Introduction to e-Symposium: The architecture of affirmative action

  • Blog Post Date 12 December, 2022
  • Symposium
  • Print Page
Author Image

Parikshit Ghosh

Editor-in-Chief, I4I; Delhi School of Economics

The Supreme Court of India recently upheld an amendment that excluded Scheduled Castes, Scheduled Tribes and Other Backwards Classes from the Economically Weaker Section (EWS) quota, restricting it only to general category applicants. However, the specifics of how this reservation policy is executed can have important social and political implications. 

Across this week, from 12-16 December, this I4I e-Symposium brings together articles that provide a theoretical basis – using principles of market design, and search and matching theory – for more efficient implementation of reservation policies. Anchored by I4I’s Editor-in-Chief Parikshit Ghosh, the e-Symposium aims to open a discussion on the architecture of affirmative action, from the mechanisms of vertical and horizontal reservations, to ensuring efficiency in meeting diversity targets.

The authors of the Indian Constitution had the wisdom to see that our tryst with destiny will be unfulfilled if we do not confront the ghosts from our past. Even as Articles 14 and 15 pronounced equal treatment for all, Article 15(4) paved the way for reservations targetted at socially disadvantaged groups. The founders of the Indian republic understood that a newly independent nation had a historic opportunity to not only break the shackles of colonialism, but also oppression in all its forms. A narrow, ahistorical notion of meritocracy did not suit this mandate. 

Still, after more than seven decades of experience, questions swirl around our reservation policy. Who deserves protection? When should it be withdrawn? Is social disadvantage synonymous with economic deprivation? Grappling with these difficult issues requires not only input from the social sciences, but also an engagement with ethics and politics. Unlike the design of airports or the sale of spectrum, this is an area where the public interest cannot entirely be left to academics and bureaucrats. 

However, affirmative action does not involve only the setting of diversity targets – which is fundamentally an expression of democratic will – but also calls for the design of concrete institutional rules to achieve these targets with the least sacrifice of the meritocratic ideal. Should general category seats be filled before the SC/ST seats or vice versa? If an OBC candidate with disability is recruited, should it count towards fulfilling both the OBC and disability quotas, or just one of them? How exactly these finer points are settled can be profoundly consequential, as economists have learnt from several decades of research on market design (Roth 2007)

While affirmative action targets have been well articulated by legislatures, the rules for implementing them have been left ill specified, requiring courts to step in time and again. Many landmark judgments of the Supreme Court are attempts to reduce the confusion and conflict arising from procedural ambiguity. 

Unfortunately, this design aspect of reservation policy, what I call the architecture of affirmative action, has not only received scant attention in the media and public debate, but its importance seems to go largely unrecognised. Our aim with this e-Symposium is to start that conversation. 

In Indra Sawhney vs. Union of India (1992), the Supreme Court mandated the earmarking of certain positions for caste-based categories (like SCs, STs and OBCs) – what has come to be known as vertical reservation – but left the fulfillment of diversity targets for other categories (such as persons with disabilities) more flexible – an arrangement referred to as horizontal reservation. In the opening article of this symposium, Ashutosh Thakur revisits this issue and provides a critique of vertical reservations. Among other things, it has no built-in sunset clause and requires legislatures to continuously revise quotas as disadvantaged groups economically catch up with others. 

The next two articles come from researchers who have studied how to devise efficient ways of meeting diversity goals, as well as matching two sides of a market (for example, assigning students to schools or colleges) in a sensible way. In the second article of the series (their first), Orhan Aygun, Bertan Turhan and Bumin Yenmez point out that though the five judge bench upheld restricting the Economically Weaker Section (EWS) quota to general category applicants, SC/ST/OBC candidates could still make themselves eligible for these positions by not declaring their caste identity, and explore the implications of such a loophole. 

The final article examines the process through which rank holders from the joint entrance examination (JEE) are assigned to the various Indian Institutes of Technology (IITs) and other technical colleges. The assignment must respect student merit ranks, their stated preference over institutions and programmes, and the quota requirements within each institution. In addition to that, the judgment in Ashok Kumar Thakur vs Union of India (2008) stipulates that unfilled OBC quota seats (but not SC/ST quotas) should be made available to general category applicants to reduce wastage. This is clearly a complex task.   

The system currently in place was designed by the government, in consultation with a group of computer scientists and market designers (Baswana et al. 2019). It is based on the celebrated Gale-Shapley algorithm1 and tries to ensure that within the constraints of the diversity requirement, the allocation is fair and efficient. Many readers may be unaware that a rare confluence of legislative will, judicial oversight and technocratic finesse has designed the staircase to success so many Indians aspire to step on. Yet, as Aygun, Turhan and Yenmez point out through simple and illuminating examples, when it comes to de-reserving unfilled OBC seats, the current system has subtle flaws that can and ought to be corrected. 

After 75 years of Independence, we can take some pride in our quest for an affluent and just society, yet be vigilant about the gaps in that attempt and strive to bridge them. 

Design choices for implementing affirmative action

Ashutosh Thakur

Ashutosh Thakur explains the various ways in which affirmative action policies can be implemented, and discusses the underlying trade-offs and issues at hand...

Challenges of executing EWS reservation efficiently

Orhan Aygün, Bertan Turhan, M. Bumin Yenmez

Aygün, Turhan, and Yenmez look at the implications of reserved category members having to choose between applying for positions on the basis of their caste or income...

Improving admissions to technical colleges in India

Orhan Aygün, Bertan Turhan, M. Bumin Yenmez

Aygün, Turhan, and Yenmez examines the process through which JEE rank holders are assigned to the various IITs and other technical colleges...


  1. The Gale–Shapley algorithm is an algorithm used for finding a solution to the stable matching problem, and has been described as solving both the college admission problem and the stable marriage problem.

Further Reading 

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