The Doctrine of “Limited Judicial Modification” under the Arbitration and Conciliation Act, 1996 – Commentary on the Constitution Bench decision in Gayatri Balasamy v. M/s ISG Novasoft Technologies Ltd. (2025 INSC 605)

1. Introduction

Following years of uncertainty over whether courts in India could alter arbitral awards, a five-judge Constitution Bench of the Supreme Court has now spoken in Gayatri Balasamy v. M/s ISG Novasoft Technologies Ltd.. The Court was confronted with diametrically opposite strands of precedent – one line holding that Section 34 of the Arbitration and Conciliation Act, 1996 (“1996 Act”) only allows an award to be set aside or remitted; the other, without much analysis, routinely “modified” awards, especially interest components. The Bench, led by CJI Sanjiv Khanna, settles the debate by carving out a narrow – yet important – power of “limited judicial modification” while affirming the core philosophy of minimal court intervention.

Key issues examined:

• Do Sections 34 and 37 empower Indian courts to modify arbitral awards?
• Can such power be exercised only where the award is severable?
• What is the impact on earlier precedent, especially the two-judge decision in Project Director, NHAI v. M. Hakeem (2021) 9 SCC 1?

2. Summary of the Judgment

The Bench (3:2) crafts a calibrated answer:

• No general merits-based “appellate” power. Courts cannot re-appreciate facts or law to substitute their view for that of the arbitrator.
• Four limited situations in which courts may vary an award instead of wholly setting it aside:
  1. Severability – where the invalid portion is distinct and its excision does not disturb the valid parts (Proviso to s.34(2)(a)(iv)).
  2. Patent clerical/computational slips – corrections that flow automatically from the record (akin to s.152 CPC).
  3. Post-award interest – the rate can be increased or reduced to comply with s.31(7)(b) or equity, without re-arbitration.
  4. Article 142 (Supreme Court only) – extraordinary power to mould relief to achieve complete justice, used sparingly.
• The doctrine enunciated in M. Hakeem is partially overruled: a Section 34 court cannot undertake merits review, but can resort to the above limited modifications.
• Section 34(4) remand remains intact and courts are encouraged to use it when deeper factual reconsideration is required.

3. Precedents Analysed

3.1 Line favouring non-modification

• Project Director, NHAI v. M. Hakeem (2021) – confined courts to set-aside/remit.
• McDermott International v. Burn Standard (2006) – courts “cannot correct errors of arbitrators”.

3.2 Line permitting modification

• Tata Hydro-electric (2003), Budhraja (2008) – interest/date of interest tweaked.
• Vedanta v. Shenzhen (2019) – LIBOR-based interest substituted.
• These were treated as fact-specific examples where the Court implicitly exercised powers similar to category (ii) and (iii) above.

3.3 Foreign statutory models

The Bench surveyed U.K., Singapore, Australia, Kenya etc., noting that they expressly confer “vary” powers on courts – India does not, but the constitutional court can legitimately imply a narrow version to avoid needless re-arbitration.

4. Legal Reasoning

4.1 Textual Analysis of Section 34

The Court re-emphasises the phrase “recourse… only by an application for setting aside” but clarifies that “setting aside” includes partial annulment or qualified correction where it inevitably follows from eliminating a Section 34 defect. Hence modification is not an independent appellate review, but a consequence of partial annulment.

4.2 Severability Doctrine

Borrowed from contract law and acknowledged in the proviso to s.34(2)(a)(iv). Tests laid down:

• Invalid part must be clearly segregable in terms of liability and quantum.
• No inter-dependence or common factual matrix with the surviving portion.
• Courts must record reasons demonstrating such independence.

4.3 Implied/incidental powers

Section 34 courts are civil courts; the maxim actus curiae neminem gravabit allows them to rectify patent slips to prevent injustice. This dovetails with the tribunal’s own power under s.33; where the tribunal fails or time limits lapse, the court can step in.

4.4 Post-award Interest

Because s.31(7)(b) statutorily hard-wires a default rate, failure to apply it is a patent illegality that requires no fact finding. Adjusting the rate is therefore permissible modification rather than merits review.

4.5 Article 142 and the Supreme Court

Reserved for the “rarest cases” to prevent multiplicity of proceedings where modification is self-evident and further arbitration futile. The concurring opinion (Viswanathan J.) cautions against over-use, but the majority leaves the door ajar.

5. Impact of the Judgment

• Provides a middle path: repudiates a blanket ban on modification, yet preserves the autonomy of arbitral tribunals.
• Reduces re-arbitration in cases of clerical errors or segregable invalid portions, saving costs and time.
• Guides High Courts and District Courts on exercising Section 34(4) remand vs. limited modification.
• Enhances India’s attractiveness as an arbitral seat by aligning with pragmatic international practice, without importing unfettered appellate review.
• Statutory amendments (as mooted by the Vishwanathan Committee) may still follow to codify the limits; until then, the precedent governs.

6. Complex Concepts Simplified

• Setting aside vs. Modification: “Setting aside” annuls (kills) the award wholly or partly; “Modification” trims or corrects it while the rest survives.
• Severability: Comparable to pruning a diseased branch without felling the tree.
• Section 34(4) Remand: Court pauses its proceedings and sends the award back to the tribunal to cure defects – used when factual reconsideration is needed.
• Patent illegality: Obvious errors visible on the face of the award (e.g., granting interest contrary to a contract clause).
• Article 142: Extraordinary constitutional power enabling the Supreme Court to do “complete justice” where statute is silent, but not to rewrite law routinely.

7. Conclusion

The Constitution Bench has delicately balanced two competing values: (a) party autonomy and finality of arbitral awards; and (b) the court’s duty to do justice and prevent multiplicity of proceedings. By recognising a doctrine of limited judicial modification, the Court closes the artificial binary of either remitting or annihilating an award, while firmly shutting the door on full-fledged appellate review. The judgment therefore ushers in clarity, coherence and efficiency in the post-award landscape of Indian arbitration.