State of Punjab v. Davinder Singh: how a seven-judge Bench permitted sub-classification within Scheduled Castes

The Supreme Court's judgment of 1 August 2024 in State of Punjab v. Davinder Singh is the most significant intervention in the architecture of Scheduled Caste reservation since M. Nagaraj v. Union of India (2006). A seven-judge Constitution Bench, led by Chandrachud, C.J., held by 6:1 that sub-classification within the Scheduled Castes for reservation purposes is constitutionally permissible — overruling E.V. Chinnaiah v. State of Andhra Pradesh (2004), which had treated the Scheduled Castes as a homogeneous class.

The judgment is doctrinally consequential on three connected propositions. The first is that Article 341 of the Constitution — which authorises the President to specify the castes that constitute Scheduled Castes — does not deem the castes within the list to be homogeneous. The second is that the State, in giving effect to the reservation architecture, can sub-classify within the Scheduled Castes to ensure that the benefit reaches the most disadvantaged sub-groups, subject to constitutional safeguards. The third — articulated by Gavai, J. in concurrence — is that the creamy-layer principle may, in time, be extended to Scheduled Castes within the reservation architecture.

The doctrinal background

The architecture of Scheduled Caste reservation operates through Articles 15(4), 16(4) and 341 of the Constitution. Article 15(4) authorises the State to make special provision for the advancement of socially and educationally backward classes of citizens and for the Scheduled Castes and Scheduled Tribes. Article 16(4) authorises reservation in public employment. Article 341 authorises the President, after consultation with the Governor of a State, to specify the castes, races, or tribes which shall, for the purposes of the Constitution, be deemed to be Scheduled Castes in relation to that State.

The question of whether the State could, within the architecture, sub-classify the Scheduled Castes had been addressed before Davinder Singh in two principal Constitution Bench engagements.

E.V. Chinnaiah v. State of Andhra Pradesh (2004) — a five-judge Constitution Bench — had struck down an Andhra Pradesh statute that had sub-classified the Scheduled Castes within the State into four groups, with separate reservation quotas for each. The Court had held that the Scheduled Castes constituted a homogeneous class for constitutional purposes; that Article 341's identification was constitutive of that homogeneity; and that the State could not, by sub-classification, displace it. The State's sub-classification, on the E.V. Chinnaiah reading, was unconstitutional.

The Punjab Government had, in the years that followed, enacted legislation that operated on the principle of sub-classification within the Scheduled Castes. The legislation came up for challenge; the matter was referred to a larger Bench in light of doctrinal doubt about whether E.V. Chinnaiah had correctly construed the Article 341 architecture.

The question before the seven-judge Bench

The question before the Davinder Singh Bench was whether E.V. Chinnaiah had correctly held that the Scheduled Castes constituted a homogeneous class incapable of sub-classification. The question had two limbs.

The first was whether Article 341 — the constitutional provision through which the Scheduled Castes are identified — operates as a constitutive declaration of homogeneity, or as a mere enumerative listing of the castes that fall within the constitutional category.

The second was whether, if Article 341 does not constitute homogeneity, the State has the constitutional competence under Articles 15(4) and 16(4) to sub-classify within the Scheduled Castes for the purposes of giving effect to the reservation architecture.

The majority position

The Bench, by 6:1, held that sub-classification is constitutionally permissible. The reasoning rested on three connected propositions.

The first was that Article 341 is enumerative rather than constitutive. The President's identification of the castes that constitute the Scheduled Castes in relation to a State does not deem those castes to be homogeneous; it identifies the constitutional category within which reservation operates. The architecture leaves open the question of whether, within that category, the State can act on the empirical reality of differing levels of disadvantage.

The second was that the constitutional purpose of reservation — bringing those who have suffered historical disadvantage into the mainstream of public participation — is served by the State's capacity to act on the empirical reality of differing levels of disadvantage within the Scheduled Castes. Some sub-groups within the constitutional category may have, on the empirical record, benefited from reservation in ways that other sub-groups have not. The constitutional purpose is best served by an architecture that permits the State to address that empirical reality.

The third was that the sub-classification must be supported by empirical data of inadequate representation of the sub-group within the broader Scheduled Caste category. The State cannot sub-classify on the basis of caste-political pressure or ad hoc administrative judgment; the constitutional standard is empirical demonstration of inadequacy of representation.

The majority opinions were authored across the seven-judge Bench. Chandrachud, C.J. wrote the leading opinion. Gavai, J. wrote a concurring opinion that addressed the creamy-layer question (discussed below). Vikram Nath, J., Pankaj Mithal, J., Manoj Misra, J. and Satish Chandra Sharma, J. each wrote separate opinions concurring in the result.

The Gavai concurrence on creamy layer

A doctrinally consequential element of the judgment is Gavai, J.'s concurring opinion on the creamy-layer question. The concurrence holds that the creamy-layer principle — articulated for Other Backward Classes in Indra Sawhney (1992) and extended to Scheduled Castes in promotion reservation in Jarnail Singh (2018) — should be considered for application to Scheduled Castes within the general reservation architecture as well.

The reasoning is that the constitutional purpose of reservation is to bring those who have suffered substantial historical disadvantage into the mainstream of public participation. Where members of the Scheduled Castes have advanced economically and socially to a point where they no longer suffer that disadvantage, the constitutional purpose is not served by extending the reservation benefit to them. The creamy-layer principle operates to ensure that the benefit reaches those for whom it is intended.

The concurrence has not, in itself, established the creamy-layer architecture for Scheduled Castes. It has — as a concurring opinion — invited the question for engagement in future cases. The doctrinal direction is suggested without being settled.

The dissent

Bela M. Trivedi, J. dissented. The dissent's principal proposition is that E.V. Chinnaiah had correctly construed the Article 341 architecture and should not be overruled. The dissent treats Article 341's identification as constitutive of homogeneity and treats the State's sub-classification as a constitutional displacement of the homogeneity that the architecture had recognised.

The dissent also engages with the practical question of whether sub-classification — even if constitutionally permissible — produces the consequences that the majority's reasoning had contemplated. The argument is that sub-classification produces fresh hierarchies within the reserved category, in a manner that the constitutional architecture had been framed to prevent.

The dissent is doctrinally significant as the minority position. It has been the subject of substantial professional commentary and represents an articulated constitutional position that the post-Davinder Singh engagement will need to engage with.

What States can now do

The judgment opens the door to State legislation and executive action sub-classifying within the Scheduled Castes for reservation purposes. The doctrinal frame — that the State acts on empirical data of inadequacy of representation — supplies the working constraint.

The operational architecture for States seeking to sub-classify includes the following elements. The State must identify the sub-groups within the Scheduled Castes that are inadequately represented in the categories of public employment or educational institutions in which reservation is provided. The identification must rest on empirical data — preferably collected through a State-level commission or similar mechanism with appropriate methodological rigour. The sub-classification must be calibrated to the inadequacy demonstrated — a sub-group that is materially under-represented may be allotted a sub-quota; the architecture must not produce arbitrary or empirically unsupported sub-classification.

The architecture must also remain within the broader constitutional constraints. The Indra Sawhney ceiling of 50 per cent reservation across all categories continues to apply. The Article 16(4A) and (4B) architecture for reservation in promotion continues to apply. The Jarnail Singh reading-down of the M. Nagaraj backwardness requirement continues to apply. The sub-classification operates within — not in displacement of — these constitutional features.

The doctrinal arc

The doctrinal arc of reservation jurisprudence, taken across the line of Constitution Bench engagements, runs through four principal nodes.

Indra Sawhney (1992) established the architecture for reservation under Article 16, including the 50 per cent ceiling and the creamy-layer principle for OBCs.

M. Nagaraj (2006) addressed reservation in promotion for Scheduled Castes and Scheduled Tribes — and imposed the three-pronged test of backwardness, inadequacy of representation, and administrative efficiency.

Jarnail Singh (2018) read down the Nagaraj backwardness requirement (the SC/ST presumption of backwardness from Indra Sawhney could not be displaced by a contrary data-collection requirement) and extended the creamy-layer principle to SC/STs in promotion reservation.

Davinder Singh (2024) overruled E.V. Chinnaiah and permitted sub-classification within the Scheduled Castes — with the Gavai concurrence suggesting that the creamy-layer principle may, in time, be extended to SCs in the general reservation architecture.

The arc, taken across these four nodes, has moved from a relatively undifferentiated treatment of the reserved categories to an architecture that engages with the empirical reality of differing levels of disadvantage within the categories.

What the judgment did not decide

Three limits should be flagged.

First, the judgment permits sub-classification but does not mandate it. States are not constitutionally required to sub-classify; the architecture leaves the decision to the State's legislative and executive discretion, subject to empirical justification.

Second, the Gavai concurrence on creamy-layer extension to Scheduled Castes is not the binding holding of the majority. The question whether the creamy-layer principle applies to Scheduled Castes within the general reservation architecture (as distinct from promotion reservation, where it has been settled in Jarnail Singh) remains formally open.

Third, the judgment does not engage with the parallel question of sub-classification within the Scheduled Tribes. The Article 342 architecture for Scheduled Tribes is doctrinally parallel to the Article 341 architecture for Scheduled Castes; the application of the Davinder Singh reasoning to the Scheduled Tribes is implied but has not been formally addressed.

What practitioners take from the judgment today

For constitutional and service-law practitioners in the reservation space, Davinder Singh is the operative authority on sub-classification within the Scheduled Castes. State legislation that operates on the principle — and challenges to such legislation — engage the Davinder Singh framework.

For policy practitioners advising State governments on reservation architecture, the judgment opens a constitutional avenue that had been closed by E.V. Chinnaiah. The empirical-data requirement supplies the discipline; the constitutional permission supplies the opportunity.

For the broader profession, the judgment is among the most significant Constitution Bench engagements of the post-Puttaswamy period — and is part of the doctrinal recalibration of reservation jurisprudence that has been the project of the Court across the Jarnail Singh – Davinder Singh line.

Related editorial pieces

• Indra Sawhney v. Union of India: the reservation architecture and the 50 per cent ceiling
• M. Nagaraj v. Union of India: the three-pronged test for SC/ST promotion reservation
• Jarnail Singh v. Lachhmi Narain Gupta: creamy layer for SC/ST promotion reservation
• Kesavananda Bharati v. State of Kerala: the basic structure doctrine