Judicial Innovation or Overreach? Indian Supreme Court’s New Take on Modification of Arbitral Awards

A five-judge bench (Constitution Bench) of the Supreme Court of India (SC) recently delivered a landmark judgment in Gayatri Balasamy v. M/s. ISG Novasoft Technologies Limited—representing a watershed moment in Indian arbitration law. By holding that courts may, under limited circumstances, modify arbitral awards under sections 34 and 37 of the Arbitration and Conciliation Act, 1996 (the 1996 Act), the SC has reopened the long-settled question of arbitral finality.

While the Constitution Bench aimed to curb delays in award enforcement, the ruling risks undermining the foundational principle that arbitral awards are final and binding, save for limited statutory exceptions. In this post, we analyse the judgment, contrast it with global norms, and evaluate its alignment with India’s goal of becoming an international arbitration hub.

Tracing the Jurisprudential Evolution: A Shift in Stance?

The Arbitration Act of 1940 (1940 Act) contained provisions (in sections 15 and 16 thereof) that explicitly empowered courts to modify, correct, or remit arbitral awards. However, with the enactment of the 1996 Act, modelled on the UNCITRAL Model Law, this explicit power was conspicuously omitted, in line with international best practices.

A significant line of judgments, including McDermott International Inc. v. Burn Standard Co. Ltd. (2006), Project Director, NHAI v. M. Hakeem (2021), and Ssangyong Engineering and Construction Company Limited v. National Highways Authority of India (NHAI) (2019), consistently held that section 34 does not grant courts the power to modify arbitral awards, limiting their role to setting aside or refusing to set aside the arbitral award. In particular, the Court in Hakeem stated that the power to modify an award was intentionally excluded in the 1996 Act, drawing a comparison with the erstwhile 1940 Act, emphasizing the autonomy of the arbitral process. While the courts in the judgments above held that awards cannot be modified under section 34, another stream of judgments, such as Hindustan Zinc Ltd. v. Friends Coal Carbonisation (2006), Oriental Structural Engineers Pvt. Ltd. v. State of Kerala (2021), and Vedanta Limited v. Shenzhen Shandong Nuclear Power Construction Company Limited (2018). invoked Article 142 of the Indian Constitution to hold otherwise, indicating that court may be able to modify awards in exceptional cases. This created a doctrinal inconsistency that led to the present reference in Gayatri Balasamy. Notably, the SC in this most recent ruling held that:

“Read in this manner, the limited and restricted power of severing an award implies a power of the court to vary or modify the award. It will be wrong to argue that silence in the 1996 Act, as projected, should be read as a complete prohibition.” (emphasis supplied)

It is argued that this interpretation effectively resurrects a power that the legislature had intentionally abandoned, unsettling the scheme of the 1996 Act and reversing the very reform it sought to accomplish.

The Supreme Court’s Holding

In a 4:1 majority decision, the Constitution Bench led by Chief Justice Sanjiv Khanna held that Indian courts possess limited powers to modify arbitral awards under sections 34 and 37 of the 1996 Act.

The majority judgment outlined four broad situations where such modification may be permissible: (i) severability, which is drawn from the proviso to section 34(2)(a)(iv), permitting setting aside a part of an award if it is separable from the rest; (ii) apparent errors like clerical, computational, or typographical errors, without entering into merits; (iii) modification of post-award interest under section 31(7)(b), especially where it is excessive or creates unjust enrichment; and (iv) exceptional cases in exercise of powers under Article 142 to do ‘complete justice’, such as severing erroneous portions, correcting minor errors, or adjusting excessive interest rates, by modifying awards, albeit with the caution that such power must not become routine.

The majority supported its interpretation by invoking the maxim omne majus continet in se minus—the greater power includes the lesser—contending that the authority to set aside an arbitral award must inherently include the power to set it aside in part, or to modify it.

Diluting Finality

The approach of the SC fundamentally misaligns with the 1996 Act’s restrictive framework, which limits courts to either upholding or setting aside awards, not rewriting them. By introducing a modification option, the majority decision risks turning courts into appellate bodies, undermining the very essence of arbitration as a private, expeditious alternative to litigation. The vague ‘exceptional circumstances’ standard not only compounds the problem, inviting inconsistent application across lower courts, but it also fuels further delays in an already burdened judicial system.

A Dissent That Reinforces Finality

In his dissent, Justice K.V. Viswanathan stressed that sections 34 and 37 limit courts to setting aside or remitting awards, not modifying them. Any modification, even on minor grounds, he argued, would require the court to ‘step into the shoes of the arbitrator,’ violating the principle of party autonomy and minimal judicial interference. His criticism even extended to the use of Article 142, arguing that constitutional powers should not be deployed to circumvent legislative intent and, if so, it should be carried out through parliamentary amendment and not judicial innovation.

The dissent cautioned that the absence of a modifying power in the 1996 Act, especially when such power existed under the previous legislation (1940 Act), is not a mere legislative silence - it is a conscious exclusion. To read modification into the power to set aside would amount to judicial legislation.

The majority judgment appears driven by the intent to minimise prolonged litigation—this intent is evident from the reasoning that excluding the power to modify would force parties into ‘an extra round of arbitration, adding to the previous four stages,’ thereby undermining the efficiency of the arbitral process. However, this interpretation sits uneasily with the precedent set in Hakeem, where the SC, while interpreting the phrase ‘recourse against the arbitral award,’ adopted a more textualist stance. It held that recourse to court is confined strictly to an application for setting aside the award under sections 34(2) and 34(3), thereby excluding any power of modification or correction. The Court further emphasized that ‘when the right is itself truncated, enforcement of such truncated right can also be only limited in nature.’ This approach suggests a narrower judicial role, one that respects the statutory limits on court intervention in arbitral proceedings, even at the cost of procedural inefficiencies.

Comparative Lens

In Hakeem, the Court also opined that a review of the arbitration legislation of England, the United States, Canada, Australia and Singapore shows that there are express provisions which permit the varying of an award, unlike section 34 of the 1996 Act.

• In the United Kingdom, the Arbitration Act 1996 permits limited modifications to awards, emphasising their finality and setting a high threshold for judicial intervention. For example, sections 67 and 69 allow challenges on law or jurisdiction only with court leave or party agreement and section 57 enables the tribunal to correct clerical errors or clarify ambiguities.
• In Singapore, while international arbitration under the International Arbitration Act does not permit court modifications, domestic arbitration allows for variation on limited grounds through appeals on questions of law, although Singaporean courts generally maintain a strict non-interventionist approach.
• Australia, on the other hand, allows variation of domestic awards with leave of the court, but its international arbitration law bars modification, aligning with the UNCITRAL Model Law.

These jurisdictions demonstrate that where modification is allowed, it is narrowly defined, consensual, or statutorily grounded and not judicially implied.

Finality, Efficiency, and the Risks of Expanded Judicial Review

The Gayatri Balasamy ruling poses significant concerns for arbitration in India. By allowing modification of awards under sections 34 and 37, it risks enabling parties to challenge the substance of awards under the guise of correction. This undermines arbitration’s core promises of speed and finality. More critically, it could erode investor confidence, as arbitral awards may no longer be seen as final and binding—casting doubt on India’s commitment to an enforcement-friendly arbitration regime.

The remedy lies in amending the 1996 Act, possibly by inserting a provision similar to section 57 of the UK Arbitration Act, providing a narrow and clearly defined window for rectification. Until then, courts must exercise strict restraint to ensure that ‘modification’ does not become a backdoor for re-litigation or judicial overreach.

Conclusion

Though intended to streamline the process of modification of awards, the judgment risks crossing into judicial rewriting. If left unchecked, it could dilute finality and compromise party autonomy—two major pillars of arbitration. India’s pro-arbitration trajectory now hinges on whether the parliament, courts, and practitioners can realign law and practice with international norms.

Deepriya Snehi is an independent practitioner before the High Court of Delhi and the Supreme Court of India.