“Group-based sub-classification does not really ensure more substantive equality”: Dr. Surendranath on the sub-classification judgement

Transcript: 

Spandana: Hello everybody and welcome to SCO’s channel. I’m Sai Spandana, and today we have a very special episode for you because we’re in conversation with Dr Anup Surendranath, Professor of Law at the National Law University, Delhi.

Anup is joining us to discuss the Supreme Court’s recent decision to uphold sub-classifications within the reserved categories in a case titled State of Punjab v. Davinder Singh. Anup teaches criminal law, constitutional law, and legal methods at NLUD and is the Executive Director of Project 39A, which is a criminal justice program at the university.

Anup has also had a long engagement with the subject of sub-classification, particularly through his doctoral thesis, which he submitted in 2013. His thesis explored the scope and challenges of sub-classification in India’s affirmative action policy.

Anup, thank you so much for joining us today.

Surendranath: Thank you. Thank you for having me.

Spandana: So, before we kind of get into the impact and challenges of the concept of sub-classification and the larger debates it throws in the affirmative action through reservation, let’s just quickly recap the decision itself. The case, of course, was heard over three days in February by a seven-judge bench led by Chief Justice DY Chandrachud. The bench reserved judgment in February, and the decision came on 1 August 2024. This decision essentially overturns the five-judge bench decision in E.V. Chinnaiah from 2004. 

So, can you briefly tell us about the decision in Davinder Singh and what the bench held, which of course included six opinions and Justice Bela Trivedi’s dissent as well?

Surendranath: Yeah, no, I mean, one of the main issues is on whether doing something like sub-classification—and we can talk of sub-classification in different ways—but does the very idea, which way you execute it, violate the constitution in terms of the constitutional provision which seems to preserve the domain of power of changing the lists of scheduled castes only to parliament, right? And the idea, the fundamental question in Chinnaiah was there, sub-classification, then, when governments do it, when state governments do it, does that amount to intruding on the power that is reserved for parliament in the constitution? Because does it amount to changing the list, the list that the 1950 presidential order that is in consequence of a constitutional provision? Does this then, if you sub-classify, does it amount to changing that list? Because certain proportions of reservation are then reserved for certain groups, and others are not entitled to that.

And Chinnaiah‘s own position was that, well, it is an infringement of that power that is preserved for parliament. And fundamentally, there is an overruling of that, that this is not tampering with that parliamentary power. Uh, it is, of course, then the, uh, it does not locate it to, uh, 341. I got the number right. Yeah. Uh, it’s not, uh, it’s not 341. This is now located within 16(4). And of course, that’s an interesting premise of the majority judgment itself, that the power is located in 16(4). And whether you can locate that power in 16(4) or not is a separate issue. But this is the legitimacy of this action is traced back in the majority’s judgment to 16(4) and does not view it in a manner that Chinnaiah views it as an infringement of parliamentary power in 341.

Yes. So. And then they say, why this is needed in terms of that the group is not homogenous in that sense, that within the scheduled caste group, there are, as if you view caste as a spectrum right from the top to the bottom. At some point in that spectrum, you’ve drawn a line and said, “this is where scheduled castes begin from.” But because of the fact that caste is on a spectrum, even within the scheduled caste, there are the more dominant and the more marginalized groups. And therefore, the majority feels that it’s in furtherance of substantive equality that you need to permit sub-classification, and 16(4) allows for that sub-classification.

Spandana: Yeah, I think the chief’s phrase exactly during arguments was “battle of backwardness within backwardness.” Do you think the judgment adequately addresses that concern, though, of unequal distribution of benefits and opportunities within the SC/ST group?

Yeah, I mean, I think it’s thin on its evidence of what is the evidence that there has been a disproportionate distribution of benefits, right? But I think it’s spot on in making the sociological argument that that distinction does exist, right? Whether that decision, how much of it is translated into disproportionate distribution of reservation benefits, I think the judgment is a little thin on it. And it’s quite the irony that it says that you have to provide empirical data to a much, much higher degree if you want to do sub-classification, that if the governments in the future want to undertake this exercise of sub-classification, you have to show significant empirical data. And it’s quite a rigorous standard that you require to demonstrate that the inadequacy of representation is due to the backwardness, that it is because, and I think there’s that word “because” is in bold and italicized in the judgment, that you have to show inadequacy because of backwardness, inadequacy of representation because of backwardness.

But currently, in the judgment’s own justification for why there is this internal disparity in distribution of benefits, I think the sociological argument is much stronger than the empirical argument within the judgment itself.

Spandana: In your thesis, “Judicial Discourse on India’s Affirmative Action Policies: The Challenges and Potential of Sub-Classification,” you had argued that the court had not yet developed a nuanced, normative framework for reservations as a whole. But with the decision in Davinder Singh, do you think the court’s decision, especially the Chief Justice’s majority opinion, kind of does that in a way?

Surendranath: Yeah, I’d say it’s, I don’t know whether to say it’s two steps forward and one step back, or one step forward and two steps back. Whatever it is, I think there are certain elements that are good about the judgment and certain elements to it that are not so great, which reflect deeper normative confusion and that lack of coherence that I viewed in my thesis. And I think the positive part is the relationship between efficiency and reservation, that the judgment has clarified efficiency and reservations in public employment to be more particular. Now, obviously, the efficiency concern that has been developed in constitutional jurisprudence is because the word finds mention in Article 335 of the Constitution, which then says that the claims of the Scheduled Castes and Schedule Tribes must be taken into account, keeping in mind the concern of efficiency, right?

Now, that, in Nagaraj, is then taken to the level of almost testing constitutional amendments on that basis. And that has been a very problematic trajectory of Indian constitutional law and reservations. It buries that, the majority, Davinder Singh buries that concern, and quite rightly so. And Justice Chandrachud’s opinion that is agreed upon by five other judges is on point that efficiency is not the focus of Article 335. It cannot be because there will be some anomaly that arises if you read it that way, because in 335, it is applicable only to Scheduled Castes and Scheduled Tribes. And you can’t read the efficiency concern from 335 because that would mean then that efficiency is a concern only when SC/STs are being provided reservations, right? 

But obviously it’s not just SC/STs that are being provided reservations within 16(4). Even OBCs are being provided with reservations in 16(4). So he’s spot on to argue that if you were trying to bring in efficiency as a controlling factor for 16(4) reservations in public employment, it leads to the anomaly then that would be able to extend efficiency concerns only to SC/ST reservations. You wouldn’t be able to do it to OBCs. And since that is not the point of 335, what then is the point of 335, according to the majority? 

The efficiency concern is only to the extent, according to Justice Chandrachud, that you can’t do away with the requirement of some qualification marks, right? That you need some qualification marks, you can’t completely do away with it. But 335 is not constitutional acceptance of the fact that higher marks means higher efficiency. And I think that is such a significant doctrinal clarification and the manner in which it is provided is something that will be lauded about this judgment for generations to come, I feel. I think it settles that position quite well, and quite rightly, it settles that position, right?

But I think the judgment gets it, unfortunately, terribly wrong on the creamy layer, right? And the reason why I say it gets it terribly wrong is because it goes back to the broader doctrinal confusions that I spoke about in my thesis that we really haven’t clarified from a constitutional perspective as to why do we provide reservations? And the answer to that question can vary depending on which sphere you’re providing reservations in.

If I ask you the question, “why do we provide reservations?” Let’s say, primary education through the RTE, the answer to that may well be very different from “why do we provide reservations in higher education?” versus “why do we provide reservations in public employment at the entry level?” “Why do we provide reservations at the promotion level?” All of these invite very different answers. It does not invite one common answer. And unfortunately, our jurisprudence has proceeded, our judicial discourse has proceeded on the assumption as though somehow the answer to all of this is clear and everybody knows it. But you know that’s not the case.

And similar questions could be asked from a group perspective. Why do we provide reservations for women? Why do we provide reservations for SC/STs? Why do we provide reservations for OBCs? All invite very different answers depending on the marginalization and the vulnerability that we’re looking to address. And in that context, I feel that the extension of the creamy layer idea to SC/STs is wrong, right? But unfortunately, it’s a longer judicial arc to it that finds approval with the majority in Davinder Singh.

This whole idea of creamy layer among the SC/STs through Nagaraj and through Jarnail Singh has a much longer trajectory, which I think, and unfortunately now it seems to have got the stamp of approval from a seven-judge bench. Though Justice Chandrachud’s opinion is silent on the creamy layer issue, he does invoke Jarnail Singh with approval to make the point that the SC/STs are not a homogeneous group. Whereas Justice Gavai’s opinion is more explicit about the approval for the creamy layer, and quite interesting that it comes from the only Dalit judge on the seven-judge bench, that creamy layer needs to be excluded.

But my difficulty with creamy layer exclusion is that why are you providing reservations to scheduled castes and scheduled tribes? I feel that it is as much a matter of representation of Dalits. And it is not necessarily a matter of just distribution. It’s not just a distribution question, right? It’s not a redistribution question. It is a question of “do we have sufficient representation of Dalits either in services or in higher education, wherever it is?” And from that perspective, I would believe that what really matters, it should be a concern of prioritization within the SC/STs, right? In the sense that people who are more, relatively more vulnerable and more marginalized and more backward within the SC/STs must certainly get preference. And if they do not get, and if they do not exercise that preference for whatever reason, right, then it must go to those who are relatively better off amongst the SC/ST.

Because I would argue that what you’re trying to do with SC/ST reservations is ensure a certain amount of representation in public employment and higher education. It’s not just a redistributive measure, right? And since it’s not just a redistributive measure, it is important to ensure that there is a fair degree of representation from the SC/ST communities. Because if you exclude all of what you call the ‘creamy layer of SC/ST,’ and if people don’t have the ability to access reservations, the result is going to be that the more marginalized and the more vulnerable don’t have the ability to make use of the reservation benefits. And as a result of that, you’re going to completely exclude SC/STs from higher education institutions or public employment. And it almost achieves the opposite effect by which why we brought in SC/ST reservations to say that a share in education institutions, a share in public employment, must stay with the SC/STs, right?

So, creamy layer, I think the judgment gets it quite wrong, but unfortunately caps off almost two decades of jurisprudence on this, which has been incremental to say that SC/ST creamy layer has to be excluded. It’s quite unfortunate, but, yeah. So I’d say a mixed bag, really, on resolving the normative incoherence in the Supreme Court’s reservation jurisprudence.

Spandana: Yes. But one thing the judgment sort of gets right across opinions is the fact that within the SC/ST groups, there are varying degrees of diversity, discrimination, marginalization, et cetera. And the chief’s opinion also calls for a qualitative representation in tune with your argument for substantive representation as well within the beneficiary groups. But what do you think are some practical issues in achieving this? Like, how can states, for instance, ensure substantive equality within their sub-classification policies?

Surendranath: Yeah, it’s a tough one, Spandana. You know, fundamentally, the challenge to the approach that the states have taken, right? That, okay, let’s say state ‘X’ has 40 SC/ST groups within that, in the 1950 presidential order, there are SC/ST groups identified for each of the states, right? Now, let’s assume that in the 1950 presidential order, state X has 40 SC/ST groups, right? What are states doing as of now? States are saying, ‘okay, let’s group these 40 into, let’s say, four subgroups,’ right? Let’s say they are divided into four subgroups, and the 15% reservation that is available for SCs is then split in different proportions amongst these four groups, right? Subgroup A, subgroup B, subgroup C, subgroup D, and the 40 scheduled caste groups are then divided in these subgroups. And then you’re saying, okay, one group will get 5%, another group will get 2%, another group will get 8%, whatever.

Now the problem here is, in adopting this group-based sub-classification approach, the same problem that afflicts SC/STs as a larger group as a whole will afflict the subgroup also, right? Just like there is a spectrum within all SC/STs as a whole, within that subgroup that you have created, that comprises different caste groups, there will again be relative backwardness and some groups being relatively well off. So you have just brought it one level down, right? So just like SC/STs as a whole have a problem because caste discrimination exists on that spectrum, and so there is internal discrimination and marginalization within SC/STs, the same will operate within the subgroups as well, right?

And therefore, that’s why I argued in my thesis that the basis for identification as a beneficiary group, right? Can be group-based, right? That it is a group-based characteristic, the common. Just like there are differences amongst the SC/ST groups, there are also commonalities within the SC/ST groups. And you use that commonality to create the legal category of SC/STs. At the end of the day, the SC/STs is a legal category. It is not a pre-existing social category, right? It is a creation of the law, right? And in the sense that you’re saying, okay, there is a connecting characteristic historically of untouchability. And because of that connecting characteristic, despite the differences, we have clubbed them into a legal category called SC/STs, right?

So my argument was, while that is absolutely correct and fine for identification as beneficiaries, you’re looking at group characteristics. There is no need to adopt the same basis, that is a group-based approach, to the internal distribution of reservations, right? I argue for a much more, and I’m not the first one to argue, my idea is from people who argued it before, that individual-based assessment of need or of marginalization should be adopted for distribution of benefits in the sense that if there are, if I’m looking to give SC/ST reservation to universities and I have a hundred applicants for 25 positions, I will say, “okay, what are the grounds of marginalization for each of these hundred applicants?” Right?

So candidate A, I mean, all of them will be belonging to SC/ST groups, but somebody might be from an extremely poor background. Somebody might be a woman, right? Somebody might have gone to a government school. So you allot. And, and JNU has a good system for this, right? In terms of adopting individual points for different axes of marginalization, right? Which district did you go to for your schooling? And where is the district with the official classification, in Indian Government data, right? So you have different axes of marginalization and you get different points for different axes of marginalization. And depending on your points, the priority to reservation depends on that, right?

So if I have 45 points arising out of different axes of marginalization and you have 30, does it mean that you are completely excluded? It only means that I have priority over you, right? It only means that I have priority over you. That it is important. It is important not to exclude what is known as creamy layer SC/STs because that non-exclusion is recognition of the fact that social discrimination affecting SC/ST groups in this country has not been cured, right? We should exclude people completely. We should exclude SC/STs, and we can start to exclude SC/STs from reservations only, and if, and only when, that kind of discrimination that Dalits face in India is no longer a reality.

As long as that discrimination, that social discrimination is a reality, creamy layer amongst the SC/STs makes no real sense, right? Because of the fundamental point: Just because I am rich does not mean that my social discrimination is cured. Economic mobility is no cure for social discrimination, right? And that’s not so difficult to grasp, even in the context of gender. A rich woman continues to face gender discrimination. It’s not that because you’re rich, your social discrimination suddenly disappears. But are you relatively better off than a poor woman? Yes, maybe. And therefore, or at the same thought, even on grounds of sexuality, right? On homosexuality. Just because you are rich doesn’t mean rich homosexuals are not socially discriminated against.

So similarly, I would say that what we really needed to provide for is not the exclusion of certain SC/ST groups, but an internal prioritization based on individual criteria, which would then decide who has more, who has higher priority to access reservation than others within the SC/ST group. And what we really needed to do was develop that metric of prioritization rather than exclusion as has been achieved by this judgment.

Spandana: Yeah, sure. Thank you for that detailed answer. I had several sub-questions, but I think your answer pretty much covers it all. So I think I want to broadly ask you, do you think that judgment largely strikes a balance between recognizing diversity within the beneficiary SC/ST groups on the one hand, and avoiding further fragmentation within the group so that those who need it most are not really excluded? What could the court have done better? Of course, one thing we just discussed is to adopt a more individual-based approach. But from the court’s perspective, what could it have avoided? What could it have done better?

Surendranath: Yeah. And I think, I mean, just before I get into that, one good thing that it does is that it takes away this cadre-based basis for saying that under 16 (4), you need to show inadequacy of representation. You refer to the text of 16 (4). Inadequacy of representation is a requirement, but it has been a long-standing controversy of where you need to show this inadequacy. Can I say that the entire state services of state X backward classes are inadequately represented? That is not the case. In Nagaraj, they said that certainly is not the case. You have to go to the cadre. You have to show cadre-level inadequacy.

What does that mean? That if I were hiring for postal delivery persons, I will have to show inadequacy of representation within that cadre of postal delivery persons. That was a position that has emerged in Nagaraj, and I think Davinder Singh now does away with that and quite rightly understands the reality that, you know, what really matters is that within your group, your relevant group services, right? Then every state service, even at the union level or at the state level, there are group A services, group B, group C, group D services. Depending on the benefits, remuneration, qualification, there is a hierarchy of services, and the judgment quite rightly notes the reality that reservations have ended up filling up the lower level services without filling up the group A and group B.

So therefore the relevant metric for assessing whether there is adequate or inadequate representation is the subgroup of the services, right? Is it a group A service, group B service, group C service, or group D service? So if you want to provide sub-classification of scheduled caste within group C services, you have to use group C as the metric to show that there is inadequacy, right?

So, what else could it have done, right? I mean, it’s… I would say it’s a sort of frustrating level of balance. While you can’t quibble with it too much, neither is it very satisfying, right? In the sense that I would have assumed that it should have gone a little deeper on the normative incoherence that it is creating by adopting a group-based approach to sub-classification. And it then raises this valid point of setting off one group against another subgroup amongst the SC/STs, right? And I think it doesn’t pay enough attention to that. Neither does it pay enough attention to the fact that is this what is best for substantive equality, right?

And therefore, and I think various authors have written about the dangerous consequences that this judgment could have towards the solidarity amongst Dalit groups, right? That it could trigger off a contestation for more internal conflict within Dalit groups. And I think, and for reasons I’ve explained earlier as well, from a substantive equality perspective as well, group-based sub-classification does not really ensure more substantive equality outcomes either.

But yes, but yes, very, very good in terms of acknowledging that there is… that despite the diversity, the group is homogeneous in certain ways. The SC/STs as a group, there are certain elements of homogeneity, commonality, and there are certain elements of heterogeneity, and that there can be elements that can be measures taken to address that heterogeneity. It’s great till then, but how to address that heterogeneity? I think it gets slightly wrong both on saying, ‘oh, we need a creamy layer to address that heterogeneity,’ and on a group-based approach to sub-classification, on the method of curing or addressing this heterogeneity. I think it gets it wrong.

Spandana: Yes, one sort of possible solution or criteria that the majority opinion lays down for sub-classification is that it should be based on empirical data. And we spoke about this briefly before, and the judgment says that it is important for this classification to be based on “rational and reasonable” basis. Now, do you think this is a standard sufficient enough to prevent potential misuses in policy and legislations?

Surendranath: Yeah. Look, this is a much broader conversation, Spandana, on the court’s approach to empiricism, right? And the court has very different approaches to the rigors of empiricism required. When we are talking of SC/STs, Prevention of Atrocities Act of 1989, or other issues of other legislations, the court seems to have a very problematic approach to empiricism in those contexts. But when it comes to reservations, the court seems to have extremely difficult and rigorous standards and unclear standards on what it wants out of this empiricism. And look, I’m not saying that it’s a good or a bad thing, and I think… but then the court must have a very clear approach to what it wants out of this. What does it want to be shown? Right.

And I think that’s not clear from the judgment as to what it exactly wants shown. It’s saying in this judgment that if you want to undertake sub-classification, you have to show two things. You have to show that the group for whom you are, or the groups for whom you are sub-classifying suffer from a significant social backwardness that is different from other subgroups within SC/ST. So you have to show that sufficient distinction in social backwardness. And not only do you have to show social backwardness, but you also have to show that social backwardness is causing inadequacy of representation in public employment, right? Within whichever group of services you are looking at, right? So it’s a twin test, and you can’t show these two things independently. You can’t, on the one hand, show that, yes, the groups for whom I’m sub-classifying are significantly more socially, and they are also inadequately represented. The judgment is clear in saying you can’t show these two things independently. You have to show that inadequacy is caused by this significant difference in social backwardness. Now, that showing of causality is again, it doesn’t tell you what it means to show this causality in empirical terms. It doesn’t tell me what it means to show this causality in empirical terms, right?

Now that will be used in judicial review when such a measure comes up at some point, or when even these measures go back to adjudication, whether these are legal, as per the terms of the judgment of Davinder Singh, that requirement of causality. Even if you say, yes, you have said, I will show group A is significantly more socially backward or that they are inadequately represented. But the burden is not just of showing those two things independently. It is by showing the causality, that backwardness, the relative backwardness of one subgroup is what is causing its inadequate representation.

Now that seems almost like a standard that is set up to fail, right? It is set up to fail in the sense that, you know, like they say about strict scrutiny, right, in the US. That strict scrutiny is strict in theory and fatal in fact. Right? It is fatal in fact, because every time you apply it, it almost seems sure shot that the measure will fail, right? Because the standard is so high, nothing can meet it, right? So I would be very interested in seeing how this causation requirement plays out, or how the judgment plays out.

Spandana: Yeah. Well, lastly, how do you think this decision broadly shapes the affirmative action discourse in our country? And what are the constitutional and political implications of this? What do you think are the constitutional and political implications of this decision to uphold sub-classification?

Surendranath: Yes. So apart from that, I’ve spoken, I think, sufficiently in detail about the efficiency question and the creamy layer question—one good, one bad. But, and I think there’s another tension that the judgment looks to resolve, is if you… I mean, again, it’s… we won’t have time to do this now. If you trace how the Supreme Court viewed the relationship over many, many decades between reservations and merit? It’s a deeply problematic, confusing one, right? Including in Indra Sawhney, right? And Justice Chandrachud’s opinion is cognizant of that, he tries to resolve that tension in India’s judicial discourse on the relationship between merit and reservations.

Has the Indian Supreme Court understood reservations as anti-merit? Justice Chandrachud has a valiant attempt at showing that, no, it doesn’t understand it as anti-merit, and he tries to reconcile that tension. But I think it’s a mountain that’s a bit too high for him to climb in a seven-judge bench. Because I think the tension in Indra Sawhney as a nine-judge bench is obvious, that if you… That there is significant material in Indra Sawhney to suggest that in certain contexts, they view reservations as anti-merit in terms of reservations and promotions, which Indra Sawhney didn’t approve of, and also reservations for certain categories of posts.

And all of that, it said, you know, reservations are a cost the society has to pay. So therefore, in these things, in these areas, reservations are not allowed. That tension is inherent in Indra Sawhney. And by virtue of that being a nine-judge bench, the seven-judge bench, it’s a mountain that’s a bit too high for it to climb. And I guess we’ll have to just wait for another bench, another matter of a higher bench strength to resolve that tension in India’s judicial discourse and reservations.

But overall, I think, and I think when the judicial legacy of Justice Chandrachud is written, a significant amount of space will be dedicated to his reservation jurisprudence. And I think this will add to that legacy—some very important moves made. And everything might not be to our liking, or we might have taken it further, but there are significant gains in this judgment that I have said. But whether this will continue is something that we just have to wait to see.

And there are many areas of reservation jurisprudence that lack this coherence. And I think it will come back to repeatedly bite us, and I think we’ll resolve some of those questions favorably. And just like this judgment has, and some questions, but this depends on those contexts. But I think it’s very important to see that. I was very interested in the political reaction to the creamy layer issue, though the judgment says creamy layer has to be excluded.

The government was quick to clarify that nothing of that sort is happening, but parties were against it, saying that that creamy layer exclusion of SC/STs is not going to happen. So for me, a very interesting dynamic, that there are spaces outside the judiciary where some of these democratic battles can be won, and that the judiciary might not be the best space to resolve this. So it was very interesting for me on the creamy layer issue. While the issue seems to have been lost in court, on the ground, in politics, it seems to have been protected to some extent. So that was a very interesting dynamic.

Spandana: Well, that brings me to the end of my questions. Thank you so much for taking the time today, Anup!

And for our viewers, Anup has also written a commentary analyzing the court’s decision on SCO’s analysis page. So if you haven’t already, please check that out. For a detailed background, hearing reports, judgment summaries, matrices, explainers, go to our cases page on scobserver.in. And as always, for more stories from the Supreme Court, stay tuned to SCObserver. Thank you!

Surendranath: Thank you!