to run out.
  • Carry-forward was permitted, but total reservation in any year may not cross 50 per cent.
  • The 10 per cent quota on purely economic criteria was struck down: Article 16(4) is about social backwardness, and poverty alone does not create the disability the clause addresses. Thirty years later the 103rd Amendment did by amendment what the Court had held could not be done by executive order.
  • Article 16(4A) territory left open: certain defence, research and specialist posts were held unsuitable for reservation, and Article 335 efficiency was affirmed as a relevant consideration.
  • A permanent statutory body was directed to be constituted to examine complaints of over-inclusion and under-inclusion in backward-class lists — the direction that produced the NCBC Act, 1993.
• The decision’s structural significance is that it converted reservation from a political concession into a constitutional regime with rules. Both the pro- and anti-reservation camps have argued inside its framework ever since.
• It is also a judgment of two halves. The Court gave the OBCs their quota and, in the same breath, handed every future government the instruments with which to limit it — the ceiling, the creamy layer, the bar on promotions.

The legislative answer on promotions

• Parliament responded to the promotion holding with four amendments in six years, each reversing a specific judicial proposition.

Amendment	Year	What it did
77th	1995	Inserted Article 16(4A) enabling reservation in promotion for SCs and STs
81st	2000	Inserted Article 16(4B): backlog vacancies treated as a separate class, outside the 50 per cent ceiling of the year they are filled
82nd	2000	Added a proviso to Article 335 permitting relaxation of qualifying marks for SC/ST promotion
85th	2001	Amended 16(4A) to add consequential seniority, retrospectively from 17 June 1995, overriding the catch-up rule

• Note what the sequence does and does not do: promotion quotas were restored for Scheduled Castes and Tribes only. The OBCs have never had reservation in promotion, and the demand for it is a live one.
• M. Nagaraj v. Union of India (2006) upheld all four amendments as consistent with the basic structure, but attached three conditions to any promotion quota: the state must show quantifiable data on backwardness, inadequacy of representation in the grade concerned, and that administrative efficiency under Article 335 is not compromised.
• Jarnail Singh v. Lachhmi Narain Gupta (2018), five judges, Justice Nariman writing, made two changes.
  • It struck down the requirement of quantifiable data on the backwardness of SCs and STs, holding that their backwardness is presumed once they are on the presidential list and that requiring proof of it contradicts Indra Sawhney.
  • It extended the creamy layer to SCs and STs in promotions, on the reasoning that the principle is one of equality rather than a rule peculiar to OBCs.
• The pairing is worth stating plainly, because it is the point at which the two categories converge: the Court removed the obligation to prove SC/ST backwardness and simultaneously imposed the obligation to exclude the advanced among them.

The ceiling breached: EWS and the 103rd Amendment

• The 103rd Amendment (2019) inserted Articles 15(6) and 16(6), creating a 10 per cent quota for economically weaker sections and expressly excluding those already covered by SC, ST and OBC reservation.
• Janhit Abhiyan v. Union of India (2022) upheld it 3:2. Justices Maheshwari, Trivedi and Pardiwala formed the majority; Chief Justice Lalit and Justice Bhat dissented.
  • The majority held that reservation on purely economic criteria is permissible, that the 50 per cent ceiling is not inviolable and in any case applies only to the Article 15(4)/16(4) categories, and that excluding SCs, STs and OBCs from EWS was reasonable because they already have a quota.
  • The dissent held that the exclusion of the poor among SCs, STs and OBCs — roughly four-fifths of the country’s poor — makes the amendment a scheme of exclusion rather than inclusion, and that breaching the ceiling opens it permanently.
• The political reading is unavoidable. The ceiling that had defeated every backward-class claim for three decades was crossed on the first occasion the beneficiary was not a backward class, and that fact has since been argued in every state seeking to exceed 50 per cent.

Who may identify the backward: the 102nd and 105th Amendments

• The 102nd Amendment (2018) gave the NCBC constitutional status under Article 338B, and inserted Article 342A providing for a central list of SEBCs notified by the President, alterable only by Parliament.
• In Jaishri Laxmanrao Patil v. Chief Minister (2021) — the Maratha reservation judgment — a five-judge bench unanimously struck down Maharashtra’s 16 per cent Maratha quota for breaching the 50 per cent ceiling without extraordinary circumstances, and declined to refer Indra Sawhney for reconsideration.
  • By 3:2 it further held that after the 102nd Amendment only the President may identify SEBCs, and states may merely make recommendations — an interpretation that stripped every state of a power it had exercised since 1950.
• Parliament reversed this within four months. The 105th Amendment (2021) clarified that states and union territories may prepare and maintain their own SEBC lists for their own purposes, restoring the federal position.
• The episode is a compressed illustration of the whole subject: a constitutional amendment produced an unintended centralisation, a court enforced the text, and Parliament amended again within months because no state government could survive the consequence.

The Creamy Layer

• The creamy layer is the rule that the socially advanced members of a backward class are excluded from its reservation benefit, on the ground that they have ceased to share the disability that justifies the preference.
• It originates in Indra Sawhney, where the majority reasoned that leaving the advanced within the group means the benefit is captured by those who need it least, and the class as a whole is never lifted. The term itself had been used earlier in argument and in Vasanth Kumar, but 1992 made it mandatory.

How the exclusion is actually drawn

• The Justice Ram Nandan Prasad Committee, appointed to give the direction content, reported in 1993 and its scheme was accepted almost unaltered. It excludes on three distinct grounds, and only one of them is income.

Ground	Test
Constitutional position	Children of the President, judges of the Supreme Court and High Courts, holders of constitutional posts, chairpersons and members of constitutional bodies
Service	Children of Group A / Class I officers, of parents both in Group B, and of equivalent ranks in the armed forces and public sector undertakings
Property and income	Holders of irrigated land above a threshold, and those whose parental income exceeds the ceiling for three consecutive years

• The income ceiling has been revised five times — from ₹1 lakh (1993) to ₹2.5 lakh (2004), ₹4.5 lakh (2008), ₹6 lakh (2013) and ₹8 lakh (September 2017).
• Salary and agricultural income are excluded from the computation, which is the source of most of the litigation: the rule is aimed at inherited advantage, not at earned wages.
• The ceiling has not been revised since 2017, against the Ram Nandan Committee’s own recommendation of revision every three years.
  • A proposal to raise it to ₹12 lakh was prepared in 2020 and shelved.
  • The government told Parliament in 2024 that no proposal was under consideration; a parliamentary standing committee has recommended the revision more than once.
• Nine years of unrevised nominal income during a period of substantial inflation is itself a policy: it silently expands the excluded band each year without any decision being taken or defended.

The administrative distortion

• In a judgment of 11 March 2026, a bench of Justices P.S. Narasimha and R. Mahadevan held that creamy-layer status cannot be decided on parental salary alone in the computation for OBC candidates.
  • The judgment is an unusually direct statement that the creamy layer has been administered so as to widen exclusion, and that the way the income test is applied was itself the mechanism.
• The pattern it exposes is general: the creamy layer is defined by executive instruction, and executive instruction has drifted steadily in the direction of shrinking the eligible pool without any of the drift being legislated.

The argument about what the creamy layer is for

• The case for. Without it, the quota is captured by the dominant OBCs’ second generation — children of officers and landholders competing against first-generation aspirants from landless service castes. The Rohini findings are the empirical vindication of the concern.
• The case against, made most forcefully by scholars of caste discrimination, is that the creamy layer converts a group remedy into an individual means test, and thereby misdescribes the injury.
  • Reservation exists because of disabilities attaching to birth, which income does not remove: a well-paid professional from a backward caste does not thereby acquire the networks, the school connections or the social capital that the reservation was meant to substitute for.
  • The exclusion applies before the group achieves proportionate representation, so it caps the very stratum most likely to sustain and mentor the next cohort.
  • There is also an asymmetry of enforcement: the creamy layer is policed rigorously through certificate verification, while the under-representation it is supposed to correct is measured almost not at all.
• Four judges in State of Punjab v. Davinder Singh (2024) extended the logic to Scheduled Castes and Tribes, holding that a creamy layer should apply to them too, on criteria different from the OBC test.
  • This fell outside the reference and is not the holding of the case; it is widely read as obiter, a signal of judicial direction rather than binding law.
  • No government has framed criteria under it, and none is obliged to.

The creamy layer asks whether an individual has escaped. Reservation was designed to ask whether a group has been let in. The two questions have never been reconciled.

Sub-categorisation: Inequality Inside the Category

• Sub-categorisation is the proposal to divide the 27 per cent among strata of OBCs, so that the extremely backward are not made to compete with the dominant.
• It is not a new idea — Kalelkar’s 837 “most backward” castes already contained it, and Balaji had rejected it in 1963 — but the evidence for it is new.

The Rohini Commission

• The Rohini Commission was constituted in October 2017 under Article 340, chaired by Justice G. Rohini, and submitted its report on 31 July 2023 after fourteen extensions.
• Its findings, drawn from 1.3 lakh central appointments and higher-education admissions over five years, are the sharpest quantitative statement of the problem.
  • 97 per cent of the reserved benefit went to roughly 25 per cent of the OBC communities.
  • Around 983 communities — nearly 37 per cent of the list — had zero representation in either jobs or admissions.
  • A further concentration within the concentration: about a tenth of the communities took the majority of the benefit.
• The Commission proposed dividing the central list into sub-categories with fixed shares, weighting the allocation towards historically unrepresented communities.
• The report has not been made public and has not been acted upon. The reason is political rather than technical: sub-categorisation redistributes from the numerically strong and organised dominant OBCs to the fragmented castes below them, and no party has been willing to pay that price.

The state precedents

State	Instrument	Outcome
Andhra Pradesh	Backward classes divided into groups A to E since the 1970s, each with a fixed share	The oldest working model of OBC sub-categorisation
Tamil Nadu	Most Backward Classes and Denotified Communities carved out of the BC quota from 1989; 69 per cent total reservation protected in the Ninth Schedule	Sub-categorisation entrenched, but the Ninth Schedule protection is itself under challenge
Tamil Nadu (Vanniyars)	10.5 per cent internal quota for Vanniyakula Kshatriyas within the MBC quota, 2021	Struck down by the Madras High Court in November 2021, affirmed by the Supreme Court in Pattali Makkal Katchi (March 2022): sub-classification is permissible in principle but requires contemporaneous data, and 1985 population figures would not do
Bihar	Annexure I / Annexure II split following the Mungeri Lal Commission, giving the EBCs a distinct quota	The organising fact of Bihar’s politics; the base of the JD(U)’s coalition
Karnataka	Internal reservation within the Scheduled Castes on the Justice Nagamohan Das Commission‘s survey; the state modified the recommended formula and the Cabinet decided on a 6-6-5 division on 20 August 2025	The first substantial post-Davinder Singh implementation

• The lesson from Pattali Makkal Katchi is precise and is now the governing standard: the objection to sub-categorisation is evidentiary, not conceptual. A state may do it; it must first count.

State of Punjab v. Davinder Singh (2024)

• Decided 1 August 2024 by seven judges, 6:1, with Chief Justice D.Y. Chandrachud writing the lead opinion. It overruled E.V. Chinnaiah v. State of Andhra Pradesh (2004), which had held that the Scheduled Castes are a homogeneous class that a state may not subdivide.
• The reasoning:
  • Article 341 governs identification, not the distribution of benefits. Including a caste in the presidential list settles who is a Scheduled Caste; it says nothing about how a state must allocate the quota among them under Articles 15 and 16.
  • The presidential list is not a class in itself — it is an aggregation of castes with materially different histories, and treating unequals as equals offends Article 14.
  • A state must justify sub-classification on quantifiable and empirical data of the relative backwardness and inadequate representation of the sub-group.
  • Adequacy is to be judged by effective representation across grades of service, not by headcount — a numerically adequate presence concentrated in the lowest grades does not satisfy Article 16(4).
  • Sub-classification may not exclude a caste altogether from competing in the general Scheduled Caste pool, and may not be based on political expediency.
• Justice Bela Trivedi dissented, holding that the presidential list is a class by constitutional fiat and that state subdivision amounts to tampering with it.
• The case is a Scheduled Caste case, but its reasoning is a backwardness-definition case, and its logic runs straight back into the OBC list: if empirical evidence of intra-category inequality justifies subdividing the Scheduled Castes, the Rohini findings make the case for the OBCs unanswerable in principle.

Sub-categorisation is what happens when a category succeeds. The winners inside it become the reason to redraw its lines.

The Caste Census

• Every claim about backwardness in India rests on a number that has not been collected since 1931. That is the whole of the caste-census question.

How the gap opened

• The colonial census enumerated caste from 1871 onwards; the 1931 census was the last to do so comprehensively, and its proportions are the base on which Mandal, and therefore the 27 per cent quota, was built.
• The 1941 census collected caste but its tabulation was abandoned to the war. In 1951 independent India took a deliberate decision to stop enumerating caste except for the Scheduled Castes and Scheduled Tribes, on the reasoning that official recognition would perpetuate what the Constitution was trying to dissolve.
• The Socio-Economic and Caste Census 2011 attempted to reopen the question and failed in execution.
  • It was conducted by the rural and urban development ministries under the Registrar General, not under the Census Act, 1948, so its returns lacked the legal status and methodological discipline of a census.
  • Enumerators recorded caste names as respondents gave them, without a pre-coded schedule. The result was roughly 46 lakh distinct caste and sub-caste entries — surnames, gotras, sect names and spelling variants — that could not be collapsed into usable categories.
  • The socio-economic data was released in 2015; the caste data never was.
• Bihar’s caste survey (2023) demonstrated that a state could do it. Conducted January to May 2023 and released on 2 October 2023, it counted 215 castes across a population of 13.07 crore.

Bihar, 2023	Share of population
Extremely Backward Classes	36.01%
Backward Classes	27.13%
Scheduled Castes	19.65%
Unreserved / general	15.52%
Scheduled Tribes	1.68%

• Two results mattered politically. OBCs and EBCs together came to 63 per cent, well above the 50 per cent generally assumed; and the Yadavs, at 14.26 per cent, turned out to be far smaller than their political weight suggested — the survey’s own data showing them under-represented in government employment relative to their numbers.
• Bihar legislated on the findings within weeks, raising caste reservation from 50 to 65 per cent in November 2023.
  • The Patna High Court struck the amendments down on 20 June 2024, holding that the ceiling had been crossed without extenuating circumstances and that the backward classes already held 68.5 per cent of state employment.
  • The Supreme Court declined to stay that judgment in July 2024; the state’s appeal is pending.

Census 2027

• The Union Cabinet decided in April 2025 to include caste enumeration in the next census — the first comprehensive caste count in the main census since 1931.
• It will be conducted under the Census Act, 1948, with census a Union subject under Entry 69 of the Union List, and will be India’s first fully digital census.
  • Houselisting and housing census across 2026; population enumeration in early 2027.
  • Reference dates 1 March 2027 for most of the country and 1 October 2026 for Ladakh, Jammu and Kashmir’s snow-bound districts and Himachal and Uttarakhand’s non-synchronous areas.
  • The Ministry of Home Affairs notified the forty questions of the population schedule in August 2026, with caste recorded alongside religion for every respondent.
• A plea seeking to halt the caste enumeration was dismissed by the Supreme Court on 20 May 2026.

What enumeration can and cannot deliver

The case for	The case against
Quotas rest on 1931 proportions and estimates; policy should rest on data	Official enumeration gives caste a statistical existence and may harden it
Indra Sawhney contemplated periodic revision of backward-class lists, which requires counting	It will generate immediate pressure on the 50 per cent ceiling, since shares will be compared to numbers
Sub-categorisation is impossible without knowing which communities actually benefit	Caste nomenclature is not uniform across regions; the SECC failure was a coding failure
Nagaraj made quantifiable data a legal precondition in some contexts	Redress for the worst-off may become proportional bargaining among already-mobilised groups
Evidence replaces assertion in a debate now conducted entirely on both sides by estimate	It strengthens the dominant OBCs, whose numbers are largest, against the EBCs it is meant to help

• The measured position is that enumeration is necessary and nowhere near sufficient. A count establishes who is where; it does not decide what follows, and the two most likely uses of it point in opposite directions.
  • Used with the Rohini findings and a revised creamy layer, a caste census makes it possible for the first time to target the genuinely excluded rather than the well-organised.
  • Used as a census of entitlement, it converts reservation into an exercise in proportional shares, in which the strongest claimants — who are already the largest — do best.
• The distinction is not academic. The SECC failed because it collected caste without a framework of use; Bihar’s survey succeeded as data and then produced a law struck down within eight months. Counting is the easy part.

A census can tell India who is where. It cannot tell India what to do about it, and that has always been the harder question.

OBC Politics as Politics

Jaffrelot on the north-south divergence

• Christophe Jaffrelot, in India’s Silent Revolution, explains why lower-caste assertion arrived in north India half a century after the south — and the explanation is structural rather than accidental.

	South and west	North
Upper-caste structure	Effectively one upper caste — Brahmins, a small minority — making the target of mobilisation single and visible	Multiple upper castes — Brahmins, Rajputs, Bhumihars, Kayasthas — holding land, arms and letters between them
Numbers	Non-Brahmin castes an overwhelming majority, so a coalition against Brahmins was arithmetically decisive	The upper castes were a larger and better-entrenched bloc, and no single lower-caste front could dislodge them
Reform tradition	Phule and the Satyashodhak Samaj, Periyar and the Self-Respect Movement, Narayana Guru and the SNDP — anti-Brahmin ideologies that were also anti-Brahminical	Arya Samaj and the Shuddhi movement channelled reformist energy into Hindu consolidation, not caste confrontation
Mode of mobility	Ethnicisation — the non-Brahmin claim recast as Dravidian identity, a separate people rather than a lower rank	Sanskritisation — mobility by emulating Brahminical practice, which accepts the hierarchy while climbing it
Party vehicle	Autonomous parties from the Justice Party (1916) through the Dravidar Kazhagam to the DMK and AIADMK	OBCs remained absorbed within Congress until the 1960s, and later within the socialist parties
State policy	The Madras Communal Government Order, 1921 reserved posts by community; reservation was routine by the 1950s	Central OBC reservation arrived only in 1990

• The consequence is a fifty-year lag, and it explains why the anti-Mandal agitation was a northern phenomenon: in the south the argument had been settled long before, and settled against the upper castes.
• It also explains a difference of temper. Southern OBC politics was ideological before it was electoral — it carried a critique of Brahminism. Northern OBC politics was electoral before it was ideological, arriving as an arithmetic of numbers in search of a programme.

The northern sequence

• Rammanohar Lohia supplied the ideology.
  • He demanded that sixty per cent of positions in party, government and administration go to backward castes, Dalits, Adivasis, minorities and women — the first serious preferential-opportunity programme in Indian socialism.
  • His slogan that “backwards get a hundred out of a hundred” made caste, not class, the socialist unit of analysis.
• Charan Singh supplied the coalition. Breaking from Congress in 1967, he built the middle-peasant bloc around the interests of cultivators against both the landlord and the city — cheap inputs, remunerative prices, low agricultural taxation.
  • The coalition was expressed as AJGAR — Ahirs, Jats, Gujjars and Rajputs — later extended to MAJGAR with Muslims, an idea traceable to Chhotu Ram in colonial Punjab.
  • Its cross-caste logic was occupational: it united castes of very different ritual rank around a common position in the agrarian economy.
  • It also fixed the substantive content of OBC politics for a generation. Alongside reservation, its standing demands are remunerative procurement prices, cheap power and credit, and loan waivers — which is why OBC politics in the north has always been simultaneously quota politics and farm politics.
• The Green Revolution made the coalition materially possible. Assured procurement and new inputs created a prosperous cultivating stratum among Jats, Yadavs, Kurmis, Patidars, Marathas, Vokkaligas, Lingayats and Reddys — the class that would fund and lead OBC politics for the next thirty years.
• The Rudolphs’ category of differential mobilisation — a bloc assembled by a party out of several castes that had not previously cooperated — describes exactly what the OBC label became in the north: a fusion engineered from above, not a community that discovered itself.
• Mandal nationalised what had been regional. After 1990 OBC identity acquired a central state benefit attached to it, which gave a category previously held together by little except a list a concrete common interest.
  • The Samajwadi Party in Uttar Pradesh and the Rashtriya Janata Dal in Bihar became explicitly OBC-anchored formations, both built on a Yadav-Muslim core.
  • The Janata Dal (United) under Nitish Kumar assembled the counter-coalition — Kurmis, Koeris and the Extremely Backward Classes, the castes below the Yadavs — and converted the EBC category into a durable electoral bloc.
  • Both are read as evidence of the emancipatory power of caste politics: the RJD made Yadavs a political force, and the JD(U) gave political voice to the castes the Yadavs had displaced.
• The result is what is called the Mandalisation of Indian politics — the condition in which caste arithmetic organises party formation, alliance-building and candidate selection across most of the Hindi belt and much of the rest of the country.
• Post-2014 consolidation has rearranged rather than reversed this. The BJP’s expansion in north India rests substantially on the non-dominant OBCs — the castes that gained least from Yadav- and Kurmi-led formations — assembled behind a Hindu majoritarian appeal.
  • The mechanism is a familiar one: Mandal politics produced its own excluded stratum, and that stratum was available to whoever addressed it.
  • The party has simultaneously claimed the Mandal legacy — constitutional status for the NCBC, the Rohini Commission, the caste census decision — while drawing its OBC support from below the dominant castes rather than from them.

Representation: the ledger

• Legislatures. OBC representation in the Lok Sabha and in the north Indian state assemblies has risen steeply since the late 1980s, and the upper-caste share has fallen from a large majority to something closer to a plurality — the change Jaffrelot calls a silent revolution.
• The Union Council of Ministers. OBC ministers are now a substantial and openly counted share of the ministry, and the count itself has become a political fact that governments publicise.
• The bureaucracy. OBC representation in central government employment rose from roughly 15 per cent in 2012-13 to about 22 per cent in 2022-23, still below the 27 per cent norm.
  • In Group A it rose from about 7 per cent to a little over 18 per cent across the same period.
  • Scheduled Caste representation has stayed at or above its benchmark; Scheduled Tribe representation has dipped below it repeatedly.
• The judiciary. The higher judiciary, which is not filled by reservation, remains the least representative institution in the constitutional order, with OBC presence in the Supreme Court and High Courts far below any proportional measure.
• The ledger reads consistently in one direction: political representation has moved fastest, administrative representation slowly and unevenly, and institutional representation where quotas do not reach has barely moved at all.

Backwardness Claimed by the Dominant

• The most striking development of the past two decades is that prosperous landowning castes have begun demanding backward status — a reversal of the Sanskritising logic under which such castes once claimed to be higher.

Community	State	Claim and outcome
Jats	Haryana, UP, Rajasthan and six other states	Added to the central OBC list in 2014; quashed in Ram Singh (2015); the 2016 Haryana agitation left about thirty dead
Patidars / Patels	Gujarat	The 2015 agitation under Hardik Patel; no OBC status; effectively absorbed into the EWS quota
Marathas	Maharashtra	16 per cent struck down in 2021; a fresh 10 per cent SEBC Act, February 2024, on the Shukre Commission finding of 84 per cent distress, under challenge in the Bombay High Court with admissions continuing; a 2025 government resolution allowing Kunbi certificates on Hyderabad-gazette evidence has been challenged by OBC organisations
Kapus	Andhra Pradesh	Legislated a 5 per cent quota in 2017 that breached the ceiling; not implemented
Gujjars	Rajasthan	Agitations from 2007-08 for ST rather than OBC status; a special backward class quota repeatedly struck down for crossing 50 per cent

• The driver is economic, not status-seeking. Agrarian distress has destroyed the profitability of the landholding these castes were built on, while liberalisation shrank secure public employment and the private sector offers no reservation at all.
  • What is left of assured, respectable, pensionable employment is the reserved public sector — so the only route to it is a claim of backwardness.
  • The claim is also a claim of relative decline: these communities are not comparing themselves to the poor but to their own past.
• The deeper implication is about the state’s vocabulary. Reservation has become the only currency in which recognition, grievance and redistribution can be denominated, so every distress finds expression as a quota demand — which is why the demands persist even where the arithmetic makes them impossible.
• The litigation record is uniform: every dominant-caste quota that crossed 50 per cent has been struck down, and the only successful accommodation has been the EWS quota, which reaches these communities without calling them backward.

The Critique and the Reply

Does the remedy reach the worst-off?

• Elite capture is the strongest empirical charge, and the Rohini findings substantiate it: a quarter of the listed communities take 97 per cent of the benefit, and over 900 take none at all.
• Ashutosh Varshney argues that caste mobilisation has produced no concrete transformation in the distribution of power in society. Indian democracy performs superbly at elections and poorly between them; the battle now is for the deepening of democracy, not for another round of representation.
• Pratap Bhanu Mehta presses a related objection: quota politics has become a politics of recognition without a politics of provision, competing over shares of a public sector that is shrinking, while the actual determinants of mobility — school quality, health, urban employment — go uncontested.
• Andre Beteille puts the constitutional version of the complaint: the Constitution abolished untouchability without abolishing caste, and then made caste a basis of public policy, so the state’s own categories keep alive what its principles condemn.
• C.P. Bhambhri adds the sequence argument: caste politics invites religious counter-mobilisation, because a politics organised by group numbers eventually prompts the largest group to organise as one.
• Kanchan Chandra‘s point closes the circle — where political patronage is distributed by caste, a caste-free politics is not available to anyone who wants to win.

The reply

• Yogendra Yadav reads the same period as two democratic upsurges — the OBCs leaving Congress in the 1960s and 1970s, the Dalits leaving in the 1990s — and treats rising participation from below as the substance of Indian democratisation rather than its distraction.
• Jaffrelot’s silent revolution and Kothari’s politicisation thesis converge on the same reply: representation was the redistribution that was actually available. Land reform failed, industrial employment never arrived at scale, and what did change was who holds office and who is feared at the district level.
• The reply has a limit its proponents concede. A revolution that stops at the dominant OBCs is not a revolution for the extremely backward classes, and the group that most needs the remedy is the one least able to claim it.
• The honest synthesis is a sequence rather than a verdict: OBC mobilisation was democratising while it was breaking a monopoly, and becomes an obstacle where the castes it empowered now defend their own share.
  • The test to apply to any claim is whether it expands access or protects a share.
  • That single test sorts the caste census, sub-categorisation, the creamy layer and the dominant-caste agitations more reliably than any position for or against reservation as such.

Conclusion

Backwardness in India is not a condition that policy discovers; it is a status the state confers, and the politics of the OBCs is the politics of that conferral. The category was left undefined, filled in by commissions, bounded by courts, and made the object of a competition no party can decline to enter.

• What the coming decade decides is narrow and consequential: whether the count beginning in 2027 is used to find the people the last thirty-five years missed, or to divide the same quota more efficiently among those who already know how to claim it.