Guest Post: Sub-Classification in Reservations

[This is a guest post by Anuna Tiwari.]


Contemporary debate around reservations has turned inwards, from reserved-versus-non-reserved, to one within the reserved communities. Several state Acts and fact-finding commissions have voiced the need for sub-classification and sub-categorization within the reserved SCs to account for their inherent gradations. One such Act by the Andhra Pradesh Government was held unconstitutional in the Supreme Court’s decision of EV Chinnaiah vs State of Andhra Pradesh (2005 1 SCC 394). Another such law has now been put to review by a larger bench of the Supreme Court by a Constitution Bench on 27th August 2020 in State of Punjab vs Dalvinder Singh.

The Supreme Court in EV Chinnaiah examined the Andhra Pradesh (Rationalization of Reservation) Act which dealt with such a conflict between sub-castes in Andhra Pradesh. The Andhra Act had bifurcated the Presidential list into four groups and had apportioned the 15% quota among them on the basis of degree of backwardness and proportional population of each sub-caste. The state argued that reservations in the state were unduly benefitting only a handful of the SCs who were relatively advanced, at the cost of the most deprived within them. The Supreme Court had held that the said sub-classification was unconstitutional as the SCs were a class by themselves, classified sufficiently (as such for the purpose of Article 14) under Article 341, and that the sub-classification could only be made by the Parliament under Article 341(2).

In Dalvinder Singh the issue of sub-classification has resurfaced. Section 4(5) of the Punjab Scheduled Caste and Backward Classes (Reservation in Services) Act 2006 provided that 50% of the total seats reserved for the Scheduled Castes be offered to Balmikis and Mazhabi Sikhs. This was challenged on the basis of the Supreme Court’s decision in EV Chinnaiah, before the Punjab & Haryana High Court which deemed it unconstitutional. In appeal, the Supreme Court has now held that (i) the state could provide preference to a class within the Scheduled Class/ Scheduled Tribe (¶ 48) ; that (ii) allotment of specific percentage out of the reserved seats  would not violate Article 341, 342 and 342A and that (iii) a federal structure put both the State Legislature and the Parliament under a “constitutional directive” for the upliftment of the SCs/ STs and backward classes (¶49). The Court has referred the matter to a 7-judge bench, endorsing the view that the EV Chinnaiah was at variance with the realities of the community.

Among the foremost critiques of EV Chinnaiah is the Supreme Court’s off-the-mark perception of the Scheduled Castes as a “homogenous” group. The Court had observed that they were a homogenous group incapable of being further sub-divided, much less by the State Legislature. Dalvinder Singh (inter-alia) by taking note of the massive anthropological and empirical data pointing to anything but homogeneity, has overcome this perception for the better.

It is a welcome move so far as the Court acknowledges that the states are entitled to account for social realities (¶38,47) and that sub-classification is just as much as facet of equality and social justice as reservations themselves (¶44) .

The “guise of homogeneity” 

In Dalvinder Singh, the Court rectifies the EV Chinnaiah Court’s reluctance to engage with the Ramachandran Raju Commission report and its findings. Justice Sinha had in EV Chinnaiah declared the redundancy of the report stating:

“It may not be necessary for us to delve deep into the question as to whether the factual foundation for enacting the said legislation being based on a report of a Court of Inquiry constituted under Section 3 of the Commission of Inquiry Act 1952 known as Justice Raju Report is otherwise laudable or not.” (¶65)  

In Dalvinder Singh, the Court comes to terms with the various factual reports carrying empirical evidence of inequality among the SCs. As they say, a problem well stated, is problem half solved. The statement of the issue itself duly acknowledges the relevance of these reports.

“Whether the action based on intelligible differentia to trickle down the benefit can be said to be violative of Articles 14 and 16 of the Constitution and whether sub-classification can be said to be an act of inclusion or exclusion particularly when various reports indicating that there is inequality inter se various castes included within the list of Scheduled Castes”. (¶29)  

Vis-à-vis the rest of the society, the degree of contempt for some sub-castes is greater than the rest. Further, internal schisms both mimic and amplify such prejudice. For instance, the Sadashiv Commission found that while some among SCs (touchable) reside alongside the upper dominant castes, some among the SCs continue to live in the peripheries of the villages and are deemed polluted by the other SCs. In Andhra, though both Malas and Madigas reside in the periphery of the village, there is hierarchy in “marginality”: the Madigas (ritually lower) are farther away from the main village than the Malas. The occupation of Malas (weaving and agriculture labour) were deemed cleaner than those of the Madigas (leather tanning, show making etc.) The notions of impurity on the basis of occupation can be seen (with somewhat equal, if not greater intensity) within the SCs. The reform movement leadership of Andhra was heavily under-representative of the Madigas: less than a dozen among the 130 Malas. In the data available from 2001 census, Adi-Andhras, another economically forward group,  showed education level of a 6.3 as against a 2.7 of Madigas, the average among SCs overall being 3.7. The 2011-Census state-wise data set on literacy and educational attainments of around 600 SC groups reveals that there is variance within the SCs. While 11% held a degree in one quarter of the SCs, only 2% of the other quarter had a degree. The Supreme Court, is itself not an absolute stranger to the nature of SCs under the Constitution and has in the past recognized the SCs as an “amalgam” rather than an immutable whole (NM Thomas vs. State of Kerala, 1976 AIR 490).

How much can be left to the “subjective satisfaction” of State

The Chinnaiah Court’s perception of homogeneity is flawed. But the greater issue is the Court’s treatment of factual reports in reservation cases. After Jarnail Singh vs Lachhmi Narayan Gupta, quantifiable data is not a pre-cursor to reservations (in promotions) (Jarnail Singh ¶21).  And apparently, after Mukesh Kumar vs State of Uttarakhand, affirmative action turns on states’ ‘subjective satisfaction’. The Supreme Court in EV Chinnaiah had looked the other way on considerable empirical evidence, thereby questioning the states’ prerogative to act on the same. However, the Court in Mukesh Kumar Gupta held that “inadequacy of representation is a matter within the subjective satisfaction of the State” and that the State could very well rely on “material it has in its possession already or [that] it may gather such material through a Commission/ Committee, person or authority(¶12). Here the Court held that while the State could choose to act on such material, it could not be compelled to do so since Article 16(4) is merely an enabling provision. This reasoning sets us back in time to the pre-NM Thomas days where Article 16(4) was considered an exception to equality. It also tells us why ‘subjective satisfaction’ is a double-edged sword, even though it may appear desirable in an expansionist context like EV Chinnaiah’s.

Clearly, what the states make of “subjective satisfaction” is not uniform. While Andhra Pradesh and Punjab expanded the scope of reservations to reflect social realities, the Uttarakhand Government instead decided to rule out reservations in all posts in public services in the overlooking the data on their backwardness. Therefore, the Court’s push to state action based on statistical data in Dalvinder Singh is desirable. The outcome of the reference should infuse greater consistency and certainty on the state’s obligations to act on such data.   

Over-broad reading of Article 341(2) by EV Chinnaiah. 

EV Chinnaiah’s discontents go beyond misplaced notion of homogeneity. Another argument against the Andhra Pradesh Act was that constitutional intention behind Article 341 forbade sub-classification by the state legislature (EV Chinnaiah ¶22). Justice Santosh Hegde observed that the Scheduled Castes come into existence only by virtue of the Constitution and the exclusive power to include or exclude from, and thereby to sub-classify the presidential list vests solely with the Parliament.

Justice Hegde cited Dr. B.R. Ambedkar who had once said that power is vested in the President to guard against the “political factors having a play in the matter of disturbance in the schedule so published”. He further read this “disturbance” widely to mean not just inclusion and exclusion from the presidential list, but also re-classification:

“Therefore any executive action or legislative enactment which interferes, disturbs, re-arranges, re-groups or re-classifies the various castes found in the Presidential List will be violative of scheme of the Constitution and will be violative of Article 341 of the Constitution.”

It is difficult to understand this extrapolation (read: re-writing) of a plain-worded provision. Article 341(2) only deals with inclusion and exclusion from the list rather than inter se sub-classification of the enlisted groups. Also, nothing in the context of his statement suggests that Dr. Ambedkar had sub-classification or re-arrangement, re-classification in mind. As Dalvinder Singh clarifies only now, sub-classification does not amend the list in any way (¶45,49).

Justice Hegde’s reading of Article 341(2) assumes greater importance since it is employed to set off a Part III claim. The doctrine of reasonable classification is conspicuous in its absence and its non-application goes unexplained. This is a limitation that the Court imposes on itself by ignoring the Commission’s findings of internal social, educational and economic divisions between SCs in Andhra Pradesh, (belittling them as ‘micro-distinctions’). The Court in Dalvinder Singh discards this reading of Article 341(2) and clarifies that sub-classification could be done within the lists without attracting Article 341(2) (¶48).

The case for sub-classification

The administrative definition of SCs under Article 366(24) states that those castes, groups, races, tribes and communities that are grouped under Article 341 by the President of India are scheduled castes for the purpose of the Constitution (The Constitution of India 1950, Article 366(24)).  Socially, SCs are categories of people who were at the bottom of the Indian society owing to their low status in the ritual hierarchy and their spatial-cultural isolation from the society. The Constitution identifies SCs on the basis of “untouchability”. The first census tests for identification revolved around incidence of disability such as debarment from using temples, polluting touch, occupational disability etc. The 1950 Scheduled Order took educational and economic criteria as indicators of inclusion into the list. Presently, “social, educational and economic backwardness” arising out of traditional untouchability is used as the primary indicator of inclusion in the presidential list. Since the criteria of distinguishing SCs (as a whole) is untouchability and relative impurity- based on occupation, residence, ritual status, eating habits etc., relative impurity among the Dalits assumes importance. If the same ritual untouchability persists within the SCs, it should adequately justify sub-classification- not as an exception but as an extension of the entitlement.

Both historical and contemporary factors indicate unequal access to education and distinct positions in the social hierarchy. Neither of these is a “micro-distinction” (as EV Chinnaiah calls them in ¶57) simply because the exact same markers of discrimination are rampantly employed in segregating the SCs from the rest as stated above. The internal segregation is not wiped out by “shared-status” feeling. If the former has pre-existed the Constitution, the latter cannot be magically conjured, notwithstanding constitutional classification. The SCs are heterogeneous and unequal. Failing to classify unequals violates Article 14 (K. T. Moopil Nair v. State of Kerala, AIR 1961 SC 552) 

Reservations emerge in response to “entrenched and cumulative nature of group inequalities” and thus accrue to the SCs on the basis of their group-identity ascribed at birth and inseparable from their beings (Marc Galanter, Competing Inequalities, Page xxv). Dalvinder Singh (barring its use of the term “largesse”)( ¶42) holds the mirror up to EV Chinnaiah on this very crucial aspect. EC Chinnaiah’s “justice to one group at the cost of injustice to another group” line of reasoning tends towards the notion of ‘justice’ championed by those opposing reservations in general. K Balagopal in his scathing critique of the judgement termed it as an argument against reservations, rather than just sub-categorizations. He laments that “at the end what we have is a judgement purportedly against subdivision of the scheduled caste reservation quota, but which is in fact replete with arguments against reservations as such.”  Dalvinder Singh’s rooting of sub-classification in the same principle as reservation itself (¶44) is thus a long overdue moral contrast.

Conclusion: Imposed homogeneity and Creamy layer exclusion must meet half-way.

EV Chinnaiah disallows sub-classification without engaging with the reasonable classification argument. Jarnail Singh on the other hand justifies creamy layer exclusion of the SCs. Both are extreme approaches. Jarnail Singh glorifies the capacity of reservations to wipe-out social disability. It conflates group-disability and an individual’s attainment of a certain advancement. Exclusion from reservation cannot wipe out persisting social disability, and cannot be the ideal. Nearly 60% reserved positions in the Central Governments and Ministries lie vacant as in March 2020. Glaring vacancies in the reserved positions across cadres and institutions indicate problems with this exclusion. If some are excluded and the rest are socially incapacitated (in the absence of sub-classification), the optimum representation would elude us. Reservation must come to the downtrodden sub-castes, not at the cost of, but alongside the rest of the sub-castes.

Special quotas can be created to counter extreme disability, like those for Arunthathiars. The Tripura High Court had in 2015 ordered filling up of vacant positions by general candidates in the absence of a SC candidate. This is undesirable if representation is to be ensured. Rather than fixed quotas, vacant reserved positions are better preserved by a preference-based layering model. There could be layering of the groups preferentially in an order of decreasing level of social disadvantage, the most disadvantaged sub-category being the most preferred. In the absence of any candidate from a sub-category higher on the preference order, only a candidate from the next preferred sub-category could fill in. Rather than the fixed apportionment-quota model (as adopted by the Andhra Pradesh Act), the layering model would ensure that the vacancy is retained within the reserved fold, while the most marginalized sub-category’s disadvantage is partly offset by the preferential order.

The EV Chinnaiah bench’s reluctance to sub-categorise closely resembles that of the central government in categorising the SCs, apprehending that the “official recognition of their low-ness” would perpetuate discrimination (Indian Statutory Commission 1930:V, 1341 as cited in Marc Galanter, p. 122). Both the fears, facially salubrious, presume homogeneity and that caste-consciousness would do more harm than caste-blindness in the law.  The courts may be limited in their power to formulate rights-based policy, they are extremely potent in mobilizing the claimants, enlarging the scope of their claims and  triggering state action: Vishakha is a case in point. Until notions of caste and impurity persist, we can neither (like EV Chinnaiah) treat distinct groups as one, nor (like Jarnail Singh) discount prejudice against group identity and exclude individuals. Extensive empirical data already at the disposal of the state must not go to waste and a middle way must be found. The Supreme Court in Dalvinder Singh thus expiates for opportunity that it lost in EV Chinnaiah and sets the debate on this middle path, rolling again (¶39).


[The author thanks Professor Aparna Chandra for comments on a draft of this piece.]

2 thoughts on “Guest Post: Sub-Classification in Reservations

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s