Sisyphus In The Arbitral Forum - Post-Award Judicial Review And The Limits Of Finality In Indian Arbitration

Author: LawSphereX

Category: Arbitration & ADR

Published: 2026-08-23T21:05:51.05503+00:00

Updated: 2026-08-30T07:08:27.411Z

SISYPHUS IN THE ARBITRAL FORUM: POST-AWARD

JUDICIAL REVIEW AND THE LIMITS OF FINALITY IN INDIAN ARBITRATION Avi Jindal, O.P Jindal Global University I. Introduction: Sisyphus in the Arbitral Forum An arbitral award is supposed to end a dispute. In India, that expectation has been broken with troubling regularity. The award moves to a Section 34 court for setting aside, then to a Section

37 appeal, then to the Supreme Court by special leave, and then potentially through review.

The Delhi Airport Metro case is the clearest proof of what this looks like in practice: after Section 34, Section 37, special leave, and review, the Supreme Court entertained a curative petition and reversed a judgment that had already restored the award. Five stages of judicial proceedings over a single arbitral award. 1 At that point, calling arbitration an alternative to litigation is aspirational rather than descriptive. Understanding why requires going back to what arbitration is, and what it is not. Arbitration derives its authority not from the State but from the agreement of parties. Yet that private authority runs out the moment the award is made. The tribunal cannot enforce its own decision. Everything that follows depends entirely on national courts and the legal system that parties were, in some sense, trying to avoid.

Redfern and Hunter describe this as a paradox: the institution exists to offer an alternative to litigation but cannot function without it. 2 Arbitration does not promise the correct answer. It promises a final one. Gary Born identifies finality and limited review as structural commitments of the system, not design preferences each jurisdiction may adjust. 3 The New York Convention and the UNCITRAL Model Law give that commitment legal form:4 grounds for setting aside and refusing enforcement are 1 The litigation proceeded through five stages of judicial proceedings over more than a decade.

For the principal Supreme Court determinations see Delhi Metro Rail Corporation Ltd v Delhi Airport Metro Express Pvt Ltd (2017) 1 SCC 52 (SC India) (setting aside the arbitral award on special leave); Delhi Airport Metro Express Pvt Ltd v Delhi Metro Rail Corporation Ltd (Curative Petition) 2022 SCC OnLine SC 275 (SC India) (reversing the

2017 judgment and restoring the award).

2 Nigel Blackaby and others, Redfern and Hunter on International Arbitration (6th edn, OUP 2015) para 1. 01. 3 Gary B Born, International Commercial Arbitration (3rd edn, Kluwer Law International 2021) 3267–3270. 4 Convention on the Recognition and Enforcement of Foreign Arbitral Awards (adopted 10 June 1958, entered narrow, exhaustive, and deliberately resistant to expansion. Courts retain a residual role because awards acquire legal force through State recognition, and that recognition carries authority to police the boundaries of the process, not to correct its outcomes. The line between those two functions is the line India has struggled to hold.

This essay argues that India needs calibrated judicial minimalism: courts that intervene for jurisdictional defects, procedural unfairness, natural justice violations, fraud, and public policy in its narrow sense, and that firmly decline to correct ordinary errors of fact, law, or contractual interpretation. The paper examines why that distinction has proved so difficult to maintain, and what it would take to make it sustain. II. The Statutory Promise: Finality Within a Supervised Framework The Arbitration and Conciliation Act 1996 made a clear legislative choice. Enacted with express reference to the UNCITRAL Model Law,5 it was designed to displace the interventionist culture of the Arbitration Act 1940,6 under which courts treated awards as provisional determinations subject to extensive judicial correction. Parliament intended a structural break, not merely a tonal shift.

Section 5 provides that no judicial authority shall intervene in matters governed by Part I except where the Act itself so provides. 7 The language is prohibitory, not permissive. Whatever authority courts exercise must be traced to a specific statutory gateway. That overriding character was deliberate: Section 5 was designed to displace any common law or equitable jurisdiction courts might otherwise claim. That gateway, at the post-award stage, is Section 348. Recourse against an award may be made only by an application for setting aside. The word ‘only’ is exclusive and exhaustive, foreclosing ordinary appeals, revision petitions, and suits.

As the Supreme Court confirmed in Project Director, NHAI v M Hakeem,9 Section 34 confers annulment jurisdiction, not appellate into force 7 June 1959) 330 UNTS 3 (‘New York Convention’); UNCITRAL Model Law on International Commercial Arbitration (adopted 21 June 1985, as amended 7 July 2006) UN Doc A/40/17, annex I (‘UNCITRAL Model Law’). 5 Arbitration and Conciliation Act 1996 (India) (‘the 1996 Act’). The Statement of Objects and Reasons expressly acknowledges that the Act was enacted with reference to the UNCITRAL Model Law.

6 Arbitration Act 1940 (India).

7 1996 Act (n 5), s 5.

8 1996 Act (n 5), s 34.

9 Project Director, National Highways Authority of India v M Hakeem (2021) 9 SCC 1 (SC India).

jurisdiction. An annulment court asks only whether the award crosses the threshold of legal impermissibility. If it does not, the court must leave it alone even if it would have decided the matter differently. That is not a counsel of judicial humility. It is a statutory command. Sections 35, 36, and 37 carry the same logic forward. 10 Section 35 declares awards final and binding. Section 36 provides for enforcement by analogy to a court decree. Section 37 permits appeal from Section 34 orders, but only on the same limited grounds. Section 48 provides the narrow Convention-based gateway for resisting enforcement of foreign awards. 11 Read together, these provisions mirror Article 34 of the UNCITRAL Model Law and Article V of the New York Convention. 12 As Lew, Mistelis and Kroll confirm, this architecture of constrained post-award review is the accepted global standard.

13 The problem was never the architecture. It was that courts found ways to operate within its language while systematically undermining its purpose. III. Public Policy and Patent Illegality: From Safety Valve to Appellate Temptation No concept has done more damage to Indian arbitration while remaining indispensable than public policy. The Supreme Court began correctly in Renusagar Power Co Ltd v General Electric Co,14 confining public policy in the foreign award enforcement context to the fundamental policy of Indian law, the interests of India, and justice or morality at the most basic level. Public policy was originally a safety valve to prevent the most egregious outcomes, not a general quality-control mechanism. The turning point was ONGC Ltd v Saw Pipes Ltd,15 which expanded public policy in the domestic context to encompass patent illegality.

The Court’s instinct that some mechanism should exist for awards plainly contrary to statute was not unreasonable. But inserting patent illegality within public policy, rather than creating a separate and bounded review mechanism, opened Section 34 to challenges having nothing to do with genuine public interest. Courts

10 1996 Act (n 5), ss 35, 36, 37.

11 1996 Act (n 5), s 48.

12 UNCITRAL Model Law (n 4) art 34; New York Convention (n 4) art V.

13 Julian DM Lew, Loukas A Mistelis and Stefan M Kröll, Comparative International Commercial Arbitration (Kluwer Law International 2003) paras 25-1–25-15. 14 Renusagar Power Co Ltd v General Electric Co AIR 1994 SC 860 (SC India). The Court held that public policy in the context of enforcement of foreign awards under the Foreign Awards (Recognition and Enforcement) Act 1961 (India) was confined to the fundamental policy of Indian law, the interests of India, and basic standards of justice or morality. 15Oil and Natural Gas Corporation Ltd v Saw Pipes Ltd (2003) 5 SCC 705 (SC India). began examining whether the tribunal had correctly interpreted the contract, properly applied the governing law, or reached conclusions the evidence could support. The safety valve was functioning as an appeal.

The erosion accelerated in ONGC v Western Geco International Ltd,16 where the Court read ‘fundamental policy of Indian law’ to include a judicial approach, compliance with natural justice, and conformity with Wednesbury reasonableness. In its public law context, the Wednesbury standard asks whether a decision is so unreasonable that no reasonable authority could have reached it. Applied to arbitral tribunals exercising contractual jurisdiction, it invited courts to examine the quality of reasoning and the proportionality of conclusions. The distinction between a perverse award and a wrong one became practically meaningless. Associated Builders v Delhi Development Authority17 attempted to systematise these grounds, but organising an expanded doctrine is not the same as constraining it.

When broad grounds of review are available, challenge ceases to be a genuine legal remedy and becomes a strategic tool: a way to buy time, impose costs on the award-creditor, and preserve the possibility of a different result in court. Broad review does not merely affect outcomes in individual cases. It changes the entire dynamic of the post-award period, and that is precisely the institutional consequence the 2015 amendments were designed to address. IV. The 2015 Recalibration: Doctrine, Enforcement, and the Limits of Statutory Reform The 2015 amendments were a direct legislative response to Saw Pipes and Western Geco. The Law Commission’s 246th Report18 documented excessive court intervention, public policy uncertainty, adjournment culture, and India’s inhospitability as a seat for high-value international disputes.

The Supplementary Report19 recommended reversing Saw Pipes and restoring the Model Law philosophy. The amended Section 34 narrowed the public policy ground substantially. 20 Explanation 2 provides that the test of contravention with the fundamental policy of Indian law shall not entail 16 Oil and Natural Gas Corporation Ltd v Western Geco International Ltd (2014) 9 SCC 263 (SC India). The Wednesbury standard of unreasonableness derives from Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1948] 1 KB 223 (CA).

17 Associated Builders v Delhi Development Authority (2015) 3 SCC 49 (SC India).

18 Law Commission of India, Review of the Arbitration and Conciliation Act, 1996 (Law Com No 246, 2014). 19 Law Commission of India, Amendments to the Arbitration and Conciliation Act, 1996 (Supplementary to Law Com No 246, 2015).

20 Arbitration and Conciliation (Amendment) Act 2015 (India), amending 1996 Act (n 5), s 34.

a review on the merits of the dispute. 21 This is a mandatory prohibition, not a counsel of restraint. Section 34(2A) introduced patent illegality as a separate domestic review mechanism,22 with a critical proviso: an award shall not be set aside merely on the ground of erroneous application of law or by reappreciation of evidence. The decisive judicial restatement came in Ssangyong Engineering and Construction Co Ltd v National Highways Authority of India,23 which held that the Western Geco interpretation of public policy could not survive the 2015 amendments. Wednesbury reasonableness was no longer a permissible test. Patent illegality had to go to the root of the matter, not appear as ordinary legal or factual error. Ssangyong is the doctrinal centrepiece of calibrated judicial minimalism. A judicial restatement is, however, only as durable as the discipline with which it is applied.

The pre-2015 expansion happened not because courts lacked the right text but because courts found ways to read existing text more broadly than its context required. That interpretive instinct has not disappeared. The 2015 amendments also addressed the enforcement problem created by the automatic stay: filing a Section 34 application had previously rendered every award commercially worthless until the challenge ran its course. The 2021 amendment introduced a complication24 by permitting an unconditional stay where the court is prima facie satisfied that the award was induced by fraud or corruption. The objective is legitimate. The risk is that fraud allegations become the new procedural language of delay if courts accept bare assertions at the prima facie threshold. The unconditional stay must not become a surrogate for the automatic stay the 2015 reform abolished. V.

Annulment Is Not Appeal: Modification and the Gayatri Balasamy Decision Whether courts may modify arbitral awards under Section 34 was, until recently, a question on which the Supreme Court’s own precedents were irreconcilable. Project Director, NHAI v M Hakeem25 and McDermott International Inc v Burn Standard Co Ltd held that Section 34 21 1996 Act (n 5), s 34, Explanation 2 (as inserted by Arbitration and Conciliation (Amendment) Act 2015 (n 20)). 22 1996 Act (n 5), s 34(2A) (as inserted by Arbitration and Conciliation (Amendment) Act 2015 (n 20)). 23 Ssangyong Engineering and Construction Co Ltd v National Highways Authority of India (2019) 15 SCC 131 (SC India).

24 Arbitration and Conciliation (Amendment) Act 2021 (India), amending 1996 Act (n 5), s 36.

25 Project Director, National Highways Authority of India v M Hakeem (n 9); McDermott International Inc v Burn Standard Co Ltd (2006) 11 SCC 181 (SC India). confers no modification power. Vedanta Ltd v Shenzhen Shandong Nuclear Power Construction Co Ltd26 held the opposite. Different courts were applying different rules to the same provision. The Constitution Bench resolved this in Gayatri Balasamy v ISG Novasoft Technologies Ltd,27 decided on 30 April 2025. By a 4:1 majority, CJI Sanjiv Khanna writing for the Court held that a limited modification power exists under Sections 34 and 37.

Its scope is confined to three categories: severing non-arbitrable or offending portions from genuinely severable valid ones; correcting computational, clerical and typographical errors apparent on the face of the record; and modifying post-award interest where compelling reasons justify it and where doing so prevents further litigation. The majority confirmed that Article 14228 may be invoked to modify an award, but only with great caution and never to rewrite it on the merits. Justice Viswanathan dissented on the ground that the Act, read with Section 5’s minimumintervention principle, does not contemplate modification at all. He accepted correction of obvious arithmetic or typographical errors but rejected anything broader. The concern is not formalistic.

The history of Section 34 is a history of bounded exceptions that courts have quietly widened: public policy began narrow in Renusagar, expanded in Saw Pipes, and swallowed merits review entirely by Western Geco. 29 The modification power carries the same structural risk, and the majority’s careful boundaries are a promise, not a guarantee. VI. The Case for Judicial Intervention: A Rebuttal and Its Limits The case for robust judicial supervision deserves to be taken seriously before it is rejected. Indian arbitration operates largely through ad hoc procedures, with limited institutional oversight, inadequate disclosure norms, and no standardised qualification requirements for arbitrators. In that environment, the risk of biased or poorly reasoned awards is real, and Section 34 review, applied conscientiously, is what catches them.

A significant proportion of highvalue Indian arbitrations also involve public sector entities, where judicial review provides a democratically accountable check on the allocation of public funds through a private process.

26 Vedanta Ltd v Shenzhen Shandong Nuclear Power Construction Co Ltd (2019) 11 SCC 465 (SC India).

27 Gayatri Balasamy v ISG Novasoft Technologies Ltd 2025 SCC OnLine SC 997 (SC India) (Constitution

Bench, CJI Sanjiv Khanna; Viswanathan J dissenting).

28 Constitution of India 1950 (India), art 142.

29 See Renusagar (n 14); Saw Pipes (n 15); Western Geco (n 16).

That concession has a limit. The appropriate response to weak arbitral institutions is to strengthen them, not to replace the tribunal with the court. If courts stand ready to correct every award they find legally or commercially unsatisfactory, arbitrators have no real decisional authority and parties have no real finality. Broad judicial supervision does not cure institutional weakness. It entrenches it, by removing the pressure on the arbitration community to reform itself. The Law Commission’s 246th Report made this point directly:30 the cultural revolution required in Indian arbitration must come from within, not from courts standing ready to correct whatever emerges. Judicial scepticism is a symptom of the problem, not a solution to it.

One stakeholder perspective the existing literature addresses inadequately is that of the repeat player: the institutional arbitration user, typically a large commercial enterprise or financial institution, that participates in multiple arbitrations over time. For the repeat player, what matters is the predictability of the system as a whole: whether the rules are stable, whether the grounds for challenge are foreseeable, and whether the costs of participation are proportionate to the value of disputes being resolved. India’s history of expanding and then contracting the grounds for post-award review has imposed particular costs on this group, because it has made the rules genuinely unpredictable over time. Restoring finality is therefore not only a matter of justice to parties in individual cases.

It is a matter of sustaining the confidence of the most commercially significant users of the system. VII. Comparative Lessons: Transplant, Design, and Institutional Credibility The limited-intervention model embedded in the 1996 Act did not emerge from Indian legal tradition. It was transplanted from a framework developed primarily in the context of European commercial practice, shaped by the UNCITRAL Model Law and the New York Convention, both of which reflect assumptions about institutional maturity, arbitrator quality, and commercial sophistication that did not straightforwardly apply to the Indian context. T T Arvind’s analysis of the transplant effect is directly relevant:31 the same statutory text can produce fundamentally different outcomes depending on the institutional and cultural soil into which it is placed. India received the letter of the Model Law.

It did not, at least not immediately, absorb its spirit. England is instructive precisely because it acknowledged this problem structurally. The

30 Law Commission of India (n 18).

31 T T Arvind, ‘The “Transplant Effect” in Harmonization’ (2012) 61 ICLQ 65.

Arbitration Act 1996 distinguishes three separate heads of court intervention: jurisdictional challenge under Section 67, serious irregularity under Section 68, and appeal on a question of law under Section 69, the last subject to restrictions and capable of contractual exclusion. 32 India had no such transparent mechanism for legal-question review. The result was that appellate behaviour was quietly absorbed into the public policy ground, which then expanded without a principled limit. The English approach does not eliminate judicial involvement; it organises it so that each type of intervention operates within defined and visible boundaries. Singapore presents the sharpest contrast with India because the two jurisdictions share the same statutory foundation, both deriving their arbitration law from the UNCITRAL Model Law.

Yet Singapore courts are consistently regarded as among the most arbitration-friendly in the world. The difference cannot be explained by the statute. It is explained by what Singapore did institutionally: it built credible arbitral institutions, developed a specialist judiciary with genuine expertise in commercial arbitration, and established a consistent pattern of decisions that gave international stakeholders the confidence to choose Singapore as a seat. That confidence, once established, became self-reinforcing. India’s challenge is that the pattern of decisions it established ran in the opposite direction, and reversing a pattern is harder than building one. VIII. The Border of Review: Foreign Awards and the Discipline of Section 48 When India refuses to enforce a foreign award on expansive public policy grounds, the damage extends beyond the individual dispute.

Section 48 governs refusal of enforcement of New York Convention awards and mirrors Article V of the Convention. 33 The grounds are exhaustive, must be construed narrowly, and do not permit an enforcement court to reopen the merits. The award may have been wrong. That is not a ground for refusal. The domestic framework retains patent illegality under Section 34(2A). 34 That ground has no counterpart in Section 48. The distinction is deliberate and must be maintained. The risk after Saw Pipes was precisely that patent illegality and Wednesbury unreasonableness, having entered the domestic vocabulary, would contaminate Section 48 proceedings. The post-2015 framework has substantially addressed that risk. Maintaining the separation is a precondition

32 Arbitration Act 1996 (UK), ss 67, 68, 69.

33 1996 Act (n 5), s 48; New York Convention (n 4) art V.

34 1996 Act (n 5), s 34(2A).

of India functioning as a reliable enforcement jurisdiction. The relationship between annulment at the seat and enforcement elsewhere is not always symmetrical: awards set aside at the seat have in some cases been enforced in other jurisdictions on the ground that the annulment itself violated the enforcement court’s public policy. India’s normative position in that transnational framework must be clear. It should apply Article V grounds in their narrowest construction, refuse to import domestic public policy standards into Section 48, and resist treating foreign award enforcement as an occasion for the correctness review that the New York Convention was specifically designed to prevent. IX. Reform, Its Limits, and What Would Actually Work The proposed Section 34A is India’s most structurally significant post-award reform since

2015. The proposal would permit parties to constitute Appellate Arbitral Tribunals within

institutional rules, channelling setting-aside challenges away from courts. A specialist appellate body that is faster, confidential, and more commercially sophisticated than a generalist court would represent genuine improvement. The problem is not the concept. It is the institutional environment into which it would be introduced. An Arbitration Council of India was established in 2019 and has still not been operationalised. If the Appellate Tribunal’s decision remains reviewable before courts under something analogous to Section 34 or Section 37, the sequence becomes original arbitration, appellate tribunal, court challenge, Section 37 appeal, and possible Supreme Court intervention. That is the ladder with an extra rung.

Section 34A is defensible only if it applies Section 34-type supervisory grounds exclusively, operates under enforceable timelines, is staffed by members independent from the original tribunal, and the statute specifies precisely the grounds on which courts may subsequently intervene. The 2024 Bill’s proposed Section 34(1B),35 requiring courts to formulate grounds of challenge precisely before proceeding to their merits, is a meaningful structural constraint that should survive whatever else the Bill does or does not achieve. The deeper conceptual risk remains: introducing a routine appellate stage implicitly concedes that the first award is not final, and once that concession is embedded in the system’s architecture, no subsequent reform can easily retrieve the expectation of finality it displaces. The MCIA’s 2024 Annual Report recorded that 91 per cent of its administered awards were

35 Arbitration and Conciliation (Amendment) Bill 2024 (India), proposed s 34(1B).

finalised within 18 months and none were set aside by courts. 36 This is the most important empirical observation in the current reform landscape. When an institution is genuinely credible, parties choose it, awards are delivered efficiently, and courts leave them alone. The institutional and the judicial move together not because a statute commanded it but because trust was earned. What the statute should instead provide is a calibrated allocation of judicial deference based on institutional conditions. Where an award is issued by an ACI-recognised institution meeting defined benchmarks of appointment transparency, disclosure compliance, case management quality, and reasoned decision-making, courts should apply a presumption of structural regularity and confine Section 34 review to its narrowest grounds.

Where an award emerges from unaccredited ad hoc proceedings, a wider but still bounded review remains available. This is not a two-tier system of substantive rights. It is a calibrated allocation of judicial deference that reflects the institutional conditions under which an award was produced, using those conditions as a proxy for the structural quality checks that Section 34 would otherwise need to perform itself. The proposal creates a direct commercial incentive for parties to choose recognised institutions, because reduced judicial scrutiny has tangible value in terms of enforcement speed and certainty. It gives courts a principled and evidence-based reason for restraint rather than asking them to exercise faith in institutions they have historically distrusted.

And it resolves the sequencing deadlock by making institutional development and judicial restraint move in the same direction at the same time, rather than each waiting for the other to move first. Its vulnerability must be acknowledged honestly: its entire value depends on the ACI functioning as a credible and independent standard-setter, and the ACI has existed for six years without operationalisation. If recognition becomes a bureaucratic formality, the framework reproduces the problem it was designed to solve. Indian arbitration law is not deficient because it lacks the right statutory text. The text is now largely adequate. Ssangyong narrowed public policy. Hakeem confirmed annulment jurisdiction. Gayatri Balasamy bounded modification. The 2015 amendments prohibited merits review by statute37 and yet the system continues to struggle because doctrine operates through

36 Mumbai Centre for International Arbitration, Annual Report 2024 (MCIA 2024).

37 See Ssangyong (n 23); Hakeem (n 9); Gayatri Balasamy (n 27); Arbitration and Conciliation (Amendment) Act 2015 (n 20). institutions, and the institutions through which Indian arbitration operates have not yet earned the level of trust that would make calibrated judicial restraint self-sustaining. In 2025, in ASF Buildtech v Shapoorji Pallonji,38 The Supreme Court criticised the 2024 Bill for failing to provide statutory recognition of the arbitral tribunal's power to implead non-signatories, and directed the judgment to the Ministry of Law and Justice for legislative consideration. The point is not the specific gap. The point is the pattern: each clarification the Court provides exposes a further question that doctrine alone cannot finally resolve, and each such question is redirected to a legislature that has not yet answered the ones before it.

The Sisyphus analogy cannot yet be retired, but the boulder has been misidentified. The problem was never the absence of the right words; the 1996 Act contained them from the start. The problem is that words require institutions to give them effect, and institutions require trust to function as designed. India has spent three decades building the law of arbitration. It has not yet built the culture of it.

38 ASF Buildtech Pvt Ltd v Shapoorji Pallonji and Co Pvt Ltd 2025 INSC 616 (SC India) (Pardiwala and

Mahadevan JJ). The Court directed the judgment be forwarded to the Ministry of Law and Justice for consideration in legislative reform of the Arbitration and Conciliation Act 1996 (India).