Reassessing Jindal Poly Films: The curious case of privatised class action and the Section 245 Dilemma

The Supreme Court's arbitration reference in Jindal Poly Films leaves many questions unanswered. A 3-paragraph consent order moved India's first admitted shareholder class action out of the Companies Act, past the logic of the Insolvency and Bankruptcy Code (IBC) and into a private forum that cannot bind the class.

On June 8, 2026, a Bench of Justices Prashant Kumar Mishra and Atul S Chandurkar disposed of Jindal Poly Films Ltd v. Monet Securities Pvt Ltd in a brief consent order. Both sides presented signed draft minutes; the Court appointed former Chief Justice Manindra Mohan Shrivastava as sole arbitrator with Delhi as the seat; the petition was disposed of; and the NCLT and NCLAT orders both confined to maintainability were set aside, with all contentions kept open.

Read alone, it is an unremarkable consensual disposal. Read in context, it quietly ended what is arguably India's first major admitted shareholder class action under Section 245 of the Companies Act, 2013, a provision notified in June 2016, dormant for 8 years and now extinguished without a word of substantive engagement from the Supreme Court.

The petition, filed in March 2024 by shareholders holding 4.99% of Jindal Poly Films, alleged that promoter-linked transactions - particularly the sale of OCPS and RPS in Jindal Powertech to the SSJ Trust and Jindal Poly Investment for roughly ₹105 crore against a claimed fair value of several thousand crore - loan write-offs and an undervalued subsidiary stake sale caused losses that FTI Consulting estimated at over ₹2,500 crore. The transactions were allegedly structured to stay below the materiality threshold for minority-shareholder approval under Regulation 23 of the SEBI (LODR) Regulations, with SEBI and the Enforcement Directorate (ED) conducting parallel investigations.

The NCLT's 61-page order is the most developed Section 245 jurisprudence yet written. It held that the provision is a deliberately Indian construct: it protects the company as well as its members, accommodates past and concluded transactions through the compensation remedy in Section 245(1)(g), and is not to be read down through the American direct versus derivative distinction in Tooley v. Donaldson, Lufkin & Jenrette. It directed publication of public notice to the entire class under Rule 87. The NCLAT affirmed across 29 pages, walking clause by clause through Section 245(4), noting the FTI report and the parallel regulatory investigations. Both orders are now set aside without engagement.

The instinctive objection that there was no arbitration agreement needs careful framing. Section 7 of the Arbitration and Conciliation Act, 1996 permits parties to agree in writing to arbitrate disputes that have already arisen. Signed consent minutes recorded by a court can constitute such an agreement between the signatories. The real difficulty lies elsewhere: in subject-matter arbitrability, a question no amount of party consent can answer, because Sections 34(2)(b) and 48(2) treat non-arbitrability as a defect going to the root of an award.

In Booz Allen & Hamilton Inc. v. SBI Home Finance Ltd. (2011), the Supreme Court established that actions in rem and disputes reserved to public fora by necessary implication are non-arbitrable, particularly listing insolvency and winding-up petitions among them. Vidya Drolia v. Durga Trading Corporation (2021) refined this into a fourfold test: disputes are non-arbitrable when they relate to rights in rem, affect third parties with erga omnes effect, involve sovereign functions, or are expressly or impliedly reserved to a statutory forum. The Bombay High Court applied parallel reasoning in Rakesh Malhotra v. Rajinder Kumar Malhotra to hold oppression petitions non-arbitrable even where an arbitration clause existed, which is a position that continues to anchor the debate on arbitrability of Sections 241-242 disputes.

Although no court has directly ruled that Section 245 claims are non-arbitrable, existing doctrine points strongly in that direction. The provision's machinery is collective by design: public notice to every class member on admission [Section 245(5)(a)], a statutory opt-out [Rule 86], mandatory consolidation of parallel actions and orders binding on the company, all its members, depositors, auditors and advisors [Section 245(6)]. An arbitral award binds only parties to the agreement and, in limited situations, certain non-signatories. What it lacks is the statutory machinery to bind the shareholder class as a class. It cannot reverse transactions against trusts and group entities not before the arbitrator, cannot reach roughly 40,000 public shareholders and cannot replicate a single feature of the erga omnes decree the petition sought. Two litigants can sign an arbitration agreement; they cannot, by signing it, convert a representative in rem proceeding into a bilateral in personam one.

The sharpest measure of the order's dissonance comes from insolvency law, administered by the very same tribunal. In Swiss Ribbons Pvt Ltd v. Union of India (2019), the Supreme Court held that once a creditor's petition is admitted, the proceeding becomes in rem - a collective proceeding no longer belonging to the initiating creditor - which is why Section 12A of the IBC conditions withdrawal on approval by 90% of the Committee of Creditors. In Indus Biotech Pvt Ltd v. Kotak India Venture (Offshore) Fund (2021), the Court completed the rule: after admission, the proceeding has erga omnes effect and reference to arbitration is impermissible even where a written arbitration clause exists.

The IBC analogy is not exact: insolvency has a moratorium, a creditor committee and a bespoke collective architecture that Section 245 does not replicate. But both regimes share the essential characteristic of transforming an individual grievance into a collective proceeding once the admission threshold is crossed. On the IBC's reasoning, the Section 245 petition had ceased to be the petitioners' private property. The NCLT had directed class-wide public notice; interveners holding a further 5% had joined. Yet, where an admitted insolvency cannot be arbitrated even with an arbitration clause, an admitted class action was sent to arbitration without one - on consent given by a substituted respondent (Monet Securities), which the Court's own order discloses had only recently replaced the original petitioners. Two collective regimes, one adjudicating institution and two irreconcilable answers to the same structural question.

Every mature class action system treats the compromise of an admitted representative proceeding as a court-supervised event, because the lead applicant litigates as a fiduciary, not an owner. Rule 23(e) of the US Federal Rules of Civil Procedure requires notice to the class and a judicial finding that any settlement is fair, reasonable and adequate. The UK Competition Appeal Tribunal must approve collective settlements; Australia bars settlement of representative proceedings without court approval; Canada demands a best-interests review.

Section 245 and Rules 84-87 of the NCLT Rules contain no settlement-approval mechanism, a lacuna this case has now exposed. But that absence was a reason for the Supreme Court to fashion one, not to proceed as though none were needed. No notice of the proposed disposal went to the class whose constitution the NCLT had set in motion. There was no fairness inquiry. Arguably, the first Indian class action to survive a maintainability challenge ended in the one manner class action law everywhere is engineered to prevent: a bilateral exit negotiated over the heads of the class.

If consensual resolution was the object, the statute book offered a structurally superior route. Section 442 of the Companies Act, read with the Companies (Mediation and Conciliation) Rules, 2016, empowers the Tribunal to refer pending proceedings to its mediation panel. The Mediation Act, 2023 now supplies an enforcement framework for mediated settlements. The decisive difference is architectural: mediation returns its product to the referring forum, which retains supervisory jurisdiction and could have tested any settlement against the interests of the class with notice before disposal. Arbitration exports the adjudication wholesale to a confidential forum with no class machinery. A substituted lead applicant inherits the representative burden, not a licence to privatise it.

The most defensible reading of the order is pragmatic: the class had not yet fully crystallised. Rule 87 notice had been directed but not published; no class had formally constituted; the proceeding was still, in structural terms, between named parties. On this view, the Court treated the dispute as sufficiently private to permit consensual referral, preferring a faster route to substantive resolution over protracted jurisdictional litigation that would serve nobody.

That reading has force but runs into two problems. First, admission is the statutory threshold: Section 245(5) is triggered ‘on admission’ and Rule 87 notice was specifically ordered, meaning the class machinery had been engaged even if publication was pending just as Swiss Ribbons and Indus Biotech peg the in rem character of an insolvency to admission, not to any later procedural step. Second, and more fundamentally, the characterisation of the dispute as ‘sufficiently private’ requires a finding on the collective versus bilateral question that the order never makes.

If the order is to be defended at all, it must implicitly rest on the Supreme Court's power under Article 142 of the Constitution to pass orders necessary for ‘complete justice’. But the five-judge bench in Supreme Court Bar Association v. Union of India (1998) held that Article 142 ‘cannot supplant the existing substantive law applicable to the case’. Complete justice for absent class members requires, at minimum, that the Court consider their interests before extinguishing the proceeding designed to protect them. An order that does neither and does not even invoke Article 142 as its basis definitely does not satisfy that standard. The power can complement statutory provisions, but it cannot override them in silence.

Because this is a consent order, it decides nothing as precedent in the formal sense. But practice travels faster than precedent and the questions it leaves behind are serious: is a Section 245 proceeding arbitrable after admission, given Booz Allen, Vidya Drolia and the Indus Biotech rule? Can a substituted applicant consent away rights the statute vests in a class? What becomes of the interveners, of the Rule 87 notice and of the bar in Section 245(5)(c) on a second-class action for the same cause? How does a confidential arbitration sit beside continuing SEBI and ED proceedings on the same transactions? Above all, what signal does it send that India's flagship Section 245 proceeding could be dissolved by dealing with the party holding the pen?

The Court was entitled to encourage settlement. It was not obliged to do so by setting aside the only two judgments India possesses on the scope of its class action remedy and by routing a collective statutory proceeding into a forum structurally incapable of delivering collective relief. Section 245 survives on the statute book. Whether it survives as a credible remedy now depends on whether the next bench treats the Jindal Poly disposal as an aberration of consent or as the way out.

Prasanth Raju is a counsel and advocate practising in the Bombay High Court.