Horizontal Reservations and the Persistence of the Myth of Merit

Editor’s Note 1Posts about the contemporary Supreme Court may be read in the context of the caveats set out in this post (link).

Editor’s Note 2: Justice is an indivisible concept. We cannot, therefore, discuss contemporary Supreme Court judgments without also acknowledging the Court’s failure – at an institutional level – to do justice in the case involving sexual harassment allegations (link) against a former Chief Justice. This editorial caveat will remain in place for all future posts on this blog dealing with the Supreme Court, until there is a material change in circumstances (e.g., the introduction of structural mechanisms to ensure accountability)].

On 18th December, a three-judge bench of the Supreme Court handed down an important judgment clarifying the law on vertical and horizontal reservations (Saurav Yadav v State of UP). The judgment is an extremely complex one, so a preliminary explanation is in order.

Horizontal and Vertical Reservations

The term vertical reservations refers to reservations enacted under Articles 15(4) and 16(4) of the Constitution, which divide a candidate pool into the General, OBC, SC, and ST categories. While this division is not stipulated by the terms of the Constitution (Articles 15(4) and 16(4) use the phrase “backward classes” or “socially and educationally backward classes in addition to SCs/STs), it has come to be encrusted in both law and judicial decisions over the decades.

Now, until the judgment of the Supreme Court in N.M. Thomas, Articles 15(4) and 16(4) were treated as the sole repositories for reservations under the Constitution. That is, the “equality code” – Articles 14-16 was believed to embody a formal vision of equality, and Articles 15(4) and 16(4) were believed to embody specifically carved out exceptions, that allowed for reservations. Sporadic attempts to justify reservations under – for example – Article 16(1)’s equality of opportunity clause, by invoking a substantive vision of equality, were struck down the courts (albeit over some important dissents). However, that changed in N.M. Thomas, when the Supreme Court clarified that Articles 15(4) and 16(4) were not exceptions to, but restatements of, the equality principle contained in Articles 15(1) and 16(1). This was a seismic shift in the jurisprudence, because in the course of one judgment, the Court went from holding that the Constitution was committed to bare formal equality, to holding that it was committed to substantive equality that must account for actually existing differences between people. While judgments after N.M. Thomas have chipped away at its scope, and blocked its transformative potential from truly taking shape, the doctrine itself has, nonetheless, been re-affirmed.

Now, a necessary corollary of Thomas is that reservations are no longer solely facilitated by Articles 15(4) and 16(4). If 16(4) (reservations), for example, is a restatement of 16(1) (equality of opportunity), it must follow that genuine equality of opportunity permits (and may even require) reservations. Consequently, government policies can provide reservations for groups that do not follow the General-OBC-SC-ST division, as long as it is in service of genuine equality of opportunity. Now this is what gives rise to horizontal reservations in favour of (for example) women (where Article 15(3) is also involved), army veterans, the transgender community (although that has not happened yet), and so on. The reason why this set of reservations is called “horizontal reservations” is because it cuts across vertical reservations, in a manner of speaking. Imagine a graph divided into four vertical bars (General, OBC, SC, ST), in accordance with their respective quotas. Now imagine a horizontal bar (“women”) that cuts across these four vertical bars (because women can belong to any of the four categories mentioned above, and the horizontal reservation in question is for “women” as a class).

The Question in Saurav Yadav

Because of the way our reservations jurisprudence has shaped up, the intersection of horizontal and vertical reservations throws up a host of technical, complicated questions. Should horizontal reservations be calculated across the board (i.e., say, 30% of women in X department), or should it be “compartmentalised” (i.e., each category – General, OBC, SC, ST must have 30% of women)? The Supreme Court has, in the past, suggested that the compartmentalised approach would be better, and in my submission, that is correct, as it takes care of inter-sectional concerns. In Saurav Kumar, however, the question was even more technical: if State policy provided for an X% horizontal quota for women, then would women candidates whose overall scores were high enough for them to be recruited without the benefit of reservation, be counted against the quota or not? Or, in other words: let us assume that State policy requires 30% reservation for women. After the selection process, it is found that 10% of the successful candidates without the application of the quota are women. Does that now mean that the quota comes down to 20%, or does it mean that it remains 10% + 30%?

In the case of vertical reservations, this issue had long been sorted, with the Supreme Court making it clear that an SC/ST/OBC candidate whose scores were high enough for the general category, would be “counted” as having made it through that category; and consequently, that candidate would not be treated as SC/ST/OBC for the purposes of reservation. In the case of horizontal reservations, however, the position was unclear, as different High Courts had taken different views on the issue. In Saurav Yadav, the Supreme Court clarified the position, and held that – like the case of vertical reservations – a candidate who belonged to a horizontally reserved category, but whose score was high enough, would not be “counted” as having come through that category for the purposes of calculating quotas.


In my submission, this decision is correct. However, I believe that the reasoning deployed by the Court to get there is not. In the majority opinion, Justice Lalit entered into a complicated analysis of various hypothetical scenarios that would arise when both methods were deployed: i.e., counting the high-scoring candidate against the quota, and not. He found that in some cases, horizontal and vertical reservations would intersect in a manner that if the high-scoring candidate was counted against the quota, a number of adjustments would follow, the end result of which might be that a higher-scoring candidate in the overall list would be bumped off in favour of a lower-scoring one (in the interests of space, I don’t want to reprise the calculations here, that run to around thirty pages – interested readers can refer to the judgment for the steps involved in the process). Justice Lalit then found that if, on the other hand, in the same situation, the high-scoring candidate from the horizontally reserved constituency was not counted against the quota, the overall list would have more higher scoring members. He then concluded that the second method enabled “merit” to a higher degree, and was therefore the constitutionally correct way of implementing horizontal reservations. In a brief concurring opinion, with a caveat that I shall later come to, Justice Bhat agreed with this analysis.

In my view, however, there was a much simpler reason for the outcome. The rationale for why, in the case of vertical reservations, a high-scoring candidate is not counted against the quota, is a simple one: the bases of reservations are the historical and structural barriers that prevent certain groups of people from gaining equal access to educational or employment opportunities. It may be the case that a combination of privilege, luck, and work enables some individual members of disadvantaged groups to overcome these barriers; however, that does not change the underlying logic of institutional disadvantage – and remedy – that structures reservation policy. For this reason, the purposes of reservations would be defeated if the numerical quota included those individuals who had managed to overcome disadvantage.

Now, this logic applies in the same fashion to (most) cases of horizontal reservations: like vertical reservations, horizontal reservations are premised on achieving genuine equality of opportunity by accounting for existing structural and institutional disadvantage. Consequently, precisely the same logic of not “counting” individuals who overcome those barriers against the quota should apply. That is all that the Court needed to arrive at its (correct) conclusion.

Unfortunately, however, the Court’s reasoning – that framed the entire question as one of “merit” – was not only unnecessary, but actively problematic at a conceptual level: effectively what the Court said was that models of reservation would be selected based on which one – in a strictly formal and technical sense – was deemed to advance “merit” more than the other (and “merit” here refers to scores in tests). “Merit”, thus, became the normative touchstone on which to adjudicate between these different models. But this took an entirely uncritical approach towards the concept of “merit” itself, something that another bench of the Court recently warned against, in B.K. Pavitra’s Case. There, Chandrachud J. had correctly observed that:

The Constitution does not define what the framers meant by the phrase efficiency of administration. Article 335 cannot be construed on the basis of a stereotypical assumption that roster point promotees drawn from the SCs and STs are not efficient or that efficiency is reduced by appointing them. This is stereotypical because it masks deep rooted social prejudice. The benchmark for the efficiency of administration is not some disembodied, abstract ideal measured by the performance of a qualified open category candidate. Efficiency of administration in the affairs of the Union or of a State must be defined in an inclusive sense, where diverse segments of society find representation as a true aspiration of governance by and for the people. If, as we hold, the Constitution mandates realisation of substantive equality in the engagement of the fundamental rights with the directive principles, inclusion together with the recognition of the plurality and diversity of the nation constitutes a valid constitutional basis for defining efficiency. Our benchmarks will define our outcomes. If this benchmark of efficiency is grounded in exclusion, it will produce a pattern of governance which is skewed against the marginalised. If this benchmark of efficiency is grounded in equal access, our outcomes will reflect the commitment of the Constitution to produce a just social order. Otherwise, our past will haunt the inability of our society to move away from being deeply unequal to one which is founded on liberty and fraternity. Hence, while interpreting Article 335, it is necessary to liberate the concept of efficiency from a one sided approach which ignores the need for and the positive effects of the inclusion of diverse segments of society on the efficiency of administration of the Union or of a State.

In other words, therefore, the concept of “merit” – here constitutionally defined as “efficiency of administration” – could not be understood in isolation from the Constitution’s social justice goals, and, in particular, could not be reduced to something as simplistic and formal as scores in an exam (which, as we well know by now, themselves reflect and entrench social privilege). The Court’s approach in B.K. Pavitra is in stark contrast to its approach in Saurav Yadav where, as we have just seen, merit-defined-as-scores-in-the-exam was uncritically treated as the normative frame within which to examine competing models of reservation.

While Justice Bhat’s concurring opinion appeared, at various points, to reflect this concern, unfortunately, he too stressed the formal conception of “merit” as a normative frame. This conflict is summed up in paragraph 15 of his judgment, where he noted that:

I would conclude by saying that reservations, both vertical and horizontal, are method of ensuring representation in public services. These are not to be seen as rigid “slots”, where a candidate’s merit, which otherwise entitles her to be shown in the open general category, is foreclosed, as the consequence would be, if the state’s argument is accepted. Doing so, would result in a communal reservation, where each social category is confined within the extent of their reservation, thus negating merit. The open category is open to all, and the only condition for a candidate to be shown in it is merit, regardless of whether reservation benefit of either type is available to her or him.

While the first sentence followed up on Pavitra’s insight, the rest of the paragraph slipped back into the rhetoric of merit that structured the majority judgment. Now, as we have seen above, in this case, the philosophical difference did not lead to a difference in outcome. However, it is easy to imagine future cases in which it will, and past cases (such as Indra Sawhney’s affirmation of the 50% cap) where it already has. The future of reservation policy as a vehicle of achieving social justice will be significantly influenced by whether and to what extent the Supreme Court is willing to critically interrogate merit as itself being a function of social privilege that the Constitution is committed to equalising, and whether – on the other hand – it continues to hold on to a view that sees “merit” and reservations locked in a war with each other, with the adjudicatory function being to “balance” the two. The difference is a subtle one, but – as this post should demonstrate – a critically important one in the broader struggle for constitutional justice.