Welcome to The Bridge! India's regulators affect markets, govern disputes, and test the rule of law through their orders, regulations, and policy implementation actions. Yet the quality of this work rarely gets the scrutiny it deserves. The Bridge is TrustBridge's window into India's regulatory state: how it decides, how it regulates, and how it could do both better.

Introduction

In a recent decision (Amazon v. Competition Commission of India, C.A. No. 4974 of 2022), the Supreme Court set aside an order of the Competition Commission of India (the “CCI”) which imposed a penalty of Rs. 202 crore on Amazon in relation to its 49% acquisition in Future Coupons Private Limited. The Court held that the CCI had exceeded its statutory powers, that the penalty provisions had been misapplied, and that the proceedings violated natural justice.

This decision has received attention not only for this outcome, but also for the Court’s statements in a non-operative part of the judgment (Part H), titled ‘Role of the regulator and standards of fair regulatory conduct’. This part does not address any of the issues involved in the dispute before the Court - the paragraph immediately preceding Part H states explicitly that all issues have been answered. As such, this part is entirely obiter, a Latin term that means ‘said in passing’, or ‘by the way’. The problem is that these passing statements may enable fallacious arguments in the future, and hold the potential to damage the core rule of law value of the separation of powers.

Enabling fallacious arguments and lack of authority

Part H makes claims that have nothing to do with the matter at hand. The preceding section Part G, offers a summary of the Court’s conclusions on the issues involved in the judgment. And yet, Part H goes on to offer the Court’s perspective on value and principles of economic regulation. It makes claims about risk premia, supply chain realignments, and the strategic value of regulatory certainty to India’s investment climate. Some might find these statements elegant and compelling - but the Court’s claims on macroeconomic consequences of competition policy are completely unrelated to the questions raised before it, and outside the scope of its powers.

Concerns over statements such as these matter because broad, rhetorically (debatably) attractive judicial language travels. It gets extracted from its context, cited in unrelated disputes, and used to support propositions which the original judgment had no authority on. Obiter pronouncements do not hold binding force; yet they are extensively relied on by lower courts for their persuasive value. The Supreme Court should therefore have acted with restraint instead of writing this essay-style exposition that encompasses its view on how regulators should behave on core policy matters. This lack of restraint opens the gates to a series of future arguments based on Part H that are logically fallacious and unsupported by authority.

Violation of the Separation of Powers

The Court has often taken the position that matters of policy are outside its remit. These matters have rightly been held to fall within the domain of the Executive. The Court’s language in this regard is apposite to our discussion. Note the Court’s self-cautionary position in this regard from the past:

In Peerless General Finance and Investment Co. Ltd. v. RBI, the Court held:

“The function of the Court is to see that lawful authority is not abused but not to appropriate to itself the task entrusted to that authority… Courts are not to interfere with economic policy which is the function of experts. It is not the function of the Courts to sit in Judgment over matters of economic policy and it must necessarily be left to the expert bodies. In such matters even experts can seriously and doubtlessly differ. Courts cannot be expected to decide them without even the aid of experts.” (Emphasis added)

Upon completing an exhaustive survey of its past decisions on this question, the Court has observed in BALCO Employees Union v. Union of India:

“It is evident from the above that it is neither within the domain of the Courts nor the scope of the judicial review to embark upon an enquiry as to whether a particular public policy is wise or whether better public policy can be evolved...

“…in the sphere of economic policy or reform the Court is not the appropriate forum...

“...In the case of a policy decision on economic matters, the Courts should be very circumspect in conducting any enquiry or investigation and must be most reluctant to impugn the judgement of the experts who may have arrived at a conclusion unless the Court is satisfied that there is illegality in the decision itself.” (Emphasis added)

Yet, in Amazon v. CCI, the Court ignores its past statements, and takes as a hook for its policy prescriptions a single phrase in the Competition Act’s preamble: “the economic development of the country.” From it, the Court derives the CCI’s statutory purpose and its warrant for everything that follows: the principles of economic regulation, the link between competition policy and investment climate. Not stopping at the CCI’s purpose and the statutory scheme, the Court makes statements about foreign investment and investors, it cites Adam Smith to support “the basic point that broader markets can sustain more specialised activity and deeper rivalry”, and Amartya Sen to claim that “global economic integration is neither new nor inherently one-sided, and that the central question is whether its gains are shared in a fair and inclusive manner”.

But this hook is too thin to bear the weight the Court hangs on it. This is because the Court misses several important points in doing so:

Preambulatory guidance may direct Executive decision-making; it does not act as a license for the judiciary to overstep the boundaries of its powers. The Preamble is an interpretive aid, not a source of substantive powers. It may assist in resolving ambiguity in enacted provisions, but does not authorise the Supreme Court to pronounce policy positions. The Court’s expertise lies in legislative interpretation, not Macroeconomic policy - and it mistakenly relies upon the Act’s Preamble to assume such expertise to prescribe what relation regulation bears to foreign investment flows.

Perhaps most egregious: in 2011, the Ministry of Corporate Affairs prepared a draft National Competition Policy, which articulated the Executive’s considered position on precisely these matters. That draft was developed through a committee process with national consultations - a much better way of forming policy than the pronouncements of a two-judge bench of the Supreme Court. The Court, in four pages of obiter, has now effectively sketched its own version of that policy without the consultation, without the coordination, and without the accountability that accompanies legitimate policy-making. The draft policy has not made it to Parliament, and remains incomplete. If the Court believed that it absolutely needed to make a comment on competition regulatory policy, it could have called on the Ministry of Corporate Affairs and the CCI to finalise that policy, through proper legislative or executive channels. Instead, it substituted its own judgment for the Executive’s. This is a massive red flag for separation of powers.

Conclusion

Part H of the Amazon v CCI joins a growing line of Supreme Court decisions that raise concerns of judicial restraint and separation of powers. The Court must check its temptation to comment on normative questions of policy, as it has always done. A failure to do so would enable future litigants to construct fallacious arguments which would compromise the integrity of precedents as sources of law.

These excesses are also illustrative of how the core rule of law value of the separation of powers may be violated. Separation of powers requires that distinct governmental functions - policy-making and adjudication, in this case - be recognised as such, and be exercised by different persons. This articulation of governmental function sits at the heart of the idea of the rule of law, and, along with the concepts such as checks and balances, are critical to the prevention of tyranny. The Court has often raised the alarm when these ideas are under threat: it would do well to check itself when tempted to violate them itself.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback and suggestions.

CITATION

Amrutha Desikan and Bhavin Patel, 2026. “Amazon v CCI: Sixteen Paragraphs Too Many”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Amazon.com NV Investment Holdings LLC v. Competition Commission of India and others, Civil Appeal No. 4974 of 2022.

BALCO Employees Union v. Union of India, AIR 2002 SC 350.

Government of India, Ministry of Corporate Affairs, National Competition Policy (2011).

Peerless General Finance and Investment Co. Ltd. v. RBI, (1992) 2 SCC 343.

Welcome to The Bridge! India's regulators affect markets, govern disputes, and test the rule of law through their orders, regulations, and policy implementation actions. Yet the quality of this work rarely gets the scrutiny it deserves. The Bridge is TrustBridge's window into India's regulatory state: how it decides, how it regulates, and how it could do both better.

Introduction

The Competition Commission of India (the “CCI”) is empowered with quasi-legislative authority, and can make regulations that it has the power to enforce. The delegation of these powers into an unelected authority raises concerns of accountability towards the delegating authority, that is, the legislature. While S. 64 of the Competition Act, 2002 (the “Act”) empowers the CCI to make regulations and provides that these regulations must be laid before Parliament, we have not noticed any evidence of legislative scrutiny of such subordinate legislation.1

The Lok Sabha and Rajya Sabha Committees on Subordinate Legislation play an important role in checking the power of delegated legislation exercised by regulators and ensuring accountability. On 21 July 2026, the Committee on Subordinate Legislation, Rajya Sabha (the “Committee”) presented its report on three sets of regulations and one set of guidelines issued by the CCI. For context, this Committee was constituted in 1964 pursuant to the Rules of Procedure and Conduct of Business in the Rajya Sabha to examine whether delegated legislative powers are being exercised in the intended manner.

We analyse the Committee’s recommendations on the CCI (Determination of Monetary Penalty) Guidelines, 2024 (the “Penalty Guidelines”) in this post. While the Committee identifies shortcomings in the CCI’s penalty determinations, we argue that the solution lies not in simply asking the CCI to enforce the Penalty Guidelines more rigorously, but in improving decision-making processes and the framework for periodic review.

The legal framework and the Committee’s recommendations

The CCI can issue regulations further to S. 64 of the Act, non-binding guidelines pursuant to S. 64B(1), and guidelines on penalty determination under S. 64B(3). S. 64B(4) mandates the CCI to take into account any guidelines issued under S. 64B(3), such as the Penalty Guidelines, when calculating penalty amounts, and to provide reasons for deviations from the Guidelines.

The Committee noted that a substantial portion of penalties imposed by the CCI are stayed or quashed on appeal, and the need for:

1. a transparent methodology for calculation of penalties,

2. strengthening investigations and penalty assessment, and

3. enforcement of the Penalty Guidelines, particularly with respect to the consideration of repetitive violations as an aggravating factor.

The Committee acknowledges an important problem: the Act requires the CCI to adopt a structured framework for penalty determination, yet such determinations do not withstand appellate scrutiny.

So what should be done?

Simply asking the CCI to enforce the Penalty Guidelines is not enough. Instead, the CCI must alter the way that it writes orders that impose penalties, and the frequency and manner of reviewing the Penalty Guidelines:

1. An order-writing template: The CCI should adopt a standardised order-writing template for every order, particularly orders that impose penalties. The template should require the CCI to discuss each aggravating and mitigating factor identified in the Penalty Guidelines, and indicate (with reasons) whether that factor is applicable or not. Such a template would structure the CCI’s orders and require adjudicators to consciously evaluate each consideration before arriving at a penalty amount. The template would also help regulated entities as well as appellate fora understand how the CCI arrived at the final penalty amount. Appellate fora are also constrained by the provisions of the Act - they may not fault the CCI for complying with its mandate under the Penalty Guidelines and S. 64B(3) of the Act. So, a clear, structured response on all the considerations in the Penalty Guidelines would mean that fewer cases are quashed in appeal.

2. Strengthening periodic review: S. 64A of the Act already requires the CCI to periodically review its regulations, however:

   • The requirement to periodically review is limited to regulations, and does not extend to the Penalty Guidelines. The legislature should subject the Penalty Guidelines to the same framework that applies to regulations. This would help ensure that the regulator has a set of relevant, updated, and applicable Guidelines to rely upon in its penalty determinations.

   • There is no timeline specified for periodic review in the Act.2 By way of contrast, the International Financial Services Centres Authority’s regulations require it to review its regulations every five years. Without a prescribed review cycle, periodic review becomes entirely discretionary: if the CCI decides not to revisit the Guidelines, they may remain unchanged despite repeated appellate criticism, changes in enforcement practice, or evolving market conditions.

   • There are no specified considerations that the CCI must take into account while conducting such reviews. For example, the Securities and Exchange Board of India’s regulations require it to account for the following in periodic reviews: (i) experience gained through surveillance, supervision and enforcement actions, and (ii) decisions by courts or tribunals.3 This creates a feedback loop whereby emerging case law feeds into regulations and guidelines, improving clarity and predictability of the regulatory regime. The CCI should similarly review its own orders and the appellate performance of those orders to identify recurring difficulties in applying the Penalty Guidelines and refine them accordingly.

Conclusion

It is insufficient to merely direct the CCI to pay closer attention to the Penalty Guidelines. Such directions must be accompanied by mechanisms that ensure the Penalty Guidelines are meaningfully implemented. We suggest two practical measures.

First, a standardised order-writing template would require the CCI to systematically consider every aggravating and mitigating factor under the Penalty Guidelines and explain, with reasons, how each factor influenced the final penalty.

Second, timely periodic review of the Penalty Guidelines should become a structured exercise rather than a discretionary one, informed by the CCI’s own enforcement experience and appellate decisions. Such periodic review must be hard-coded into the Act or, alternatively, included in a set of regulations on issuance of regulations and subsidiary instructions.

While Parliament’s track record on examining delegated legislation laid before it has been sub-optimal, the requirements of accountability may be satisfied somewhat if it takes proactive measures such as these to hold its delegatees accountable.

- The authors are researchers at TrustBridge, and would like to thank Pratik Datta and Renuka Sane for their feedback.

CITATION

Natasha Aggarwal and Bhavin Patel, 2026. “Strengthening the CCI's penalty determination framework”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Anirudh, 'How Well Does Parliament Examine Rules Framed under Various Laws?' (The PRS Blog, PRS Legislative Research, 10 January 2011) https://prsindia.org/theprsblog/how-well-does-parliament-examine-rules-framed-under-various-laws.

Read on to see what’s kept us busy.

Policy inputs and submissions

Comments on the Ministry of Power’s Draft National Electricity Data Sharing Framework

On July 21, Chitrakshi Jain, Renuka Sane, and Rakesh Kacker submitted comments on the Ministry of Power’s draft National Electricity Data Sharing Framework, which proposes a voluntary data portal with public and access-controlled data tiers. Key recommendations include:

• the framework does not identify the source of statutory authority for the proposed data portal, and the Central Electricity Authority should instead use its powers under the Electricity Act to mandate disclosure through amended or fresh disclosure regulations;

• “public data” should be interpreted broadly to include datasets covering the adjudicatory work of Electricity Regulatory Commissions; and

• outage-related data should be published in more granular form and at greater frequency.

Comments on the RBI’s Draft Guidance on Regulatory Principles for Model Risk Management, 2026

On July 22, Renuka Sane, Pratik Datta, Aditya Sai Phutane, and Amol Kulkarni submitted comments on the Reserve Bank of India’s Draft Guidance on Regulatory Principles for Model Risk Management, 2026. Key recommendations include:

• define “material impact,” or exclude simple tools like basic spreadsheets from the definition of “Model”;

• scale institutional duties (three lines of defence, RMCB oversight) by size or complexity, not by each entity’s own judgment;

• clarify how third-party/AI models can be validated when providers won’t share technical details;

• replace a fixed explainability threshold with guidance on who needs to understand the explanation, and what makes it adequate for them; and

• link reporting frequency and retention periods to a model’s risk tier, not a single fixed period for all.

Papers

On July 27, Prashant Narang and Ishika Tanwar published 'Disortho v. Meril: Error Costs and the Architecture of Arbitration' in the Journal of International Arbitration, Vol. 43, Issue 3/4 (2026). The paper critically examines the Supreme Court's decision in Disortho S.A.S. v. Meril Life Sciences Pvt. Ltd. and, using an error-cost framework, argues for a clearer seat-based default for determining the law governing international arbitration agreements. Below is the abstract of the paper.

The article critically examines Disortho S.A.S. v. Meril Life Sciences Pvt. Ltd., a recent Indian Supreme Court ruling that reshapes how governing law, seat, and venue interact in international arbitration. The dispute involved a hybrid clause combining Indian governing law and Gujarat courts with Bogotá as the arbitration venue under Colombian rules — and the Court applied a presumption based on the law of the contract, treating Indian law as governing the arbitration agreement while reading Bogotá as merely a venue rather than the legal seat. The authors place this decision within India's post-BALCO seat-centric case law and compare it with Sulamérica, Enka, and the UK's codified seat-based default rule. Using an error-cost and institutional-efficiency lens, they argue that a seat-default presumption — treating the named place as the legal seat and applying that seat's law to the arbitration agreement — reduces systemic error, discourages opportunistic behavior, and improves administrative clarity. The article concludes by proposing a combined approach: courts reaffirming seat-based presumptions, paired with potential legislative clarification for cross-border arbitration.

On August 3, Natasha Aggarwal, Amol Kulkarni, Bhavin Patel, and Vishnu Suresh, along with Shruti Aji Murali (Axiom5 Law Chambers), co-authored two papers on the Competition Commission of India’s order-writing at the threshold stage:

• ‘Order writing and statutory discretion in the CCI’s Section 26(2) determinations’, which examines the completeness of the CCI’s order-writing under Section 26(2) of the Competition Act, 2002, and the adequacy of the statutory framework governing the CCI’s discretion at the threshold stage. Here is the abstract of the paper —

The Competition Commission of India (CCI) exercises significant discretion in its preliminary screening function under Section 26(2) of the Competition Act, 2002. These orders close proceedings at the threshold stage without directing investigation, are appealable, and determine the rights of the informant. This paper assesses the completeness of the CCI’s order-writing in S. 26(2) matters and the adequacy of the statutory framework governing this discretionary power. We evaluate a random sample of 111 S. 26(2) orders issued between 2014 and 2024 using customised Good Order Writing indicators and examine appellate outcomes.

This was supplemented with semi-structured interviews with practitioners and CCI officials. We find that orders are generally complete, with an average score of 59.51%, and perform well in appeal. However, we find that the statutory factors under Ss. 19(3) and 19(4) are applied inconsistently. We recommend incremental improvements in order-writing alongside the development of a dedicated framework for threshold stage determinations under S. 26(2).

• The CCI's allocation of scarce resources in Section 26(2) matters', which examines how the CCI allocates its limited regulatory capacity in its threshold-stage screening function under Section 26(2).

The Competition Commission of India (CCI) operates under significant capacity constraints, with nearly half its sanctioned posts vacant and a growing backlog of pending cases. The efficiency with which it allocates scarce resources in its threshold stage screening function under Section 26(2) of the Competition Act, 2002 therefore assumes considerable importance. We analyse a sample of 111 Section 26(2) orders issued between 2014 and 2024, interview practitioners and CCI officials, and survey screening practices across six jurisdictions.

We find that approximately 42% of the orders in our dataset arise from “peripheral matters”, that is, complaints falling outside competition law’s scope or unsupported by evidence, which are nonetheless processed through the same institutional procedures as substantive complaints, with a median disposal time of 30 days. We recommend a layered approach, including clearer public guidance, guided digital filing mechanisms, AI-assisted review tools, and statutory prioritisation frameworks, to improve resource allocation while preserving access to competition law enforcement.

On August 5, Prashant Narang, Aryan Pandey, and Vijay Chauhan published ‘India's Customs Advance Rulings Regime and the Limits of Legal Certainty’ in the Global Trade and Customs Journal, Vol. 21, Issue 7/8 (2026). The paper evaluates the performance and institutional design of India's Customs Authority for Advance Rulings (CAAR), drawing on rulings issued between 2021 and 2024, and proposes reforms to improve timeliness, nationwide applicability, and legal certainty.

This paper evaluates whether India's Customs Authority for Advance Rulings delivers the legal certainty and predictability required of an advance ruling regime under Article 3 of the WTO Trade Facilitation Agreement. An analysis of CAAR decisions issued between January 2021 and December 2024 shows that fewer than 53 per cent of rulings are delivered within the statutory three-month time frame prescribed under Indian customs law. CAAR's under-performance is best understood as a design failure, not a capacity constraint: port-specific applicability and a three-year validity limit undermine the very certainty that advance rulings are meant to provide, even where individual officers act diligently.

On August 12, Laveesh Bhandari (CSEP) along with Prashant Narang and Aryan Pandey co-authored a working paper, ‘Reassessment after the Finance Act 2021: From "Reason to Believe" to jurisdictional gates’, examining how the post-2021 reassessment framework under Sections 147–151 of the Income-tax Act, 1961 has worked in practice, through doctrinal analysis and a review of appellate outcomes. Given below is an abstract—

This paper argues that the 2021 reassessment reforms reflect a conscious legislative move toward rule-bound administration, even at the cost of occasional substantive under-enforcement in individual cases. By privileging record-verifiable statutory preconditions over open-ended evaluative standards, the framework deliberately constrains administrative discretion and shifts the focus of legality from ex post justification to jurisdictional compliance. From the standpoint of taxpayer certainty and the doctrine of repose, the stricter enforcement of jurisdictional preconditions appears to be functioning as intended. However, from the perspective of tax administration and revenue protection, the data points to a persistent institutional inability to comply with statutory requirements that are, on their face, neither complex nor ambiguous.

This paper makes two contributions to the literature on tax reassessment. First, it provides a structured empirical comparison of ITAT reassessment decisions before and after the Finance Act, 2021. Second, it shows that, in the sampled post-2021 decisions, reassessment proceedings are now rarely upheld, with tribunal review increasingly turning on compliance with statutory preconditions relating to limitation, sanction, and institutional authority rather 4 than on close scrutiny of the Assessing Officer’s reasons for reopening.

On August 13, Chitrakshi Jain and Akshay Jaitly published a working paper, 'Risks in electricity contracts'. The paper empirically studies how the Central Electricity Regulatory Commission adjudicates force majeure disputes in electricity contracts, examining 52 orders covering 77 party-level observations between 2018 and 2023.

The restructuring of India’s power sector marked a shift from negotiated contracts to competitive bidding as the primary mechanism for procuring generation capacity. As a result, private participation in the sector increased significantly. Despite the increase in participation of private parties in generation, the offtakers remain largely state owned. Standard bidding documents issued by the Ministry of Power and Ministry of New and Renewable Energy, now govern the contractual relationships between state-owned distribution companies (discoms) and private generators, with power purchase agreements (PPAs) typically spanning 20-25 years.

The length and inherent incompleteness of these contracts makes them more vulnerable to contingencies and risks. Contracts, in addition to increasing joint gains, also serve the purpose of allocating risks between parties. These risks range from the foreseeable which include price and counterparty risk, bureaucratic delays, transmission constraints, to the unexpected which include pandemics, natural disasters, changes in law and policies.

Force majeure clauses are one of the ways in which parties provision for risks and plan for the future. These clauses make express provision for contingencies where performance of the parties may be affected.

Op-eds

On July 17, Chitrakshi Jain and Rakesh Kacker wrote an article for The Indian Express, ‘A power verdict with consequences beyond Railways’, examining the Supreme Court's dismissal of a batch of appeals concerning whether the Indian Railways qualifies as a deemed distribution licensee under the Electricity Act, with implications for other public and private entities seeking the same status.

A division bench of the Supreme Court recently dismissed a batch of appeals involving the Indian Railways. The judgment has implications for public and private entities seeking the status of a deemed distribution licensee under the Electricity Act. The APTEL judgment from which the appeals arose had rejected the claim of the Indian Railways to be treated as a deemed distribution licensee, a view that has been upheld by the Supreme Court.

The Court was tasked with deciding first whether the Indian Railways qualifies as a deemed distribution licensee. It held that a distribution licensee must not only operate and maintain a distribution system, but it is also required to supply electricity to consumers.

On July 28, Renuka Sane, Rakesh Kacker, and Upasa Borah wrote an article for The Print, ‘India's solar boom masks a worrying dip in utility-scale projects’, examining why FY2026 saw a sharp jump in renewable capacity additions — 55 GW, of which 44.6 GW was solar, well above FY2025's numbers — and identifying three policy features likely behind the surge.

According to data from the Central Electricity Authority, solar capacity additions fell from 6.7 GW in March 2026 to around 4 GW in April and 3.2 GW in May. In June, however, solar capacity was 5.1 GW, driven by the boom in rooftop solar, without which the overall growth would be less. In the first quarter of FY2027, rooftop solar added about 5 GW, almost matching ground-mounted solar additions over the same period. Ground-mounted solar panels, which are built directly on open land rather than on rooftops, have declined as compared to last year.

A similar trend is visible when we look at project developments. In 2025, solar tenders declined by 45 per cent year-on-year, while solar auctions fell by 57 per cent. The slowdown continued in 2026, with solar tenders in the first quarter amounting to only around 6 GW, a 68 per cent fall compared to the same period in the previous year. Another source for understanding longer-term trends of projects in the pipeline is the data on electricity generation projects from the CapEx database maintained by the Centre for Monitoring Indian Economy (CMIE).

On August 11, Renuka Sane wrote an article for The Print, ‘Free UPI was a policy success and an economic distortion’, examining the recent reversal of the government’s January 2020 mandate setting the Merchant Discount Rate on UPI and RuPay debit card transactions to zero, after six years of shaping how India adopted digital payments.

The Taxation and Other Laws (Amendment) Bill, 2026 removes the legal bar that prevented banks and payment service providers from levying MDR on notified payment modes. While the Union government retains the authority to designate fee-bearing channels, immediate pricing power now reverts to the NPCI’s steering committee.

One may sympathise with the government’s intention of wanting to increase digital payments. But any aspiration has to meet economic reality. If financial inclusion via free transactions was the ultimate goal, the government should have fully paid for the underlying infrastructure. By instead enforcing a blunt statutory ban, the state distorted the market. Unwinding this policy is a welcome correction. Depending on what pricing structure is adopted, commercial sanity may finally be restored to India’s digital payment rails, allowing the market to accurately price risk, fund critical infrastructure, and build a resilient payments ecosystem.

The Bridge

The Bridge is TrustBridge's series looking closely at the quasi-judicial work of India's regulators — how they decide, and how they could decide better. On July 24, Natasha Aggarwal and Bhavin Patel published ‘Communication breakdown: SEBI's unreasoned actions and the SAT's half-measure’. The piece examines the Securities Appellate Tribunal's July 2026 decision in Marwadi Chandarana Intermediaries Brokers Ltd. v. SEBI, where SEBI had rejected the appellant's request to withdraw an open offer through an unreasoned communication rather than a formal order. Here is an abstract from the piece:

In July 2026, the Securities Appellate Tribunal (the “SAT”) was presented the opportunity to answer this question in Marwadi Chandarana Intermediaries Brokers Ltd. v. SEBI (the “Marwadi Order”). The controversy in the case involved the appellant’s request to withdraw an open offer, which the Securities and Exchange Board of India (“SEBI”) rejected through a ‘communication’. Notably, the communication did not provide any reasons for SEBI’s decision.

The appellant argued that SEBI’s communication “meets the appellant with serious civil consequences”, was “bereft of cogent reasons and without application of mind”, and that the markets regulator “ought to have passed a reasoned order”.

SEBI’s counsel argued, on the other hand, that the communication could not be considered an “order”. This matters because a conjoint reading of Section 15T of the Securities and Exchange Board of India Act, 1992 and National Securities Depository Limited v Securities and Exchange Board of India² clarifies that only SEBI’s ‘orders’ are appealable.

On August 1, Natasha Aggarwal and Vishnu Suresh published ‘Destination set, route unknown: Evaluating the CCI's IndiGo–Air India order’. The piece evaluates the Competition Commission of India's order declining to investigate allegations concerning cancellation charges imposed by IndiGo and Air India, against the standard for reasoned Section 26(2) orders set out by the Supreme Court in CCI v. SAIL.

CCI v. SAIL establishes the legal standard for Section 26(2) orders, but provides little guidance on how compliance with that standard should be assessed. We therefore use a structured set of indicators to evaluate whether the CCI Order satisfies the requirements identified by the Supreme Court.

TrustBridge has developed a set of Good Order Writing (“GOW”) indicators to measure the completeness and quality of regulators’ adjudicatory orders. We customised these GOW indicators for the assessment of the CCI’s orders under Section 26(2) of the Act. The resulting framework comprises 89 indicators, including 41 competition-specific indicators (the “CCI-GOW indicators”).

Using the CCI-GOW indicators, we assessed the standards set in CCI v. SAIL. For example, application of mind may be demonstrated by satisfaction of the indicator on whether order precisely demonstrates the applicability of the law to the facts. A detailed mapping of the requirements recognised in CCI v. SAIL to the relevant CCI-GOW indicators is provided in the Annexure.

On August 7, Natasha Aggarwal, Amrutha Desikan, and Bhavin Patel published ‘Non-existent citations, real consequences’. The piece examines an NCLT matter in which J&K Bank's 2023 CIRP application led to a 2024 order that, after three levels of adjudication, must now be reconsidered afresh because the reasoning relied on citations that do not exist.

On 2 July 2026, the Supreme Court remanded an order to the National Company Law Tribunal (“NCLT”)¹ because it relied on hallucinated citations to arrive at its conclusion. The impugned decision admitted a petition filed by Jammu and Kashmir Bank Limited (“J&K Bank”) to initiate corporate insolvency resolution proceedings (“CIRP”) in respect of Essel Infraprojects Limited (“Essel”) under the Insolvency and Bankruptcy Code, 2016. Some citations in the order referred to judgments that do not exist, while others correctly identified reported decisions but attributed non-existent paragraphs to them: there were three non-existent citations, two correct citations with non-existent paragraphs, and one misattributed citation with a non-existent paragraph.

Events and Trainings

On August 13, Natasha Aggarwal, Amrutha Desikan, Bhavin Patel, and Khushi Singh organised and attended a convening on the “application of mind” project at the National Law School of India University, Bangalore. Appellate courts routinely invoke a lack of "application of mind" to set aside administrative and adjudicatory orders, yet the phrase itself remains legally undefined, with its exact origins and nature unclear. The team presented their ongoing research exploring what “application of mind” means under Indian law, its impact on challenges to quasi-judicial orders, and whether it is best understood as a concept of administrative law or constitutional law, to the administrative and constitutional law faculty cluster at NLSIU.

From Paper to Practice

Our research doesn't just sit on a page. It's finding its way into the rooms where regulatory practice is shaped. Dr. M. S. Sahoo, one of India's most respected voices in financial markets regulation, used TrustBridge's research on IFSCA's order-writing practices in training sessions for officers of the International Financial Services Centres Authority (IFSCA) and the Insurance Regulatory and Development Authority of India (IRDAI). A small win that our work is landing where it matters most.

Work with us

We’re looking for interns and collaborators to join us in research and communications. If you’re passionate about impactful policy work and want to contribute to improving India’s business environment through the rule of law, we’d love to hear from you. Reach out to us at info@trustbridge.in.

Thanks for reading The TrustBridge Newsletter! Subscribe for free to receive new posts and support our work.

          Website  |  LinkedIn  |  YouTube 

Welcome to The Bridge! India's regulators affect markets, govern disputes, and test the rule of law through their orders, regulations, and policy implementation actions. Yet the quality of this work rarely gets the scrutiny it deserves. The Bridge is TrustBridge's window into India's regulatory state: how it decides, how it regulates, and how it could do both better.

The Competition Commission of India’s (the “CCI’s”) orders under Section 26(2) of the Competition Act, 2002 (the “Act”) are an important part of its adjudicatory function: they close proceedings at the threshold without directing an investigation, are appealable, and impact informants’ rights. They also utilise significant regulatory capacity. This is especially significant given the CCI’s capacity constraints: 42% of sanctioned posts are vacant, grants from the Ministry of Corporate Affairs were insufficient relative to expenditure, and 130 antitrust cases were pending at different stages.1 Therefore, how the CCI decides which matters do or don’t warrant investigation, and what resources it allocates to these decisions, becomes important.

In two new working papers, we examine the CCI’s Section 26(2) decision-making and ask what guardrails constrain its discretion, and how it allocates scarce regulatory resources.

In “Order writing and statutory discretion in the CCI’s Section 26(2) determinations”, we use a customised set of indicators based on the Good Order Writing framework developed by Aggarwal et al. (2025) in “A guide to writing good regulatory orders”, to assess the completeness of a random sample of 111 Section 26(2) orders issued between 2014 and 2024 (the “Dataset”). We find that the orders are generally complete, with an average CCI-GOW score of 59.51%, and perform well in appeal, but that the statutory factors under Sections 19(3) and 19(4) are applied inconsistently. We recommend incremental improvements in order-writing alongside a dedicated framework for threshold-stage determinations under Section 26(2).

Our second paper, “The CCI’s allocation of scarce resources in Section 26(2) matters”, examines how the CCI allocates limited regulatory capacity in such matters. We find that approximately 42% of the orders in the Dataset arise from “peripheral matters”, that is, complaints falling outside the scope of competition law or unsupported by evidence. These matters are processed through the same institutional procedures as substantive complaints, with a median disposal time of 30 days. We recommend a layered approach to improve screening, including clearer public guidance, guided filing mechanisms, AI-assisted review tools and statutory prioritisation frameworks, while preserving access to competition law enforcement.

We are pleased to invite you to a presentation of both papers on Friday, 21 August 2026 at 4:00 PM IST, followed by a discussion with competition law practitioners and academics. If you would like to join the discussion virtually, please register here.

- Natasha Aggarwal, Amol Kulkarni, Bhavin Patel, and Vishnu Suresh are researchers at TrustBridge and Shruti Aji Murali is KM Counsel at Axiom5.

CITATION

Natasha Aggarwal, Amol Kulkarni, Shruti Aji Murali, Bhavin Patel, and Vishnu Suresh, 2026. “What happens when the CCI decides not to investigate?”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Natasha Aggarwal, Bhavin Patel, and Karan Singh, “A Guide to Writing Good Regulatory Orders” [2025] Trustbridge Rule of Law Foundation Working Papers (https://trustbridge.in/publications/a-guide-to-writing-good-regulatory-orders/).

Standing Committee on Finance, Ministry of Corporate Affairs Demand for Grants (2026–27) (Thirty First Report, Lok Sabha Secretariat 2026) (https://elibrary.sansad.in/server/api/core/bitstreams/418570d6-7f80-4a04-9eee-6d03ab9b9f03/view)

Standing Committee on Finance, Evolving Role of Competition Commission of India in the Economy, Particularly the Digital Landscape (Twenty Fifth Report, Lok Sabha Secretariat 2025) (https://elibrary.sansad.in/items/39840bff-3981-4e4d-9eda-28fba9cfc1a1).

Welcome to The Bridge! India’s regulators affect markets, govern disputes, and test the rule of law through their orders, regulations, and policy implementation actions. Yet the quality of this work rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into India’s regulatory state: how it decides, how it regulates, and how it could do both better.

Introduction

There does not appear to be consensus on whether governments should regulate how much flight tickets cost: While some countries like the United States and Canada have chosen to completely deregulate air transport,1 the Chinese government retains some control over air fares.2 Even where regulation is accepted, methods vary, from price floors and ceilings and route-specific interventions to mandatory disclosures. What we can agree on, however, is that choosing whether and how to regulate air fares is essentially a matter of economic policy, which falls within the domain of the executive.

In India, airlines can decide the airfares on their own as per dynamic pricing and operational needs; the law requires that airlines publish these airfares, and if the Directorate General of Civil Aviation (the “DCGA”) believes that they are excessive or predatory, it can issue “directions” to airlines.3 Some suggest this means that the DCGA can also regulate airfares in exceptional situations.4 As of now, DGCA’s engagement has been limited to supervising fare declarations through its Tariff Monitoring Unit, which analyses airfare trends on key routes, and flags potential cases of exploitative pricing.5

In a recent Public Interest Litigation (“PIL”) filed before it, the Supreme Court will likely have to address whether these measures are adequate. In S. Laxminarayanan v. Union of India & Ors,6 (“S. Laxminarayanan”) the petitioner argued that the “DGCA’s oversight is inadequate” and that the regulatory vacuum necessitates guidelines to control airfares, and the establishment of a new regulator to address price exploitation. In a curious move on 13 July 2026, the Court directed the government to present its unpublished rules framed under the Bhartiya Vayuyan Adhiniyam, 2024 (the “BVA”) inside a sealed envelope, “irrespective of the fact whether they are placed before the Houses of Parliament or not.”7

What might the Court be trying to achieve with this order? The one-page order does not provide a clear explanation. So, in this post, we postulate two potential objectives of this order and explore their implications on the separation of powers.

What is the Supreme Court order trying to achieve?

From the order, it appears that the Central Government argued that the Supreme Court should wait to intervene until they have published rules under the BVA. Under the BVA, the Central Government has the discretion to make rules on “approval, disapproval or revision on tariff of operators of air transport services.”8 The Central Government informed the Court that the rules are currently being translated, after which they must be presented before Parliament. The Court however sided with the petitioner, who argued that placing the rules before Parliament is an ex post check and that the Court need not wait until that happens. Curiously, the Court seems to have ignored the fact that the rules are not even published in the Official Gazette yet.

In the absence of any reasons for why the Supreme Court wants to look at unpublished rules in a sealed envelope, we suggest this action might stem from one of two potential motives. Both bode ill for the principle of separation of powers—a “basic feature of the Constitution and an essential constituent of the rule of law”.9

The first potential motive could be to review the validity of the unpublished rules under the BVA before they are publicised. This would be an unprecedented encroachment into the Central Government’s domain. Nothing in the law allows the Supreme Court to review rules before their publication in the official gazette. The function of the courts is to adjudicate and resolve disputes, interpreting the law as necessary in the process. A review of unpublished rules would essentially mean evaluating Central Government proposals which do not yet have the force of law. In the Supreme Court’s own words, “[t]he constitutional scheme does not permit the judiciary to prescribe the text of a law or to mandate that Parliament adopt a specific policy choice.”10

If the Supreme Court makes any pronouncement on the shape and form of the unpublished regulations, the Central Government will most likely have no option but to go along with that decision. When Parliament enacted the Tribunals Reform Act, 2021 in disregard of the Supreme Court’s judgment in Madras Bar Association v. Union of India (2021),11 the Court reminded Parliament that it could not override or contradict judicial decisions.12 The Supreme Court’s interpretation of the Constitution constitutes a binding declaration of law.13 At best, Parliament can only cure the identified defect identified in a manner consistent with the Court’s pronouncements.14

While the Madras Bar Association ruling was with respect to Parliament, a similarly placed Central Government would also be bound by the Court’s decision. The possibility of the Court’s judicial review of the unpublished rules raises other concerns as well: if the Central Government amends these rules in line with the Court’s decision prior to publication, would individuals later be able challenge those rules? Or would the Court’s pre-publication review and the Government’s amendments, if any, effectively shield the rules from any further challenge? We are not aware of any constitutional mechanism that permits such an ‘advance blessing’ system for delegated legislation.15

The second possibility is that the Court wants to nudge the government into publishing the rules. Even this counts as tip-toeing across the executive’s boundary. The BVA says that the Central Government retains the discretion over whether to publish rules on air fares. The choice to not make any new rules on air fares is as legitimate a policy choice as the choice to make such rules. Although indirect, the Court exerting pressure on the Central Government to act constitutes interference at best, and at worst, amounts to deciding the matter of whether to regulate air fares, an out-and-out policy decision, completely outside the Court’s remit. Perhaps it would have been better for the Court to restrict itself to questioning the Central Government on whether they intend to make rules on airfares, instead of rendering the matter moot by asking for unpublished rules to be placed before it.

Regardless of the motive, the method itself is problematic. The Supreme Court has itself in the past expressed contempt towards review of material in sealed envelopes because of the opacity of the process.16 Judicial propriety demands the Court clarify why such an exceptional process is both necessary and urgent in this instance, particularly when it precedes parliamentary review. Surely this should not happen in the form of a one-page unreasoned order.

Conclusion

Whether driven by a desire to pre-screen unpublished rules or compel executive action, the Supreme Court’s order in S. Laxminarayanan seems a troubling example of judicial overreach. The Court should stick to its established principle of judicial restraint in economic policy.17 Its prior jurisprudence wisely acknowledges a lack of competence and expertise to decide upon such economic issues, as well as its inability to assess the impact of a particular action.18 As the case moves forward, we hope that the Supreme Court does not escalate and adjudicate a broad policy agenda simply because an individual petitioner framed an economic policy concern in terms of fundamental rights. If the Court’s intention does not extend to stepping into the Executive’s policy-making domain, the least it should do is clearly articulate reasons for orders as perplexing as the one it made in its single-page pronouncement.

P.S. The Supreme Court issued another order in this matter on 7 August 2026: that order grants the Central Government more time to produce the unpublished rules. On 17 August 2026, after this post was written, the Court opened the envelope and looked at the unpublished rules. Counsel for the Union sought time to finalise the draft rules. Reportedly, Counsel informed the Court that the rules were still under discussion.

- The authors are researchers at TrustBridge, and would like to thank Pratik Datta for his feedback.

CITATION

Bhavin Patel and Khushi Singh, 2026. “Should the Supreme Court regulate your flight ticket fares?”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Aircraft Rules, 1937, Rule 137.

Balco Employees’ Union v. Union of India, (2002) 2 SCC 333.

Bharatiya Vayuyan Adhiniyam, 2024, section 10(2)(c).

Bhavesh D. Parish v. Union of India, (2000) 5 SCC 471.

Douglas McDonald-Norman and Arun K. Thiruvengadam, “Supreme Court holds the line against executive-legislative attempts to control tribunals” (Supreme Court Observer, 22 December 2025)

In Re: Kartikeya Rawal and InterGlobe Aviation Ltd, Case No. 44 of 2025 (CCI, 4 February 2026).

Kalinga Mining Corporation v. Union of India, Civil Appeal No. 1013/2013 (order dated 07.02.2013)

Madhyamam Broadcasting Limited v. Union of India & Ors, 2023 INSC 324.

Madras Bar Association v. Union of India (2021), (2022) 12 SCC 455.

Madras Bar Assn. (6) v. Union of India, (2026) 2 SCC 1.

Prag Ice & Oil Mills v. Union of India, (1978) 3 SCC 459.

Rajya Sabha’s Department-related Parliamentary Standing Committee on Transport, Tourism and Culture, ‘Issue of Fixing of Airfare’ (353rd Report) (10 August 2023).

R.K. Garg v. Union of India, (1981) 4 SCC 675.

S. Laxminarayanan v. Union of India & Ors, WP(C) 1124/2025 (order dated 13 July 2026).

State of T.N. v. State of Kerala, (2014) 12 SCC 696.

Su Wu, “China Says It Is Not Restricting Airlines From Selling Low Fares” (Simple Flying, 23 February 2023).

Vivek Narayan Sharma (Demonetisation Case-5 J.) v. Union of India, (2023) 3 SCC 1.

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

On 2 July 2026, the Supreme Court remanded an order to the National Company Law Tribunal (“NCLT”)1 because it relied on hallucinated citations to arrive at its conclusion. The impugned decision admitted a petition filed by Jammu and Kashmir Bank Limited (“J&K Bank”) to initiate corporate insolvency resolution proceedings (“CIRP”) in respect of Essel Infraprojects Limited (“Essel”) under the Insolvency and Bankruptcy Code, 2016. Some citations in the order referred to judgments that do not exist, while others correctly identified reported decisions but attributed non-existent paragraphs to them: there were three non-existent citations, two correct citations with non-existent paragraphs, and one misattributed citation with a non-existent paragraph.

One of Essel’s directors filed an appeal before the National Company Law Appellate Tribunal (“NCLAT”). The NCLAT stated “we do not find error in the impugned order”, and dismissed the appeal.2 Notably, neither side’s lawyers nor the NCLAT detected the imaginary citations in the impugned decision.

The fact that the NCLT relied on hallucinated citations was finally raised in second appeal before the Supreme Court. In response, the Supreme Court recorded zero-tolerance for “producing, citing or using AI-generated precedents without verification”, and stated that any decisions that rely on such material are to be set aside.

In this post, we explore why hallucinated citations went undetected across three levels of adjudication, and argue that the solutions lie in awareness, training, and process and technological safeguards. These include: training adjudicators on the limitations of LLMs, developing clear institutional guidance on appropriate use, and deploying technical safeguards.

Why did this happen?

It’s tempting to blame technology for this mess, but hallucinations are a risk inherent to large language models (“LLMs”) - and users of this technology need to be aware of this ex ante.3 LLMs are not built to retrieve authoritative legal materials from verified databases. They are designed to generate text by predicting the most statistically likely sequence of words. As a result, it is highly probable that they will produce citations that appear perfectly plausible but do not exist.

The real problem is that adjudicators and other actors in this ecosystem are increasingly relying on LLM-generated citations without independent verification.4 Large pendency volumes and crushing workloads create incentives to let through research that looks half-decent without strict scrutiny.

This has consequences: in the instant case, J&K Bank filed the CIRP application in 2023, the NCLT passed the impugned decision in 2024 and now, after the parties have made their way through three levels of adjudication, the matter must be considered afresh by the NCLT. While these delays are a worry, the greater concern is jurisprudence developed off imaginary precedent.

The root cause of this problem is a lack of awareness of how these systems work, and lack of training in verification processes. This is dangerous, because these systems and tools are remarkably useful and produce responses that are polished and perfectly plausible, and have the potential to reduce the time taken to adjudicate a matter, which is tempting for an NCLT adjudicator burdened with an overflowing case docket.

How do we resolve this?

We can think of three ways:

First, we suggest that adjudicators and other actors involved in the adjudicatory workflow be trained on the use of LLMs. Such training should cover LLMs’ inherent limitations, methods to mitigate the impact of such limitations, and appropriate use cases in the adjudicatory workflow.

Second, adjudicatory bodies should develop process manuals that govern how AI is to be used in this ecosystem. Such manuals must be foregrounded in design principles such as human oversight and transparency and explainability,5 and could include material such as verification checklists.

Third, adjudicatory workflows will benefit from a number of technical solutions. Examples of such solutions include the use of retrieval augmented generation, integration of legal knowledge graphs, implementation of processes for post-generation verifications, and mandatory reviewer verification.6 Such solutions will augment an adjudicator’s capacity to use LLMs responsibly. Some of these technical solutions and safeguards are not necessarily complex; one can imagine simple solutions such as mandating that all filings include a list of citations. Once citations are presented in a standardised format, simple deterministic programs, such as Python scripts, can automatically check whether the cited cases actually exist in a verified database, and whether quoted extracts correspond to the correct judgments. Any discrepancy can be flagged for human review.

Conclusion

Hallucinations are not a bug waiting to be fixed, they are an inherent feature of how LLMs work. This means that no future version of this technology will make the problem disappear, and that the burden of managing it falls on institutions. LLMs are genuinely useful tools for overburdened adjudicators. That utility creates the precise conditions that make them so tempting to use without safeguards. The appropriate response is not to retreat from AI, but to build workforces and institutions capable of using it responsibly.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback.

CITATION

Natasha Aggarwal, Amrutha Desikan and Bhavin Patel, 2026. “Non-existent citations, real consequences”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Aakriti Bansal, ‘10 Cases that Show Indian Courts Have an AI Hallucination Problem’ (MediaNama, 3 July 2026) <https://www.medianama.com/2026/07/223-10-cases-ai-hallucination-cases-in-indian-courts/>

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

An aggrieved customer1 of InterGlobe Aviation Limited (“IndiGo”) and Air India Limited (“Air India”) recently challenged an order of the Competition Commission of India (“CCI”) before the National Company Law Appellate Tribunal (“NCLAT”).2 In the underlying order dated 11 March 2026 (the “CCI Order”), the CCI decided not to investigate the customer’s allegations that the two airlines imposed “unconscionable and illegal” cancellation charges, on the grounds that:3

1. There was no evidence of agreement or concerted conduct between the airlines under Section 3(3) of the Competition Act, 2002 (the “Act”),

2. Collective dominance is not recognised under Section 4 of the Act, and

3. The dispute is contractual in nature and outside the scope of the regulator’s remit under the Act.

The appeal has brought renewed attention to the CCI Order. In this post, we evaluate whether the CCI Order satisfies the standards laid down in Competition Commission of India v. Steel Authority of India Ltd. (“CCI v. SAIL”),4 in which the Supreme Court held that orders passed by the Commission in its adjudicatory and determinative capacity, such as those under Section 26(2) of the Act, must be speaking orders, with reasons, reflecting due application of mind. CCI v. SAIL establishes a qualitative benchmark that the CCI’s orders must match, and serves as a useful benchmark for evaluating the quality and completeness of the CCI’s orders. We propose a set of customised metrics that help determine whether a CCI order meets the standards prescribed by CCI v. SAIL.

In our evaluation, we find that while the outcome of the CCI Order is clear, its reasoning is not.

Data and methods

CCI v. SAIL establishes the legal standard for Section 26(2) orders, but provides little guidance on how compliance with that standard should be assessed. We therefore use a structured set of indicators to evaluate whether the CCI Order satisfies the requirements identified by the Supreme Court.

TrustBridge has developed a set of Good Order Writing (“GOW”) indicators to measure the completeness and quality of regulators’ adjudicatory orders.5 We customised these GOW indicators for the assessment of the CCI’s orders under Section 26(2) of the Act.6 The resulting framework comprises 89 indicators, including 41 competition-specific indicators (the “CCI-GOW indicators”).7

Using the CCI-GOW indicators, we assessed the standards set in CCI v. SAIL. For example, application of mind may be demonstrated by satisfaction of the indicator on whether order precisely demonstrates the applicability of the law to the facts. A detailed mapping of the requirements recognised in CCI v. SAIL to the relevant CCI-GOW indicators is provided in the Annexure.

Two researchers assessed the CCI Order against the CCI-GOW indicators: one conducted the initial assessment, and a second subsequently reviewed and verified the findings. The CCI Order scored 40% on the CCI-GOW indicators.

What the CCI Order gets right

The CCI Order satisfies several CCI-GOW indicators, including indicators that capture some of the requirements identified in CCI v. SAIL. The CCI Order identifies the statutory and factual basis of the proceedings, presents the material facts clearly, evaluates the informant’s evidence, records the arguments advanced and clearly states its conclusions. It also avoids mere reproduction of statutory provisions and does not rely on extensive quotations from precedents or the informant’s submissions.

Moreover, the CCI Order attempts to anchor its decision within the statutory framework of the Act, and precedent. This is demonstrable from the following:

• In relation to the allegations under Section 3(3) of the Act, the Commission identifies the applicable statutory thresholds, examines the material placed before it and concludes that the informant has not produced evidence from which a concerted conduct between the airlines may be inferred.

• Similarly, while dealing with the allegations under Section 4, the Commission relies on its previous decision8 to reiterate its established position that the concept of collective dominance does not form part of the present statutory scheme of the Act. The CCI Order briefly examines the impugned cancellation and refund terms before concluding that the allegation is not a competition issue under Section 4. To that extent, the order situates its conclusions within both the statutory framework and the Commission’s existing jurisprudence.

Together, this indicates that the Commission engaged with the material placed before it meaningfully, identified the legal basis of its conclusions and expressed those conclusions in its own reasoning rather than through mechanical reproduction.

What the CCI Order gets wrong

The CCI Order contains two weaknesses:

First, the CCI Order does not provide (i) a concise statement of the main question, (ii) a well-reasoned analysis or an analysis that demonstrates applicability of the law to the facts, or (iii) the issues under consideration. As a result, its analysis is not organised in a step-by-step manner: it identifies the informant’s allegations and the relevant statutory provisions, but then jumps to the Commission’s conclusions. Consequently, we find that the CCI Order fails to meet some of the indicators associated with requirements outlined in CCI v. SAIL.

Second, the order satisfies only 6 of the 36 applicable competition-specific indicators.9 Most notably, it does not satisfy any of the six indicators related to Section 19(3) of the Act or any of the thirteen indicators related to Section 19(4) of the Act. These sub-sections specify factors that the CCI must consider in arriving at its decision. These deficiencies are also relevant to the conditions set in CCI v. SAIL, particularly the requirements that the order identify and apply the relevant statutory grounds and provide reasons connecting the material considered to the conclusions reached.

These two weaknesses are demonstrable from the following:

• While assessing Section 3(3) allegation, the Commission found that the informant had not produced material showing an agreement or concerted conduct between the airlines. In those circumstances, it may not have been necessary to consider the Section 19(3) factors. However, the order does not explain this. Instead, the reader is left to infer why the Commission’s analysis ended at that stage.

• In ascertaining whether Section 4 of the Act has been violated, the CCI identifies the form of abusive conduct alleged and considers the terms of cancellation and refund. However, it does not undertake the preceding steps of analysing whether IndiGo holds a dominant position in the relevant market or even determine what the relevant market is. The CCI Order should have organised its analysis around the following three questions: the relevant market, dominance in that market, and abuse of such dominance.10 If the Commission considered it unnecessary to proceed through each step, it could have identified the stage at which the informant’s allegation failed and explained why the remaining stages did not require consideration.

Together, these weaknesses affect the traceability and coherence of reasoning. The CCI order tells the reader what the Commission concluded, i.e., that there was no evidence indicating coordination between the airlines and that the dispute arose from contractual terms and is not a competition issue. What is less visible is how the Commission used the statutory framework to reach those conclusions, or why it considered certain stages of the inquiry unnecessary.

Conclusion

Our evaluation of the CCI Order against the standards in CCI v. SAIL shows that the Commission’s conclusions are identifiable, but the intermediate step of reasoning supporting those conclusions is only partially visible. The issue is not whether the Commission reached the correct outcome, but whether the Commission has been able to sufficiently demonstrate how the applicable law, the arguments and materials placed before it were applied to arrive at this outcome.

The Commission, therefore, may consider adopting standardised templates or an internal order-review framework for Section 26(2) orders. This minor but effective shift in practice could help ensure that the reasoning underlying the Commission’s conclusions is consistently reflected in its written orders.

The present CCI Order makes the Commission’s destination clear. The challenge, and the opportunity, is to make the route to that destination equally visible.

- The authors are researchers at TrustBridge, and would like to thank Bhavin Patel and Renuka Sane for their feedback and suggestions.

CITATION

Natasha Aggarwal and Vishnu Suresh, 2026. “Destination set, route unknown: Evaluating the CCI’s IndiGo–Air India order”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

Annexure 1

Relying on the SC’s judgements, we have distilled the elements indicating conditions set in CCI v. SAIL to the following:

1. Identification specific, relevant statutory ground and explanation of its application to the specific case,11

2. Consideration of relevant materials and facts,12

3. Independent decision, i.e., no reliance on extraneous factors,13

4. Description of non-ambiguous reasons14 which links the materials which are considered and the conclusions which are reached,15 and

5. Absence of mechanical reproduction.16

Our CCI-GOW indicators assess the conditions set by CCI v. SAIL through the indicators as categorised below.

References

Airen Metals Private Limited v. Hindalco Industries Limited (Case No. 31 of 2024, Competition Commission of India)

Ajith K. v. Aneesh K.S., 2019 SCC OnLine SC 1074

Balram Garg v. Securities and Exchange Board of India, 2022 SCC OnLine SC 472

The Competition Act, 2002 (12 of 2003)

Competition Commission of India v. Steel Authority of India Ltd. (2010) 10 SCC 744

ICICI Lombard General Insurance Co. Ltd. v. Ajay Kumar Mohanty, 2018 SCC OnLine SC 180

In Re: Kannadiputhur Sundararaman Suresh v. Interglobe Aviation Ltd and Air India Ltd (Case No 42 of 2025, Competition Commission of India)

Indian National Shipowners’ Association (INSA) v. ONGC, (Case No. 01/2018, Competition Commission of India)

Kannadiputhur Sundararaman Suresh v. CCI (Competition Appeal (AT) No. 10 of 2026, National Company Law Appellate Tribunal)

Natasha Aggarwal and Bhavin Patel, “Evaluating regulators’ orders: The good order writing indicators explained”, (2026) The Bridge, TrustBridge Rule of Law Foundation

Punjab State Civil Supplies Corporation Ltd. v. M/s Atwal Rice and General Mills, 2017 SCC OnLine SC 726

Rajeev Suri v. Delhi Development Authority, 2021 SCC OnLine SC 7

Telefonaktiebolaget Lm Ericsson v. CCI, (W.P.(C) 464/2014, Delhi High Court)

Union of India v. Mohan Lal Capoor, 1973 SCC OnLine SC 292

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

In December 2025, the Securities Market Code Bill 2025 (“SMC”) was introduced in the Lok Sabha by the Ministry of Finance, and subsequently referred to the Parliamentary Standing Committee on Finance (Committee). The SMC seeks to consolidate India’s securities laws by replacing several existing statutes with a single legislative framework. Following its introduction, the Bill was referred to the Committee for a detailed examination. The Committee invited comments; a group of us, comprising Natasha Aggarwal, Pratik Datta, K. P. Krishnan, Bhavin Patel, M. S. Sahoo, Renuka Sane, Ajay Shah and Bhargavi Zaveri-Shah, submitted comments in March 2026.

Our comments sought to strengthen legislative clarity, enhance predictability and proportionality in enforcement, and better align the SMC with rule of law principles and the efficient functioning of markets.

The Committee published its report on 23 July, and we’re pleased to share that several of our recommendations have been accepted. These are mentioned in the table below:

Some of our other recommendations, such as the removal of all criminal liabilities and prescribing a maximum timeline for the adjudicatory process, were not accepted or acknowledged.

We hope that these recommendations are reflected in the SMC, and are encouraged by the fact that several of our recommendations will contribute to a securities law framework that better reflects rule of law principles by promoting legislative clarity, certainty, transparent decision-making, and institutional accountability. For a complete mapping of our recommendations against the Committee’s recommendations, see here.

- The author is a researcher at TrustBridge, and would like to thank Bhavin Patel and Renuka Sane for their feedback.

CITATION

Natasha Aggarwal, 2026. “Our suggestions on Securities Market Code reflected in the Parliamentary Standing Committee Report”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

Is executive action immune from the requirement to give reasons?

In July 2026, the Securities Appellate Tribunal (the “SAT”) was presented the opportunity to answer this question in Marwadi Chandarana Intermediaries Brokers Ltd. v. SEBI (the “Marwadi Order”).1 The controversy in the case involved the appellant’s request to withdraw an open offer, which the Securities and Exchange Board of India (“SEBI”) rejected through a ‘communication’. Notably, the communication did not provide any reasons for SEBI’s decision.

The appellant argued that SEBI’s communication “meets the appellant with serious civil consequences”, was “bereft of cogent reasons and without application of mind”, and that the markets regulator “ought to have passed a reasoned order”.

SEBI’s counsel argued, on the other hand, that the communication could not be considered an “order”. This matters because a conjoint reading of Section 15T of the Securities and Exchange Board of India Act, 1992 and National Securities Depository Limited v Securities and Exchange Board of India2 clarifies that only SEBI’s ‘orders’ are appealable.

In the Marwadi Order, the SAT held that it is “imperative that the Regulator conveys its decision with reasons”, and the impugned SEBI ‘communication’ was “akin to an order”. The appeal was allowed, and the matter remitted to the SEBI.

This seems like a good decision at first blush - but it conceals a devil of doctrinal confusion in the details, and is worth unpacking.

Why does the Marwadi Order matter?

For at least a couple of reasons:

First, the Marwadi Order states that since the ‘communication’ could result in “several civil consequences” for the appellant, it should have been reasoned. This makes tremendous sense, and ensures that regulators can be held accountable for their actions. Any decision that affects the rights of parties should be accompanied by reasons, ideally in writing. This speaks to the core rule of law requirements of the knowability and clarity of the law, and is crucial information for appellants, appellate and review fora, and market participants generally.

The Marwadi Order is certainly a step in the right direction in this regard, but on closer scrutiny, we discern a somewhat troubling gap:

The SAT says that the ‘communication’ was “akin to an order”, and should therefore have been reasoned. It does not say that ‘communications’ or other instruments that are not ‘orders’ must also be reasoned. This statement can be interpreted in two ways, each of which lead to differing results:

On one interpretation, this limits the applicability of reason-giving to SEBI’s quasi-judicial order-writing, and leaves its other functions exempt from this requirement. If a ‘communication’ or other instrument is not ‘akin to an order’, the SAT suggests, no reasons need be provided for the decision it contains. So, reason-giving does not apply to executive instruments, only to quasi-judicial ‘orders’. But this is troubling because it goes against the requirements for reason-giving that we have already described - for regulatory accountability, enabling review, and for general awareness of market participants.

On another interpretation, doctrinal confusions compound: are we to think that any regulatory communication or instrument that results in ‘severe civil consequences’ for the affected party is ‘akin to an order’? If yes, then can we think of any regulatory instrument that does not result in such consequences? If not, that means all regulatory instruments are ‘akin to orders’, and so, the requirement of reason-giving extends to all regulatory actions, including purely executive instruments. This collapses the distinction between various regulatory functions, a position that the Supreme Court took as far back as 1969 in A. K. Kraipak & Ors. v. Union Of India,3 and reconfirmed in 2023 in AERA v. Delhi International Airport Ltd.4

The definitions and boundaries of various functions (quasi-legislative, executive, and quasi-judicial) are blurry, at best, in the Indian jurisprudence.5 The Marwadi Order points to an instance in which this distinction is rendered meaningless - that is, if you agree that the use of regulatory power must be reasoned, accountable, and transparent.

Conclusion

We argue that all regulatory action must be accompanied by reasons. This is critical for several reasons - it demonstrates application of mind, and is a safeguard against the arbitrary exercise of power. Distinguishing between executive ‘communications’ and quasi-judicial ‘orders’ is unnecessary and counter-productive in this respect. All regulatory action can result in severe consequences for regulated entities, and hence it is important that the regulator explain why it is taking a particular action or how it arrives at a decision.

Given the lack of doctrinal clarity about the differences between various regulatory functions, and their specific requirements in Indian case law, codifying the requirement for reason-giving as a default position might be a good idea. This is a good time to consider such a move, given that the Securities Markets Code Bill, 2025 is still under consideration. The Marwadi Order is constrained by what the SAT can do as an appellate regulatory tribunal, but it points the way to an important and necessary set of changes in Indian regulatory law.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback.

CITATION

Natasha Aggarwal and Bhavin Patel, 2026. “Communication breakdown: SEBI’s unreasoned actions and the SAT’s half-measure”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

AERA v. Delhi International Airport Ltd. 2024 SCC OnLine SC 2923.

A. K. Kraipak & Ors. v. Union Of India AIR 1970 SC 150.

Bhavin Patel, Renuka Sane, 2026. “Quasi-judicial function at regulators,” Working Papers, Trustbridge Rule of Law Foundation.

Marwadi Chandarana Intermediaries Brokers Ltd. v. SEBI, Order dated 8 July 2026 in Appeal No. 37 of 2026.

National Securities Depository Limited v Securities and Exchange Board of India (2017) 5 SCC 517.

Read on to see what’s kept us busy.

Policy inputs and submissions

Comments on the Regulations for Use of AI in Courts, 2026

On June 20, Amrutha Desikan and Khushi Singh submitted comments on the Supreme Court’s draft Regulations for Use of Artificial Intelligence (AI) in Courts, 2026, which govern the use of AI in courts, tribunals, and statutory Commissions by setting out core governing principles, permissible and prohibited use cases, and mechanisms for vendor procurement, oversight, audits, and incident management. Key recommendations include:

• the draft Regulations should clarify their scope of application to eliminate jurisdictional ambiguity;

• a graded approach should be considered, with a separate regulatory tier for high-risk applications that fall between permissible and prohibited uses, and the pro-innovation presumption restricted strictly to low-risk administrative tasks;

• the composition of both the AI Committee and the AI Secretariat should be reconsidered to mandate the inclusion of technical experts;

• the draft Regulations should introduce objective risk boundaries to restrict the subjective waiver of human verification duties; and

• the framework should clarify the consequences for breaches by codifying explicit disciplinary measures for internal personnel and defining specific remedial orders.

Comments on the CERC Staff Paper on Capacity Market for Electricity in India

On July 13, Upasa Borah, Chitrakshi Jain, Akshay Jaitly, Renuka Sane, and Subodh Singh submitted comments on the Central Electricity Regulatory Commission’s (CERC) 2026 staff paper proposing a capacity market for India. Key recommendations include:

• near-term reform should focus on ancillary services, particularly market-based procurement of primary reserves, rather than a capacity market;

• the case for introducing a capacity market remains premature and unproven on the present evidence;

• before any such market is introduced, CERC needs stronger empirical grounding on resource adequacy assessments and demand curve construction;

• net CONE (Cost of New Entry) estimation methodology needs to be substantiated rather than assumed; and

• the staff paper does not adequately address technology eligibility criteria or the market power risks a capacity market could introduce.

Blogs

On June 25, Prashant Narang and Renuka Sane published their blog on IndiaCorpLaw, ‘When Arbitration Procedure Becomes Litigation’. The piece shows that routing routine procedural tasks— constituting a tribunal under Section 11, resetting a deadline under Section 29A through adversarial court proceedings makes litigation a built-in stage of arbitration rather than an exception to it, and sketches an administrative alternative for both.

Reform debates on Indian arbitration tend to focus on the ad hoc versus institutional divide. Should parties default to institutional arbitration to reduce court intervention?

The recurring court exposure under sections 11 and 29A is a consequence of how certain procedural tasks are classified. Judicial proceedings are designed to resolve contested legal questions — to hear both sides, apply legal standards, and produce findings reviewable on appeal. Administrative mechanisms are designed to manage routine process tasks: schedule-setting, appointment, coordination.

When the latter are channelled through the former, the costs and formality of adjudication attach to tasks that do not require them. Appointments after a deadlock, deadline extensions by mutual consent, uncontested timeline resets — these are process-management tasks. Routing them through adversarial court hearings before constitutional courts is a category error, and an expensive one.

Op-eds

On June 16, Renuka Sane and Prashant Narang wrote an article for The Print, ‘Arbitration in India has become luxury litigation. Let retd judges go, hire a private institution’, analysing that while criticism of arbitration’s cost and delay is valid, abandoning it altogether without considering alternatives like institutional arbitration would be premature.

While institutional arbitration does hold promise, it can have its own pitfalls. It can reproduce every vice of ad hoc arbitration if it is governed badly. For example, it could have a closed panel dominated by the same retired judges, its fee-setting could be opaque, and it could give up its neutrality, losing credibility in its processes. The verdict, in short, is still out, on whether institutions build a serious model of governance. That said, there is a real opportunity for it to offer private dispute resolution that is transparent, fast and credible.

The task ahead for the institutions is clear. They must earn trust the way any credible forum does — by publishing their numbers, holding their tribunals to fixed timelines and value-based fees. They should draw arbitrators from a wider and more transparent pool than the retired bench. For the government, it is worth remembering that arbitration, which was introduced to avoid court delays in the first place, hasn’t moved cases any faster. It may be time to give the more promising version of arbitration a real chance.

On June 16, Prashant Narang co-authored an article for Mint, ‘Quasi-22: India must sort out how institutions are picked for arbitration once and for all’, examining whether courts’ ad hoc designation of arbitral institutions is procedurally sound, given that institutions themselves are not quasi-courts and have no independent judicial sanctity without party consent.

Institutional appointments hinge on party autonomy, as per the court’s ‘appointment of arbitrators’ scheme. The law stipulates that the concerned chief justice can request an institution to be designated in a particular case only once parties to the dispute make a formal request along with requisite documents. Correspondence in this regard is to be recorded and preserved by the registrar. The Supreme Court examined this procedure over two decades ago.

Yet, India lacks a consistent procedure, which is worrisome for the business environment. Not only does it add to the burden of costs, it can also unsettle previously adjudicated cases. Moreover, courts sometimes pick certain institutions over others of their own accord without explanation.

One also wonders whether recent observations by some high courts endorsing the validity of blanket designations of specific institutions through a one-time notification are remiss in their evaluation.

On July 7, Renuka Sane wrote an article for The Print, ‘India needs to prepare a strategy document that charts coal's decline, year on year’, examining the gap between India's installed renewable capacity, now over 50% of the mix and its continued reliance on coal for close to three-quarters of actual electricity generation.

That a complete switch is not possible in the near future, does not mean we should not aspire to it over time. Take the example of OECD countries. They, too, did not make the transition overnight. Many relied on, and continue to rely on, natural gas as a bridge fuel, but in a way that it does not dominate the energy landscape. The United Kingdom closed its last coal-fired station in September 2024. Their energy mix is dominated by renewables at 47 per cent, but natural gas comes second at 28 per cent. The United States, which has a relatively higher share of fossil fuels than countries in Europe, generates 41 per cent of its electricity from natural gas and 17 per cent from coal — far below India’s 70 per cent share from coal. Even in China, coal-based electricity generation has fallen to around 54 per cent. The point is simple — India needs to start planning an explicit glide path such that over the next decade coal does not dominate our electricity system.

What would such a plan look like? India needs to set a declining ceiling on coal’s generation share. It has to stop approving new coal-fired power plants, put a price on coal’s health and carbon costs, and treat storage and transmission with urgency. This runs counter to the government’s current policy which aims to add 80 GW of thermal power capacity by 2031-32. It is true that India needs far more electricity, and that demand is rising fast. But meeting that demand need not require more coal.

The Bridge

The Bridge is TrustBridge's series looking closely at the quasi-judicial work of India's regulators — how they decide, and how they could decide better. On July 1, Natasha Aggarwal, Amol Kulkarni, and Bhavin Patel published ‘SEBI learns fast, but has the Supreme Court been a good teacher? The Reliance judgment and how it shaped the Rajesh Exports order’. The piece examines the Supreme Court's May 2026 judgment in Reliance Industries Limited v. SEBI, which sharpened the definition of "fraud" under Regulation 2(1)(c) of the PFUTP Regulations, and tests SEBI's ex-parte interim order against Rajesh Exports Limited, issued five days later against this newly clarified standard.

The Rajesh Exports order alleges that REL, Rajesh Mehta, and certain related entities engaged in a scheme that materially misrepresented the company’s financial position and operations, and prima facie finds violations of several legal provisions, including the PFUTP Regulations.

The Rajesh Exports order responds to the Reliance judgment in Parts C (Dealing in securities and price volume impact) and D (Prima facie findings against Noticees). Part D specifically calls out the Reliance judgment, describes how it has clarified the interpretation of fraud, and goes on to argue that the order has satisfied the threshold prescribed in the Reliance judgment. Prior to the Reliance judgment, SEBI might have proceeded directly from allegations to a conclusion of fraud under the PFUTP Regulations.

On July 15, Amrutha Desikan published ‘Lessons from Balram Garg for SEBI's adjudicatory process’. Tracing the SEBI order, the SAT appeal, and the Supreme Court decision in Balram Garg v. SEBI — an insider trading case built entirely on circumstantial evidence of family ties the piece identifies four specific reasoning failures the Supreme Court found in the original order, and converts each into a systemic recommendation for SEBI's adjudicatory process.

Systematically reading appellate decisions helps identify systemic gaps and reduce appeals. Because complete lifecycle datasets are currently non-public and hard to map, SEBI should maintain a structured database linking orders to their appellate outcomes to loop feedback back into its processes. This feedback should be codified into a public procedural manual covering proceedings, evidentiary standards, and record-keeping to standardise institutional adjudication and help stakeholders navigate the system. SEBI should build a precedent bank supported by responsibly governed LLMs to verify case usage, ensuring AI aids, rather than replaces, human adjudicatory reasoning. Finally, adjudicators require continuous access to legal resources to prevent errors of law. Our synthesis of appellate feedback in the SEBI Lifecycle Datasets aims to provide precisely this support.

Events

On June 29, Subodh Singh co-presented a paper titled ‘The Price of Reliability – Energy Justice Evaluation of CRMs in Ireland and Poland’ at the Groningen Centre of Energy Law and Sustainability (GCELS) conference on ‘Contested Transitions: Law for Energy Justice’ at the University of Groningen.

From July 3–6, Renuka Sane presented on the panel ‘Enclave vs Mainland’ and Pratik Datta was a discussant on the panel ‘Capital Controls Policy’ at Cross-border flows and frictions in India conference, organised by XKDR Forum in Goa.

On July 7, TrustBridge contributed to the AI Knowledge Consortium's comments on the Supreme Court's Regulations on the Use of AI in Courts. Amrutha Desikan attended the AIKC Huddle at the Shardul Amarchand Mangaldas office in New Delhi for a discussion on the regulations.

Trainings

On June 30, Prashant Narang conducted a two-hour session for the inaugural cohort of The 1991 Fellowship at the National YMCA Hostel, New Delhi, as part of the Fellowship’s Delhi orientation. Using case studies of specific reform initiatives, the session explained how policy problems are identified, how reform is negotiated through institutions, and the factors that enable implementation in practice, for a cohort of 12 fellows working on state-level policy reform across ten states.

Work with us

We're recruiting a Research Fellow for our Financial Markets function and law interns in New Delhi. If you're passionate about impactful policy research, email us at info@trustbridge.in.

Thanks for reading The TrustBridge Newsletter! Subscribe for free to receive new posts and support our work.

          Website  |  LinkedIn  |  YouTube 

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

Appellate feedback about the regulator’s processes and order quality provides key insights for regulators for avenues of improvement. However, this feedback dispersed across individual decisions and rarely synthesised. For the adjudicatory processes at the Securities and Exchange Board of India (“SEBI”), systematically reading order documents from every stage in matters that have been appealed through both the Securities Appellate Tribunal (“SAT”) and the Supreme Court1 can help with this.

This piece details findings from a study of the orders leading up to the Supreme Court decision in Balram Garg v. SEBI,2 a widely discussed case involving allegations of insider trading in the scrip of PC Jeweller Ltd. By tracing the SEBI Order,3 the SAT,4 and Supreme Court decisions, we identify specific failures in the originating order, highlighting system-level improvements for SEBI’s adjudicatory process.

The facts and the originating order

PC Jeweller Ltd. (“PCJ”) is a leading organised retail jeweller in India. Between April and July 2018, it made two corporate announcements: a board-approved share buyback on 10 May, and its withdrawal on 13 July after State Bank of India refused a no-objection certificate.

During this period, Balram Garg was the Managing Director of PCJ (Noticee 5). He lived in the same building (though a different unit) as his nephews, Sachin Gupta (Noticee 2) and Amit Garg (Noticee 4), and Shivani Gupta (Noticee 1, Sachin’s wife). Shortly before this, Amit Garg had a controlling interest in Quick Developers Pvt. Ltd. (Noticee 3).

SEBI initiated an investigation against these parties (collectively, the “Noticees”), on allegations of insider trading in PCJ’s scrip, treating the Noticees’ knowledge of the future announcements as Unpublished Price Sensitive Information (“UPSI”). It charged Noticees 1-4 with trading during this window while in possession of inside information, and Noticee 5 with communicating it.

The SEBI Order concluded all Noticees engaged in insider trading based entirely on circumstantial evidence: chiefly their shared address, family ties, trading patterns, and bank transactions. In their arguments, the Noticees pointed to a ‘Family Arrangement’ document as evidence that their family ties had broken down. Dismissing the ‘Family Arrangement’ document as a mere internal business division rather than proof of estranged ties, SEBI assumed family proximity and inferred they must have shared and acted upon the UPSI.

The SAT decision

SAT affirmed SEBI’s conclusion without material reasoning. Relying entirely on SEBI’s trading and bank analyses, it concluded on a “preponderance of probability” that UPSI had been shared within the family, despite acknowledging there was “no direct evidence as to who had disseminated” it.

The Supreme Court’s position

The Supreme Court set aside the orders of both lower fora. The inferential chain, it held, broke at its starting point: the foundational facts required to raise any presumption of communication had never been established. This judgment points out four specific reasoning failures in the SEBI order:

Relying on evidence not in the show-cause notice. The SEBI Order relied on rent transactions between Noticee 2 and PCJ. Noticee 2 was also a nominee to PCJ’s former Chairperson’s demat account. However, these facts were not included in the show cause notice (“SCN”). The Supreme Court cited Tarlochan Dev Sharma v. State of Punjab5 to hold that an order cannot be founded on grounds not put before the noticee, as this deprives the noticee of the opportunity to defend themselves, violating natural justice principles.6

Sufficiency of evidence. The SEBI Order had drawn inferences from trading patterns and bank statements, instead of identifying actual instruments of communication, to prove that UPSI had been communicated. The Supreme Court held that without material showing frequent communication (letters, emails, witnesses etc.), no presumption of communication could arise.7

Erroneous citation of precedent. The Noticees argued that Chintalapati Srinivasa Raju v. SEBI8 precluded any finding of insider trading based on family ties alone. To support its contrary conclusion, SEBI cited Kishore R. Ajmera v. SEBI9 (referenced within Chintalapati) for the proposition that circumstantial evidence suffices. The Supreme Court rejected this because Kishore Ajmera concerned Fraudulent and Unfair Trade Practices rather than insider trading, and the law was already settled by Chintalapati. SEBI misquoted the very case the Noticees correctly cited, extracting a subsidiary reference while ignoring its actual holding.

Shifting the burden of proof. SEBI unlawfully shifted the burden onto the Noticees to prove their estrangement. SAT upheld the SEBI Order’s reasoning on the ground that the Noticees were unable to demonstrate the fact of their estrangement, concluding that they must have communicated with each other.10 Rejecting this, the Supreme Court held that SAT’s approach “turns the SEBI Act on its head,” placing the burden of proving a breakdown of ties on the Appellants while ignoring that the onus was actually on SEBI to prove possession of or access to UPSI.

What lessons can SEBI learn from Balram Garg?

The appellate lifecycle in Balram Garg has several points of feedback for the adjudicatory process and for order writing at SEBI. Each of the Supreme Court’s critiques can be organised into insights. Table 1 summarises the Supreme Court’s critique of the SEBI Order and the corresponding feedback for SEBI’s adjudicatory process.

Relying on evidence not in the SCN. SEBI’s analysis must be confined to the grounds put to the noticee in the show-cause notice. This is a requirement of natural justice, not just good order writing. The SCN is downstream of the investigation report (“IR”), and the IR must contain every piece of relevant information that grounds the SCN. Investigating officers must examine its completeness rigorously before it proceeds further. At the adjudication stage, the adjudicating officer must separately verify that the evidence and facts being relied upon were disclosed in the SCN and that noticees had a meaningful opportunity to respond to them. These are two distinct checkpoints (one at investigation, one at adjudication) and both must hold. Balram Garg surfaces the second checkpoint; the first is equally important, even if it did not arise here. This requires robustness and co-ordination throughout the investigation-adjudication axis at SEBI.

Sufficiency of evidence. Balram Garg establishes that when relying on circumstantial evidence, SEBI must logically stress-test each inferential link before drawing conclusions from it. Per the Supreme Court, communication of UPSI can only be proved through cogent evidence (call records, messages, emails) and cannot be inferred from trading patterns alone. SEBI should codify evidentiary standards like this in a procedural manual: one that sets out internal rules of evidence, steps of proceedings, and settled questions of law, made publicly available to provide a standing reference and improve consistency.

Erroneous citation of precedent. One of the more serious failures in the SEBI order was the misapplication of precedent. The order extracted a single paragraph from Chintalapati Srinivasa Raju to support its conclusion, while ignoring the holding that immediately followed, which explicitly rejected the inferential link SEBI was relying on. Adjudicators must therefore read cited judgments in full and not selectively rely on passages that support their conclusions while overlooking those that constrain them. LLMs offer a practical tool here: they can summarise precedential judgments, identify the ratio, and verify whether an order’s characterisation of a case matches its actual holding.

Burden of proof. According to the Supreme Court, shifting the burden to disprove their insider relationship onto the Noticees was a fundamental error of substantive law. It was SEBI which was legally required to back its allegation up with evidence and reasoning. While appellate courts exist to correct such misapplications of the law, SEBI can curb their recurrence by supporting their adjudicators through targeted and frequent legal training, better research resources, and an expanded procedural manual codifying settled statutory standards.

Conclusion

Systematically reading appellate decisions helps identify systemic gaps and reduce appeals. Because complete lifecycle datasets are currently non-public and hard to map, SEBI should maintain a structured database linking orders to their appellate outcomes to loop feedback back into its processes. This feedback should be codified into a public procedural manual covering proceedings, evidentiary standards, and record-keeping to standardise institutional adjudication and help stakeholders navigate the system. SEBI should build a precedent bank supported by responsibly governed LLMs to verify case usage, ensuring AI aids, rather than replaces, human adjudicatory reasoning. Finally, adjudicators require continuous access to legal resources to prevent errors of law. Our synthesis of appellate feedback in the SEBI Lifecycle Datasets aims to provide precisely this support.

- The author is a researcher at TrustBridge, and would like to thank Pratik Datta, Bhavin Patel, and Renuka Sane for their feedback and suggestions.

CITATION

Amrutha Desikan, 2026. “Lessons from Balram Garg for SEBI’s adjudicatory process”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Balram Garg v. Securities and Exchange Board of India [2022] 4 S.C.R. 888.

Balram Garg v. Securities and Exchange Board of India SAT Appeal No. 375 of 2021.

Chintalapati Srinivasa Raju v. SEBI (2018) 7 SCC 443.

Final Order in the matter of Insider Trading in the scrip of PC Jeweller Ltd, (SEBI Order, 11 May 2021) WTM/AB/IVD/ID6/11662/2021-22.

Kishore R. Ajmera v. Securities and Exchange Board of India (2016) 6 SCC 368.

Shivani Gupta and ors. v. Securities and Exchange Board of India SAT Appeal No. Appeal No. 376 of 2021.

Tarlochan Dev Sharma v. State of Punjab (2001) 6 SCC 260.

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

On 29 May this year, the Supreme Court issued its judgment in Reliance Industries Limited v. Securities and Exchange Board of India (the “Reliance judgment”).1 The judgment is significant for many reasons (for example, it distinguishes between valid hedging and fraud). In this post, we focus on its sharpening of the definition of “fraud” in Regulation 2(1)(c) of the Prohibition of Fraudulent and Unfair Trade Practices Regulations, 2003 (“PFUTP Regulations”). We choose this point of focus since this definition has long been criticised as being too open-ended and over-inclusive,2 which is a matter of concern as it sits at the core of the market regulator’s enforcement actions in market abuse matters.

The Reliance judgment’s impact is already visible: the Securities and Exchange Board of India’s (“SEBI’s”) ex-parte interim order dated 3 June in respect of Rajesh Exports Limited (“REL”) and Rajesh Mehta (the “Rajesh Exports order”) offers an early glimpse into how securities enforcement on market abuse may evolve.

First, we examine how the Supreme Court has reinterpreted the concept of fraud under Regulation 2(1)(c). Then, we analyse the Rajesh Exports order as an early example of how SEBI may operationalise the framework articulated in the Reliance judgment.

Fraud under the PFUTP Regulations: What did the Supreme Court say it means?

Under the Securities and Exchange Board of India (Prohibition of Fraudulent and Unfair Trade Practices relating to Securities Market) Regulations, 1995 (the “1995 Regulations”), fraud was defined largely along the lines of Section 17 of the Indian Contract Act, 1872: that is, acts committed with intent to deceive another person or induce them to enter into a contract.

In 2003, SEBI significantly expanded the definition of fraud through Regulation 2(1)(c) of the PFUTP Regulations (which replaced the 1995 Regulations). The term now came to include “any act, expression, omission or concealment” committed while dealing in securities, “whether in a deceitful manner or not”, in order to induce another person to deal in securities. This definition has not been changed since 2003.

In the Reliance judgment, the Court identified three features of this definition:

• First, a deceitful or mala fide intention is not required,

• Second, inducing another person to deal in securities is required, and

• Third, proof of injury is not required.

Therefore, a bare reading of Regulation 2(1)(c) would suggest that inducement is the only mandatory element of the definition of fraud. Sandeep Parekh notes that this definition is “meaningless, as it requires neither deceit nor harm.”3

The Court quoted Mr Parekh’s statement, and agreed with its own observations in SEBI v. Kanhaiyalal Baldevbhai Patel,4 that fraud may be very difficult to define, but went on to state “such difficulty should not result in such a legislation that would cover every act, expression, omission or concealment under the sky.”5 It criticised Regulation 2(1)(c) as an “illustration of inelegant legislative drafting” and found it necessary to ask the question: “[W]hat exactly is the basis for someone to fall under the definition of fraud because at the moment, anything and everything in the stock market that may induce someone to deal in securities, could very well be termed as fraud by the respondent.”6

The Court then wades into the weeds of purposively interpreting Regulation 2(1)(c), and articulates the following requirements to establish “fraud” under the PFUTP Regulations:

• An injury must be established:

  • The injury can be established by showing ‘inducement to deal in securities’,

  • The inducement must be such as “...has caused the other person to be adversely affected and allowed the party accused of fraud to gain unlawful profits or avert ordinary losses at the former’s expense”,

  • If inducement of this sort is proved, then “there would be no requirement on the respondent authority to prove deceitful intention”,

  • Alternatively, injury can also be established by showing ‘manipulation’: inducement need not be shown “once the factum of manipulation” is established, although this requires a higher standard of ‘preponderance of probabilities’ than inducement;

• But where “injury is impossible to be proved, the requirement of wrongful intention becomes mandatory”:

  • Where “intention to defraud and manipulate the securities market is clear from the blatant misconduct or attending circumstances that cogently establish wrongful intention, then proving the injury would not be required.”

Finally, the Court carves out an exception to this approach for the limited set of circumstances covered by the case of SEBI v. Rakhi Trading (P) Ltd (“Rakhi Trading case”).7

We now examine whether the Rajesh Exports order, published a mere five days after the Reliance judgment, satisfies this newly-minted re-defined version of fraud.

The Rajesh Exports order: How has SEBI responded to the Reliance judgment?

The Rajesh Exports order alleges that REL, Rajesh Mehta, and certain related entities engaged in a scheme that materially misrepresented the company’s financial position and operations, and prima facie finds violations of several legal provisions, including the PFUTP Regulations.

The Rajesh Exports order responds to the Reliance judgment in Parts C (Dealing in securities and price volume impact) and D (Prima facie findings against Noticees). Part D specifically calls out the Reliance judgment, describes how it has clarified the interpretation of fraud, and goes on to argue that the order has satisfied the threshold prescribed in the Reliance judgment. Prior to the Reliance judgment, SEBI might have proceeded directly from allegations to a conclusion of fraud under the PFUTP Regulations.

We examine whether the Rajesh Exports order does indeed meet the requirements of the Reliance judgment. Note that this is an interim order, based on SEBI’s prima facie findings. As a result, the Rajesh Exports order need not conclusively establish that fraud has taken place. It just has to establish that there is enough evidence to justify interim measures.

While the Rajesh Exports order seems to fulfil the requirements of the Reliance judgment on the face of things, it can be argued that it does not show that the inducement “allowed the party accused of fraud” (i.e., REL and Rajesh Mehta), “to gain unlawful profits or avert ordinary losses at the former’s expense.” SEBI argues that the increase in the number of shareholders and the fall in the share price shows that investors were adversely affected; but it does not show how REL or Rajesh Mehta profited or avoided losses at the investors’ expense.

This may perhaps be expected, since this is an interim order based only on prima facie findings.

Conclusion

The significance of the Reliance judgment extends beyond the facts of a dispute that began nearly two decades ago. At one level, the judgment represents an important attempt to impose clarity in the previously ‘meaningless’ definition of fraud under the PFUTP Regulations: the Court has raised the analytical burden that must be satisfied before a finding of fraud can be established. SEBI must now prove either injury or intention in the manner the judgment sets out. In that sense, the Reliance judgment is a welcome development. However, we suggest the Court has missed two important aspects:

• It has left open the question of how SEBI must demonstrate that the ‘induced person’ was ‘adversely affected’ or how the accused person gained “unlawful profits or avert ordinary losses at the former’s expense”; SEBI has often argued that concepts such as ‘amount of loss or gain caused to investors’ are impossible to establish, and the possibility exists that it will use similar logic to argue that it cannot make such calculations in the case of PFUTP violations; and

• It has left the definition of ‘manipulation’ open; consequently, future litigation is likely to focus arguments over whether ‘manipulation’ has occurred, which could have been avoided had the Court defined this term clearly in the Reliance judgment. One way in which this could have been done would be to set down that ‘manipulation’ means ‘market manipulation’, that is, broadly speaking, some actions that demonstrably affect price discovery in the market. Not doing this leaves wiggle room for the regulator to get away with self-concocted definitions of manipulation.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback.

CITATION

Natasha Aggarwal, Amol Kulkarni, and Bhavin Patel, 2026. “SEBI learns fast, but has the Supreme Court been a good teacher? The Reliance judgment and how it shaped the Rajesh Exports order”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Reliance Industries Limited v. Securities and Exchange Board of India (Civil Appeal No 4015 of 2020).

Sandeep Parekh, Fraud, Manipulation and Insider Trading in the Indian Securities Markets (5th edn, LexisNexis 2025).

Securities and Exchange Board of India v. Kanhaiyalal Baldevbhai Patel, (2017) 15 SCC 1.

Securities and Exchange Board of India v. Rakhi Trading (P) Ltd., (2018) 13 SCC 753.

Read on to see what’s kept us busy.

Events

From May 22-23, 2026, researchers from TrustBridge participated in the final convening of Regulation in India: Theory, Practice and Reform, held at BASE University, Bengaluru. The initiative is a joint project of the Isaac Centre for Public Policy (ICPP) at Ashoka University, TrustBridge, and XKDR Forum, and has evolved through a series of conferences across India exploring regulatory theory, its application across sectors, and the place of regulation in India's growth journey. The Bengaluru convening brought together former and current regulators, civil servants, academics, and practitioners to reflect on the evolution of regulatory practice in India. TrustBridge researchers contributed to panel sessions on strengthening India's regulatory institutions through improvements in design, capacity, governance, and accountability — discussions that will inform the themes and recommendations of the forthcoming book. Learn more about the project here.

Blogs

On June 9, 2026, Prashant Narang and Aryan Panday co-authored an article on the Leap Blog, 'When remedies become regulation: The Karnataka High Court's intervention in food licensing and street vending'. The article examines the Karnataka High Court's September 2025 judgment in Karnataka Pradesh Hotel & Restaurants Association v. Union of India, in which the Court directed the Union Government to classify restaurants by size and asked the State to introduce separate health and safety rules for street vendors and food trucks. The authors note that the directions stepped beyond statutory adjudication into regulatory design, ignoring the Street Vendors Act, 2014, and assuming a level of state capacity that does not exist on the ground. The result, they suggest, is selective enforcement and compliance costs borne disproportionately by informal workers.

The Street Vendors (Protection of Livelihood and Regulation of Street Vending) Act, 2014 (”SVA”) was specifically enacted to balance livelihoods against congestion, public health and urban order. It overrides inconsistent municipal laws and works through town vending committees (”TVCs”), surveys, and certificates of vending. The SVA is not merely a procedural architecture; it embodies a considered normative choice by Parliament, that street vendors are rights-holders, entitled to livelihood protection, meaningful participation through TVCs, and procedural safeguards before any restriction on their vending.

By directing new health and safety rules for vendors without engaging with this framework, the Court implicitly undoes that normative settlement. It treats vendors not as participants with protected rights but as subjects of fresh regulation – inverting the very premise of the statute Parliament enacted for them.

Papers

Natasha Aggarwal, Amrutha Desikan, Bhavin Patel, Satyavrat Bondre, and Dipyaman Sanyal (of Dono Consulting) co-authored a paper titled 'Can technology augment order writing capacity at regulators?', published on June 4, 2026 as an advance article in Volume 22, Issue 1 of NLSIU's Indian Journal of Law and Technology. The paper examines the opportunities and challenges of using Large Language Models (LLMs) to assist regulatory order writing in quasi-judicial settings in India. It identifies the core principles of administrative law at stake, examines how the inherent limitations of LLMs may undermine them, and draws on international frameworks and case studies to propose a Problem-Solution-Evaluation (PSE) framework for responsibly integrating LLMs into order writing — augmenting rather than replacing human decision-makers.

We suggest that there is room for improvement as regards the current state of Indian regulatory order writing: such orders are written in inconsistent ways and often do not withstand challenge in appeal. As an example, the Securities and Exchange Board of India’s (”SEBI”) adjudicatory orders take an inconsistent position on what standard of proof should be applied in insider trading matters: while some SEBI orders cite ‘beyond reasonable doubt’ as the applicable standard, others cite ‘preponderance of probabilities’ as the relevant standard. This, despite the Supreme Court having established the latter as the relevant standard in SEBI’s quasi-judicial proceedings. We also note that between 2009 and 2023, the Securities Appellate Tribunal (”SAT”) allowed, partly allowed, or remanded 52% of the appeals before it that arose from SEBI’s orders in insider trading matters. This problem is not unique to SEBI; in 2023-24, half the appeals preferred against orders of the Competition Commission of India were allowed.

Op-eds

On May 20, 2026, Prashant Narang co-authored an opinion piece in The Print, 'Court must stop playing theologian. Approach Sabarimala through religious freedom lens', examining the Essential Religious Practices (ERP) test that Indian courts use to determine constitutional protection for religious practices.

Religious freedom should remain a shield against interference, not a licence granted after theological cross-examination. India does not need courts or governments to reorganise faith. For decades, Indian courts have dealt with the Essential Religious Practices (ERP) test. Under this doctrine, a practice receives constitutional protection only if the court finds it essential to a religion. What began as a way of protecting religious autonomy has slowly become a method of judicial certification.

On May 22, 2026, Atul Raman co-authored an article in DT Next, 'Fiscal discipline: Why welfare promises need sunset clauses', examining Tamil Nadu's fiscal trajectory as outstanding liabilities cross Rs 9.5 lakh crore. The piece traces how welfare schemes in the state are rarely dismantled but repeatedly renamed and rebranded across successive governments, and makes the case for mandatory sunset clauses to evaluate welfare spending, ensure long-term fiscal sustainability, and strengthen accountability.

In a State with outstanding liabilities exceeding Rs 9.56 lakh crore — among the highest in the country according to RBI data, and an expanding welfare architecture that now includes promises such as monthly stipends for women and subsidised LPG cylinders, what institutional mechanism prevents welfare from turning into permanent expenditure without evaluation? This is where India’s policy discourse urgently needs to consider the idea of sunset clauses, a policy mechanism through which a scheme automatically comes up for review after a fixed period unless formally renewed following evaluation.

The principle is simple: no programme should continue indefinitely without periodic assessment of whether it is achieving its intended outcomes. Increasingly, institutions financing state development are also demanding measurable outcomes rather than perpetual expenditure.

On May 27, 2026, Renuka Sane authored an opinion piece in The Print, 'Rainfall derivatives have arrived in India. We need 3 steps to make them work', examining the NCDEX's launch of a Liquidity Enhancement Scheme for Mumbai Rainfall Futures as the beginning of a market for monsoon risk in India. The piece notes that unlike insurance, which compensates for actual loss, weather derivatives pay out based on the occurrence and magnitude of predefined weather conditions, and sets out the steps needed to make this nascent market work.

The idea of derivatives to deal with weather-related risks started around 1999. The Chicago Mercantile Exchange listed the first Heating Degree Day futures in September that year. The derivative essentially prices how cold a winter will be, measured by how many degrees each day’s temperature falls below 18 degrees Celsius, the point at which people start turning on their heaters. If it is a warm winter, people won’t turn on their heaters, and the gas utility would lose revenues.

Cooling Degree Day contracts, which measure how many degrees each day’s temperature rises above 18 degrees Celsius, followed in January 2000. While these were initially tools for gas utilities and energy traders, they gradually came to be used by businesses from breweries to construction companies. Temperature gave way to rainfall, wind speed, frost days, and snowfall.

Outstanding weather derivatives are estimated at $25 billion. However, the vast majority of these are bilateral contracts between banks, energy traders, insurers, and corporates, away from any exchange, without transparent pricing or centralised clearing. India has enormous potential to become a sophisticated user of weather derivatives, given its agricultural dependence, its growing urban economy, and its acute climate vulnerability.

On June 2, 2026, Renuka Sane co-authored an opinion piece in The Print, 'Why India needs a dedicated financial conduct regulator', examining how India's rapidly expanding financial system has outpaced its consumer protection architecture. The piece suggests that while all major financial regulators recognise consumer protection as an objective, it remains subordinate to their primary mandates of financial stability, prudential supervision, and market development, and makes the case for a dedicated conduct regulator with specialised expertise and clear accountability.

All major financial regulators in India—the Reserve Bank of India (RBI), Securities and Exchange Board of India (SEBI), Insurance Regulatory and Development Authority of India (IRDAI) and Pension Fund Regulatory and Development Authority (PFRDA)—recognise consumer protection as an important regulatory objective. Yet, in practice, it remains subordinate to their primary mandates of financial stability, prudential supervision and market development. Ombudsman and grievance-redressal arrangements exist, but they are frequently perceived as slow, fragmented and lacking sufficient authority to provide effective remedies.

This is further complicated by many issues that fall into the cracks of India’s fragmented regulatory architecture. For example, a customer who has been mis-sold an insurance product through a bank branch may find the issue touching the jurisdictions of multiple regulators. Similar complexities arise in the distribution of mutual funds, pension products and hybrid financial instruments. Too often, consumers find themselves negotiating directly with the very institutions against whom they seek redress.

The Bridge

The Bridge is TrustBridge's series looking closely at the quasi-judicial work of India's regulators — how they decide, and how they could decide better. On June 1, 2026, Natasha Aggarwal, Amrutha Desikan, and Bhavin Patel published 'The invisible goalpost: LLMs, implicit assumptions, and the limits of prompt engineering'. The piece tests whether commercial LLMs can review a regulator's adjudicatory order as reliably as a legally trained human reviewer, using SEBI's March 2026 order against Coffee Day Enterprises Limited evaluated against TrustBridge's Good Order Writing indicators. The authors compare the LLM's findings with those of a human reviewer to assess how well present-day models replicate the implicit assumptions and heuristics that shape legal order-reading.

TrustBridge is interested in the potential use of technology, particularly large language models (LLMs), to improve order review and writing practices at Indian regulators. To be clear, we do not suggest that LLMs 'write' or 'draft' orders. We think of 'order writing practices' as comprising several steps, which include a review of draft orders. Such a review could be conducted when the order is at a draft stage, to alert the human order-writer to any deficiencies or errors. It is in that context that we examine the feasibility of using LLMs — to conduct such a preliminary review of orders as a part of the order-writing process.

On June 10, 2026, Natasha Aggarwal and Bhavin Patel published 'Evaluating regulators' orders: The good order writing indicators explained' for The Bridge. The piece explains the Good Order Writing (GOW) indicators that TrustBridge has been developing and applying over the past three years to evaluate the completeness and quality of regulators' adjudicatory orders, setting out what the indicators are, why they were developed, and how they are adapted to different regulators and proceedings.

Our GOW indicators support a structured methodology to measure the completeness of adjudicatory orders. The indicators are flexible by design, and the choice of indicators, weights, response formats, and sufficiency standards all depend on context. Each regulator operates under a different statutory framework, exercises different forms of adjudicatory power, and deals with different types of violations. A framework suitable for analysing SEBI orders may not be appropriate for studying tariff orders of electricity regulators or preliminary closure orders under Section 26(2) of the Competition Act, 2002. This also means that the process of developing and applying GOW indicators is iterative, and our work on the GOW indicators is constantly evolving.

Trainings

On May 26, 2026, Prashant Narang delivered an online lecture on 'Right to Services and E-Governance' as part of the Continuing Development Programme on Applied Administrative Law, coordinated by Dr. Ram Manohar Lohiya National Law University. The programme brought together around 30 law professors alongside senior academics and practitioners working on administrative law.

On June 4 and June 13, Prashant Narang delivered sessions on 'Rule of Law and Generality' and 'Rule of Law and Regulation' respectively at the Indian School of Public Policy as part of the iPolicy for Young Leaders programme. The programme engaged 32 scholars spending eight weeks on the theme of Economics of Education.

Work with us

We wish to recruit a Research Fellow for our Financial Markets function in the New Delhi office. If you are passionate about translating rigorous research into national impact, we invite you to explore the opportunity. Reach out to us at info@trustbridge.in, we’re always open to working with like-minded individuals.

Thanks for reading The TrustBridge Newsletter! Subscribe for free to receive new posts and support our work.

          Website  |  LinkedIn  |  YouTube 

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

TrustBridge has been studying Indian regulators’ adjudicatory orders and practices for about three years.1 As a part of this work, we use a set of “good order writing” (“GOW”) indicators that were developed to evaluate the completeness and quality of regulatory orders. The GOW indicators are context-dependent, which means that the specific indicators to be used in a study are adapted depending on the regulator, the statutory framework, the type of violation, and the type of proceedings involved. This post explains what these GOW indicators are, why we developed them, and how we use them.

The GOW indicators are based on our working paper, A guide to writing good regulatory orders (the “GOW Paper”).2 The paper surveys Indian administrative law requirements, comparative guidance from other jurisdictions, and existing scholarship on adjudicatory writing by regulators to identify the core principles that orders should satisfy. It argues that good orders must do more than merely announce conclusions: they must clearly establish the legal basis of action, identify the relevant facts and issues, demonstrate application of mind, explain the reasoning connecting facts to law, justify sanctions, and communicate decisions in language that is accessible and comprehensible to affected stakeholders.

The starting point: what makes a “good” regulatory order?

We postulate a set of principles of good order writing in the GOW Paper; well-written orders must:

• Clarify the basis for the exercise of regulatory power,

• Clearly identify the legal provisions that are purportedly violated,

• Provide detailed background and contextual information,

• Provide well-reasoned justifications for findings of fact and law,

• Specify the regulatory action being taken and correlate it with the findings of fact and law, and

• Describe the rights of review or appeal of the persons involved

To satisfy these principles, we suggest that regulatory orders should satisfy four broad sets of requirements:

1. Informational requirements: Orders must include certain information (such as the names of the decision-makers).

2. Structural requirements: Orders should be structured in a logical manner. We suggest the following sequence: prefatory information, introductory statement, procedural history, issues, findings of fact, analysis, decision, sanctions, directions for compliance, and appeal and review processes.

3. Substantive requirements: Orders must demonstrate that the requirements of statutory and administrative law have been satisfied.

4. Stylistic requirements: Orders must be written in clear, concise, and comprehensible language.

The GOW indicators operationalise these into measurable variables.

Developing the GOW indicators

We have developed the GOW indicators based on the framework in the GOW Paper. One of the first use cases for these indicators was a set of 82 GOW indicators to study orders of the Securities and Exchange Board of India (SEBI) (available here). Some examples, and how they relate to the principles and requirement categories identified in the GOW Paper, are in the table below.

These GOW indicators are intended to measure the completeness and quality of orders. They are not designed to determine whether a regulator reached the “correct” decision on merits.

Moreover, these GOW indicators are not intended to function as a fixed checklist applicable across all regulators and types of proceedings. While some GOW indicators are broadly applicable across regulators, the GOW indicators may also be modified to suit the regulator, the statutory framework, the type of violation, and the type of proceedings involved. Some examples are in the table below.

Weights

Not all deficiencies in an order are equally consequential. An order that omits the names of counsel is incomplete but not necessarily problematic; on the other hand, an order that fails to explain why sanctions were imposed raises a much more serious problem. To reflect such differences, we assign weighted scores to GOW indicators on a scale from 0 to 1. For example, whether the names of counsel are included is weighted at 0.25, while whether the reasons for imposing sanctions are explained is weighted at 1.

Recording findings

Recording responses for the indicators raise various methodological challenges: recording responses for some is relatively straightforward, while others require contextual interpretation and qualitative judgment. Similarly, different studies may require response formats and different thresholds for recording findings.

Accordingly, when applying GOW indicators, we also develop:

• An objective/subjective classification of the GOW indicators,

• Response formats for recording findings, and

• Sufficiency standards that specify what level of information is required for a finding to be recorded against an indicator.

Objective/ subjective

The GOW indicators can be classified into two categories: ‘objective’ and ‘subjective’:

• An objective indicator can be satisfied by the explicit presence of the required information, with little room for interpretation or disagreement. For example, “Are the names of decision-makers present?” is objective, since the information either is or is not present in the order.

• A subjective indicator requires contextual interpretation and qualitative assessment. For example, “Is there a clear, concise, and comprehensive summary of uncontested material facts?” is subjective because evaluators may reasonably differ on what constitutes a “material” fact, or whether the discussion is sufficiently clear and comprehensive.

This distinction helps identify which aspects of adjudicatory quality could be measured through relatively straightforward coding, and which require more interpretive evaluation.

Response formats

Response formats are the formats that may be adopted to record findings for GOW indicators. Some response formats, and the context in which they may be appropriate, are:

Closed responses

These responses may comprise values such as:

• ‘1’ or ‘Yes’ for a positive response;

• ‘0’ or ‘No’ for a negative response; and

• ‘99’ or ‘NA’ for situations where the indicator is inapplicable

These may be appropriate where:

• Responses can be objectively provided with a high degree of confidence, such as for objective indicators; and/ or

• Downstream analysis depends on simple values that can be analysed using quantitative methods.

Open responses

These responses do not use a closed set of values. They may:

• Be descriptive in nature, and non-conclusive;

• Provide an affirmative or negative response, with supporting reasons and source references; or

• Use terms such as ‘likely’, ‘possible’, or ‘probable’, or their inverse, that is, terms such as ‘unlikely’, to indicate a broad range, rather than a precise response.

These may be appropriate where:

• It is hard or impossible to draw definitive conclusions because of information insufficiency;

• Data is qualitative, unstructured, and open to interpretation; and/ or

• A final response requires interpretation and decision-making by an authority such as a court or tribunal.

Sufficiency standards

The level, type, extent, and adequacy of information that must be present in an order for a particular finding to be recorded against an indicator. While the choice of sufficiency standards, like the choice of response format, is contingent on context, it is also contingent on the choice of response format: closed response formats may require the articulation of very precise sufficiency standards, while open responses may not require the identification of very precise sufficiency standards.

Some examples of the sufficiency standards that may be chosen for various GOW indicators, and when they may be appropriate are:

• For the indicator ‘Are the names of decision-makers present?’, the chosen standard may be:

‘As long as names appear towards the very beginning or ending of the order, record a positive finding.’

This may be appropriate where:

• The names of decision-makers consistently appear in the designated locations.

OR

‘As long as names appear towards the very beginning or ending of the order, along with the designations of the decision-makers, record a positive finding.

This may be appropriate where:

• The names and designations of decision-makers consistently appear in the designated locations; and/ or

• Obtaining the names and/or designations is important for analysis.

• For the indicator “Is there a clear, concise, and comprehensive summary of uncontested material facts?”, the chosen standard may be:

‘As long as the uncontested material facts can be identified from statements scattered across the order, record a positive finding.’

This may be appropriate where:

• The focus is only on whether the facts are present somewhere in the order, rather than how they are presented;

• The downstream use of the information does not require the facts to be consolidated in one location; and/or

• The proceeding is preliminary in nature, such as an order under Section 26(2), where the expectation on the adjudicator to demonstrate detailed application of mind is comparatively lower.

OR

‘Record a positive finding only if the order contains a consolidated summary of uncontested material facts.’

This may be appropriate where the focus is on readability and structural coherence.

OR

‘Record a positive finding only if the order contains a clear, concise, and comprehensive summary of uncontested material facts in a designated location, such as the factual background section of the order.’

This may be appropriate where:

• The focus is on readability and structural coherence;

• The proceeding culminates in a definitive determination, such as an adjudication order imposing penalties, where a higher expectation of application of mind and structured reasoning is warranted.

Conclusion

Our GOW indicators support a structured methodology to measure the completeness of adjudicatory orders. The indicators are flexible by design, and the choice of indicators, weights, response formats, and sufficiency standards all depend on context. Each regulator operates under a different statutory framework, exercises different forms of adjudicatory power, and deals with different types of violations. A framework suitable for analysing SEBI orders may not be appropriate for studying tariff orders of electricity regulators or preliminary closure orders under Section 26(2) of the Competition Act, 2002. This also means that the process of developing and applying GOW indicators is iterative, and our work on the GOW indicators is constantly evolving.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback.

CITATION

Natasha Aggarwal and Bhavin Patel, 2026. “Evaluating regulators’ orders: The good order writing indicators explained”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Chitrakshi Jain and others, 2025, “Examining the Performance of ERCs at APTEL,” The Leap Blog.

Natasha Aggarwal and others, 2025. “A guide to writing good regulatory orders,” Working Papers, Trustbridge Rule of Law Foundation.

Natasha Aggarwal and others, 2025, “Balancing Power and Accountability: An Evaluation of SEBI’s Adjudication of Insider Trading,” Working Paper, Trustbridge Rule of Law Foundation.

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

Can an LLM review a regulator’s adjudicatory order as reliably as a human lawyer? Can it replicate implicit assumptions and heuristics that are an intrinsic part of human order-reading?

TrustBridge is interested in the potential use of technology, particularly large language models (LLMs), to improve order review and writing practices at Indian regulators.1 To be clear, we do not suggest that LLMs ‘write’ or ‘draft’ orders. We think of ‘order writing practices’ as comprising several steps, which include a review of draft orders. Such a review could be conducted when the order is at a draft stage, to alert the human order-writer to any deficiencies or errors. It is in that context that we examine the feasibility of using LLMs - to conduct such a preliminary review of orders as a part of the order-writing process.

To test the readiness of commercial LLMs for this purpose, we design an experiment to evaluate whether an LLM can review an order as accurately as legally trained human reviewers.

This experiment can provide insight on two questions:

1. How reliably can an LLM replicate or approximate human evaluation in applying standards of good order writing?

2. How well can LLMs assess the completeness of adjudicatory orders against the same standards?

Answers to these questions can help determine the feasibility of using present-day LLMs as an order-writing-assistive tool: if the experiment shows that the technology is reliable, a pathway exists to using this technology to support order writing and review processes at regulators. If the tech doesn’t come out looking great, the experiment can tell us where exactly it falls short, so that potential solutions can be developed. We acknowledge that a true experiment would require several observations, spanning a reasonably large sample of orders. That said, we suggest that looking at even one order is useful since it helps us understand the question better, and can help set up a more reliable experiment for a larger study.

We use the Securities and Exchange Board of India’s (the “SEBI’s”) March 2, 2026 adjudicatory order against Coffee Day Enterprises Limited (the “SEBI Order”) for this experiment. For the standards expected of good order writing, we use the “Good Order Writing” indicators developed by Trustbridge (the “GOW Indicators”). We then compare the findings of our human reviewer with those of the LLM, and postulate reasons for differences between them.

Methods

TrustBridge has developed a set of “good order writing” indicators to measure the completeness and quality of adjudicatory orders issued by Indian regulators.2 We customised these GOW Indicators for SEBI orders,3 and the resultant final framework comprises 82 indicators (the “SEBI-GOW Indicators”).4

Indicators

The SEBI-GOW Indicators were classified into two categories: ‘objective’ (53/82 indicators) and ‘subjective’ (29/82 indicators):

• An objective indicator can be satisfied by the explicit presence of the required information within the text; further, there is little room for interpretation or disagreement on such text. For example, “Is the information that led to the initiation of proceedings mentioned?” is objective; the answer is binary, since the information either is or is not present in the order.

• A subjective indicator demands contextual interpretation and relies on qualitative judgment. For example, “Does the order avoid repetition of non-material facts and/or inclusion of irrelevant details?” is subjective, as human reviewers may disagree on what qualifies as “non-material” or “irrelevant.”

Review

Our ‘human reviewer’ holds a Master’s qualification in law, and is part of the team that developed the GOW Indicators. Their work was double-checked by another researcher, who also holds a Master’s qualification in law. (Notice that we studiously avoid calling a Master’s qualification in law an ‘LLM’ to avoid confusion. You’re welcome.) This helped us establish a robust baseline for this experiment.

We chose to use Anthropic’s Claude Sonnet 4.6 as the LLM reviewer. We use this LLM frequently in our work on order review and analysis, and find it generally more suitable for this work than other commercial LLMs: Sonnet’s context window is sufficiently large for most regulatory orders; at the same time, it is more cost-efficient in terms of token usage than larger models such as Anthropic’s Opus.

Recording data

We recorded descriptions and the manner of recording responses for all the SEBI-GOW Indicators and fashioned them into a prompt manual. We provided the LLM the following material:

• the list of SEBI-GOW Indicators,

• our descriptions for each Indicator, and

• the prompt manual.5

If an indicator is satisfied, the human and the LLM record a ‘1’. If an indicator is not satisfied, they record a ‘0’. If the indicator is inapplicable, they record ‘99’. The human reviewer and the LLM were both instructed to score the order for each indicator, and to record reasons for their position.6 The findings of both are available here. We examined the findings to identify mismatches in positions. Finally, we analysed the reasons for disagreement between the human and the LLM.7

What the human and the LLM agree on

The human and the LLM agreed upon answers for 72/82 indicators (a staggering 87.8%). Of these, 48 were objective indicators (out of a total of 53 objective indicators), and 24 subjective (out of a total of 29 subjective indicators).

This indicates that when trained on a benchmarking framework like the SEBI-GOW Indicators and provided precise instructions, an LLM can conduct preliminary order review exercises with near human-level accuracy. This also suggests that the SEBI-GOW Indicators are well-designed for translation into prompt manuals that can be used to train LLMs. We propose to release a detailed prompt manual based on the SEBI-GOW Indicators shortly.

What they don’t see eye-to-circuit on

The human and the LLM disagreed on responses to five objective and five subjective indicators. These are:

And why they disagree

Our review of the indicators for which there was human-LLM disagreement suggests that these can be divided into two broad categories:

1. Indicators for which better prompt engineering can improve results, and

2. Indicators for which prompt engineering may not increase human-LLM agreement.

Indicators for which better prompt engineering can improve results

For some indicators, it is objectively possible to identify ‘correct’ responses. In such situations, it would also be possible to ensure consistency across various human reviewers. For example:

For the indicator: ‘Is there a unique order number?’, we can specify the format typically used by the regulator whose orders are under study, and it is highly improbable that two human reviewers would disagree, in the case of almost any order, whether the order number appears in the order or not.

Where prompts are framed narrowly, the LLM adheres to those limits, but in some cases, human reviewers apply additional, implicit expectations. These may include differentiating between reproduction of parties’ arguments and the adjudicator’s own analysis, based on the language used. These expectations may arise from conventions of legal reasoning and adjudicatory practice that human reviewers internalise through legal training and experience. To the extent that such expectations can be articulated, better prompting can lead to higher human-LLM agreement in responses.

For such indicators, it is possible to ensure a high degree of human-LLM concordance through careful prompt engineering. This could be achieved by ensuring that:

• Prompts are detailed, specific, and provide clear instructions;

• Implicit assumptions and heuristics that human reviewers consistently rely upon are made explicit in the prompts; and

• The general structure and flow of orders are well-explained to the LLM, so that they are not too dependent on the presence of clear headings and sub-headings in the order.8

Accordingly, we employed these methods and revised the prompts for two objective indicators.9 This resulted in an agreement rate of approximately 90%, as opposed to the previous 87%.

Indicators for which prompt engineering may not lead to better human-LLM agreement

Human reviewers may not always agree on the responses for some indicators, or may have different ways of interpreting them. In such cases, there is a high likelihood that various human reviewers may provide different responses to the indicator. For example:

For the indicator: ‘Does the order avoid mere repetition of precedent?’, we instructed the LLM that precedent should not be extensively quoted. Accordingly, the LLM concluded that precedent cited in the SEBI Order were not extensively reproduced, as only short paragraphs were extracted. The human was dissatisfied with this because, in their view, an order does not meaningfully use precedent unless it explains how those extracted paragraphs and the reasoning in them has been applied to the present facts. Another human reviewer might argue that, in addition to the explanation on how precedent is applied to the facts under consideration, the order must explicitly identify whether precedent was followed, distinguished, or overruled.

In such situations, it is hard to identify a singular ‘right’ answer or method of answering the indicator that would be satisfactory to all human reviewers. And so, it is not possible to achieve greater human-LLM concordance on responses to such indicators.

In the case of some indicators, evaluating adjudicatory quality is inherently subjective. In such cases, disagreements are not necessarily errors, but reflect legitimate differences in interpretive standards. In this context, the value of the LLM lies less in replacing human judgment and more in making evaluative assumptions explicit. Requiring the LLM to provide reasons for its assessment, including the threshold it applied in determining whether an indicator was satisfied, enables human reviewers to critically engage with the model’s reasoning and use it as an input into their own assessment.

This also illustrates the limitations of using benchmarks like ‘as good as a human’ in such evaluations: this manner of judging the accuracy of an LLM assumes there is an objective, valid response for all situations which all humans always agree upon. As we have seen, that is not the case. We therefore suggest that the perspective one should adopt in such situations is determining the possibility of a singular, objectively verifiable answer in any particular situation, rather than adopting a comparison with subjective human evaluation. LLM outputs bear strong correlation to human prompting quality. If the goalpost is invisible, one cannot prompt an LLM to see it.

Conclusion

The fact that the human and LLM reviewers agreed on findings for 87% of our SEBI-GOW Indicators (90% with the improvements we made) suggests that it is indeed feasible to use such technologies as order-writing-assistive, or review, tools. Further, we are able to postulate the reasons for the remaining 10% mismatch in findings. This offers us a pathway towards further experimentation that may help further improve the degree of agreement between humans and LLMs in reviewing orders.

It is obvious that there would be a high correlation between prompt crafting and human-LLM agreement. Prompts that seek to optimise for accuracy must be highly specific; assuming that the LLM would infer or apply implicit or unstated instructions leads to mismatches. Where prompts are incomplete, ambiguous, or fail to capture implicit assumptions, the LLM will reproduce those deficiencies. A rigorous prompt engineering process for order review can provide an additional benefit: it forces human teams to standardise their own metrics for what constitutes a well-written order, requiring them to articulate and codify their assumptions. This boosts replicability of the order review process. TrustBridge has developed expertise in this regard, and we aim to release detailed prompts that can be used for order review, shortly.

We suggest our experiment demonstrates three things clearly:

• First, LLMs, with the appropriate training and prompting, may be used as preliminary review tools that can help order writers standardise order writing practices and improve order quality.

• Second, their effective deployment requires an iterative process, in which mismatches are treated as diagnostic signals to refine both prompts and underlying indicators.

• Third, while LLMs can be used to assist the order review process, by themselves, they are insufficient. A human review is indispensable. This is particularly the case for indicators on which it is hard to obtain widespread agreement across human reviewers.

Any stakeholder - whether a regulator, appellate forum or parties - using LLMs in the order review process must address the limitations of their prompting efforts by consistently iterating on the quality of their instructions and verifying subjective and inferential reasoning. We also suggest that the ‘as good as a human’ standard may not be appropriate in all instances, and that the review process must also recognise the inherent subjectivity of some indicators. In such cases, reasoning and source citation would be appropriate responses from the LLM, rather than binary ‘yes’ or ‘no’ responses.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback.

CITATION

Natasha Aggarwal, Amrutha Desikan, and Bhavin Patel, 2026. “The invisible goalpost: LLMs, implicit assumptions, and the limits of prompt engineering”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

Annexure 1

The following prompts were used in an Anthropic Claude Project. The Project was configured with a system prompt (reproduced below) that provided Claude with the evaluation framework and instructions. The order was then uploaded and assessed using the user prompt below.

System prompt

You have been provided with a spreadsheet titled ‘20260316_GOW’ with the following columns: Column B contains a list of indicators, column C has notes corresponding to each indicator, column E has the possible responses corresponding to each indicator, and column F has the weights corresponding to each indicator. You have also been provided with a document titled ‘20260316_GOW_prompts,’ which contains instructions for some of the indicators listed in the spreadsheet and, in some cases, sample orders. The sample orders have been provided to you as PDFs.

I will upload orders to this project, and ask you to evaluate them against the good order writing metrics. Please return the responses in a spreadsheet format, with the reasons for each response recorded in a separate column.

User prompt

Here is a SEBI order. Evaluate this order against the good order writing metrics. Please return the responses in a spreadsheet format, with the reasons for each response recorded in a separate column.

References

• Natasha Aggarwal and others, 2025, “A Guide to Writing Good Regulatory Orders,” Working Paper, Trustbridge Rule of Law Foundation.

• Natasha Aggarwal and others, 2025, “Can Technology Augment Order Writing Capacity at Regulators?,” Working Paper, Trustbridge Rule of Law Foundation.

• Natasha Aggarwal and Bhavin Patel, 2026, “Stitch in time: An early evaluation of the IFSCA’s adjudicatory orders”, The Bridge, TrustBridge Rule of Law Foundation.

• Adjudication Order in the matter of financial mis-statements of Coffee Day Enterprises Limited, (SEBI, 2 March 2026) Adjudication Order No: ORDER/AK/RK/2025-26/ 32161-32170.

Read on to see what’s kept us busy.

Policy inputs and submissions

Inputs on safeguards in digital payments to curb frauds

Renuka Sane, Pratik Datta, Amol Kulkarni, and Aastha Asthana submitted comments to the Reserve Bank of India (RBI) on its Discussion Paper released on April 9, 2026 titled 'Exploring safeguards in digital payments to curb frauds'. Using the Regulatory Impact Analysis framework, the submission analyses two of the four proposed options: a one-hour lag for person-to-person APP transactions above INR 10,000 (Option 1) and a default credit ceiling on all bank accounts (Option 3) and finds both to be over-inclusive, imposing universal friction to address a failure concentrated in a small subset of transactions. Key recommendations include:

• the one-hour delay under Option 1 should be triggered only where a transaction crosses a calibrated risk-score threshold, rather than applying automatically to all transactions above INR 10,000;

• all accounts should default to normal credit access, with credit ceilings applied selectively based on identified risk indicators rather than imposed on every account;

• the RBI’s own tools such as the Digital Payment Intelligence Platform and MuleHunter.AI, capable of real-time transaction-level beneficiary risk scoring and detecting mule accounts respectively, offer the targeted approach this problem demands; and

• the Discussion Paper quantifies neither the expected fraud reduction nor the direct and indirect costs of the proposed measures, leaving stakeholders unable to meaningfully assess proportionality.

Blogs

On May 11, 2026, Arjun Gupta, Sonam Patel, and Renuka Sane authored an article on the Leap Blog, 'Market Reaction to Insider Trading: Evidence from Regulatory Orders in India'. The article examines whether Indian stock markets react to SEBI insider trading enforcement actions and SAT appellate decisions, finding no statistically significant market response in sharp contrast to negative abnormal returns documented in the US and UK. The authors attribute this null result to high appeal and reversal rates, long enforcement delays, and persistently low penalty amounts, raising questions about whether SEBI's enforcement actions are achieving the goals that justify the expenditure.

Indian stock markets exhibit no statistically significant response to insider trading enforcement, in contrast to the negative abnormal returns documented in the US and UK. This result is robust across SEBI final orders and SAT appellate decisions, and persists even for high-severity violations involving senior insiders and large monetary outflows. The functioning of SEBI entails considerable public expenditure, and the Board has, over time, sought progressively wider powers — including expanded surveillance capabilities. Given this, the question of what is actually being achieved warrants serious scrutiny. A stock price reaction to an enforcement order is one observable signal of whether the market believes the enforcement actions carry some significance. A null result across many orders suggests the market does not view these actions as conveying meaningful new information. It is, therefore, worth questioning if enforcement actions are advancing the goal that justified the expenditure in the first place.

Op-eds

On April 17, 2026, Aastha Asthana and Bhavin Patel authored an article in SCC Online Blog, 'SEBI's Conflicts of Interest Framework: Analysis', examining gaps in SEBI's existing conflicts of interest framework across three dimensions: disclosure obligations, investment restrictions, and post-employment rules. Drawing on international comparators such as the SEC and FCA, the piece argues for a tiered, role-based approach that balances institutional integrity with reasonable market access.

On three key questions, we critically examined the existing framework, and conducted a comparative study of leading international agencies such as the OECD and the World Bank, and regulators like the SEC and FCA. Our recommendations employ insights from these comparisons, and emphasise adopting a tiered approach to disclosures, easing investment restrictions for junior personnel in non-sensitive roles, and implementing a risk-based assessment for post-employment restrictions for all personnel.

We suggest that SEBI’s conflicts of interest framework needs changes to address gaps and inconsistencies, and to bring it in line with prevailing international standards. This need for change is based on the requirements of consistency and comprehensiveness.

Renuka Sane authored an opinion piece on April 28 in The Print, 'Want to save Alphonso mango from heatwave? Start with open data', examining the climate crisis facing Alphonso mango farmers in the Konkan region — where production is down 60-90 per cent this season, as a warning sign for Indian agriculture more broadly. The piece suggests that the standard policy response of compensation is unsustainable when climate change is structural rather than cyclical, and makes the case for treating farm and weather data as public infrastructure essential to any serious adaptation strategy.

If we keep following the same practices under different climatic conditions, we will keep debating compensation amounts every year. In the new normal, compensation does not make for a sustainable strategy. We need a shift in the way we have been organising our production, infrastructure, and consumption. This is the central objective of climate adaptation. Any serious plan to adapt to climate change has to start by treating farm and weather data the same way we treat roads, electricity, or the internet: as essential infrastructure the nation depends on.

The Bridge

This month, TrustBridge kickstarted The Bridge — a new series dedicated to scrutinising the quasi-judicial work of India's regulators. Regulatory orders shape markets, govern disputes, and test the rule of law. Yet how regulators decide, and whether they could decide better, rarely receives the attention it deserves. We launched the series with two pieces.

On April 30, 2026, Natasha Aggarwal and Bhavin Patel authored their first piece ‘Stitch in time: An early evaluation of the IFSCA’s adjudicatory orders’, evaluating the International Financial Services Centres Authority’s seven adjudicatory orders against TrustBridge’s Good Order Writing indicators. The piece finds that while the IFSCA has successfully established the architecture of good decision-making — with consistent structural clarity and procedural completeness — its orders show recurring weaknesses in sanction justification and treatment of disputed facts, gaps that are best addressed before they become institutionalised.

As one of India’s newer SRAs, the IFSCA has a short history of issuing adjudicatory orders. Their website currently lists eight documents in the Enforcement section (seven orders, the earliest dated 29 July 2024 and the latest, 6 April 2026, and one warning letter). The IFSCA made headlines recently when, in a first-of-its-kind action for the new SRA, it cancelled the broker-dealer registration of String AI IFSC Private Limited.¹ This is seen as a significant move on the regulator’s part, potentially signalling an escalation in its enforcement actions, and moving away from the ‘marketing mode’ stance that market participants suggest it had hitherto adopted. An examination of the regulator’s enforcement practices would therefore be timely.

On May 8, Natasha Aggarwal and Bhavin Patel authored ‘How much does a poorly-written order cost?’, examining the opportunity costs imposed by the Competition Commission of India’s 2020 order against Grasim Industries, which was set aside and remanded by the NCLAT in May 2026 after over six years. The piece demonstrates how applying Good Order Writing indicators at the drafting stage could have avoided the controversy, and the unproductive lock-up of Rs. 30.16 crores over that period.

We point out the opportunity costs of poor order writing in one instance; but consider what the aggregate value of such practices across regulators would be over the period of a single year, over a few years, or over a decade. And how little in comparison it would take to avoid such costs - LLMs and Generative AI make solutions like the Order Review Tool accessible, and regulators and researchers have the domain expertise to make such tools capable and effective. But first we need to acknowledge the scale of the problem, and open our minds to the possibilities that imaginative solutions can offer.

Events

On May 9, Prashant Narang moderated Session I of the multidisciplinary webinar on 'Multidisciplinary Approach to Academic Integrity: Challenges and Remedies'. The session brought together distinguished speakers including Justice V. Gopal Gowda, former Judge of the Supreme Court of India; Prof. Upendra Baxi; Prof. N. L. Mitra; and Prof. Jason M. Stephens. The discussion examined academic integrity across education, policy, and research, covering questions around AI use, plagiarism, institutional culture, and legal education reform.

On May 13, Pratik Datta participated as a panelist in the session on 'Bankruptcy across jurisdictions: Frameworks and Cultural Contexts' at the MDI Gurgaon – IBBI Conference on Powering Viksit Bharat 2047 through Financial Sector Reforms. The panel discussed comparative insolvency frameworks and the cultural contexts that shape their design and functioning across jurisdictions.

Podcasts

On April 21, Chitrakshi Jain featured on the podcast 'Law, Laughter and Beyond', hosted by the Kautilya Society at NUJS, in a conversation on court reforms. She spoke about her co-authored book Tareekh Pe Justice, which offers a window into why district courts and their judges behave the way they do, and discussed how the National Judicial Data Grid overestimates pendency, among other findings. Chitrakshi and Prashant Reddy also shared their theory of change on judicial reforms, their instincts on why High Court judges behave the way they do, and how legal education could evolve in light of their findings.

Collaborate

If our work resonates with you, we would love to collaborate. We focus on financial markets regulation, energy transition, and contracts and dispute resolution. Reach out at info@trustbridge.in.

Thanks for reading The TrustBridge Newsletter! Subscribe for free to receive new posts and support our work.

          Website  |  LinkedIn  |  YouTube 

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

On May 13, 2026, a whole-time member (“WTM”) of the Securities and Exchange Board of India (the “SEBI”) issued an order (the “Order”) against Dr. Rashmi Saluja, the former Executive Chairperson of Religare Enterprises Limited (“Religare”), in which it held that Dr Saluja had traded in the shares of Religare while in possession of unpublished price sensitive information.1 It ordered that Dr Saluja disgorge an amount of INR 1.99 crore, and also imposed a penalty of INR 40 lakh.

The Burmans’ takeover of Religare was one of the more acrimonious corporate control battles in recent Indian corporate history. Conflicts raged across three regulatory battlefields, those of the SEBI, the Insurance Regulatory and Development Authority of India (the “IRDAI”), and the Reserve Bank of India (the “RBI”). Pleas were also filed in court, and Dr Saluja at one point alleged before the Delhi High Court that SEBI had “failed to discharge its responsibilities as a regulator”.2 Among various other shots fired, the Burmans accused Dr Saluja of insider trading.3 The SEBI acted upon those allegations through the proceedings that led to the present decision. The Order would appear to be the latest episode in the takeover saga that never ends.

The Order spans 87 pages and relies upon evidence such as Call Data Records and hotel invoices, and includes several WhatsApp screenshots to establish the veracity of the allegations against Dr Saluja by M.B. Finmart Private Limited, Puran Associates Private Limited, VIC Enterprises Private Limited, and Milky Investment & Trading Company (whom the order collectively refers to as the ‘Burman Group’). (Yes, Milky Investment.)

We examine the Order against the ‘Good Order Writing’ indicators (the “GOW Indicators”) developed by TrustBridge to measure the completeness and quality of adjudicatory orders issued by Indian regulators.4 While we do not examine the correctness of the Order, or second-guess the regulator’s decision, we are interested in the question of order quality as a matter of compliance with procedural rule of law requirements.

The TrustBridge working paper “Balancing Power and Accountability: An Evaluation of SEBI’s Adjudication of Insider Trading”5 (the “Insider Trading Paper”) evaluates SEBI’s orders on insider trading matters over a 15-year period and the performance of those orders in appeal before the Securities Appellate Tribunal. The paper develops an evaluation framework based on elements of the rule of law applicable to regulatory adjudication. It finds that SEBI orders often fall short of standards laid down in its own laws. The paper also indicates several areas in which SEBI’s orders do not provide clear reasoning or justification, particularly as regards the choice and quantum of sanction imposed.

The dataset for the Insider Trading Paper included orders issued until July 2023. Nearly three years later, does the Religare Order improve on the areas of weakness, or is it just one step up and two steps back?

One step up…

The Insider Trading Paper finds that, on average, it takes almost six years, from the end of the UPSI Period, for SEBI to complete its enforcement proceedings.

The table below compares the overall adjudicatory timelines for the orders in the Insider Trading Paper with the time that SEBI took in the case of the Order:

Across all three stages, the Order demonstrates substantially shorter timelines than the averages reported in the Insider Trading Paper.

What may be most relevant in the case of the Order, however, is the duration between the date that SEBI received the allegations against Dr Saluja and the date it passed the Order; information of this nature was not available to us in the case of the Insider Trading Paper, but is available in the text of the Order. SEBI received an email from the Burman Group on November 8, 2023. The Order was passed on May 16, 2026. This means that the entire proceedings, from allegation to adjudication, took a mere two-and-a-half years, around the time SEBI typically takes between issuing an SCN and passing a final order.

And two steps back

The Order perpetuates some of the deficiencies observed in the Insider Trading Paper, particularly in relation to citing precedent. The Insider Trading Paper finds that 87% of orders do not cite any previous SEBI order. The Order follows the same pattern, as it does not cite any previous SEBI order. In the absence of engagement with prior decisions, it becomes difficult to determine whether the Order reflects an established position, departs from past practice, or applies standards differently from cases involving similar facts and circumstances. Consistency in adjudicatory practices is of particular concern when decisions of such large import are taken on the basis of circumstantial evidence, and on the standard of ‘preponderance of probabilities’, as in this case.

The Order also reflects continuing weaknesses in the reasoning relating to the choice and quantum of sanctions. The Insider Trading Paper examines the use of the factors set out in Section 15J of the Securities and Exchange Board of India Act, 1992, and finds that no WTM orders identify the loss to investors, one of the factors listed in that Section. While the Order directs disgorgement and imposes a penalty, it does not identify any loss to investors.

Moreover, our evaluation of this Order against the GOW Indicators shows the Order fails to:

• Explain why disgorgement and penalty were preferred over other sanctions, such as debarment;

• Show the method used to calculate the penalty amount;

• Demonstrate proportionality of the penalty amount with the violation.

This last is of particular concern, since the Order finds the loss averted, thereby justifying the call for disgorgement, but states clearly that it is unable to calculate the loss caused to investors. If this is the case, then two questions remain open:

1. Why was a penalty ordered in addition to the disgorgement?

2. On what basis was the penalty amount determined to be INR 40 lakhs? Why not INR 10 lakhs, the stipulated statutory minimum? Why not INR 1 crore, if the violation was egregious and an example had to be set to deter potential wrongdoers?

Conclusion

The Order demonstrates a marked improvement in enforcement timelines – SEBI acted with unusual speed by completing these proceedings in two-and-a-half years from receipt of allegations to issuance of the Order. At the same time, the Order lacks citations of precedent and sufficient reasoning on sanctions, indicating that substantive order quality on these issues has not improved since 2023.

What explains this unusually short timeline? If we are to give the regulator the benefit of the doubt, we might say that this was part of a very public controversy that was closely tracked by many observers, that the informants, the Burman Group, made SEBI’s job easier by providing all the information and evidence required to substantiate their allegations against Dr Saluja, and that therefore, the investigation department had it easy in this case.

But it is precisely because this was such a public matter that the standards of reasoning and justification expected of the regulator should be higher. Recall the series of events that lead up to this order: the alleged violator makes very public allegations against SEBI before the Delhi High Court, saying “it failed to discharge its duties”. The counterparty in the very public, very loud takeover battle royale emails allegations against Dr Saluja to SEBI, who then proceeds post haste and finds Dr Saluja has committed insider trading - taking far less time than it usually does in such instances. Such a situation requires that we shine a very bright light on SEBI’s decision-making process, particularly as regards its choice of sanctions. Unfortunately, the Order doesn’t reflect much of this light back, continuing the opaque practices of the past in not providing clear reasoning on this front.

So yes, the shorter timeline seems like one step up, but the lack of reasoning on sanction choice and quantum, especially under the circumstances that led up to this Order, pushes us two steps back.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her feedback.

CITATION

Natasha Aggarwal and Bhavin Patel, 2026. “SEBI cracks the Religare case in record time — now, about that sanctions reasoning…”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

Khushboo Tiwari, ‘Religare saga: Saluja moves HC against Sebi, seeks to quash Burmans’ offer’ (Business Standard, 11 February 2025) https://www.business-standard.com/companies/start-ups/saluja-moves-delhi-hc-against-sebi-seeks-quashing-of-burmans-open-offer-125021101397_1.html

Krishna Gopalan, ‘Burmans vs Religare Board: The High-Stakes Battle for Control of the Rs 6,300-Crore Rashmi Saluja-Led Firm’ (Business Today, 25 July 2024) https://www.businesstoday.in/magazine/cover-story/story/burmans-vs-religare-board-the-high-stakes-battle-for-control-of-the-rs-6300-crore-rashmi-saluja-led-firm-438708-2024-07-25

Natasha Aggarwal and Bhavin Patel, 2026. “Stitch in time: An early evaluation of the IFSCA’s adjudicatory orders”, The Bridge, TrustBridge Rule of Law Foundation.

Natasha Aggarwal and others, 2025, “A Guide to Writing Good Regulatory Orders,” Working Paper, Trustbridge Rule of Law Foundation.

Natasha Aggarwal and others, 2025. “Balancing Power and Accountability: An Evaluation of SEBI’s Adjudication of Insider Trading,” Working Papers 13, Trustbridge Rule of Law Foundation.

SEBI final order in the matter of suspected insider trading by the Noticee in the scrip of Religare Enterprises Ltd. (13 May 2026) https://www.sebi.gov.in/enforcement/orders/may-2026/final-order-in-the-matter-of-suspected-insider-trading-by-certain-entities-in-the-scrip-of-religare-enterprises-ltd-_101405.html

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

What could Grasim Industries Limited (NSE: GRASIM) (“Grasim”) do with Rs. 30.16 crores over 6.14 years?

At a risk-free rate of about 7%, this money would have amounted to about Rs. 45 crore over that duration. Even a boring old fixed deposit with the State Bank of India would have yielded Rs. 13.3 crores in interest. These calculations do not even take into account the legal fees and associated costs of prosecuting a high-stakes regulatory dispute, and the time of the personnel involved. This is time and money that can’t be used for productive purposes.

That’s the price we pay for a poorly drafted regulatory order.

Let that sink in.

Why do we ask such a question? And why do we say this cost is attributable to poor order writing?

Because on May 5, 2026, the National Company Law Appellate Tribunal (the “NCLAT”) set aside an order of the Competition Commission of India (the “CCI”) dated March 16, 2020, in which the CCI had imposed a penalty of Rs. 301.61 crores on Grasim. The NCLAT also directed the CCI to consider the matter afresh.1 According to Grasim’s Annual Report,2 the company had obtained a stay on the operation of the CCI’s order by depositing Rs. 30.16 crores with the NCLAT. This means that for the 6.14 years between the original CCI order and the NCLAT order that remanded it, Grasim would effectively have been unable to productively deploy this sum of Rs. 30.16 crores. Instead of using it to produce more goods, or hire more workers, this amount was locked up so that the penalty might be paid on short notice, if required. Note that the matter has not reached a conclusion yet - the CCI will reconsider the matter, it may well impose a penalty again, and a fresh appeal against that order (and a second appeal, and several interim applications) might be waiting in the wings. But for the moment, let’s think of this as a ‘one order, one appeal’ story.

Could this unproductive, undesirable outcome have been avoided? We apply a subset of ‘Good Order Writing’ indicators developed by TrustBridge to examine whether the initial CCI order might have been better written, and how such problems can be avoided. This subset of indicators forms part of a set of 68 indicators to measure the completeness and quality of regulators’ adjudicatory orders based on our paper.3 These indicators are classified into four overlapping categories: structural, informational, substantive and stylistic, and are assigned weighted scores on a scale of 0 to 1.

We first provide a summary overview of the CCI and NCLAT orders, and then demonstrate how the subset of indicators could have helped avoid this situation. Finally, we examine whether the adoption of such indicators, as well as other tools and practices to assist good order writing, could lead to positive results.

Summary overview of orders

The CCI order4 resulted from an information filing against Grasim and three other entities, alleging that Grasim had abused its dominant position as the country’s largest producer and seller of Viscose Staple Fibre, a violation punishable under S. 4 of the Competition Act, 2002 (the “Act”).

The date when the information was filed before the CCI is not available in the CCI’s order. However, the CCI passed an order on November 10, 2016 under S. 26(1) of the Act, directing the Director General (“DG”) to investigate the matter. The DG submitted its investigation report on March 27, 2018. This means that while the gap between the CCI’s and the NCLAT’s order is over six years, the matter as a whole has been in controversy for about a decade.

The CCI order includes a detailed description of the DG’s investigation, and the findings included in its investigation report. It notes that the DG’s report found violations of Ss. 4(2)(a)(ii) and 4(2)(d) read with S. 4(1) of the Act, but did not find any information or material in its investigation to indicate that a case was made out under S. 4(2)(e) of the Act against Grasim. Upon a reading of the NCLAT’s order, however, it would appear that the CCI order does not provide a comprehensive, accurate overview of the DG’s findings.

The CCI considered the DG’s report, and forwarded a non-confidential version to the Informant and Grasim so that they could file their suggestions or objections. Grasim’s objections to the DG’s report are described in detail in the CCI order. However, there is no information in the CCI order that suggests Grasim was provided the opportunity for a hearing before the CCI took its decision.

The CCI order responds to Grasim’s submissions, and, on the face of matters, vigorously defends the DG’s report. It eventually agrees with the DG’s findings that Grasim violated Ss. 4(2)(a)(ii) and 4(2)(d) read with S. 4(1) of the Act, and imposes a penalty of Rs. 301.61 crores against Grasim.

In its appeal against the CCI’s order, Grasim argued that the CCI had deviated from the findings in the DG’s report, but did not provide Grasim a notice declaring that it proposed to do so, nor did it provide Grasim the opportunity for a fair hearing in which it might present its case. Grasim sought to have the CCI order set aside on this ground of failure to adhere to the principles of natural justice.

The NCLAT considered the effect of two prior decisions of the erstwhile Competition Appellate Tribunal (the “COMPAT”),5 and accepted Grasim’s contention that the CCI could not have passed an order in which it deviated from the DG’s report, without offering Grasim the opportunity to present its case at a fair hearing. It set aside the CCI’s order, and directed it to consider the matter afresh, with the direction that it provide Grasim an opportunity for a hearing “wherever the Commission differs with the findings of the DG”.

Applying ‘Good Order Writing’ indicators

Could this reversal, and the lost opportunity costs of the penalty imposition have been avoided? Our analysis of the order against our benchmarks suggests that there were multiple missing pieces. So yes, these problems could have been avoided had the CCI considered the following indicators at the time of drafting its order:

1. Is there a list of dates of hearings or a link to this and other procedural information?

2. Does the order summarise parties’ oral arguments at hearings?

3. Are the results of the investigation summarised?

4. Is the analysis well-reasoned?

5. Does the order cite previous COMPAT, NCLAT, or SC orders?

We expect that the resultant order would have been more complete and more likely to withstand challenge in appeal. The benefits of these indicators are described below.

Conclusion and recommendations

The use of order-writing-assistance methods, such as TrustBridge’s Good Order Writing indicators, could have helped avoid the problem created by the CCI’s order in the Grasim matter, and its remand by the NCLAT. These indicators can be used as part of an ‘Order Review Tool’ (see video below), or even as a simple checklist. We suggest that regulators consider how they could address the demands of order writing as carefully as they consider the requirements of the law applicable to their sector.

Another way to avoid controversies like this would be the development and use of a ‘precedent bank’ that order writers may refer to. Such a precedent bank would have brought the applicable COMPAT orders to the adjudicator’s attention. The CCI may then have conducted proceedings in a different manner, by issuing Grasim a notice and providing them an opportunity for a hearing. Alternatively, the order could have explained why those decisions were not applicable in the present situation, giving it a stronger chance of withstanding challenge in appeal.

We point out the opportunity costs of poor order writing in one instance; but consider what the aggregate value of such practices across regulators would be over the period of a single year, over a few years, or over a decade. And how little in comparison it would take to avoid such costs - LLMs and Generative AI make solutions like the Order Review Tool accessible, and regulators and researchers have the domain expertise to make such tools capable and effective. But first we need to acknowledge the scale of the problem, and open our minds to the possibilities that imaginative solutions can offer.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her suggestions and guidance, and Upasa Borah for her assistance in reviewing the calculations of opportunity costs.

CITATION

Natasha Aggarwal and Bhavin Patel, 2026. “How much does a poorly-written order cost?”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

References

1. Natasha Aggarwal, Bhavin Patel & Karan Singh, 2025. “A guide to writing good regulatory orders,” Working Papers, Trustbridge Rule of Law Foundation.

2. Board of Control for Cricket in India v. Competition Commission of India, 2015 SCC OnLine Comp. AT 238.

3. Grasim Industries Limited, Annual Report, available at: https://www.grasim.com/Upload/PDF/grasim-integrated-annual-report-2024-25.pdf

4. In Re: XYZ and Association of Man Made Fibre Industry of India and others (Case No. 62 of 2016) (CCI Order dated 16 March 2020).

5. Interglobe Aviation Ltd. (IndiGoAirlines) v. Competition Commission of India, Appeal No.07/2016.

Welcome to The Bridge! Regulatory orders shape markets, govern disputes, and test the rule of law, yet their quality rarely gets the scrutiny it deserves. The Bridge is TrustBridge’s window into the quasi-judicial work of India’s regulators: how they decide, and how they could decide better.

Introduction

The International Financial Services Centres Authority (the “IFSCA”) is the statutory regulatory authority (“SRA”) in charge of developing and regulating financial products, services and institutions in India’s Gujarat International Finance Tec-City (“GIFT City”). The IFSCA is a unified regulator, in that the powers of the Reserve Bank of India (“RBI”), the Securities and Exchange Board of India (“SEBI”), the Insurance Regulatory Development Authority of India (“IRDAI”), and the Pension Fund Regulatory Development Authority of India (“PFRDA”) with respect to financial institutions, services, and products in IFSCs are all vested in the IFSCA. This suggests that the IFSCA has a tougher role to play than sector-specific regulators, since it has to evolve and execute a quasi-judicial function that combines the needs of the various complex sectors under its control, imposing greater demands on the rigour of its order-writing processes.

As one of India’s newer SRAs, the IFSCA has a short history of issuing adjudicatory orders. Their website currently lists eight documents in the Enforcement section (seven orders, the earliest dated 29 July 2024 and the latest, 6 April 2026, and one warning letter). The IFSCA made headlines recently when, in a first-of-its-kind action for the new SRA, it cancelled the broker-dealer registration of String AI IFSC Private Limited.1 This is seen as a significant move on the regulator’s part, potentially signalling an escalation in its enforcement actions, and moving away from the ‘marketing mode’ stance that market participants suggest it had hitherto adopted. An examination of the regulator’s enforcement practices would therefore be timely.

Data and methods

TrustBridge has developed rigorous methods to evaluate the effectiveness and quality of SRAs’ quasi-judicial process through an examination of their regulatory orders.2 This may also be the only external way to examine these functions, since these are the only documents related to the adjudication process published by Indian SRAs - they typically do not make their enforcement manuals or conventions of procedure (where these exist) publicly available.3 We apply these methods to the IFSCA’s adjudicatory orders, with a view to pointing out early opportunities for course correction, and with the hope that this will prevent incipient problems from cascading into larger systemic issues.

We have developed a set of 68 indicators to measure the completeness and quality of regulators’ adjudicatory orders based on our paper on ‘good order writing’.4 These are classified into four overlapping categories: structural, informational, substantive and stylistic, and are assigned weighted scores on a scale of 0 to 1. Examples of this classification and weighted scores are:

We collected the seven adjudicatory orders available on the IFSCA’s website, and processed them using Anthropic’s Claude Sonnet 4.6 Large Language Model. Claude was prompted to review each order and record its responses, along with the reasoning for its responses. The prompts are available in Annexure 1. The results (available here) were then verified by a human reviewer with expertise in reviewing Indian SRAs’ regulatory orders.

Results

Notably, orders issued in 2024 and 2025 were authored by a single officer, while all orders in 2026 were issued by three-member benches of the Quasi-Judicial Authority for Enforcement (QJAE).

The average score for all seven orders across the years is 74.65%. By way of context, in a previous study, we created a sample set of orders issued in 2024 comprising:

• 16 orders (5% sample) by SEBI, for which the average weighted score was 65.1%;

• 5 of 17 orders by IRDAI for which the average weighted score was 58%; and

• A 2019 order by PFRDA (no orders were issued in 2024), for which the weighted score was 77.3%

At an aggregate level, all seven orders satisfy 39 of our 68 indicators, suggesting a consistent baseline of compliance with core elements of good order-writing. The average over the years is represented below. The figure shows that the score in 2024 for one order is 76.4%; the average score for two orders in 2025 is 79.2%, and the average score for four orders in 2026 has reduced to 71.8%.

The score for each order is represented below. Order 1, issued in 2024, scored 76.4% on our indicators. Orders 2 and 3, both issued in 2025, scored 76.1% and 82.2%, respectively. This shows that the performance of orders peaked in May 2025, followed by a gradual decline in 2026. Orders 4, 5, 6 and 7 were all issued in 2026, and the scores range from 69.5% to 75.6%. The lower scores in 2026 are primarily because those orders do not clearly state the issues involved or summarise the uncontested material facts or the disputed material facts.

What do the IFSCA orders do well?

The IFSCA’s orders perform well on structural organisation and prefatory information. They are generally well-structured, identify applicable legal provisions, and set out procedural history in a clear and traceable manner. They also perform well in setting out parties’ written submissions and engaging with the applicable statutory framework. Each order records the salient arguments and provides a reasoned analysis.

What could be better?

Certain indicators are not satisfied by any order. These include:

1. A concise statement of the main question,

2. Names of counsel (or a statement regarding absence of representation),

3. A summary of disputed material facts,

4. A description of remedies available to affected parties (such as an appeal), and

5. Publication in a machine-readable format.

While most orders summarise procedural history adequately, three do not record or summarise parties’ oral arguments at the hearing. Moreover, there are recurring limitations in the treatment of facts: orders typically do not provide a clear summary of uncontested material facts and, as mentioned above, none summarise disputed material facts.

Engagement with Supreme Court precedent is minimal - only one order imposed a penalty and therefore cited the Supreme Court’s decision in Adjudicating Officer, SEBI v Bhavesh Pabari (Civil appeal 11311/2013). This may be because the remaining orders did not involve issues that require precedential application, and a more effective evaluation of the IFSCA’s performance on this indicator requires a larger sample of orders.

Performance is weakest in relation to justifications for sanction determination and quantification. Only one order explains the harm caused by the violation, and none explain the method used to determine the quantum or duration of the sanction. Similarly, only one order sets out the factors considered in determining the sanction, and only one addresses whether the imposed sanction is proportional to the violation.

Recommendations

Across all seven orders, there is consistency in structural clarity and procedural completeness. Orders are generally well-organised, identify the relevant statutory framework, and engage with parties’ submissions. The analysis sections meet a basic threshold of reasoning. This is non-trivial, particularly for a relatively new, unified regulator operating across complex sectors. Many established regulators fail to maintain comparable consistency.

The fact that the IFSCA satisfies a substantial proportion of the indicators across all orders suggests that a foundational template for good order-writing has taken hold. However, the IFSCA’s orders demonstrate the need for improvement in some aspects:

Conclusion

Overall, the IFSCA’s adjudicatory orders suggest a regulator that has successfully established the architecture of good decision-making. The trajectory is, in many respects, a positive one: the core elements are in place, and the gaps that remain are identifiable. The fact that these issues are visible at an early stage presents a valuable opportunity to embed stronger practices before such gaps in order-writing practices are institutionalised.

The fact that the IFSCA has only passed seven orders thus far, and that matters are at an early stage, also allows the opportunity of creating a well-indexed, searchable, and accessible database of orders that can be used as a precedent bank in later decision-making. This will help ensure consistency across orders and over time, and help the regulator present a consistent, coherent narrative in its orders that are better able to withstand challenge in appeal.

- The authors are researchers at TrustBridge, and would like to thank Renuka Sane for her suggestions and guidance.

CITATION

Natasha Aggarwal and Bhavin Patel, 2026. “Stitch in time: An early evaluation of the IFSCA’s adjudicatory orders”, The Bridge, TrustBridge Rule of Law Foundation

Thanks for reading The Bridge! Subscribe for free to receive new posts and support our work.

Website | LinkedIn | YouTube

Subscribe now

Annexure 1

The following prompts were used in an Anthropic Claude Project. The Project was configured with a system prompt (reproduced below) that provided Claude with the evaluation framework and instructions. Individual orders were then uploaded and assessed one at a time using the user prompts below.

System prompt

You have been provided with a spreadsheet titled ‘20260316_GOW’ with the following columns: Column B contains a list of indicators, column C has notes corresponding to each indicator, column E has the possible responses corresponding to each indicator, and column F has the weights corresponding to each indicator. You have also been provided with a document titled ‘20260316_GOW_prompts,’ which contains instructions for some of the indicators listed in the spreadsheet and, in some cases, sample orders. The sample orders have been provided to you as PDFs.

I will upload orders to this project, and ask you to evaluate them against the good order writing metrics. Please return the responses in a spreadsheet format, with the reasons for each response recorded in a separate column.

User prompts

Prompt 1: Now, I will provide you with an order by the International Financial Services Centres Authority and you will assess this order against the good order writing metrics. Got it?

Prompt 2: Now evaluate this order by IFSCA.

Prompt 2 was repeated for each order.

References

Natasha Aggarwal and others, 2025, “Balancing Power and Accountability: An Evaluation of SEBI’s Adjudication of Insider Trading,” Working Paper, Trustbridge Rule of Law Foundation.

Natasha Aggarwal and others, 2025, “A Guide to Writing Good Regulatory Orders,” Working Paper, Trustbridge Rule of Law Foundation.

Natasha Aggarwal and others, 2025, “Can Technology Augment Order Writing Capacity at Regulators?,” Working Paper, Trustbridge Rule of Law Foundation.

Chitrakshi Jain and others, 2025, “Examining the Performance of ERCs at APTEL,” The Leap Blog.

Read on to see what’s kept us busy.

Policy inputs and submissions

Inputs on Securities Markets Code Bill

Natasha Aggarwal, Pratik Datta, K.P. Krishnan, Bhavin Patel, M.S. Sahoo, Renuka Sane, Ajay Shah, and Bhargavi Zaveri-Shah submitted detailed comments to the Parliamentary Standing Committee on Finance on the Securities Market Code Bill, 2025. The submission welcomes the Bill's advances on governance and regulation-making, while noting areas that merit further attention, particularly around the concentration of powers at SEBI. Key recommendations include:

• a strong structural separation of investigative and adjudicatory functions,

• the removal of criminal liability in favour of punitive civil penalties,

• codified methodologies for calculating unlawful gains,

• and the deletion of provisions that allow SEBI to extend investigation timelines indefinitely or make regulations governing its own adjudication proceedings.

Inputs on liability framework for digital transactions

Renuka Sane, Pratik Datta, Amol Kulkarni, and Aastha Asthana submitted comments to the Reserve Bank of India on its proposed revision to the customer liability framework for digital transactions. The submission examines the adequacy of the draft directions in balancing consumer protection with regulatory clarity for banks and payment service providers.

Trainings

On March 26, Prashant Narang delivered a talk titled 'Professional Regulation in India: Rethinking Reform Beyond the State vs Self-Regulation Debate' at the Indian School of Public Policy (ISPP). He examined the challenges of regulating professions such as law, medicine, and accountancy, particularly in contexts where quality is difficult to assess and enforcement capacity is uneven. He made the case for a reform agenda centred on transparency, user-facing enforcement, and stronger internal checks—moving beyond the traditional binary of state versus self-regulation.

On April 6, Salonie Chawla delivered a talk at Sri Guru Gobind Singh College of Commerce, University of Delhi, on corporate career pathways for graduates of Public Policy and Economics in advocacy and government engagement. She offered students a candid look at how applications are assessed in practice by consulting firms and the private sector, and why authenticity matters in the recruitment process. She also spoke to how core skills developed in academic training translate fluidly across corporate, research, and policy roles.

On April 11, Prashant Narang delivered a session for high school students at iCitizen (Centre for Civil Society, in collaboration with Athena Education), introducing key concepts including the rule of law versus rule of men, the generality principle, and discretion. He drew on pop culture references and landmark cases to illustrate how legal rules shape the relationship between citizens and the state.

Working papers

Prashant Narang and Renuka Sane published their working paper titled ‘Clock versus consequence: How Delhi High Court operationalises Section 29A’s time discipline’. The paper examines how Section 29A of the Arbitration and Conciliation Act which imposes strict timelines on arbitral proceedings works in practice.

This paper studies how Section 29A of the Arbitration and Conciliation Act, 1996 works in practice. The provision sets a time limit for making arbitral awards and allows courts to extend time after the limit expires. Using a dataset of 202 reported orders of the Delhi High Court between 2015 and 2024, the paper examines how extension petitions are decided and whether courts use their powers to impose sanctions for delay. The findings show a clear pattern. Courts grant extensions in almost all cases, often within a few days and usually in a single hearing. By contrast, sanctions are rare. There are no instances of fee reduction, and only a small number of cases involve costs or replacement of arbitrators. Even repeat extension requests are routinely allowed without penalties. The paper argues that this pattern reflects the structure of Section 29A. Granting an extension is simple, especially when parties agree. Imposing sanctions requires courts to identify responsibility for delay and follow procedural safeguards, which takes more time and effort. As a result, the provision operates mainly as a mechanism to regularise delay rather than deter it. The study suggests that deadlines are effective only when consequences are predictable and easy to apply. Where sanctions depend on detailed judicial findings, they are used sparingly. This raises questions about whether court-based enforcement is the right tool for ensuring timely arbitration.

Prashant Narang published a paper titled ‘Judicial Abdication in a Rights-Based Constitution: Scrutinising the Indian Supreme Court’s Approach to Reasonable Restrictions on Trade and Business’, in the Indian Journal of Constitutional Law, Volume XIII. The paper analyses 187 Supreme Court cases from 1950 to 2015 to examine how judicial scrutiny of restrictions on trade and business has evolved over time. Below is the abstract:

Despite the Indian Constitution’s rights-based architecture, the Supreme Court has long declined, in a subset of cases, to rigorously test legislative constraints on economic liberty. This article recovers that under-examined doctrinal history and shows that what recent commentary decries as new-found ‘judicial abdication’ is in fact a recurring mode of review rooted in Article 19(1)(g) jurisprudence. Blending close doctrinal analysis with a descriptive survey of 187 reported Supreme Court decisions between 1950 and 2015 that substantively engage Article 19(1)(g), it demonstrates that the Court expressly avoided applying the governing proportionality or reasonableness framework in approximately 9.6 per cent of cases. Crucially, clearly identifiable instances of abdication increase from about five per cent before Garg (1981) to roughly thirteen per cent thereafter, with around eight per cent of post-Garg Article 19(1)(g) decisions explicitly relying on Garg’s ‘greater latitude’ formulation. The article first reconstructs the genealogy of abdication, moving from early instances in Khyerbari Tea and Pyarali Tejani through Srinivasa Enterprises to Bhagwati J’s majority in Garg, which transposed a U.S. dissent into Indian economic rights adjudication. It then traces how later benches—Peerless, Dalmia Cement, SIEL and others—invoked this rationale to justify reduced scrutiny.

Upasa Borah, Akshay Jaitly and Renuka Sane published titled ‘Evaluating India’s Energy Ambitions: Evidence from Electricity Generation Project-Level Data’. The authors have used project-level data from the CMIE CapEx database to assess whether India is on track to meet the capacity targets set by the Central Electricity Authority for 2030.

India’s electricity demand has been growing rapidly, at 9% per annum since 2021. Meeting this demand by 2030 would require around 777 GW of installed capacity, as estimated by the Central Electricity Authority (CEA). At the same time, India has committed to achieving 500 GW of installed non-fossil capacity by 2030. A study by CEEW (2025) finds that meeting this target would require adding around 56 GW of non-fossil capacity every year between 2025 and 2030, failing which India would need an additional 10 GW of coal-based capacity to meet future demand. There is little doubt that renewable energy in India has seen a sharp growth, with 74 GW in 2018 to 162 GW by the end of 2024 (excluding large hydro and nuclear projects), driven by falling renewable energy prices, and policy support like subsidies for developers, waivers on inter-state transmission charges, Green Energy Corridor investments, changes in Green Open Access Rules and various state-level initiatives that signal policy commitment to the sector. In 2025 alone, the country added 45 GW of renewable capacity.

Op-eds and Articles

Articles in Puliyabaazi magazine

Prashant Narang authored an article in Hindi for the March edition of the magazine, Puliyabaazi.in, ‘Dharmo Rakshati Rakshitah’, reflecting on the idea of the rule of law through Indian political thought, literature, economics, and cinema. He highlights that good governance rests on principles such as transparency, generality, predictability, and restraint in the use of power, while cautioning that excessive or arbitrary regulation may lead citizens toward evasion rather than trust in the law.

The policy lesson is simple: over-regulation makes the honest feel like a criminal and creates a playground for the dishonest. Therefore, the goal should be that the law is light, clear and reason-based, so that people follow it out of respect, not fear, and in proportion to their capacity.

Akshay Jaitly authored an article in Hindi for the March edition of the magazine Puliyabaazi.in ‘The Compliance Trap: Why Business Freedom is Fading‘, examining how India’s economic freedom has long been constrained by a labyrinth of regulations, excessive compliances, and arbitrary government overreach. The piece makes the case for a fundamental rethink of the compliance architecture that governs Indian businesses.

The Indian Republic has now turned 75 years old. From the adoption of the Constitution in 1950 until today, we have covered a long distance. However, India has still not been able to achieve its economic potential to a large extent. Even today, the hands of India’s entrepreneurs and businessmen are shackled by those same old ideologies and systems that were created after independence to keep a check on the economy.

Opinion pieces in The Print

On March 31, 2026, Renuka Sane authored an opinion piece in The Print, 'Why shielding consumers from rising fuel prices can backfire', examining the political economy of fuel price subsidies. The piece suggests that insulating consumers from rising prices distorts incentives, suppresses conservation, burdens public finances, and ultimately defers rather than resolves the adjustment.

If international prices have risen, and retail prices are not reflecting that increase, who is bearing the brunt? The major cost is being borne by India’s state-owned oil marketing companies, with their marketing margins turning negative owing to the rise in international prices. Reports suggest that IOC, BPCL, and HPCL have lost about $2.25 billion (about Rs 20,000 crore) due to the fuel price freeze. These losses will accumulate on balance sheets, raise borrowing costs, and ultimately circle back to the government as contingent liabilities. The government recently slashed the Special Additional Excise Duty (SAED) on petrol and diesel by Rs 10 per litre each, with the intention of keeping the retail price in check. This gives some relief to the oil companies, but takes away the government’s spending kitty, which will probably have to be made up by increasing other taxes.

On April 14, 2026, Renuka Sane and Pratik Datta authored an opinion piece in The Print, 'RBI is going out of its way to compensate fraud victims. It doesn't have the mandate', examining the RBI's evolving approach to compensating victims of digital payment fraud. While sympathetic to the intent, the piece suggests that the RBI risks overstepping its regulatory mandate by expanding liability without a clear statutory basis. The authors recommend that the RBI may instead consider evaluating the structural reasons why banks do not invest adequately in fraud protection.

The deeper question that the RBI must confront is whether this is really the mandate of a central bank? The RBI’s mandate is inflation targeting and financial stability, not consumer insurance. When the RBI steps in as a direct compensation source, it conflates two entirely separate functions: systemic oversight and retail-loss socialisation. These should be kept apart.

The RBI’s own sunset clause acknowledges the awkwardness: it states that compensation will be reviewed after one year to ‘reduce or eliminate’ the RBI’s share. If the design intention is ultimately to remove the RBI from this equation, why put it there in the first place?

Collaborate

If our work resonates with you, we would love to collaborate. We focus on financial markets regulation, energy transition, and contracts and dispute resolution.

Reach out at info@trustbridge.in.

Thanks for reading The TrustBridge Newsletter! Subscribe for free to receive new posts and support our work.

                                                Website  |  LinkedIn  |  YouTube