(Yet Another) Troubling Stay Order: The Supreme Court on the UGC Regulations, 2026

The Supreme Court’s order “keeping in abeyance” the 2026 University Grants Commission (Promotion of Equity in Higher Education Institutions) Regulations is yet another unfortunate example of hugely consequential judicial decisions that take the form of unreasoned interim or “stay” orders. The UGC Regulations – themselves the outcome of public interest litigation – had been issued earlier this month with a view to tackling discrimination (and specifically, caste-based discrimination) in higher education institutions (HEIs). These Regulations were challenged, primarily on the basis that the definition of “caste-based discrimination” was asymmetrical, in that it only envisioned Scheduled Castes, Scheduled Tribes, and Other Backward Classes being the victims of discrimination, while excluding Upper (“General”) Castes from protection.

To anyone with even a passing acquaintance of Indian constitutional doctrine, this argument is a non-starter (and that’s putting it kindly). Ever since the judgment in N.M. Thomas, delivered fifty years ago, substantive equality has been the bedrock of Indian equality and non-discrimination jurisprudence, especially when it comes to caste. For example, after N.M. Thomas, reservations are not thought of as exceptions to the rule of equality, but as integral elements of it, because their purpose is to address systemic and structural barriers that exist and operate asymmetrically and to the specific detriment of SC/ST/OBCs. Similarly, the Prevention of Atrocities Act operates asymmetrically – extending protection only to SCs and STs – because it understands the straightforward reality that caste atrocities are asymmetric. Nor is India alone in this: across the world, discrimination codes and hate speech codes are asymmetric, because they are built upon the understanding that discrimination is a function of power relations within society, and the ability of some to use their power to the detriment of others. Indian judgments themselves have built upon N.M. Thomas for the last fifty years, and it is too late in the day to suggest that a caste-discrimination code operating asymmetrically is unconstitutional.

Perhaps unsurprisingly, therefore, the Court records this submission of the petitioners, but does not base its stay order upon it; indeed, doing so would have entailed undermining by stealth (or what lawyers call sub silentio) the last five decades of Indian equality jurisprudence. That is not something the Court appears to be willing to do, or at least not yet. But then, on what basis does it take the very consequential decision of staying the Regulations? Significantly, the Court does not even advert to the well-established rules for granting a stay: that of (a) a prima facie case, (b) irreparable injury should a stay not be granted, and (c) the balance of convenience weighing in favour of the petitioners. As I have noted in the first sentence of this post, this is unfortunately part of a dominant trend now where the Court grants or refuses a stay without applying the relevant law in support of its determination (see here). This assumes even greater significance because of the delays endemic in our system, where a final decision is not going to be forthcoming for quite a while (for example, the next date of this case is now in March, and you can rest assured there will be no adjudication on that date either). This being the case, stays assume the character of permanence, or of a final judgment, thus making the need for a reasoned, in-depth order even greater. Unfortunately, however, the Court appears to be going in the opposite direction – or, to quote its own language in the order – “regressing.”

In the absence of legal reasoning, we must extrapolate. In paragraph 5, the Court says that some provisions of the Regulations “suffer from certain ambiguities, and the possibility of their misuse cannot be ruled out.” Some might argue that this is the basis of the stay. But it cannot be so: time after time after time – and especially in civil rights cases – the Court has said that the “possibility of misuse” has nothing to do with the constitutionality of a provision, and the remedy for abuse is to approach the authorities, instead of a declaration of unconstitutionality. Assuming that the Court is acting in good faith, it cannot be that the “possibility of misuse” is not a ground in civil liberties cases, but suddenly becomes a ground in this case.

The Court then says that based on its prima facie analysis, there are certain “substantial questions of law” that have arisen, and which must be examined. Alright. However, as noted above, the mere existence of substantial questions of law is no ground for a stay; if that was the case, then – just to take one example – all of the Aadhaar Act and Rules should have been stayed for all the years that it took the Court to adjudicate their constitutionality. Of course, the Court did no such thing. It would also mean – in future – that if a petitioner can demonstrate that a constitutional challenge raises “substantial questions of law,” there should be a stay on the law. As everyone knows, that is not going to happen.

Furthermore, a close look at the four questions framed by the Court reveals that even taking them at face value, they reveal no case for a stay. Let’s deal with them in reverse order. The Court’s fourth question is whether the absence of “ragging” in the Regulations amounts to a “regressive and exclusionary legislative omission.” With respect, this is a non-starter. “Ragging” is dealt with under separate UGC Regulations, which have not been excluded by the 2026 Regulations (the 2026 Regulations only replace the 2012 Regulations of the same name, not every UGC Regulation in existence). The Court’s third question is whether the use of the word “segregation” in Regulation 7(d) amounts to a “separate but equal” measure. This is also, with respect, a non-starter, for several reasons, the first of which is that the Regulation does not mandate “segregation”, but only notes that if there is “any selection, segregation, or allocation for the purpose of hostels, classrooms, mentorship groups, or any other academic purposes,” then it has to be done on a transparent basis. Secondly – and more importantly – the use of the words “selection, segregation, or allocation” together make it very clear what this Regulation is referring to: for instance, something like specialised SC/ST scholarships, or minority mentoring programmes, and the like. To get from here to late 19th century American Jim Crow laws (which is where the phrase “separate but equal” comes from) is – to put it very politely – something of a stretch. Thirdly, even if there is a problem with the word “segregation”, the answer is surely to stay any segregation in pursuance of the Rules, and not the entirety of the Rules themselves. That is how it has always worked.

The Court’s second question is whether the definition of “caste-based discrimination” will have an effect on sub-classification. This is, with respect, a red herring, because sub-classification is an element of reservations jurisprudence, and – to repeat a point above – the doctrine of sub-classification does not mandate governments to sub-classify for the purposes of reservation, it only permits them to do so. Therefore, it is unclear how sub-classification has any relevance to the UGC Regulations, which are about discrimination. The Court goes on to question whether the Regulations effectively protect extremely backward classes (EBCs); but – assuming for a moment that they do not – its answer defies logic, because what it does is to stay these Rules and revive the 2012 Rules, which also have nothing to say about sub-classification and the EBCs!

Finally, the Court’s first question is whether there is a need for a separate definition of “caste-based discrimination”, given that the Regulations already define “discrimination” exhaustively, and given that there is no separate redressal mechanism for caste-based discrimination in the Regulations. Once again, the answer to this question is to be found in the Court’s own previous jurisprudence where laws and rules have been upheld, which it appears not to have engaged with here: which is that it is well within the competence of the legislature to recognise a specific, and particularly serious, subset of harm within a broader category of legislatively-defined harm. The Regulations define “discrimination” in an exhaustive and inclusive sense. They also define “caste-based discrimination” as a specific instance of discrimination that is marked by whom it targets: that is, SCs, STs, and OBCs. The Court’s objection here seems to be that the Regulations only define, and do not go beyond. That might be a policy call for strengthening the Regulations; it is certainly not a legal argument for unconstitutionality.

Ideally, after having framed these four questions, the Court would have explained why, on an application of the three-pronged test for stay to each question, a case was made out for keeping the Rules in abeyance. Because it did not do so, in this blog post, we have been forced to carry out a somewhat strange exercise, where we have set out what might be the arguments if they were to be made, and then countered them ourselves. It’s a bit like being both proposition and opposition in a debate: not ideal, but nor have we been left with much of a choice. The point, however, remains: increasingly, a significant part of the Court’s most consequential activity is taking place without the fundamentally important element of public reasoning. The fact that many of these orders are passed in the context of highly politicised cases (such as the present one) only heightens the need for public reasoning that can withstand scrutiny; its absence is a significant loss, both for constitutional adjudication, and for the ability of the public to hold the Court accountable for its decisions.

Active Democracy and a Culture of Justification: the Kenyan High Court’s Presidential Advisors Judgment

One of the paradoxes of contemporary representative democracy is how public power is often wielded by unelected individuals. It is argued the the complexities of modern-day administration require elected officials to rely on the unique skills and expertise of non-elected – and sometimes, specifically, technocratic – individuals. The civil service can – but does not always – provide such individuals. Consequently, elected officials must reach beyond, and into the private sector, to recruit. But equally, on the other hand, it is argued that such appointments serve as sources of political patronage, they are the “spoils” of political power, they undermine public accountability, and that they contribute to the “revolving door” between politics and big business. An instance of this is the “Special Advisor” [“SpAD“] position in the United Kingdom, which has been critiqued on precisely these grounds.

In this context, the recent judgment of the High Court of Kenya – Katiba Institute vs The Hon’ble Attorney-General – which struck down the appointment of certain “Presidential Advisors,” is a fascinating contribution to the global jurisprudence on this issue. The High Court, here, deployed to great effect what I call here “procedural constitutionalism”: that is, while upholding the power of the relevant constitutional bodies to take certain actions, it insisted on strict compliance with procedural standards of transparency, probity, and institutional independence. Procedural constitutionalism rests upon the faith that, to a large extent, institutional mala fides can be constrained if institutional authorities are required to publicly explain themselves on the touchstone of the Constitution. Procedural constitutionalism – as I shall subsequently argue – is interwoven with other constitutional values such as the culture of justification and popular checks upon elected authority, which the Kenyan courts have keenly protected over the years.

The key issue before the High Court was whether the creation – and appointment – of the Presidential Advisors was compliant with the Constitution. The most relevant constitutional provision was Article 132(4)(a), which authorises the President to establish “an office in the public service [only] in accordance with the recommendation of the Public Service Commission.” (emphasis supplied) The Respondents argued that the PSC had signed off on both the creation of the office of the Presidential Advisor, as well as the specific appointees. Not so, the Petitioners contended: while it was true that the PSC had, in a formal sense, ratified these appointments, a close look at the records showed that these ratifications had come at the behest of the President, demonstrating a fatal non-application of independent reason.

The High Court examined the record. It noted that the record…

… reveals a pattern that significantly undermines the assertion of a substantive and independent recommendation. Several letters from the Office of the President are framed not as requests to establish an office based on demonstrated need, but as communications that the President has identified, selected, or appointed a specific individual and seeks the PSC’s concurrence or its action to hasten the appointment. (para 49)

And:

The PSC’s responses often compound this error by purporting to approve the creation of the office and the appointment in a single sentence, even where the request letter did not explicitly seek creation. (para 50)

To add to this, the High Court found that there were no internal PSC deliberations that demonstrated the need for these positions, that the PSC’s approval letters were cryptic and unreasoned, and that often, they were issued within a matter of days. All of this raised “a legitimate doubt as to whether a genuine, independent consideration, as envisioned by the Constitution, took place.” (para 52) This, the Court held, was a “fundamental procedural defect” that violated the requirements of Article 132. The Court buttressed this constitutional argument by examining statutory violations: the Public Service Commission Act required a written statement verifying the conditions for establishing a new public office had been met (Section 27(2)). Importantly, the Court treated the condition of verification as requiring public reason-giving: that is, “verification” could not simply be a pro-forma recording of satisfaction (as the PSC had done) that the conditions had been met, but required substantive explanation. The PSC’s letters on the other hand “are brief, rarely exceeding a page, and focus almost exclusively on recommending a named individual. They do not attach workload analyses, financial implications beyond perhaps a generic grading if at all, or comparative data on existing posts.” (para 58) Indeed, the PSC had not even complied with secondary legislation – the PSC Regulations – that required a determination of capping and rationalising the number of posts based on objective needs, and an explanation that “the proposed advisor’s technical competencies do not exist in the public service and that they match the job requirements.”

Having gone into the weeds of technical statutory interpretation, the Court then wove these findings into the warp and weft of constitutional analysis. Two crucial constitutional norms were attracted. The first was that of public participation. Given the nature and importance of these advisory functions, the creation of these new posts had to go through the process of public participation. The Court here specifically noted how the 2010 Constitution ushered in the era of active democracy, which envisaged the involvement of the People between, and beyond, simply election cycles and the ballot box (what Roberto Gargarella calls “external checks” on public power). As the Court eloquently observed:

Pre-2010, Kenyans were largely spectators to governance. Their role was that of electoral automatons whose participation in public life was limited to one day, every five years (barring by-elections), to cast their vote and thereafter be silent and uninvolved until the next elections. Under this new and current order, the citizens are individually and collectively active participants in the governance of this nation, albeit having delegated much of the day-to-day responsibilities of steering the nation across three branches and two levels of government. (para 88)

Closely linked to this is a second idea, that of a culture of justification, which we have discussed extensively on this blog. The culture of justification – as distinct from a culture of authority – envisages a constitutional culture where the exercise of public power is opened up to public scrutiny, and must justify itself to the People. The opaque character of these appointments ensured that this had manifestly not been done. As the Court noted: “for the people of Kenya to fully be the Sovereign that they are, they need to participate as appropriate in as much of public life and public affairs as is possible and practical.” (para 92) Breach of this, coupled with the specific statutory and constitutional breaches, meant that the entire process was null and void.

A close analysis of the judgment reveals, however, that even these lofty statements of constitutional principle – that of active democracy and the culture of justification – were rooted in constitutional procedure. At no point did the Court say that Presidential Advisors couldn’t exist. What it did say was that if they were to exist, their existence would have to depend on a good-faith, independent evaluation of need by a fourth branch institution, the PSC, and – crucially – that the process for arriving at this evaluation had to be public, transparent, and demonstrably independent.

I suggest that this iteration of procedural constitutionalism (which bears comparisons both to newer glosses on John Hart Ely’s representation reinforcement theory, and to Ros Dixon’s idea of responsive judicial review) has long been practiced by the Kenyan courts, although they do not explicitly call it such. For example, in the BBI Judgment, when considering the issue of constitutional amendments and popular initiatives, the Court declined to place a substantive check on the amending power, but insisted on rigorous procedural standards, one of which – indeed – was specifically designed to ensure that the President could not bypass what was meant to be an independent, People-driven process with a top-down executive-driven process (see here). If you replace “the People” in the BBI Judgment with the PSC in this judgment, you come to a very similar idea of procedural constitutionalism as a check upon the misuse of public power.

There are, of course, limits to proceduralism: a sufficiently clever, determined, and committed executive can still subvert the spirit of the constitution. What proceduralism does do, however, is to make this substantially harder, while also insulating the courts from executive backlash. An example I like to think of, for instance, is how Latin American constitutionalism approaches the idea of free, informed, and prior consultation before indigenous lands can be taken over for public projects. This “consultation” is not equivalent to a veto, yes. However, think of an example where a railway line is passing through a fragile, indigenous habitat, and it is demonstrated that shifting that line five kilometres to the north will achieve the same goal with minimal financial disruption. Procedural constitutionalism – encased within the principles of active democracy and the culture of justification – will require the state to publicly justify why the alternative is not being taken up. And the very need to justify will rule out extraneous reasons and – indeed – often demonstrate that there is no valid reason for the action (thus forcing a change). We can think of the High Court’s decision in these terms.

Katiba Institute vs Attorney-General, therefore, is both a valuable contribution to the global debate on the increasing role of unelected individuals in governance, as well as to the conversation about the part that constitutional principles such as active democracy and the culture of justification have to play in resolving concrete disputes. The High Court’s seamless shift between the abstract and the particular serves as a case study for how this is to be done in the context of judicial adjudication.

---

Postscript: The Demonetisation Judgment

Reading the High Court’s judgment, one is struck by an example from India. In early 2023, the Indian Supreme Court upheld demonetisation (six years after the fact). The legal framework was very similar, in that what was required was a recommendation by the central bank (much like the PSC) One of the most astonishing features of the Justice B.R. Gavai (as he then was)-led majority judgment was its observation that the fact that the Reserve Bank of India had taken 24 hours from the government’s communication to decide on something as complex as demonetisation did not mean that its “independence” had not been respected (see e.g., my piece on Schrodinger’s Central Bank). On very similar facts, the Kenyan High Court’s application of common sense and an understanding of basic reality is quite the contrast; indeed, it reminds one of Nagarathna J.’s dissenting judgment in that case, which did employ a similar iteration of procedural constitutionalism. Indeed, the similarities between that judgment and the Kenyan High Court’s PSC judgment are striking. As I wrote at the time:

It is worth briefly flagging Nagarathna J’s analysis of process. In a section titled “Affidavits and Records of the Case”, Nagarathna J observes that the RBI’s own documents show that it was the government that “recommended” demonetisation, that a draft scheme was being made “as desired” by the government, and that a recommendation had been “obtained” from the bank for this (paragraph 17). She also notes that the documents show a conflict between the bank’s goals for demonetisation and the government’s, showing that the so-called “consultation” was illusory (paragraph 17.8); and that furthermore, the records showed that the hurry in which these communications were exchanged (in under 24 hours) – clearly pointing to an absence of application of mind by the bank (paragraph 19).

Both Nagarathna J.’s dissent, and the Kenyan High Court’s PSC judgment, therefore, show how procedural constitutionalism allied with a dose of common sense can serve as a powerful tool of constitutional accountability.

The Constitutional Status of Inaction: Towards a Doctrine of Default in Gubernatorial Assent [Guest Post]

[This is a guest post by Arihant Jain.]

---

Introduction

In recent years, a notable pattern of constitutional tension can be seen across multiple Indian States: Governors withholding assent to legislation not by an express refusal but through prolonged silence. In State of Tamil Nadu v. Governor of Tamil Nadu, the apex court recorded that multiple bills had remained pending with Raj Bhavan for an extended period without decision

While some delay may arise from complex federal questions or the requirement of consultation or legal scrutiny, prolonged silence without stated justification enables an unelected authority to suspend legislative will without producing any reviewable decision prompting the need for a Doctrine of Constitutional Default.

Article 200 And the Constitutional Framework of Assent

Article 200 establishes the Governor’s role in the legislative process as a constitutional checkpoint, not as a parallel veto-wielding authority. Upon receiving a bill for assent, the Governor has only four textually specified options: to assent, to withhold assent, to return the Bill once for reconsideration, or to reserve it for the consideration of the President. The provision does not offer a fifth option of indefinite action. This intentional textual design reflects the framers’ insistence to preserve democratic legitimacy within the elected assembly, positioning the governor as a constitutional guardian rather than a political rival.

There is judicial precedent to support this understanding. In Shamsher Singh v. State of Punjab (1974), it was held that the Governor is formal head bound by the aid and advice of council of ministers, the area of his discretion being narrowly defined within the constitutional circumstances. Subsequently, in Nabam Rebia v, Deputy Speaker (2016) the Court further emphasized that the gubernatorial discretion cannot be applied to influence political outcomes or hinder the operations of a democratically elected government. As Article 200 is without procedural consequences for inaction, this omission has allowed silence has turned into an informal and constitutionally unscrupulous form of veto, unbalancing the dynamic between accountable government and constitutional oversight.

The 2025 Advisory Opinion: Clarity, Restraint And New Uncertainty

The Supreme Court’s response to the growing issue of gubernatorial delay cumulated in presidential reference of 2025 which addressed the scope of the Governor’s and President’s powers. The court held that constitutional authorities must take action within a reasonable period of time and that silence cannot be used to obstruct the functioning of an elected legislature. At the same time, it refused to accept any notion of “deemed assent”, on the basis that the courts could not judicially manufacture consequences which were not present in the constitutional text. However, significant institutional limits were also reflected in the opinion. The court carefully avoided defining what would constitute unreasonable delay and it avoided prescribing enforceable consequences for unconstitutional inaction.  This restraint can be defended on the grounds that the advisory jurisdiction under Article 143 is intended to elucidate constitutional meaning and not to build up remedial regimes or hard legal tests. This reference could also be understood as a pragmatic response to a nation-wide constitutional deadlock.

At the same time, this reference mooted issues with close links to the current legal disputes pending before various high courts. Here, the approach adopted by the court blurred the traditional line between advisory interpretation and adversarial adjudication. Article 143 was intended to be a non-adversarial forum for abstract clarification but here it was being used to sort out issues with immediate political and federal implications. The net result is a partial resolution: the silence of the governor is now recognized as constitutionally problematic yet there is no principled work in existence for determining when it becomes unconstitutional and what follows once it does.

Why India Needs A Doctrine Of Constitutional Default

The various repeated instances of gubernatorial inaction reveal a more fundamental structural lacuna in the Indian constitutional law. While the Constitution contains detailed mechanisms for reviewing unconstitutional action, it requires almost no doctrinal framework for responding against the failure to perform constitutional duties. As a result, legislative paralysis can be accomplished without triggering any of the accountability mechanisms that ordinarily are associated with the exercise of public power.

This deficit matters because Articles 200 and 201 presuppose constitutional choice rather than silence, because the colonial design of the Governor’s office already risks federal asymmetry, and because democratic legitimacy suffers when legislative outcomes are negated through administrative silence rather than constitutional determination.

One may argue that Constitution’s use of “reasonable time” preserves flexibility for consultation. Legal clarification and inter-governmental coordination. Also, it can be argued that doctrine of inaction risks judicialising political disputes better left to constitutional convention. These concerns are real but they reinforce a case for narrow, principled doctrine one that distinguishes ordinary administrative delay from strategic non-performance and operating at margins of justiciability.

Towards A Doctrine of Constitutional Default 

For the Constitution to function as a framework of enforceable obligations, it must develop principles to address not only unconstitutional action but also unconstitutional inaction. A doctrine of Constitutional default would offer a limited, structured method to identify when silence becomes constitutionally impermissible and what minimal consequences should follow.

The first pillar of such framework is a duty to decide. Comparative constitutional law reflects this widely accepted proposition that where public power is granted for constitutional purposes, its exercise cannot be indefinitely withheld. For instance, in the United States prolonged administrative inaction is reviewable as failure to act under the Administrative Procedure Act. In Westminster systems, executive delay in the grant of legislative assent is disciplined primarily through constitutional convention rather than judicial enforcement with the Sovereign being constitutionally bound to act on ministerial advice.

Building on this, unconstitutional inaction may be identified through a calibrated, effects-based two-part test. First, delay becomes constitutionally impermissible only when it produces demonstrable institutional paralysis that is, where (i) a coordinate organ has completed its prior constitutional act, including repassage where required; (ii) the next constitutional step in the governance chain is rendered legally impossible; and (iii) the resulting standstill disables the State from operationalising policy, incurring expenditure, constituting statutory bodies, or discharging time-sensitive constitutional obligations. This framework confines “frustration” to cases of constitutional disablement, not simply political inconvenience. Second, such paralysis must be without any recorded and intelligible constitution justification such as unresolved questions of legislative competence, bona fide requests for legal opinion or demonstrable inter-governmental consultation. The test thus turns on institutional obstruction without constitutional explanation rather than on judicial speculation about political motive. It is calibrated to exclude routine administrative delay while identifying deliberately prolonged non-performance.

Implementing the doctrine necessitates restrained procedural mechanism, not expansive remedies. Where considerable delay remains unexplained, a requirement of brief written reasons would operate as a minimal transparency safeguard, enabling courts and legislatures alike to distinguish bona fide deliberation from non-performance. If silence nevertheless persists beyond what constitutional purpose can justify, a default transmission rule should permit the Bill to move to next constitutional stage, typically consideration by the President. Unlike “deemed assent”, which judicially attributes legal validity to a Bill, this rule does not validate the legislation at all, it merely reallocates the constitutional responsibility from a non-performing authority to the next designated constitutional actor. The law would still come into force only through a fresh, express constitutional act, not by judicial implication.

A legitimate concern is that such procedural transmission may replicate the problem of inaction at a higher constitutional level, particularly given the functional alignment between the Governor and the Union executive. This framework does not deny that risk. Its purpose is not to presume greater independence at the Union level, but to ensure that no single constitutional authority can indefinitely monopolise constitutional paralysis through silence. Once delay migrates to the President, the character of the impasse shifts from inter-governmental federal friction to national-level constitutional responsibility, at which point continued inaction becomes institutionally attributable, politically visible, and far more amenable to principled judicial scrutiny.

Finally, constitutional morality supplies the normative discipline for this framework, not as a free-floating ethical appeal, but as a standard of institutional fidelity. The inquiry is not whether the Governor’s conduct is politically agreeable but whether it represents loyalty to the neutral constitutional role of the office, the principles of responsible government and the equality of federal units. Delay that advances partisan preference under the guise of constitution form would fail this test even where there was no express textual prohibition In this context, constitutional morality is the source of a standard for judging intent and institutional discipline, ensuring that constitutional silence is not made an instrument for partisan manipulation.

Conclusion

The recent series of disputes between Governors and elected State governments prove that constitutional paralysis in India has increasingly been emerging not through over confrontations, but through institutional non-performance. The Supreme Court has now recognised that prolonged silence in the assent process is constitutionally suspect. No doctrine can eliminate political friction completely, but constitutional silence cannot continue as a sphere of unreviewable authority. A doctrine of constitutional default would not invite intrusive judicial supervision. It would operate at the margins only in those exceptional cases in which inaction is frustrating the functioning of elected institutions and destroys federal balance. Whether its contours ultimately develop through future litigation, legislative clarification or constitutional convention, the task of governing constitutional silence is no longer possible to defer.

Notes from a Foreign Field: The Speaker, the Judicial Service Commission, and Judicial Independence in Sri Lanka [Guest Post]

[This is a guest post by Sanjit Dias.]

---

The speaker of the parliament of Sri Lanka recently made a ruling which addressed issues of the independence of the judiciary vis-à-vis the judicial service commission (JSC), and the separation of powers. In this post, I argue that the ruling (a) is incorrect in characterising the functions of the commission as ‘judicial’ functions, and fails to regard the JSC as a guarantor institution; (b) misconceives the issue of separation of powers, though perhaps erring to preserve the right optics; (c) raises questions on the composition and functions of the JSC; and (d) is significant for inter-branch collaboration in Sri Lanka.

Background and ruling

The context of this ruling is briefly as follows. The present chief justice (CJ) assumed office in July 2025. The chief justice and two other supreme court judges constitute the judicial service commission of Sri Lanka, an independent commission / guarantor institution tasked with the appointment, transfer, disciplinary control and removal of judges of the lower courts. In the months after the CJ assumed office, many judges were interdicted and / or removed from office; others resigned. In parliament, the ruling National People’s Power (NPP) regime holds a 2/3 majority, and several opposition groups have adopted a very combative strategy to oppose the government. In light of the recent actions of the JSC, some opposition MPs presented a motion to set up a parliamentary select committee (PSC) to review the functions of the JSC and its conduct in the course of these events, among other things. Under the standing orders of parliament, the speaker must decide whether to place such a resolution before the house.

In his ruling, Speaker Jagath Wickramasinghe determined that the motion was out of order as it violated judicial independence by seeking to interfere in the judicial branch, and because the committee envisioned by the motion would exercise an oversight function over the JSC, thereby breaching the separation of powers.

All functions performed by judges are not an exercise of ‘judicial’ power

The speaker states: “From both its composition and its conferred functions, it is my opinion that the operations of the JSC are attributable to an exercise of judicial power.”

I contend that this is incorrect. The JSC is an independent commission which comprises three judges of the supreme court, and it was designed to enhance judicial independence by limiting direct governmental control in judicial administration. It may properly be regarded as a guarantor institution, similar to other independent commissions such as the election commission and human rights commission. The JSC is established by the constitution, and is created to guarantee the non-self-enforcing constitutional norm of judicial independence. However, its primary functions are executive and / or administrative in nature, not judicial. As described above, the JSC is responsible for the recruitment, transfer, disciplinary control and removal of junior judges, all of which are administrative tasks. By analogy, one could look to Sri Lanka’s independent public service commission, which performs a similar role in respect of public servants. There would be no controversy in stating that such functions are executive or administrative in nature when performed by that commission. Therefore, just because three judges perform these administrative tasks in the JSC, it does not become ‘an exercise of judicial power.’

The speaker’s ruling in fact contradicts this position. It later (correctly) points out that the only avenue to review a decision of the JSC is to bring a fundamental rights (FR) action under article 126 of the constitution. Article 126 provides redress for the infringement of fundamental rights by ‘executive or administrative action’; legislative and judicial acts are excluded. Therefore, the speaker’s statement that redress would be by way of an FR application concedes that the JSC’s functions are executive or administrative in nature. In the celebrated case of The Queen v Liyanage (1962), the supreme court of Sri Lanka held that the JSC as it existed under the independence (‘Soulbury’) constitution performed ‘executive functions.’ Another commentator has critiqued the speaker’s ruling on this basis.

The present constitutional text complicates the picture. Other independent commissions typically have an immunity provision, with an exception for the supreme court’s jurisdiction to review their actions under article 126. The JSC does not. Instead, it has a more open-ended provision (article 111K) which states that no suit may be brought against its members for lawful actions done in good faith. In any event, the supreme court’s jurisprudence suggests that the JSC’s actions are liable to review by way of a FR application.

But how meaningful would this remedy be? Such a mechanism may appear to breach principles of natural justice, as a different panel of supreme court judges would – in essence – be reviewing a decision of their fellow judges (including that of the chief justice) in their capacity as members of the JSC. This does not necessarily mean that such a hearing would be biased, but important questions arise as to whether a litigant and / or the public would perceive such a hearing to be fair and impartial.

Separation of powers: correct on optics, wrong on constitutional theory?

The second principle on which the speaker’s ruling rests is the separation of powers. Tracing several Sri Lankan supreme court precedents, and parliamentary practice and constitutional convention in the UK, the speaker determines that it would breach the separation of powers for a parliamentary committee to supervise the functions of the JSC. He states, “Placing the judiciary before a Select Committee — a political body — would subject judges or judicial administrators to political scrutiny, thereby undermining the essential independence required of the judicial branch.” (emphasis mine)

Some nuance is warranted here. As the JSC is presently constituted, to bring the JSC before parliament may be perceived as bringing the judiciary before parliament. However, strictly speaking, the relevant judges would be answering in their capacity as commissioners of the JSC, not as judges of the supreme court. As far as constitutional theory goes, this would not raise an issue of separation of powers; it would be the same as any of the other independent commissions being responsible to parliament. Regardless, if the JSC – as a guarantor institution proper –  were constituted more like other guarantor institutions (more on this below), it would be less controversial for the JSC to be responsible and answerable to parliament. The speaker correctly notes that the constitution presently requires the other independent commissions to be answerable to parliament, but not the JSC (article 41B (6)). As a practice, however, the JSC submits annual reports to parliament, just as the other independent commissions do. 

Crucially, the speaker notes that parliament’s custody of the public purse does not connote a hierarchical power over the judiciary, but a fiduciary responsibility to give effect to the public trust reposed in the legislature. In effect, the speaker rules out any possibility of influencing or interfering with the judiciary by curtailing its funds. Therefore, the speaker may be incorrect in constitutional theory in ruling that bringing the JSC before parliament would breach separation of powers. He may however have erred in favour of avoiding the negative optics of judges being questioned by politicians.

Sri Lanka’s JSC: an outlier?

Where does that leave the accountability and answerability of the JSC? Transparent, accountable, and efficient judicial administration is crucial to retaining public faith in the judiciary. It is also essential to attracting excellent candidates as judges, and to retaining them within the system. On the one hand, political answerability of the current JSC may not be the solution, as that may threaten the public perception of judicial independence. On the other hand, accountability solely within the judicial branch (to judges comprising the same court) perhaps extends too much independence, to the detriment of public confidence and accountability.

Sri Lanka’s JSC may be an outlier in terms of comparative practice. In the UK, judicial recruitments are by a 15-member judicial appointments commission (JAC), which comprises a mix of judges, non-lawyer professionals, and lay people. This increases the diversity of experience and perspectives in the overall recruitment process. In South Africa, the judicial service commission comprises up to 23 members, including the chief justice and judges of different courts, law teachers, and members of parliament, with rules on which members may be involved in different types of decisions. In neighbouring India, judges of the high courts and the supreme court are selected through the collegium system, while district judges are recruited through the state public service commissions.

The JSC in Sri Lanka is smaller and more narrowly constituted than any of these mechanisms. The past abuse of the JSC has been well documented, demonstrating how it can be captured and weaponised to undermine judicial independence. The time may be ripe to begin a discussion on broadening the composition of the JSC to ensure diverse perspectives and interdisciplinary expertise to bolster judicial efficiency and independence in Sri Lanka. A call for such reforms has been made in the past. In such an arrangement, having the JSC answerable to parliament would not raise the same concerns as at present. This would be particularly true if non-judge members of a future commission perform the public-facing roles, including answering to parliament.

Concluding thoughts: inter-branch collaboration

Regardless of its theoretical weaknesses, the ruling is noteworthy for its institutional restraint. Legislatures are generally more powerful and wield more public legitimacy than the judiciary. The trend in the recent past has been for Sri Lanka’s parliament to try to assert its power over the judiciary. Therefore to have the speaker of the house making a good faith effort to engage with the court’s jurisprudence to uphold the principles of judicial independence and separation of powers is a welcome change for democratic and constitutional governance. It may represent an attempt at collaborative constitutionalism as theorised in comparative constitutional law scholarship.

Presidential Reference No. 1 of 2025 and the New Afterlife of Judgements

[This is a guest post by Rajyavardhan Singh.]

---

On 20 November 2025, the Supreme Court delivered its advisory opinion in Re: Assent, Withholding, or Reservation of Bills by the Governor and President of India (“Re:Bills”), answering eleven of the fourteen questions referred under Article 143. The very act of making the reference itself drew sharp objections on maintainability (see here, here, here, and here), particularly because only months earlier – in State of Tamil Nadu v. Governor of Tamil Nadu – the Court had already addressed indefinite gubernatorial inaction by recognising a “deemed assent,” imposing timelines, and expanding judicial review so as to enable state governments to seek mandamus.

Against this background, a subsequent presidential reference made little sense. It appeared less a genuine request for clarification of unsettled law, and more an attempt to revisit – if not dilute – a final judicial determination through executive initiative. These objections were considered by the Court in Chapter III-B of the opinion, where challenges grounded on malafide intent and abuse of form were addressed.

The Court’s answer on the maintainability of the reference has already been ably examined on this blog (see here). This post does not intend to rehearse those claims. Instead, it advances a briefer and narrower discussion, examining how the Court seeks to escape the charge that the reference was, in substance, an “appeal in disguise” – namely, through its reliance on the idea of an untouched decree.

The Court’s Defence of The Untouched Decree

Borrowing wholesale from Cauvery (2), the Court’s defence separates the reliefgranted in State of Tamil Nadu from the law that produced it. It holds:

“The questions do not require this Court to vacate, amend, or modify the final relief granted by this Court in State of Tamil Nadu (supra), and instead, at the most, seek clarification on the propositions of law laid down in it, which have ramifications for the governance of all States, i.e., beyond the parties that were before it in that lis.” (paragraph 26)

It is on this idea of an untouched decree that the Court constructs its justification for why the reference does not amount to sitting in appeal:

… (s)itting in appeal would mean the variation, or vacation of the operative order in a concluded lis. Article 143 cannot be invoked to overturn a concluded adjudication inter se parties; but that cannot be conflated with the authority of this Court to answer general questions of law referred to it by the Hon’ble President… (paragraph 27)

The underlying idea of the Court’s reasoning is built around a distinction taken from Natural Resources Allocation. Therein, every judgement is said to consist of two components: first, a binding decree between the parties; and second, the general propositions of law that led to it. Only the former is insulated from reopening under Article 143. The latter remains, in the Court’s words, “malleable.”

And so, even if the reference were to hollow out – or altogether dismantle State of Tamil Nadu’s holdings on deemed assent, timelines, or the scope of gubernatorial discretion (as in many respects it did), no impropriety would follow. So long as yesterday’s winners get to keep their trophies – the Court has not, in its telling – sat in appeal.

Put simply, the Court claims a power to express a different view of law, but not to disturb the relief. And because Cauvery (2) is read as barring only the reopening of operative decrees, the reference is said to lie on the permissible side of Article 143.

Relief Preserved – and the Law Reopened

To be clear, the problem is not the Court changing its mind on the law. Courts do that all the time, and often for good reason. In Re: Bills change is triggered not by subsequent litigation or a bench confronting the question afresh. Instead, it is done at the instance of the very executive that lost under that law.

The Court, however, appears to think it has struck a clever equilibrium. Tamil Nadu keeps its relief. The Union gets the law changed. In effect saying that the executive may not undo what Tamil Nadu secured – but it may invite the Court to reconsider everything that made Tamil Nadu’s victory possible. Taken to its logical end – this reasoning produces nothing but absurdity. For this is the same as saying that a future Bench, acting under Article 143, could very well just repudiate Maneka Gandhi’s understanding of due process while defending that no overruling has occurred so long as her passport remains untouched.

This becomes particularly troubling because of the nature of the decision sought to be revisited. The real stakes in State of Tamil Nadu were never just confined to the ten sitting Bills that were granted deemed assent. They were about whether Governors may sit indefinitely on legislation, whether courts can impose timelines, and whether deemed assent exists at all.

At issue, therefore, was a recurring problem in Indian federalism (i.e. the use of gubernatorial inaction as a tool of political obstruction) and as to how constitutional functionaries must act within a scheme of responsibility and accountability. Besides, the Court itself expressly distinguishes this reference from the previous fifteen and repeatedly characterises it as ‘functional’ (paras 10, 11, 30, 51) acknowledging that the questions posed directly implicate the operational behaviour of constitutional authorities rather than abstract points of law.

Naturally, the premise that only general propositions of law are being revisited makes little sense. Functional questions structure the future exercise of public power. They are not detachable from the settlement they produce. To treat these holdings as malleable propositions at executive request, would thus be conceptualising finality that is unmistakably private-law in orientation – one that treats adjudication as complete once relief is granted.

And so, to conclude.

The foundational difficulty with the Court’s approach lies in its redefinition of finality. Constitutional adjudication does not end with relief. It ends with rules that bind power going forward. By treating decrees as sacrosanct and the proposition of law as freely revisable at executive request, the reference empties adversarial victories in constitutional law of their future value.

Advisory jurisdiction exists to clarify uncertainty. It does not exist to manage the afterlife of inconvenient judgements. For if precedents can be unsettled without overruling, then finality survives only in name.

Bail granted, but freedom is not: The Supreme Court’s erroneous bail conditions in Gulfisha Fatima [Guest Post]

[This is a blog post by Mihir TejaKalle.]

---

*(This article does not pass comment on the denial of bail to Sharjeel Imam and Umar Khalid)

On 5^th January, 2026, the Apex Court in Gulfisha Fatima v. State (Govt. of NCT of Delhi) granted bail to five individuals – namely Gulfisha Fatima, Meeran Haider, Shifa-ur-Rehman, Mohd. Saleem Khan, and Shadab Ahmed. This would to many be a sigh of relief and a moment of joy (in times of despair), but the judgement still proves to befuddle our minds. In granting bail, the Court listed out eleven conditions that the accused would have to follow, including two very puzzling conditions –

viii. The appellants shall not participate in any programme or address or attend any gathering, rally or meeting, whether physically or virtually till conclusion of the trial.

ix. The appellants shall not circulate any post either in electronic form of physical form or circulate any hand bills, posters, banners, etc in any form whatsoever.

At first glance, one might argue that these conditions are relatively minimal when viewed against the gravity of the accusations involved. However, a deeper introspection reveals a far starker picture, one where such conditions violate the basic and inviolable fundamental rights of the individual and signal a troubling regression in bail jurisprudence in India. Bail is no longer the rule, and even when granted, it increasingly arrives burdened with harsh restrictions on fundamental liberties.

A deviation from past jurisprudence

Bail was conceived as a mechanism to preserve the liberty of the accused while simultaneously securing their presence during investigation and trial. The Supreme Court has consistently emphasised that onerous bail conditions should not be imposed if they are unduly burdensome or infringe upon fundamental rights. Any condition imposed must bear a clear nexus with the administration of justice, securing the presence of the accused, preventing interference with the investigation, ensuring witnesses are not intimidated, and safeguarding the trial process.

The Apex Court has previously held in the case of Siba Shankar Das v. State of Odisha, that a bail condition that restricted a politician from being involved in any political activities is a breach of their fundamental rights and hence, liable to be withdrawn. In Frank Vitus v. Narcotics Control Bureau, (discussed on this blog here) the Court cautioned that restraint must be exercised while imposing bail conditions and that such conditions must minimally impair the accused’s fundamental freedoms and rights.

These decisions reflect a consistent judicial philosophy, one that seeks to protect individual liberty while allowing narrowly tailored restrictions strictly necessary for the smooth administration of justice.

The lack of rationale in Gulfisha Fatima

In Gulfisha Fatima, the Court appears to have imposed bail conditions mechanically, without providing concrete reasoning beyond a broad invocation of national security, public order, and the integrity of the trial process. The judgment places a blanket prohibition on the accused from participating in, or even addressing, any programme, rally, or meeting. This restriction surprisingly extends to virtual platforms as well. Such an expansive condition directly impinges upon the fundamental rights to freedom of speech, expression, and assembly.

If it was impermissible to impose a similar condition upon a politician, who notably had multiple criminal cases registered against him, it is difficult to see how such a restriction becomes permissible in the present case.

A similar condition was proposed by the State in Mohammed Zubair v. State of NCT of Delhi which called for a blanket bar on the accused ‘tweeting’ i.e posting on social media, X (Twitter). The Court deemed this to be a blanket ban on expression and held that it would be impermissible to impose it as a condition for the liberty of bail. History may not look favourably upon Gulfisha in terms of its bail conditions.

Moreover, it is difficult to discern the necessity of such sweeping restrictions when condition (vi) already prohibits the accused from associating with or participating in the activities of any group or organisation linked to the subject matter of the FIR. The judgment effectively operates as a gag order, rendering the accused unable to express their views, participate in ordinary social and civic life, or engage on social media platforms. This mirrors the concerns raised in Frank Vitus, where the Court frowned upon bail conditions that effectively placed the accused in a state of confinement outside prison walls.

Even assuming that the impugned bail conditions pursue a legitimate aim, such as the preservation of public order and trial integrity, they miserably fail the test of proportionality. A blanket prohibition on all forms of public participation and expression, including virtually, is neither necessary, nor the least restrictive means available. There were multiple alternatives that the Court could have explored that were less intrusive, such as restrictions limited to specific organisations (read condition vi.), or speech directly linked to the alleged offences (read condition vii.), which were plainly available and satisfactory to ensure the goal of the bail conditions. The additional and onerous conditions therefore collapse at this stage of necessity and balancing, hence, rendering them constitutionally disproportionate.

It is difficult to come to the conclusion that merely participating in any event or circulating any post either physically or electronically can have a reasonable nexus to the investigation or the trial itself. The Supreme Court has to bear in mind that it is the guardian of the Constitution. If the guardian itself offends the Constitution, then the remedial promise of constitutionalism stands gravely in danger.

Especially when the Court has repeatedly warned against the use of law in ways that chill the freedom of speech and expression, it bears a heightened responsibility to ensure that its own orders do not negate that very principle through overbroad and unjustified restrictions.

Breaching the Trojan Horse: Examining Section 17A’s Pre-Investigative Sanction Regime through the Lens of Manifest Arbitrariness [Guest Post]

[This is a guest post by Krishaan Doctor and Nitya Singhania.]

---

A division bench of the Supreme Court, comprising Justices Viswanathan and Nagarathna, recently delivered a split decision regarding the constitutionality of Section 17A of the Prevention of Corruption Act, 1988. Section 17A requires prior governmental approval to be obtained before the commencement of any enquiry, inquiry or investigation into an offence under the scope of the PCA Act, where such offence is alleged to have been committed by a public servant in discharge of official functions or duties. The statutory definition of official functions used is broad, covering any recommendations made or decisions taken in the discharge of a public servant’s official duties. The Court was called upon to decide whether Section 17A’s sanction regime was consistent with Article 14 and the object of the Prevention of Corruption Act, or merely a resurrection of prior approval regimes already struck down by the Court.

On this question, Viswanathan J. held Section 17A was “qualitatively different” from prior sanction regimes because it did not specifically create a status-based distinction between senior and junior officers, and argued that it could be cured simply by reading down “Government” to refer to an independent authority such as the Lokpal/Lokayukta instead of the executive branch. He further argued that a pre-sanction requirement was imperative to protect ‘honest’ officers and ensure that they could carry out their functions efficiently and fearlessly. On the other hand, Nagarathna J. held that in practice, Section 17A created the same status based distinction struck down by the Court in  Vineet Narain and Subramanian Swamy respectively, since  “decisions” and “recommendations” were disproportionately made by senior officers.

While Nagarathna J may be correct in the latter assertion, a larger bench may not readily accept the latter characterisation. This is because the word decision and (especially) recommendation are sufficiently vague to be interpreted to include numerous official functions carried by officers of a junior designation. Moreover, confining the analysis to the test of reasonable classification risks shifting attention from the constitutionality of a pre-investigative sanction regime simpliciter, thereby allowing the State to escape judicial scrutiny merely by changing s17A’s sanctioning authority. Placing reliance solely on the potential status-based distinction created by Section 17A as a ground to challenge its validity is therefore unlikely to be very fruitful.

We argue that the correct, and more robust ground to challenge Section 17A is through the lens of manifest arbitrariness, which would not require proving that the provision creates two separate classes of senior and junior officers. Despite its significance to an Article 14 challenge, the doctrine of manifest arbitrariness finds no express engagement in either judicial opinion. While Nagarathna J.’s opinion correctly observes that a pre-investigation sanction regime is arbitrary in a broader sense, this observation does not form the central part of her reasoning and remains largely underdeveloped. This blog undertakes this missing analysis, arguing Section 17A is manifestly arbitrary because (a) a pre-investigative sanction regime necessarily lacks adequate determining principle, (b) the overbreadth inherent to the ‘recommendation’ can protect all public servants for virtually all official acts and (c) its real purpose is to shield corrupt public servants from accountability, rather than protect honest officers from frivolous complaints.

Putting the Cart Before the Horse: Pre-Investigative Sanction and Lack of Determining Principle

In Shayara Bano, the Court defined manifest arbitrariness as something done by the legislature capriciously, irrationally and/or without adequate determining principle. This section argues that the requirement of a pre-investigative sanction is without adequate determining principle, since there is no material to base such a decision upon, necessarily opening the gates for irrational and extraneous considerations to influence the decision to sanction.

The purported purpose of Section 17A’s pre-investigation sanction regime is to filter out false, frivolous, and vexatious investigations at the threshold, thereby protecting well-meaning public servants from the abuse of the criminal process. While this purpose is laudable, the question that necessarily arises is – on what basis is the government coming to the conclusion that a particular complaint is frivolous or vexatious versus legitimate/bona fide? In her opinion, Nagarathna J. addressed the anomalous, catch-22-esque situation created by Section 17A as follows (para 20.14–20.15) :

According to learned Solicitor General, Section 17A of the Act has been inserted precisely to scrutinize a request made by a police officer for enquiry, inquiry or investigation in order to ascertain whether it is a genuine complaint or a frivolous one. This, in my view, is like putting the cart before a horse. If a complaint is enquired into, the truth will unravel. If approval is not granted to even make a preliminary enquiry, the truth and genuineness of the complaint would not be known.

This absurdity identified by Nagarathna J. in the nature of a pre-investigation sanction fits precisely within the scope of manifest arbitrariness. It is impossible for the state to determine the validity of a police complaint, because only the process of a preliminary inquiry that follows such a complaint can supply the material to make such a determination. This problem would persist even if, as Viswanathan J. has suggested, the sanctioning authority is transferred to an independent body such as the Lokayukta.

Moreover, the absence of objective material at the pre-investigative stage creates a real risk that the competent authority’s decision is influenced by extralegal and irrational factors such as the seniority of an officer, anticipated reputational or political fallout, or even profiling of the complainant. The lack of objective material and the potential influence of these factors supply no adequate determining principle to distinguish vexatious complaints from genuine ones, and is therefore manifestly arbitrary. In any case, Section 19 of the PCA – which provides for additional government sanction to prosecute a  public servant – provides a rational basis to identify vexatious complaints since, at that time, a preliminary investigation has generated at least some relevant material.

The (In)Determinate Boundary of Protection : The Problem of “Recommendation”

A key point of disagreement between the Justices was regarding the nature of officials that would be protected by Section 17A. While Nagarathna J. held that officers involved in making recommendations or taking decisions would necessarily correspond to the senior most public officials, Vishwanathan J. rebutted this contention by arguing the following [para 108]

The assumption that lower-level officers merely perform clerical tasks and do not make recommendations is factually incorrect and inconsistent with administrative practice. In governance and administration … File notings, scrutiny reports, technical evaluations, and compliance assessments constitute recommendations in law … Numerous statutory and regulatory frameworks vest recommendatory power in officers below senior ranks…

He specifically noted that the expression “recommendation” is deliberately broad and is not confined to final policy decisions alone.  We are inclined to agree with this proposition, and fail to see why ‘recommendation’ would be interpreted narrowly as Nagarathna J. suggests. The statute does not specifically define the term or indicate that it be interpreted to mean only ‘further recommendations’ or any ‘final recommendations’, which may be made by a more senior officer. The ordinary and popular meaning of the term also favours an expansive reading, with Webster defining the term as ‘to suggest (an act or course of action) as advisable’ and Black’s Law Dictionary as ‘to advise or counsel’. Further, the Act’s expansive definition of ‘public servant’ (including employees of service commissions, public universities, government-aided cooperative societies) indicates that Parliament contemplated recommendations being made across a wide range of institutional settings.

However, while Viswanathan J. recognises this ‘deliberate broadness’ as a reason to uphold the constitutionality of the provision (para 108), we argue that the doctrine of manifest arbitrariness would require the exact opposite. The overbreadth inherent in the ordinary meaning of recommendation, and the lack of legislative guidance with respect to its interpretation, creates an amorphous category that could encompass virtually all official acts. There is an absence of an adequate determining principle as to what would come under the term itself, injecting significant uncertainty into the entire legal process. First, if almost every official act can be redescribed as a “recommendation”, then investigability itself becomes indeterminate and it becomes impossible for the police to know which acts they are even allowed to look at in the first place. Second, given this indeterminacy, the executive branch would in effect acquire the power to unilaterally determine whether a particular action qualifies as a “recommendation” at the instance of a complaint.

Tilting the Scales: Minimal Harm, Maximum Restraint

In Association of Democratic Reforms vs Union of India, the court laid down a two-part test for when the doctrine of manifest arbitrariness can be used to strike down a provision. First, where the legislature fails to make a classification by recognizing the degrees of harm; andsecond, where its purpose is not in consonance with constitutional values. In applying this test, Courts are required to make a distinction between the  “ostensible purpose” (purpose claimed by the State) and the “real purpose” (purpose identified by Courts based on the available material) of a given provision/law. We argue that upon examination of Section 17A, its real purpose is to shield corrupt public servants from criminal prosecution, rather than protect honest officers from vexatious/false investigations.

Ostensible harm addressed by Section 17A

The State’s claimed objective  (accepted by Viswanathan J.) is:

• Protection of honest public servants from frivolous, vexatious, or motivated complaints
• Avoidance of chilling effect, “policy paralysis”, and risk-averse administration [para 41]

Nature and probability of the ostensible harm

Such an inquiry would require answering two questions. Namely, (a) what is the harm that a ‘frivolous’ complaint has on an individual public servant, and (b) what is the likelihood that a public servant is subjected to such a complaint.

Addressing (a), the harm of a frivolous complaint is largely non-coercive and primarily reputational in character. A preliminary enquiry may carry significant negative stigma, generate adverse perceptions among colleagues, and even induce professional anxiety. However, it is unlikely to have any permanent legal consequences because (i) In any case, the State can deny sanction to prosecute Section 19 of the PCA and (ii) If the case was truly frivolous, i.e completely lacking merit, it would not result in conviction at trial.

Turning to (b), the likelihood of a public servant being subjected to such a complaint appears limited on the State’s own record (para 5). Of the 2,395 cases in which prior approval under Section 17A was sought (presumably since the provision’s introduction in 2018), sanction was refused in approximately 41.3% (989 cases). Even assuming that all such refusals correspond to frivolous complaints (which is highly generous given the lack of determining principle at the stage), this figure is marginal when assessed against the exceptionally wide class of persons covered by the PCA’s definition of ‘public servant’. The breadth of the definition is evident from its inclusion, inter alia, of “any person who holds an office by virtue of which he is authorised or required to perform any public duty” and “any person in the service or pay of the Government.”

The exceedingly low percentage of so called frivolous complaints filed, alongside the limited harm caused by them to an individual civil servant is unlikely to cause a ‘policy paralysis’ in the sense that Viswanathan J. describes it.

Countervailing harm caused by Section 17A                                                                                  

The real or countervailing harm here is of an entirely different order than the ostensible harm  sought to be prevented by the state. Section 17A creates a chokehold on the possibility of a preliminary inquiry, and by logical corollary, eventual prosecution and conviction of a public servant at a stage where no material evidence relating to the merit of the complaint exists. The most clear harm that would follow is a systemic under-enforcement of anti-corruption law, and de facto legal immunity for corrupt public servants. Most corruption offences rarely present themselves through self-evident facts on the face of a complaint, and require the construction of complicated evidentiary links to prove, needing access to (among other things) documentary records, witness testimony and forensic analysis. By enabling even the most basic inquiry to be forestalled by the executive branch, who is both an interested actor in the outcome of a given case and faces no negative consequences for immunising guilty officers because (a) such officers will not be found out in the first place owing to lack of sanction; and (b) the State could merely claim that it was unaware at the pre-investigative stage. The provision therefore makes it exceptionally simple for corrupt officials to get off without consequence.

Real Purpose

Weighing the state’s stated objective with how it is has been operationalised through Section 17A paints a very interesting picture. The supposed harm caused by a frivolous complaint is fairly minimal (and self-resolving through Section 19) and is clearly overstated – as the empirical data the State provides shows. On the contrary, the harm caused by Section 17A sanction requirement lies in the systematic under enforcement of anti-corruption law, thereby   eroding  both the rule of law and public accountability (undoubtedly important constitutional values). A comparison of these harms inevitably leads to the conclusion that the ‘real purpose’ of the provision does not lie in protecting ‘honest’ officers from vexatious investigations.

Conclusion

This blog has aimed to evaluate Section 17A through the lens of manifest arbitrariness, given the relevance such an approach might have to a larger bench tasked with clarifying the law on this point. We have discussed (a) how a pre-investigative sanction regime necessarily lacks adequate determining principle, (b) the problems created by the indeterminate scope of ‘recommendations’ and (c) how the law employs such an excessive, irrational and disproportionate means to protect honest officers from vexatious investigations that it ends up primarily shielding corrupt officers from public accountability.

Before concluding, we find it worth recalling the slogan of the Party in George Orwell’s 1984:

“War is peace.
Freedom is slavery.
Ignorance is strength.”

And similarly; 17A protects honest officers.

While the provision is framed with the stated purpose of protecting honest public servants, it unfortunately constructs an architecture that systematically protects corrupt ones. We hope that a future bench of the Court breaches the trojan horse of Section 17A, and holds it to be wholly unconstitutional.

Eyes Wide Shut: The Supreme Court’s Bail Order in the Delhi Riots Cases

This post analyses the Supreme Court’s bail order in the Delhi Riots cases, delivered today. Seven appellants had approached the Court against the Delhi High Court’s September 2025 order denying them bail. The Supreme Court allowed the appeals of five, and granted them bail. The appeals of Umar Khalid and Sharjeel Imam were rejected.

This blog has examined the proceedings in this case in some detail, dating back to 2022; we have examined the 2022 trial court order denying bail (here), the 2022 High Court order denying bail (here), and the 2025 High Court order denying bail (here and here). An analysis of these orders has revealed a consistent pattern: at each stage, the courts have adopted what this blog has called an “eyes wide shut approach” to the question of bail under the UAPA: that is, they have invoked Section 43(D)(5) of the UAPA to avoid any meaningful scrutiny of the material presented by the prosecution, and have filled in the obvious gaps in the prosecution’s case through inferences and assumptions. When it comes to Umar Khalid and Sharjeel Imam, today’s judgment, as we shall see, follows the same pattern.

The Question of Delay

The Court frames two preliminary questions, which we shall address. The first is the question of delay. At this point, all the appellants had spent upwards of five and a half years in jail, without trial. They invoked the right to life under Article 21, which also guarantees the right to a speedy trial. In addition, the Supreme Court’s prior judgment in Najeeb, which holds that a delay in a trial under the UAPA will attract Article 21, was also cited.

How does the Supreme Court justify the continued incarceration of Umar Khalid and Sharjeel Imam, even after five and a half years in jail without trial? It does so by deploying two arguments: first, that the delay in the trial is not “solely” attributable to the prosecution or to the court; and secondly, that the “seriousness of the offence” is a relevant factor in deciding the question of delay. Neither reason stands up to scrutiny.

On the first reason: the Court observes that from the record, it appears that despite the prosecution’s readiness to proceed, “objections, requests for deferment, and issues relating to sequencing of arguments were raised on behalf of the accused.” The Court does not actually engage in any analysis with respect to how much of the delay was actually caused by this, which is a rather glaring omission in the analysis. However, let us assume, for the purposes of argument, that some part of the delay is attributable to “objections, requests for deferment, and issues relating to sequencing of arguments” by the accused. Two points follow. First, “objections,” “requests for deferments,” and “sequencing issues” are not the same. It would be remiss of defence counsel to not raise all the objections that criminal law entitles them to raise, on behalf of their client. Pinning a delayed trial on an accused individual’s lawyer invoking legal remedies cannot, therefore, stand. This leaves only “requests for deferment.” But, ultimately, it is the court that decides whether or not to accede to a “request for deferment.” Defence counsel are not in charge of a trial. Defence counsel do not control the courtroom. The court does so. The court can always decline a request for an adjournment and compel the defence to argue. The responsibility for a delayed trial, therefore, lies only on the shoulders of the court conducting the trial. This point is obscured by the Supreme Court’s eliding of “objections”, “requests for deferment” and “sequencing issues”, as if they were all the same, and that they all pointed to the defence’s culpability.

On the second reason: the Supreme Court states that the weight accorded to delay must be “balanced” with the gravity of the alleged offence. This is incoherent. A delayed trial breaches Article 21 because every individual has the right to a speedy trial. How serious an offence is may have a bearing upon consideration of bail on merits, but it has no bearing on the right to a speedy trial. Drawing this link, in fact, opens a dangerous Pandora’s Box where fundamental human rights are made contingent upon what the State chooses to accuse an individual of. In fact, the Supreme Court itself has recognised the untenability of such distinctions, across contexts: for example, the Court has not accepted the proposition that the its rule for hearing death penalty reviews in “open court” be curtailed for particularly serious cases, such as terrorism convictions – because the right to an open court review itself flows from Article 21.

In fact, the Supreme Court’s order here is contrary to its own prior bail jurisprudence under the UAPA: in Sheikh Javed Iqbal vs State of Uttar Pradesh, the Court specifically noted that, in fact, the more serious the offence, the greater the imperative that the trial be concluded expeditiously, and that the seriousness of the offence cannot be invoked in order to make trial delays constitutionally acceptable. Today’s judgment makes no mention of Javed Iqbal, and advances a proposition directly contrary to it. This is not something that co-ordinate benches of the Court can do.

Finally, it is notable that when courts examine the question of delay, they specifically look at the number of witnesses that have to be examined, and the probable further time that it will take for the trial to be completed. None of this is done in the present case.

Definition of Terrorism

As a second preliminary point, the Court considers the definition of “terrorism” under Section 15 of the UAPA. This is relevant because – as this blog has pointed out before – the test for bail under Section 43(D)(5) of the UAPA is whether a “prima facie” case exists against the accused. The question of whether or not a “prima facie case” exists depends on what the ingredients of the offence are. These, therefore, must be outlined with care and specificity.

Unfortunately, the Court is neither careful nor specific. It notes that in addition to weaponry, Section 15 uses the term “by any other means of whatever nature” which – according to the Court – includes “conduct that destabilises civic life or societal functioning, even in the absence of immediate physical violence.” There are two problems with this definition. The first is that it is entirely vague. Every protest that blocks a road “destabilises civic life or societal functioning.” Every such protest is not – obviously – statutory terrorism. Secondly, this definition conflates a number of concepts that decades of Supreme Court jurisprudence have carefully articulated – and separated – from each other. Every law student is familiar with the Supreme Court’s judgment in Ram Manohar Lohia, which advanced the famous “concentric circles” image to distinguish between a disruption of law and order, public order, and of State security. Today’s judgment does not engage in any analysis of these concepts, and indeed, appears to conflate the three circles in the umbrella term “destabilises civic life.”

Why is this a problem? It is a problem, in general, because such a vague definition of “terrorism” ensures that pretty much every act of civic disobedience can invite a UAPA case, and endless incarceration. It is also a problem in this specific case, because the foundational reason for the Court’s decision to deny Sharjeel Imam and Umar Khalid bail is – in its view – the difference between a “dharna” and a “chakka jam.” In my post on the High Court’s judgment, I wrote:

On the chakka jam, it is worthwhile to remember, for a moment, that we live in a country that owes its existence, as a nation-state, to forms of mass mobilisation and protest that were, by their very nature, designed to be disruptive (and the chakka-jam is one among them). Nor is this restricted to the freedom struggle: in a book titled Hailing the State, Lisa Mitchell examines a range of unconventional ways in which Indians have communicated their claims to the State outside of the electoral process: from the “rail roko” to the “chakka jam.” These forms of protest have a specific history and a specific vintage, and have been practiced by Indians of different persuasions, across time and space. It may be argued that the State reserves the right to respond to such methods through the legal form; while that it is a separate debate, what the State has done here is to use anti-terror laws to deal with what was a political protest (recall, once again, that the link between the protesters and the violence remains unproven). That is where the problem lies.

The Supreme Court’s catch-all definition of “terrorism,” therefore, invites selective prosecution – followed by selective incarceration – under the UAPA.

It is important to note one more thing: two years ago, in Asif Iqbal Tanha, the High Court of Delhi embarked on a very detailed and very careful examination of the phrase “terrorism” under the UAPA (see here). In that case, the High Court of Delhi noted the following:

… the extent and reach of terrorist activity must travel beyond the effect of an ordinary crime and must not arise merely by causing disturbance of law and order or even public order; and must be such that it travels beyond the capacity of the ordinary law enforcement agencies to deal with it under the ordinary penal law. (para 49)

The High Court also noted:

The making of inflammatory speeches, organising chakkajams, and such like actions are not uncommon when there is widespread opposition to Governmental or Parliamentary actions. Even if we assume for the sake of argument, without expressing any view thereon, that in the present case inflammatory speeches, chakkajams, instigation of women protesters and other actions, to which the appellant is alleged to have been party, crossed the line of peaceful protests permissible under our Constitutional guarantee, that however would yet not amount to commission of a ‘terrorist act’ or a ‘conspiracy’ or an ‘act preparatory’ to the commission of a terrorist act as understood under the UAPA. (para 47)

Now, when this case was appealed to the Supreme Court, instead of engaging with the High Court’s reasoning, the Supreme Court simply declared that it would “not be treated as precedent” (see here). The result is what we have today, which is a Section 15 definition that is so wide and vague that more or less any disruptive act can come within it – and whether or not a particular act will come within it depends entirely on how the State, in a given case, decides to frame it. This is not the rule of law.

The Facts

This blog has gone into the details of the factual allegations against Sharjeel Imam and Umar Khalid – and why they do not stand up to UAPA scrutiny – in some detail, on three previous occasions. We must now do it a fourth time. Let us first consider Umar Khalid. As has been noted before, the prosecution’s case has always had a fundamental problem: that Umar Khalid never called for violence, was not involved in any violence (he was not even in Delhi when the riots broke out), and no material was recovered from him. To get over this rather substantial hurdle, the prosecution has alleged the existence of a conspiracy, and that Umar Khalid was one of the two masterminds (Sharjeel Imam being the other).

But the allegation of a conspiracy must be based on something. Here, we see a rather interesting evolution in how the courts have adjudicated this case over the years. In the beginning, the courts have tried to nonetheless find something concrete or specific in the facts (this was taken to the point of absurdity in the High Court’s first order denying bail, where Umar Khalid’s reference to “revolutionary greetings” in a speech was taken as potentially violent because he did not specifically clarify that was talking about a peaceful revolution). However, judgment by judgment – perhaps noting the untenability of these claims – courts have begun to frame Umar Khalid’s role in more and more abstract language, abandoning specificity altogether.

In the Supreme Court’s order, this abstraction is taken to its vanishing point. Note the following: in paragraph 98, the Court says this: “the prosecution material comprises direct, corroborative, and contemporaneous evidence, including recoveries, digital communication trails, and statements indicative of managerial responsibility.” First, evidence of what, precisely? Later on, we are told that the only evidence is that of starting a “chakka jam” (see above), but that to go into this question in more detail would involve scrutiny beyond what is permitted by Section 43(D)(5). Secondly, in paragraph 215, the Court itself admits that there have been no “recoveries” – so it is unclear why it mentions “recovery” in paragraph 98, when this is – quite literally – not true. In that same paragraph 215, the Court says that the absence of recovery may have weight in “ordinary IPC bail jurisprudence” but not under a “statute alleging a wider conspiracy.” But “conspiracy” is not a concept invented by the UAPA – it is, quite literally, a concept defined – and interpreted – under the IPC!

In paragraph 211, the Court states that the prosecution’s narrative is “not episodic; it is architectural.” What on earth does this mean? The Court then says: “It asserts a phased progression: mobilisation and indoctrination; institutionalisation through committees and digital platforms; expansion of protest sites into permanent blockades; preparation for escalation; and culmination in coordinated chakka jams and widespread violence.” Note that none of this except the last two words constitute an offence under the UAPA (I have dealt with the chakka jam point above). What we would therefore need from the Court is the following: a plausible causal link between every word used before the words “widespread violence,” and the “widespread violence” itself. The moment the Court is asked to do this, however, it repeats that such a level of scrutiny is barred under Section 43(D)(5).

To put it schematically: the UAPA requires (conspiracy to) commit “widespread violence.” The actual evidence is well short of establishing any such thing. The Prosecution asserts that this gap between the evidence and the act is filled by “conspiracy.” The Court states that it is barred from examining this (evidence-free) assertion. To take an example, consider para 218:

At this stage, the Court does not decide whether each meeting was conspiratorial. But where multiple meetings across weeks and months are alleged, and where witness material and electronic records are pressed to place the same accused at several of these junctures, the Court is entitled to view continuity itself as a relevant circumstance. Continuity is the difference between a participant and an organiser in the prosecution narrative.

But, for the umpteenth time, we have to ask: continuity of what? Continuity to do what? In the paragraph before, the Court has referred to the following “decision points” (decision of what?):

... the Jangpura meeting of 08.12.2019; the Jamia phase of mid-December; the formation and functioning of JCC/JACT as mobilisation structures; the creation and operation of DPSG as a dissemination mechanism; the meetings at ISI, Gandhi Peace Foundation and Shaheen Bagh; and the January Seelampur phase alleged to involve preparation for escalation.

But once again, it is not in these meetings that the violence took place. There is always a permanent gap between the actual evidence against Umar Khalid and the allegation against him, and no matter how many multi-syllabled words are used in the judgment (“architectural”, “conceptual leadership”, “architect of escalation”), this is a gap that will remain unfilled, because it cannot be filled except through a effort of prosecutorial and judicial imagination.

The final giveaway comes in para 239. After insisting throughout the judgment that its task is to scrutinise the material to see if there is a reasonable nexus with the ingredients of the offence, the Court finally concludes its analysis of Umar Khalid’s case with this: “at the bail stage, the Court does not determine whether these allegations ultimately satisfy the ingredients of a “terrorist act” under Section 15.”

But then, what are we doing here? If the Court will not even ask if the allegations satisfy the ingredients of the offence, then what will it do? What is left for it to do?

In fact, this same issue arises in the Sharjeel Imam bail denial. I have written at some length before about the facts in Imam’s case, and how his speeches – in particular – do not meet the standard of criminalisation under the Constitution (see here). Under the Constitution, speech is protected unless it constitutes incitement to imminent violence; Imam’s speech does not fall within this judicial test. Before the Supreme Court, Imam’s counsel made this argument. The Court says, in para 162:

The defence has urged that the speech contains no direct incitement to violence and is protected. That contention cannot be adjudicated in the manner the defence invites at this stage. The statutory enquiry is not whether the Court, after a full trial, would accept the prosecution’s interpretation. The enquiry is whether the prosecution’s reading is prima facie plausible on the face of the material and whether, read cumulatively with the other links, it contributes to a coherent narrative of planned disruption.

With due respect, this gets the law upside down and back to front. Consider the argument schematically:

Proposition 1: The Constitution prohibits criminalising speech that is short of incitement to imminent violence.

Proposition 2: The UAPA must be interpreted consistently with the Constitution. Consequently, speech that does not incite violence does not fall within Section 15.

Proposition 3: The question of whether or not a speech constitutes incitement, therefore, is essential to determine whether the ingredients of Section 15 have been satisfied, and therefore, whether a prima facie case exists against the accused.

This should make it crystal clear that the question of incitement is fundamental to the question of bail under UAPA. If there is no incitement, then there is no Section 15 offence, and – therefore – no prima facie case. The Court does not engage in any of this analysis.

Miscellaneous

Reading the orders granting bail to the other appellants, a rather curious picture emerges: it is as if the other individuals are automatons, mechanically carrying out the orders delivered from on high by Sharjeel Imam or Umar Khalid. We are treated to phrases such as absence of “autonomous decision-making authority,” “limited to coordination,” “layered participation,” “site level participant whose presence and conduct derive meaning primarily from directives purportedly issued by others,” “conceptual command,” “multi-layered mobilisation architecture,” “conduit for information,” “operational execution,” and so on.

These phrases appear to have been borrowed from the domain of corporate governance, so let’s also use a corporate law phrase: when you pierce the veil that has been drawn by these multi-syllabled words, and when you go back to check the actual evidence that distinguishes what Sharjeel Imam and Umar Khalid have done vis-a-vis what the others have done, there is absolutely nothing on the record that justifies this scary image of two individuals sitting at the node of a vast conspiracy, issuing directions to subordinates who then act unthinkingly on them. This image is built entirely upon a set of assertions that exist in the gap between the actual pieces of evidence.

The tragedy of the case, ultimately, lies in this: that instead of scrutinising these gaps, the courts – at all levels – have chosen to fill them up. As we have seen before, the UAPA does not compel courts to do this. In this case, however, the Courts have chosen an eyes wide-shut approach. The result is that only measured in an indefinite, open-ended, and continuing imprisonment, without trial or conviction.

Orders

A final point. In its order granting bail, the Court has set twelve conditions. Conditions (vii) and (ix) are:

viii. The appellants shall not participate in any programme or address or attend any gathering, rally or meeting, whether physically or virtually till conclusion of the trial.

ix. The appellants shall not circulate any post either in electronic form or physical form or circulate any hand bills, posters, banners, etc in any form whatsoever.

It is one thing to limit rights in relation to the case at hand. But what justifies a sweeping judicial abrogation of the appellants’ Article 19 rights to speech, assembly, and association altogether? How does any of this have any relation to the case actually before the Court – one in which, it is worth remembering, the trial is yet to commence? As with the rest of the judgment, we ask these questions in vain, because no answer is forthcoming.

Conclusion

With respect to Umar Khalid and Sharjeel Imam, the Court states that they are entitled to re-apply for bail after the “protected witnesses” have been examined, or after one year from the date of today’s order, whichever is earlier. An optimistic reading of this direction would suggest that the Court is of the view that at the most, the “balance” between delay and the “gravity” of the offence would – after six and a half years – finally tip in favour of the former. To quote a piece written by two other individuals who have spent years in jail without trial, perhaps then the question – “how long is too long?” – will, at last, be answered that “now, at last, it is too long.” Until then, we can only wait.  

---

[Thank you to Paras Nath Singh for pointing out the SC’s coordinate bench judgments on delay.]