INTRODUCTION
Arbitration’s promise of speedy, final, and largely court-free resolution of commercial disputes has been central to India’s effort to attract domestic and international parties to its forums. Yet, that promise has repeatedly come up against a persistent question: when and how should domestic courts intervene in arbitral awards? The recent debate, crystallised by the Supreme Court’s Constitution-Bench decision in Gayatri Balasamy and the concurrent Draft Arbitration & Conciliation (Amendment) Bill, 2024, exposes a deep fault line between preserving arbitral finality and allowing narrowly tailored judicial correction when awards contain clear legal defects. This paper revisits that tension and asks whether doctrinal coherence and legislative drafting can preserve finality while permitting surgical remedies where strictly necessary.
The problem is both doctrinal and practical. Under the 1996 Act, courts traditionally exercised a limited annulment power under Section 34; courts and tribunals diverged, however, on whether they could go beyond annulment to correct or sever portions of an award. Gayatri Balasamy marks a turning point: the majority recognised a constrained power of “partial setting aside” or severance in narrowly defined situations, while a vigorous dissent warned against opening a backdoor to appellate review. At the same time, the Draft Bill seeks to codify mechanisms of partial set-aside, express remittal procedures, and an optional Appellate Arbitral Tribunal that would institutionalise many of the majority’s practical solutions. These developments raise immediate questions about consistency, enforcement, and incentives for post-award litigation.
This article examines those questions. It identifies which draft provisions most directly shift the balance toward judicial correction, analyses the scope and limits of the Gayatri ruling, and maps doctrinal tensions between the Court’s approach and the Bill’s text. Using a comparative doctrinal method grounded in statute, precedent, and policy, the paper proposes concrete drafting choices and judicial tests that could reconcile finality with a narrowly confined corrective power, minimising re-arbitration while preventing courts from assuming an appellate role.
By situating the Supreme Court’s reasoning alongside proposed legislative fixes, the study aims to offer practitioners and policymakers a practical roadmap: how to preserve the twin goals of predictability and justice in post-award review, and how to draft rules that curb tactical litigation without denying relief where an award is plainly tainted.
COMPARATIVE ANALYSIS OF THE ARBITRATION & CONCILIATION ACT, 1996 AND THE DRAFT AMENDMENT BILL, 2024 IN THE CONTEXT OF GAYATRI BALASAMY JUDGMENT”
Setting aside an award: Under the 1996 Act, Section 34 provides the narrow grounds on which a court can set aside an arbitral award (e.g., incapacity, invalid arbitration agreement, lack of notice, tribunal composition, matters not capable of settlement by arbitration, public policy, etc.). The statute does not expressly mention partial set-asides or remittal mechanics.1
Draft Bill (2024): The draft expressly allows an award to be set aside “in whole or in part”, clarifies the grounds and adds explanations on public policy and patent illegality; it also inserts provisos that an award shall not be set aside merely for an erroneous application of law or by re-appreciation of evidence.2
Balasamy implication: This change codifies the majority’s practical holding in Gayatri Balasamy courts (or a statutory appellate-arbitral forum) may excise clearly separable infirm parts of an award without annulling the whole, while still preserving the prohibition on merits re-appraisal.
Partial set-aside / Severance and remittal mechanics: The statute permitted setting aside but lacked an express, structured remittal procedure for the tribunal to decide only the issues set aside or to bind it to earlier findings.3
Draft Bill (2024): The draft adds an express provision: where an award is set aside in part, the Court or appellate arbitral tribunal may direct the original arbitral tribunal to decide, within a fixed time, only the issues set aside using the existing record and the tribunal must be bound by findings that were not set aside.4
Balasamy’s implication: This inserts precisely the “surgical” tool the majority approved a statutory remittal route that reduces needless re-arbitration and preserves tribunal findings that are not tainted by the legal defect.
Appellate Arbitral Tribunal: There is no statutory “appellate arbitral tribunal” in the 1996 Act; Section 34 challenges run to courts.5
Draft Bill (2024): The Bill creates an optional Appellate Arbitral Tribunal that arbitral institutions may provide. Where parties agree to that route, no setting-aside application may be brought to the Court. The appellate tribunal will follow procedures to be framed by the Council.6
Balasamy’s implication: The new mechanism responds to the majority’s caution about courts acting like appellate bodies by creating a specialist, party-driven forum for narrow post-award challenges — a middle ground between pure judicial supervision and unrestricted appeal. It also raises the dissent’s worry that an institutional appellate layer could, in practice, resemble an appeal mechanism if widely used.
Form and contents of award clearer checklist: Section 31 prescribes form and contents (reasons, date, place, signatures) but has a relatively terse checklist.7
Draft Bill (2024): The draft requires awards to expressly state that certain safeguards were observed (e.g., no incapacity, validity of arbitration clause, proper notice, tribunal composition, procedure followed, subject-matter within submission). It also substitutes “seat” for “place” and permits digitally signed copies.8
Balasamy implication: Strengthening the on-face contents reduces ambiguous challenges: a clearer award record makes it easier to identify defects that are “apparent on the face” (the narrow class of cases where surgical correction is appropriate), thereby limiting frivolous Section-34 collateral attacks.
Post-award interest: The award may include interest; by default a money award carried interest at two per cent higher than the current rate (as explained in the Act).9
Draft Bill (2024): The draft raises the statutory default to three per cent higher than the prevailing RBI repo rate, omits an earlier explanation tied to the Interest Act, and tightens the drafting requirement for awards on interest.10
Balasamy implication: The majority in Balasamy accepted limited recalibration of post-award interest when awards are silent, unconscionably high, or contrary to agreement. The Draft’s clearer statutory benchmark reduces ad-hoc judicial tinkering while preserving the limited, principled corrections the Court accepted.
Emergency/arbitral interim measures and institutional fixes Courts and tribunals could grant interim measures; there was not a statutory scheme for emergency arbitrators at institutional level.11
Draft Bill (2024): The draft recognises emergency arbitrators (appointment prior to constituting the tribunal), makes emergency orders enforceable like tribunal orders, and allows such orders to be confirmed, modified, or vacated by the tribunal; it also clarifies tribunal interim-measure powers.12
Balasamy implication: These institutional avenues reduce routine recourse to courts for urgent relief — consistent with the majority’s theme of keeping ordinary, time-sensitive matters within the arbitral machinery rather than the courts.
Procedural safeguards: disclosures, formulation of grounds, timelines, and records: Section 34 had general procedures (notice, time limits, adjournment power), and Section 32 dealt with termination and record-keeping; there were no express obligations to disclose related challenges.13
Draft Bill (2024): The draft inserts (i) an obligation for applicants to disclose any pending/decided challenges arising from the same legal relationship; (ii) a requirement that the Court or appellate tribunal formulate the specific grounds to be argued before hearing (to focus litigation); (iii) timelines for disposal (expeditious, one year target retained) and (iv) a new duty for tribunals to return arbitration records to the institution or parties after termination.14
Balasamy’s implication: These procedural checks aim to reduce serial or tactical challenges (the very concern Justice Viswanathan flagged). Formulation of grounds and mandatory disclosure push parties and courts toward focused, efficient resolution helping ensure the new power of limited correction is not used as a backdoor appeals strategy.
COMPARATIVE PROVISIONS – 1996 ACT VS. DRAFT BILL 2024 IN LIGHT OF GAYATRI BALASAMY JUDGMENT
| Provision | 1996 Act | Draft Bill 2024 | Implication in light of Gayatri Balasamy |
|---|---|---|---|
| Section 34 — Setting aside an award | Allows setting aside on limited grounds; no express provision for partial set-aside or remittal. | Expressly allows award to be set aside ‘in whole or in part’; adds explanations limiting merits review. | Codifies majority view — enables courts to excise defective parts while prohibiting merits reappraisal. |
| Partial set-aside / Severance | No structured remittal procedure; unclear binding effect of prior findings. | Provides that only set-aside parts are re-decided within fixed time; tribunal bound by unaffected findings. | Implements ‘surgical correction’ approach approved by majority — avoids unnecessary re-arbitration. |
| Appellate Arbitral Tribunal (new s.34A) | No appellate arbitral forum; challenges heard by courts. | Creates optional appellate arbitral tribunal; excludes court setting-aside where adopted. | Specialist forum for limited review — aligns with majority’s caution, but may echo dissent’s appellate-layer concern. |
| Form & contents of award (s.31 changes) | Basic requirements (reasons, date, place, signatures). | Adds checklist of safeguards (valid agreement, notice, jurisdiction, etc.); ‘seat’ replaces ‘place’; digital signatures allowed. | Clearer records support quick defect identification; helps confine intervention to defects apparent on face. |
| Post-award interest (s.31(7)) | Default: two per cent higher than current rate. | Default: three per cent above prevailing RBI repo rate; omits old explanation. | Statutory benchmark reduces need for ad-hoc judicial recalibration — consistent with majority’s limited correction. |
| Emergency arbitrator & interim measures | No statutory emergency arbitrator scheme; interim measures from courts/tribunals. | Recognises emergency arbitrators; emergency orders enforceable; tribunal may confirm/modify. | Institutionalises urgent relief — consistent with keeping routine matters out of court. |
| Procedural safeguards & disclosures | No obligation to disclose related challenges; grounds argued as per pleadings. | Mandatory disclosure of related challenges; formulation of grounds before hearing; timelines; record return duty. | Reduces serial/tactical challenges; supports dissent’s concern about discipline while enabling focused use of correction power. |
Pre-Balasamy’s understanding of Section 34
Section 34 has long been at the centre of debate over the proper extent of judicial intervention in arbitration. Courts have taken varied approaches, but the dominant trend has been caution: judges traditionally resisted remodelling arbitral awards so as to protect the autonomy and finality of tribunals.15 Sub-section 34(2)(a) itself specifies the limited grounds on which a court may set aside an award.16
In several leading decisions, for example, McDermott International Inc. v. Burn Standard Co. Ltd., Project Director, NHAI v. M. Hakeem & MMTC Ltd. and Sangyoung Construction Ltd. v. NHAI, the courts confronted questions about modification but concluded that their power was confined to setting aside awards. Those courts held that no implicit authority to alter the substance of an award could be read into the statutory power to annul it.
That said, a line of authority went the other way. Decisions such as Vedanta Ltd. v. Shenzden Shandong Nuclear Power Construction Co. Ltd.,17 Oriental Structural Engineers Pvt. Ltd. v. State of Kerala18 and Tata Hydroelectric Power Supply Co. Ltd. v. Union of India19 took a broader view of Section 34, endorsing the idea that courts could, in appropriate circumstances, modify an arbitral award.
The Supreme Court’s ruling in Gayatri Balasamy marks a watershed. Heard by a Constitution Bench after an earlier division-bench decision, the Court resolved the long-running disagreement by holding that modification is permissible where the parts of the award that require correction can be severed from the valid portions.
Gayatri Balasamy v. ISG Novasoft Technologies Ltd.,20
This Supreme Court ruling marks a careful recalibration of post-award review in India, carving out a narrow space for courts to excise or correct clearly severable defects in arbitral awards while upholding the core principle of arbitral finality; it guides when clerical fixes, interest adjustments, or surgical severance are appropriate without permitting courts to re-decide merits or assume an appellate role.
BACKGROUND AND JUDICIAL INTERVENTION IN ARBITRATION
The Arbitration and Conciliation Act, 1996 was introduced to align India’s arbitration regime with the UNCITRAL Model Law and to promote speedy, final, and largely court-free resolution of disputes. Section 34, which lists the limited grounds on which an arbitral award may be set aside, embodies this philosophy. It restricts judicial interference to discrete situations — for example, where a party was denied proper notice, where the tribunal’s procedure was unfair, where the award offends public policy, or where there is a manifest illegality. The scheme of Section 34 does not contemplate a re-hearing of facts or merits, nor does it expressly give courts a power to alter the substance of awards.
In practice, however, Indian jurisprudence has shown a lack of uniformity on the question whether courts may do more than annul an award — whether they can, in suitable cases, mould or modify it. This patchwork approach has arisen from two sources: a judicial impulse to deliver complete justice in individual cases and an absence of explicit statutory guidance about modification. Some judgments, such as Vedanta Ltd. v. Shenzhen Shandong Nuclear Power Construction Co. Ltd., permitted limited adjustments often to address interest or other incidental matters on equitable grounds. Other rulings, most notably McDermott International Inc. v. Burn Standard Co. Ltd., firmly rejected any role for courts as substitute tribunals, insisting that judges should confine themselves to the narrow remit that Parliament intended.
The resulting uncertainty has been problematic for parties and practitioners alike, undermining predictability and the core aims of the 1996 Act. When only a part of an award is tainted, the lack of a settled approach to partial modification whether to excise, correct, or remit that portion created both doctrinal confusion and practical difficulty.
It was this unsettled state of the law that led the Supreme Court to refer the matter to a Constitution Bench in Gayatri Balasamy v. ISG Novasoft Technologies Ltd.. Confronted with conflicting precedents, the Court had the task of harmonising competing principles and clarifying the permissible scope of judicial review under Section 34.21
Facts and Procedural History of the Present Case
The dispute in Gayatri Balasamy v. ISG Novasoft Technologies Ltd. originated from an employment conflict between Ms. Gayatri Balasamy, a former employee, and ISG Novasoft, a private software firm. Ms. Balasamy initiated arbitration claiming wrongful termination and seeking monetary relief and other dues. The arbitral tribunal delivered an award that granted her only part of the relief she sought.
Dissatisfied, both sides challenged elements of the award. Ms. Balasamy moved the Madras High Court under Section 34 of the Arbitration and Conciliation Act, 1996, arguing that the award contained legal defects and failed to address several of her claims; she asked the court not simply to annul the award but to modify it to provide fuller relief. A Single Judge agreed in part and altered the award to increase her entitlement. ISG Novasoft appealed to a Division Bench under Section 37, which upheld the Single Judge’s order.
ISG Novasoft then appealed to the Supreme Court. During proceedings, it emerged that courts across the country had taken differing views on whether a court hearing a Section 34 challenge may legitimately vary an arbitral award. Given the wider significance of this conflict, the Supreme Court referred the matter to a Constitution Bench to resolve the question: does a court’s jurisdiction under Section 34 permit modification or variation of an award, or is it confined to setting the award aside on the statutory grounds? The five-judge Bench’s subsequent decision provided a definitive statement on the permissible scope of judicial intervention after an award is rendered.22
Core Issues before the Supreme Court
When the matter was placed before the Constitution Bench, the Supreme Court had to settle a long-standing and fundamental question in arbitration law: whether courts, while hearing challenges under Section 34 of the Arbitration and Conciliation Act, 1996, may actually modify the terms of an arbitral award.
To resolve this, the Bench identified several central issues:
1. Scope of Section 34: Does Section 34 authorize courts to alter, vary or otherwise modify an arbitral award, or is the court’s role confined strictly to setting the award aside? The Court examined the statute’s wording and structure to determine whether any power to change an award can be implied from the provision.
2. Partial setting aside and severability: Can a court partially annul an award by separating and removing the invalid portions while leaving the rest intact? The Bench explored whether the doctrine of severability permits partial annulment and whether that amounts — in effect — to modifying the award.
3. Correction and moulding of awards: Are there statutory or equitable bases on which a court may correct or reshape an award — for example, by adjusting post-award interest, fixing computational or clerical errors, or granting relief the tribunal omitted? This required the Court to draw a line between permissible judicial review and impermissible appellate-style correction, including consideration of Section 31(7) on interest and other incidental powers.
4. Interaction between Sections 34 and 37: How do the powers under Section 34 (setting aside) interact with the appellate jurisdiction under Section 37? The Bench questioned whether Section 37 could be read more broadly than Section 34 and whether any such reading would fit the overall design of the Act.
5. Role of constitutional powers: To what extent can constitutional remedies notably powers under Article 142 be used to reshape arbitral awards in the interests of justice, and what limits should govern such use? Although not the primary focus, the Bench addressed this because courts have occasionally relied on constitutional powers to grant relief beyond statutory remedies, with implications for predictability in arbitration law.
ANALYSIS OF THE SUPREME COURT’S JUDGMENT (MAJORITY OPINION)
The majority opinion, authored by Chief Justice Sanjiv Khanna and joined by Justices B.R. Gavai, P.V. Sanjay Kumar and A.G. Masih, marks a careful shift away from the strict “no modification” rule while remaining true to the Arbitration Act’s purpose. The Court rejected a literal, rigid reading of Section 34 and preferred a contextual and purposive approach. Although the statute does not explicitly empower courts to modify awards, the majority recognised that certain limited interventions — especially where the objectionable portion is severable — amount to narrow, justifiable corrections rather than an impermissible encroachment on arbitral autonomy.
The judgment’s principal innovation is to validate what it calls “partial setting aside” or severance: when an award contains a discrete, separable defect, courts may excise that part and leave the remainder intact. The Court framed this as a pragmatic way to avoid needlessly nullifying entire awards and forcing re-arbitration on issues that are otherwise settled. Far from creating an appellate gateway, the majority emphasised that these powers are procedural and tightly circumscribed.
The majority outlined four principal categories in which limited judicial intervention is appropriate:
• Correction of manifest clerical or arithmetic mistakes: Courts may correct obvious errors on the face of the award arithmetic slips, typographical mistakes or clerical inaccuracies so long as the correction does not alter the tribunal’s substantive reasoning or findings.
• Severance of an invalid portion of an award: Where the award contains both valid and invalid elements that are clearly separable, courts may set aside only the invalid segment and uphold the rest. This severance preserves the effective parts of an award and avoids unnecessary re-arbitration.
• Adjustment of post-award interest under Section 31(7)(b): The Court held that judicial recalibration of post-award interest is permissible where the award omits to fix interest, awards interest that is unconscionably high, or otherwise conflicts with the parties’ agreement or law. Such limited intervention is an exercise of oversight, not a substitution of the tribunal’s role.
• Use of Article 142 powers in exceptional cases: In truly exceptional circumstances, the Supreme Court may invoke Article 142 to mould relief and secure complete justice. The Court stressed that this is an extraordinary remedy and must not be used as a routine tool to rewrite awards.
Throughout, the majority repeatedly warned against treating these doctrines as license for courts to re-decide facts or to second-guess the arbitrator’s assessment of evidence. Intervention is confined to narrow, judicially manageable defects that are apparent on the face of the award; courts must not assume appellate functions, substitute contractual interpretation, or reappreciate evidence.
On precedent, the Court distinguished Project Director, NHAI v. M. Hakeem (2021), explaining that Hakeem concerned statutory interest under a special statute and should not be read as a blanket prohibition on modification in all contexts. The majority also relied on earlier judgments where courts had moulded relief or retained parts of awards, treating those practices as established and legitimate.
In short, the judgment preserves the central aim of finality in arbitration while permitting surgical, narrowly-defined corrections where necessary to make awards lawful and effective. The majority’s approach seeks a balance: protect arbitral autonomy, but allow limited judicial remedies so that justice is not defeated by avoidable formal defects.23
MINORITY OPINION BY JUSTICE K. V. VISWANATHAN: A CLEAR REAFFIRMATION OF “NO MODIFICATION”
Justice K. V. Viswanathan filed a reasoned dissent that firmly rejects the majority’s conclusion that courts may exercise limited power to modify arbitral awards under Section 34. His judgment rests on a close textual and structural reading of the Arbitration and Conciliation Act, 1996, alignment with the UNCITRAL Model Law, and a strong commitment to judicial restraint in arbitration. In his view, Section 34 is designed to annul awards that suffer from specific vitiating defects fraud, denial of natural justice, lack of jurisdiction, or patent illegality not to give courts a remedial power to alter them.
Viswanathan begins by insisting that Section 34 functions as an annulment provision. Allowing courts to tinker with awards, even in limited ways, would blur the carefully drawn line between the tribunal’s role as fact-finder and contract-interpreter and the court’s role as a narrow supervisory forum. He warns that permitting modification risks turning an exceptional corrective power into an ordinary appellate-like review.
That said, Justice Viswanathan does accept narrow, conventional interventions: manifest clerical or computational mistakes and certain post-award interest questions under Section 31(7). But he qualifies these exceptions tightly. Corrections should be confined to clear, on-the-face errors or situations where the arbitrator has plainly exceeded statutory bounds. He expressly rejects any notion of “reasonableness-based recalibration” of awards a subjective standard, he argues, that would erode deference to arbitral decision-making.
A recurrent practical concern in the dissent is the slippery slope: once courts are permitted to modify in a few cases, the boundary between supervision and review will rapidly erode. As Justice Viswanathan warns, “Once we allow courts to modify in some cases, we lose the discipline that arbitration demands. The exception will soon become the rule.” That caution informs his insistence on bright-line limits.
Drawing on comparative practice, he notes that most jurisdictions following the UNCITRAL framework do not empower courts to rewrite awards; enforcement or annulment remains the accepted binary. He further contends that the New York Convention presumes the same binary outcome, recognition or annulment, rather than judicial rewriting.
Finally, he rejects the broad use of Article 142 to justify routine modification. Article 142, he says, is reserved for truly extraordinary constitutional exigencies and cannot be turned into a tool for commonplace award correction.
Although in dissent, Justice Viswanathan’s opinion serves as a doctrinally disciplined counterweight to the majority, urging fidelity to the statutory text, protecting arbitral finality, and cautioning against judicial overreach.24
IMPACT AND IMPLICATIONS OF THE JUDGMENT
This judgment will reshape India’s arbitration landscape not by upending it, but by fine-tuning the balance between finality and limited judicial intervention.
At a doctrinal level, the Court has narrowed and clarified the role of courts under Section 34. It accepts that interference with awards is permissible in tightly defined situations for example, where an illegal segment can be severed or where post-award interest is plainly erroneous so that only the defective portion is excised and the rest of the award remains enforceable. That approach removes the all-or-nothing outcome that previously led to unnecessary re-arbitration on settled questions.
The ruling also brings coherence to earlier, divergent decisions. By distinguishing the facts and reasoning in precedents like McDermott, Hakeem and Vedanta, the Court supplies a practical framework: courts cannot generally “modify” awards, yet they may grant targeted relief through severance or limited corrections where statutory or doctrinal grounds exist. This should reduce inconsistent treatment across benches and give lower courts and practitioners a clearer roadmap for post-award challenges.
For litigators and parties, the practical consequences are immediate. Claimants who fear losing on technical grounds can pursue narrower remedies that preserve favourable findings instead of risking a complete annulment. Respondents must now be prepared for opponents seeking surgical relief rather than wholesale reversal. The decision also raises the premium on precise drafting, especially on interest clauses and the articulation of remedies, because awards are now more likely to be preserved in part rather than set aside in whole.
Institutionally, the judgment strengthens India’s standing as a forum that respects arbitral autonomy while providing responsible judicial oversight. By avoiding extremes, it harmonises domestic practice with international expectations of predictability and limited court intervention. That balance should reassure domestic and foreign parties alike, improving confidence in India as an arbitration-friendly seat and lowering the perceived risk of protracted post-award litigation.25
The Severability Debate: Limits of Modification under Section 34
The Constitution Bench examined how far a court may go in altering an arbitral award, focusing on when and how parts of an award can be excised or corrected without disturbing the rest. Central to this discussion was Section 34(2)(a)(iv), which contemplates the court’s ability to separate non-arbitrable or invalid elements of an award from those that are valid. Put simply, a court may remove, remit, or declare void only those portions that can be cleanly detached from the award’s valid remainder.
The Court held that the proviso to Section 34(2)(a)(iv) should be read as clarificatory: the power to sever an “invalid” segment from a “valid” one is implicit in the court’s jurisdiction when it considers setting aside an award. In support it relied on precedent such as Pratap Chamaria v. Durga Prasad Chamaria, which recognised the possibility of partially setting aside an award where valid and invalid parts are separable.
Earlier authorities had been inconsistent on this point. For example, ONGC Ltd. v. Western GECO International Ltd. accepted modification where severability existed, while the line of decisions stemming from NHAI v. M.A. Hakeem denied the courts any power to modify awards. The five-judge bench in Gayatri Balasamy reconciled these conflicting strands by holding that modification is permissible to the extent that the offending portion is severable from the remainder of the award.
The Bench also considered the roles of Sections 33 and 34. It was observed that Section 33 gives the tribunal limited corrective powers, mainly to fix calculation or typographical mistakes and other obvious, manifest errors, but does not permit substantive alteration. Section 34, by contrast, does not deal with clerical corrections; it can, however, operate to sever and set aside parts of an award on the statutory grounds that Section 34 supplies.
Invoking the maxim omne majus continet in se minus (the greater includes the lesser), the Court read into Section 34 the narrower authority to modify those parts of an award that are invalid while preserving the valid portions. In doing so, the Court sought to thread a needle, preserving party autonomy and tribunal finality while allowing limited judicial correction where strictly necessary. Although the ruling settles a long-running controversy, its practical consequences will only become clear with time.
In a powerful dissent, Justice Viswanathan disagreed. He argued that the Arbitration Act — especially after the post-2015 and 2019 reforms reflects a deliberate move away from broad judicial intervention, and that the statute does not authorise modification of awards. He warned that even a limited modification power risks eroding arbitral finality and undermining tribunal autonomy. According to the dissent, existing remedies (including those under Section 33 and Section 34(4)) are adequate, and courts should be cautious not to cross the boundary between interpreting an award and effectively rewriting it.
Critique and Practical Considerations
The Judgment brings useful clarity and a practical middle path, but it also raises important questions about consistency, judicial discretion, and long-term predictability. This section looks at those concerns and the practical problems that could surface now that the Court has authorised limited post-award interventions.
At the centre of the majority’s approach is severability: the idea that a court can excise a discrete unlawful part of an award while leaving the rest intact. The doctrine is defensible in principle, but applying it depends heavily on judicial judgment. What one judge treats as “severable” another may not. Absent legislative guidance or binding tests, High Courts and commercial benches could reach divergent outcomes, reintroducing the uncertainty the judgment aims to eliminate.
A related problem is the blurred line between permissible correction and impermissible substitution. The Court allows correction of obvious clerical or computational errors and limited adjustment of post-award interest, but it does not define precise limits. If a court alters the damages awarded after removing an unlawful component, that could look very much like rewriting the award. Without clear guardrails, lower courts may unintentionally encroach on the tribunal’s role under the guise of correction or severance.
There is also a practical, behavioural risk. By opening a narrow route for modification, the decision may encourage losing parties to pursue partial annulments or “surgical” corrections as fallback litigation strategies. That would increase post-award litigation, slow down enforcement, and shift commercial focus from finality to procedural second-guessing precisely the outcome the dissent warned against.
Thus, critics may say the Court stepped into a policy gap better filled by Parliament. Where arbitration law and practice are evolving, a statutory amendment could have provided clearer, democratically grounded rules about modification and severance. The Court’s purposive solution resolves the immediate conflict in precedent, but it leaves room for future debate and possibly legislative clarification.
CONCLUSION
The analysis shows that Gayatri Balasamy attempts a pragmatic middle path: preserving the core value of arbitral finality while permitting narrowly-tailored judicial corrections where an award is plainly tainted on its face. The decision recognises limited tools severance of separable illegal portions, correction of clerical or calculational errors, and extremely narrow post-award adjustments that allow courts to remove clear defects without converting themselves into appellate tribunals. This approach brings greater doctrinal clarity compared with earlier, conflicting rulings, but it is not a complete solution on its own.
At the same time, both the judgment and the Draft Amendment expose real risks. Without precise legislative language and firm judicial tests, “surgical” correction can drift into merits re-examination or become an avenue for tactical litigation. To prevent that slippage, reform should focus on bright-line rules: clear severability criteria, a strict “face-apparent” threshold for intervention, limited timelines, mandatory disclosure of issues relied on for correction, and a narrowly circumscribed remittal mechanism that confines reconsideration to identified portions only.
Policy and institutional responses can further reduce uncertainty. If the legislature adopts precise drafting that mirrors these safeguards, and if institutional options (for example, an optional Appellate Arbitral Tribunal) are accompanied by transparent procedures and expertise requirements, parties will gain predictable, specialist routes for limited review without overburdening courts. Courts, for their part, should treat extraordinary remedies (including relief under Article 142) as genuine exceptions, not defaults.
Therefore, the path forward is one of calibrated harmonisation: preserve arbitral autonomy and finality as the baseline, but allow tightly defined corrective mechanisms where justice would otherwise be defeated by clear, visible errors. Achieving that balance will require coordinated action by the judiciary, legislature and arbitral institutions so that corrections remain surgical, predictable and confined and never a substitute for appellate review.
Suggestions
1. Statutory recognition of partial set-aside with clear guardrails: Amend Section 34 to expressly permit an award to be set aside “in part” and to provide a statutory remittal route (limited re-decidion of only the set-aside issues), while also narrowing the grounds that permit intervention so courts cannot re-appraise merits.
2. Judicial practice direction on severability: Issue a practice direction or binding guidelines that list factors courts should apply when deciding severability — for example: whether the impugned portion is legally and economically independent, whether excision alters the award’s operative relief, the parties’ autonomy and expectations, and whether the tribunal’s reasoning on the remaining heads is intact. This reduces ad-hoc, bench-to-bench variation.
3. Award drafting standards for tribunals: Encourage institutions to adopt model award templates and training so awards are modular: distinct heads of claim, separate reasoning for each relief item, explicit findings on jurisdiction and interest, and clear calculation sheets. A well-structured award makes face-apparent defects easier to spot and limits unnecessary court interference.
4. Bright-line limits on judicial correction: Require courts to confine corrections to narrow categories (manifest clerical/arithmetic errors, clearly severable illegal segments, and strictly circumscribed post-award interest adjustments). Mandate that courts formulate the precise grounds they will address before hearing and set short, fixed timelines for disposal to curb tactical delay.
5. Institutional alternatives and emergency remedies: Promote optional institutional mechanisms — an appellate-arbitral route where parties opt in, and a reliable emergency-arbitrator scheme — with transparent rules and time limits. These parallel tracks can handle specialist, time-sensitive problems without dragging ordinary matters into courts.
6. Procedural safeguards to deter tactical litigation: Introduce mandatory disclosure of any pending or decided challenges arising from the same legal relationship, a pre-trial formulation of issues by the court or tribunal, and strict record-return/retention rules so remittal uses the original file. These steps limit serial or ambush challenges and focus hearings on genuinely contested legal defects.
7. Capacity building and specialization at first instance: Invest in targeted judicial training for Commercial Court benches on the new doctrine, provide model bench orders, and where possible assign Section-34 matters to specialist or trained judges. Parallelly, run institutional training for arbitrators so awards meet the clarity standards judges will expect.
*****
Footnotes
1. The Arbitration and Conciliation Act s 34.
2. The Arbitration and Conciliation (Amendment Bill) 2004.
3. The Arbitration and Conciliation Act.
4. Supra n 2.
5. The Arbitration and Conciliation Act s 35 A.
6. Supra n 2.
7. The Arbitration and Conciliation Act s 31.
8. Supra n 2.
9. The Arbitration and Conciliation Act s 31(7)
10. Supra n 2.
11. The Arbitration and Conciliation Act s 9A & s 17
12. Supra n 2.
13. The Arbitration and Conciliation Act s 34 (1A/1B) & s 32.
14. Supra n 2.
15. The Arbitration and Conciliation Act s 34.
16. The Arbitration and Conciliation Act s s 34(2)(a).
17. AIR 2018 SC 4773
18. (2021) 6 SCC 150
19. (2003) 4 SCC 172
20. 2025 SCC OnLine SC 986
21. Gayatri Balasamy v. ISG Novasoft Technologies Ltd., 2025 SCC OnLine SC 986
22. Gayatri Balasamy v. ISG Novasoft Technologies Ltd., 2025 SCC OnLine SC 986
23. Gayatri Balasamy v. ISG Novasoft Technologies Ltd., 2025 SCC OnLine SC 986
24. Gayatri Balasamy v. ISG Novasoft Technologies Ltd., 2025 SCC OnLine SC 986
25. Gayatri Balasamy v. ISG Novasoft Technologies Ltd., 2025 SCC OnLine SC 986