Challenges of executing EWS reservation efficiently

  • Blog Post Date 14 December, 2022
  • Perspectives
  • Print Page
Author Image

Orhan Aygün

Boğaziçi University

Author Image

Bertan Turhan

Iowa State University

A recent Supreme Court judgement excluded members of reserved categories from the scope of EWS. Aygün, Turhan, and Yenmez look at the implications of this decision, including reserved category members having to choose between applying for positions on the basis of their caste or income, and the ambiguity about how EWS reservations are defined. They corroborate their findings with examples of recent court cases, and highlight the issues likely to arise with implementation, especially de-reservation. 

India has been implementing a sophisticated affirmative action programme through vertical and horizontal reservations for decades, in admissions to publicly funded educational institutions and recruitment for central government jobs. Prior to the 103rd Constitution Amendment Act in 2019, vertical reservations were provided only to historically discriminated Scheduled Castes (SC), Scheduled Tribes (ST), and Other Backward Classes (OBC), by earmarking 15, 7.5, and 27% of positions to these groups, respectively. SC, ST, and OBC are collectively referred to as reserved categories, while individuals who do not belong to reserved categories are called General Category (GC) members.

A crucial feature of vertical reservations was that no individual qualified for more than one reserved category – that is, vertical reservation categories were non-overlapping – until the 2019 amendment. The remaining unreserved positions (referred to as open category) were available to everyone, including reserved category members. As mandated by the Supreme Court’s judgement in Indra Sawhney and Others vs Union of India (1992), vertical reservations have been implemented as over-and-above. That is, positions taken by the reserved category members based on merit from open category are not counted against their quotas.

Meanwhile, horizontal reservations are considered secondary and implemented within each vertical category on a minimum guarantee basis. That is, individuals eligible for horizontal reservation who secure positions based on merit are counted against horizontally reserved positions.Since almost all of the policy-related debates related to the 103rd Amendment revolve around vertical reservations, we do not discuss horizontal reservations in this article.

Reservation based on economic criteria

The 103rd Amendment introduced 10% percent reservations for the ‘Economically Weaker Sections’ (EWS) of people who are not covered by any of the existing vertical reservations. The amendment was announced on 14 January, 2019. Shortly after, it was challenged by different groups on the grounds of violations of the Constitution’s basic structure and the doctrine of equality. The five-judge Constitution Bench of the Supreme Court of India announced its verdict on 7 November, 2022, on the following issues, among others:

  1. Whether the 103rd Constitution Amendment can be said to breach the basic structure of the Constitution by permitting the State to make special provisions, including reservation, based on economic criteria?
  2. Whether the 103rd Constitution Amendment can be said to be in breach of the basic structure of the Constitution by excluding reserved categories from the scope of EWS reservation?
  3. Whether the cap of 50% referred to in earlier decisions of the Supreme Court is considered to be a part of the basic structure of the Constitution? If so, can the 103rd Constitution Amendment be said to breach the basic structure of the Constitution?

The judges unanimously agreed that reservations could solely be based on economic criteria. In a controversial verdict, with a 3:2 split, the Supreme Court approved the exclusion of reserved categories from the scope of EWS. The majority of judges stated that low-income GC applicants form their own disadvantaged group. The dissenting judges, on the other hand, noted that reserved category members form the majority of the poorest in India, and therefore should qualify for EWS. Moreover, the majority judges considered the 50% limit to be flexible and apply only to reserved categories, whereas the dissenting two worried that breaching the limit here would be a gateway for further categorisation. 

Choosing between exclusive vertical reserved categories

The Supreme Court’s controversial decision to exclude reserved category members from the scope of EWS will potentially lead to widespread confusion and ambiguity. We identify two important issues that can (and will likely) arise with the implementation of the EWS reservations.

The first issue stems from the fact that reporting reserved category membership is optional. When reserved category members do not reveal their caste status, they are considered GC members. Once they are deemed GC members, they can be eligible for the EWS quota if they obtain their income eligibility certificates. Deshpande and Ramachandran (2019) report that, with the current income eligibility level, more than 98% of the Indian population qualifies for the EWS reservations.

The majority judges’ statement in the court’s verdict defines the EWS as an exclusive vertical category, whose beneficiaries are the low-income GC members. By excluding the reserved category members from the EWS, the court aims to preserve the non-overlapping structure of vertical reservations. However, because the declaration of caste status is optional, there may still be an overlap unless every applicant considered for an EWS position is checked further to see whether they may belong to a reserved category. Such a checking mechanism would be burdensome.

As a result, members of reserved categories may face a choice between obtaining an income certificate and a caste certificate under the current status of the amendment. Reserved category members may even hurt their chances by declaring their caste status, because they may not be receiving a position on the basis of their caste; but instead, if they get an income certificate, they may be able to receive an EWS position.A similar issue with the implementation of horizontal reservations persisted for decades, and was only recently resolved in the case of Saurav Yadav vs The State of Uttar Pradesh (2020).3

Even when obtaining EWS positions is more competitive than reserved category positions, a reserved category member may still prefer obtaining a low-income certificate to receive an EWS position over a caste-based reserved category position. Aygün and Turhan (2022a) present ample evidence on reserved category members expressing preferences between open category and reserved category positions. Many reserved category members opt out of less competitive reserved category positions by not disclosing their caste, and compete only for open-category positions. This behaviour is well-documented in the prior literature (see Gille 2013), as well as in court cases.

In a recent Bombay High Court case (Shilpa Sahebrao Kadam And Another vs. The State of Maharastra (2019)), it was highlighted that the petitioners did not report their membership to reserved categories, and instead sought assignment through open category positions:

The petitioners contend that though they belong to reserved category, they have filled in their application forms as a general category candidates and have not claimed any benefit as a member belonging to reserved category.”

In another case from Bombay High Court (Vinod Kadubal Rathod And Another vs. Maharashtra State Electricity (2017)), two petitioners, who are members of reserved categories, applied to technician positions without reporting their caste status:

The petitioners belong to VJ (A) Category. However, they applied for the said post from General Category and as such, did not claim the benefits of reservation meant for VJ (A) Category.”

The strategic non-disclosure of reserved category membership may lead to unfair and less meritorious outcomes. To see this, consider an OBC member Ms. A. Suppose Ms. A does not disclose her OBC membership, but instead obtains a low-income certificate to receive an EWS position. Her non-disclosure of OBC membership may help a lower merit score OBC member Mr. B to receive an OBC position. If Ms. A disclosed her OBC membership, she would become ineligible for EWS and receive an OBC position instead. Therefore, the EWS position she previously held would be taken by a GC member Ms. C who may have a higher merit score than Mr. B.

One possible solution is to ask individuals to report their preferences over institution-vertical category pairs.4 When applicants are asked to submit their rankings over institution vertical category pairs, one can design an allocation procedure where each applicant finds it optimal to submit their rankings truthfully. Moreover, eliciting individuals’ preferences in this expanded domain makes the order in which categories are filled immaterial.

Implementation of EWS reservations: hard or soft cap?

The second issue of the 103rd Amendment is that it is unclear whether the EWS is a hard or soft reservation.5 In the case of hard reservations for a category, leftover vacancies are not provided to individuals who do not belong to that category. According to Indra Sawhney’s 1992 judgement, SC and ST reservations are hard reservations.

On the other hand, OBC reservation in admissions to publicly funded higher educational institutions is a soft reservation. In its historic judgment in Ashoka Kumar Thakur vs Union of India (2008), the Supreme Court required that unfilled OBC vacancies be made open category positions, without specifying a well-defined procedure. The lack of specific guidance leads to ad-hoc procedures to disentangle the OBC de-reservation. One such procedure was implemented in admissions to technical colleges, including the prestigious Indian Institutes of Technology. In a recent study, Aygün and Turhan (2022c) identified that this particular de-reservation scheme may lead to inefficiencies, unfairness, and manipulability.6

Policy lacuna and implications

If the EWS reservation is a soft reserve – which we believe it is, even though it was not clearly stated – then how EWS de-reservation is implemented will be consequential. Especially in centralised admissions (such as the admissions to technical universities in India, where the celebrated Gale and Shapley (1962) algorithm7 is used) a well-defined procedure to handle de-reservation is necessary. Otherwise, this ambiguity in implementing the EWS reservations may be susceptible to favouritism, as different institutions could implement them differently to favour members of different categories.

Given the current status of the EWS amendment, the two issues we outlined above may lead to widespread implementation issues, most likely to result in judicial inquiries. Therefore, it is crucial to address these issues while the issue of reservations is still highly debated and divisive. For instance, market designers have been studying assignment problems under affirmative action constraints for more than a decade.They can inform policymakers about the implementation issues, and provide well-defined procedures to alleviate these concerns.

The authors are grateful to Parikshit Ghosh, the Editor-in-Chief of Ideas for India, for his extensive comments and suggestions. They would also like to thank Jenna M. Blochowicz, Aditya Kuvalekar, Kriti Manocha, and Arunava Sen for their helpful comments.

This post is the second in a three-part series on 'The architecture of affirmative action'.


  1. Hafalir et al. (2013) was the first to introduce reservation in market design literature. Choice rules based on the minimum guarantee principle were first introduced in Echenique and Yenmez (2015).
  2. Aygün and Bó (2021) show a similar problem with the implementation of the affirmative action policies in Brazil.
  3. Sönmez and Yenmez (2022) identified this shortcoming before the Supreme Court’s decision and provided a solution based on the joint implementation of horizontal and vertical reservations.
  4. Aygün and Turhan (2022a) introduce the idea of enlarging the language in which applicants express their preferences.
  5. See Hafalir et al. (2013) for details of hard and soft quotas.
  6. The authors proposed a better way to jointly implement reservations and de-reservations – in Aygün and Turhan (2022b), they introduce multiple de-reservation procedures with desired theoretical and practical properties.
  7. The Gale–Shapley algorithm finds a stable matching between students and colleges, or men and women.
  8. For instance, a group of market designers and engineers designed and implemented school assignment procedure under affirmative action procedures in Chile.

Further Reading

  • Aygün, Orhan and InácioBó (2021), “College admission with multidimensional privileges: The Brazilian affirmative action case”, American Economic Journal: Microeconomics, 13(3): 1-28.
  • Aygün, O and B Turhan (2022a), ‘Affirmative action in India: Restricted strategy space, complex constraints, and direct mechanism design’, Iowa State University Working Paper.
  • Aygün, O and B Turhan (2022b), ‘Affirmative action in India via backward and forward transfers’, Iowa State University Working Paper.
  • Aygün, Orhan and Bertan Turhan (2022c), “How to de-reserve reserves: Admissions to technical colleges in India”, Management Science, forthcoming.
  • Deshpande, Ashwini and Rajesh Ramachandran (2019), “The 10% quota: Is caste still an indicator of backwardness?”, Economic & Political Weekly, 54(13): 27-31.
  • Echenique, Federico and M. Bumin Yenmez (2015), “How to control controlled school choice”, American Economic Review, 105(8): 2679-2694.
  • Gale, David and Lloyd Shapley (1962), “College Admissions and the Stability of Marriage”, American Mathematical Monthly, 9(1), 9-15.
  • Gille, V (2013), ‘Stigma in positive discrimination application? Evidence from quotas in education in India’. Available here.
  • Hafalir, Isa E, M. Bumin Yenmez and Muhammed A Yildirim (2013), “Effective affirmative action in school choice”, Theoretical Economics, 8(2): 325-363.
  • Sonmez, Tayfun and M. Bumin Yenmez (2022), “Affirmative action in India via vertical, horizontal, and overlapping reservations”, Econometrica, 90(3): 1143-1176.
No comments yet
Join the conversation
Captcha Captcha Reload

Comments will be held for moderation. Your contact information will not be made public.

Related content

Sign up to our newsletter