From Clause to Court | Part III of III
The Last Objection: What Survives Section 48(2)(b), and What It Now Costs
| Authored by R & D Law Chambers LLP | Practice led by Ravish Bhatt — Advocate, Bar Council of Gujarat (Enrolment G/504/2008) | Solicitor of the Senior Courts of England and Wales (SRA No. 492 477) | ADIT, Chartered Institute of Taxation, London
Published: 2 July 2026 | Last reviewed: 2 July 2026 |
Index of Topics
- The Progression: Renusagar to Nagaraj
- Nagaraj Mylandla: A Gate That Relocates the Problem
- The Missing Filter: Why Exclusion-Only Tests Cannot Hold
- The Filter, Applied
- The Practitioner’s Protocol
- Services We Provide
| Statutory references are to the Arbitration and Conciliation Act, 1996, as amended with retrospective effect from 23 October 2015, unless otherwise stated. References to Section 7(1)(b)(ii) of the Foreign Awards (Recognition and Enforcement) Act, 1961 concern the pre-1996 regime; that provision is the direct predecessor to Section 48(2)(b). This Part concludes the series; Part I addressed clause drafting and Part II the enforcement route for foreign awards in India. |
Introduction
| One defence survives when an award-debtor has exhausted the merits: that enforcement would offend the “fundamental policy of Indian law” under Section 48(2)(b). This Part argues that the positive content of that undefined phrase is already latent in binding authority — and that its name is private international law. |
Parts I and II carried the foreign-seated award from the drafting of its clause to the door of the Indian enforcement court. This part is about the last thing standing at that door. When the award-debtor has lost on the merits before the tribunal and failed to set the award aside at the seat, one objection still survives under Section 48(2)(b) of the Arbitration and Conciliation Act, 1996: that enforcement would offend the “fundamental policy of Indian law.” It is the defence that remains when every other defence is spent — and it is defined almost entirely by what it is not.
Two developments now bear on it. Nagaraj Mylandla v PI Opportunities Fund-I (2026 INSC 298; Supreme Court, 25 March 2026) has placed a procedural gate in front of the defence: a party cannot recolour a merits point already decided at the seat as a public-policy objection. Beneath that gate lies an older, unanswered question — what positively fills “fundamental policy of Indian law”? This part argues that the content is already latent in binding authority, and that its name is private international law.
- The Progression: Renusagar to Nagaraj
| Since Renusagar (1994), the Supreme Court has defined “fundamental policy of Indian law” only by exclusion — not merits review, not patent illegality, not every statutory breach — while reasoning, from the outset, in the language of private international law. Three decades of narrowing have policed the categories without ever defining the term. |
The phrase has never been positively defined. Across three decades the Supreme Court has said what “fundamental policy of Indian law” is not — not a licence to review the merits, not patent illegality, not every contravention of statute — while the same line of authority has reasoned, from the outset, in the language of private international law. The starting point is Renusagar.
In Renusagar Power Co Ltd v General Electric Co (Supreme Court, three-judge bench; 1994 Supp (1) SCC 644 / AIR 1994 SC 860), the Court construed “public policy” in Section 7(1)(b)(ii) of the Foreign Awards (Recognition and Enforcement) Act, 1961 — the direct predecessor to Section 48. Its reasoning ran in two steps. First, it fixed the register: because the enforcement of foreign awards is “governed by the principles of private international law,” the expression “public policy” had to be “construed in the sense the doctrine of public policy as applied in the field of private international law” (para 66) — the narrower sense that conflict of laws reserves for matters carrying a foreign element, not the wider sense in which Section 23 of the Contract Act treats public policy in a purely domestic dispute. Second, it drew the consequence: “contravention of law alone will not attract the bar of public policy and something more than contravention of law is required” (para 65).
From that narrower register the Court derived the test that still governs. Enforcement may be refused only where it would be contrary to “(i) fundamental policy of Indian law; (ii) the interests of India; or (iii) justice or morality” (para 66). Those three limbs are the ratio, and it is para 66 that the later cases cite. And they are avowedly India-specific: the Court held that a violation of FERA — a statute enacted to “safeguard the economic interests of India” — would engage the bar (para 76), though on the facts it found no such violation and enforced the award (para 84). A purely domestic statute could therefore qualify. Renusagar thus borrowed private international law’s instinct for narrowness — the reason public policy is read down when a foreign award is in issue — without adopting international public policy (ordre public international) as a limiting label. It took the field and left the narrower concept untaken: a point Section 4 returns to.
What follows Renusagar is best read not as a search for the missing definition — none of these decisions supplies one — but as a narrowing of the gate on two separate axes. One is substance: which categories of objection can ever pass. The other, reached only in 2026, is procedure: whether a party may raise the objection at all. On neither axis does a court say what “fundamental policy of Indian law” positively is.
ONGC v Saw Pipes Ltd (Supreme Court, 2003) read “patent illegality” into the public-policy ground: an award contrary to the substantive law of India could be set aside as offending public policy. That came close to the very equation Renusagar had refused — contrary to law, therefore contrary to public policy — but Saw Pipes was a Section 34 decision about a domestic award, and its breadth never really governed foreign-award enforcement. Shri Lal Mahal Ltd v Progetto Grano SpA (Supreme Court, three-judge bench, 2014) said so directly: overruling Phulchand Exports, it held that “public policy of India” in Section 48(2)(b) bears the narrow Renusagar meaning, that patent illegality is unavailable against a foreign award, and that the enforcement court may take no “second look” at the merits. (Shri Lal Mahal also adopted a passage of conflict-of-laws reasoning — that public policy is applied more sparingly where a foreign element is present — which Section 4 deploys; the verbatim passage is flagged below for verification.)
The 2015 Amendment then entrenched the narrow line and pruned it further. Explanation 1 to Section 48(2)(b) provides that an award conflicts with public policy “only if” its making was induced by fraud or corruption or violated Sections 75 or 81; or it contravenes the fundamental policy of Indian law; or it conflicts with the most basic notions of morality or justice. Explanation 2 forbids any review on the merits. Two things happened in that drafting. The open-ended standard became a closed list — the unruly horse was stabled. And the “interests of India” limb, present in Renusagar, was deleted; on the Law Commission’s own account the amendment was meant to restore Renusagar and undo Saw Pipes.
The pruning continued case by case. Vijay Karia v Prysmian Cavi e Sistemi SRL ((2020) 11 SCC 1; three-judge bench) shut the exchange-control route Renusagar had seemed to open: a rectifiable FEMA breach can never amount to a violation of fundamental policy, FEMA being a “management” regime whose contraventions are curable, as distinct from FERA’s prohibitory “policing” regime. Fundamental policy, the Court said, reaches only the core values so basic to Indian law that they cannot be compromised.
The subtractions, however, left one thing untouched. Because the term is undefined, a party that had lost a factual or contractual point at the seat could still dress that point as “fundamental policy” and compel the enforcement court to hear it out — the categories were policed, but the term itself remained undefined.
- Nagaraj Mylandla: A Gate That Relocates the Problem
| Nagaraj Mylandla (2026) adopts transnational issue estoppel: a party cannot recolour a merits point decided at the seat as a Section 48(2)(b) objection. But “forum-connected” is a sorting test, not a substantive standard — Nagaraj narrows access to the defence without supplying its content. |
Nagaraj V. Mylandla v PI Opportunities Fund-I supplies the procedural gate Section 1 anticipated. The Court adopted the four-condition test for transnational issue estoppel from Good Challenger Navegante SA v Metalexportimport SA [2003] EWCA Civ 1668 — a foreign court of competent jurisdiction; a judgment final, conclusive and on the merits; identity of parties; identity of subject matter — subject to the overriding qualification that estoppel “must work justice and not injustice.” Where the seat court has already decided an issue, the enforcement court will not permit it to resurface “by giving a different colour” to bring it within Section 48(2)(b).
The Court distinguished — adopting the taxonomy of the Singapore High Court in Sacofa Sdn Bhd v Super Sea Cable Networks Pte Ltd [2024] SGHC 54 — “forum-neutral” issues — contractual interpretation, factual findings, commercial characterisation: once decided at the seat, closed — from “forum-connected” issues — public policy, arbitrability, and basic notions of justice — which remain open regardless of the seat’s findings.
Nagaraj therefore does real work: it stops a losing party from relabelling a decided factual or contractual point as “public policy” merely to reopen it. But look at what “forum-connected” is actually doing. It is not a substantive account of what fundamental policy contains — it is a sorting test, and a negative one: an issue is forum-connected because it is not one of the forum-neutral categories Sacofa lists. Nagaraj polices who may raise the objection. It says nothing about what the objection must contain to succeed once raised.
This matters because the two axes are easy to conflate. A court could ask “is this forum-connected?”, answer yes, and stop — treating survival of the estoppel gate as though it were satisfaction of the fundamental-policy standard itself. It is not. Nagaraj has narrowed access to Section 48(2)(b); it has not supplied content for it. Absent some further filter, “forum-connected” will accumulate the same defect as “fundamental policy” before it — a category whose boundary is drawn case by case, by exclusion, because nothing tells a court in advance which domestic rules qualify as connected to the forum in the relevant sense and which do not. The gate has moved. The gap behind it has not closed.
That gap is what Section 3 turns to next.
- The Missing Filter: Why Exclusion-Only Tests Cannot Hold
| No Indian test separates statutes that genuinely engage fundamental policy from those that merely exist. Private international law supplies one: the overriding mandatory rules doctrine, codified at Article 9(1) of the Rome I Regulation — offered here as a comparative model, not as authority any Indian court has adopted. |
An objection can be forum-connected and still not tell a court whether it should succeed. Nothing in Sacofa’s taxonomy, or in Renusagar’s list of exclusions, supplies a rule for separating the Indian statutes that genuinely engage “fundamental policy” from the ones that merely exist. Left unfiltered, the category runs in one direction only: every domestic mandatory provision — tax, exchange control, companies law, sector regulation — becomes a candidate, and each enforcement court must reinvent the boundary on the facts before it. Renusagar’s “something more than mere contravention of law” (para 65) names the requirement without saying what supplies it. That is the gap this section identifies — filling it is Section 4’s task.
France supplies the oldest developed answer. Loi de police is a French private international law doctrine, not a legal maxim — French courts distinguished these laws from ordinary mandatory rules as early as the nineteenth century, on the ground that some domestic provisions are so bound up with a state’s own political or economic order that they apply regardless of what law would otherwise govern the contract. The European Union later codified an analogous test at Article 9(1) of the Rome I Regulation (Regulation (EC) No 593/2008): a provision qualifies as “overriding” where a state regards compliance as crucial to safeguarding its public interest — its political, social or economic organisation — to the extent that the provision applies irrespective of the otherwise-applicable law.
Rome I itself has no bearing on arbitral enforcement: it governs choice of law for contractual disputes before EU member state courts, expressly excludes arbitration agreements from its scope, and does not bind arbitral tribunals at all. What is borrowed here is narrower and more precise — Article 9(1)’s definitional test for what makes a domestic rule “overriding” — transplanted from a choice-of-law context into a recognition-and-enforcement one. The transplant is the article’s proposed move, not an application of Rome I’s own field of operation.
Two further points bear on how the definition is used here. First, Rome I is EU law, directly applicable only within the Union; it is not a convention India has signed or could accede to, and no Indian court has considered it. Second, the definition is deliberately non-exhaustive — the drafters qualified the listed interests with “such as,” and courts applying Article 9(1) since have exercised case-by-case discretion rather than working from a closed checklist. That openness appears to be a designed feature, not an oversight: public, economic and social interests vary across states and change over time, and a rigid list would foreclose recognition of interests the drafters could not have anticipated. This is inference from the drafting choice and its subsequent application, not a quoted statement of legislative intent, and is flagged as such.
The clearest settled illustrations of the category are competition law, capital-markets regulation (insider-trading and market-integrity rules), exchange-control legislation protecting a state’s currency, and trade embargoes pursued for reasons of political or security policy. Consumer-protection and labour-law provisions are sometimes discussed in the same breath, but their status is contested even within EU jurisprudence — the European Court of Justice’s extension of the doctrine to commercial-agency compensation in Ingmar has been criticised by national courts and commentators as a doubtful characterisation, and the relationship between Rome I’s separate consumer/employment protections and Article 9 itself remains unsettled. They are omitted here for that reason.
The distinction is instructive for India not because it transplants but because it demonstrates a working method: flexibility can coexist with structure. The EU’s openness operates inside a codified test with a decade and a half of judicial application behind it. India currently has neither the structure nor the openness — only Renusagar’s undischarged “something more,” with no framework built around it at all. Article 9(1)’s design is offered here as a comparative model for what such a framework could look like, not as authority India has adopted or that Indian courts have left open. No Indian judgment has cited Rome I, the loi de police doctrine, or the ILA’s parallel 2002 Resolution in a Section 48 context. The synthesis that follows in Section 4 is this article’s own proposed answer to that gap, argued by analogy — not a report of settled or contested Indian law.
- The Filter, Applied
| The proposed filter runs two stages: first, is the objection forum-connected, or a repackaged merits point closed by estoppel; second, if forum-connected, is the rule invoked genuinely overriding, or an ordinary mandatory rule whose breach Renusagar held insufficient? Only three fact patterns credibly pass both stages. |
The test the last three sections have been building toward is now stated: does the objection survive Nagaraj’s estoppel gate, and if it does, does it survive private international law’s substantive filter for what actually counts as “fundamental policy”?
Stage 1 — is the objection forum-connected? Per Section 2’s Nagaraj/Sacofa taxonomy: an issue the seat court already decided on the merits — contractual interpretation, factual findings, commercial characterisation — is forum-neutral and closed. Public policy, arbitrability, and basic notions of justice are forum-connected and remain open regardless of what the seat decided. An objection that fails Stage 1 is stopped there; it never reaches Stage 2.
Stage 2 — is the rule invoked an overriding mandatory rule? An objection that clears Stage 1 still faces a question Nagaraj never reaches: is the Indian rule behind it one a court regards as so bound up with India’s political, economic, or social order that it applies irrespective of the otherwise-governing law — a test of the kind Article 9(1) supplies, discussed in Section 3 — or is it an ordinary domestic mandatory rule, breach of which Renusagar already held insufficient absent “something more” (para 65)? The same line separates fundamental from merely domestic policy in the international commentary: the ILA’s 2002 Resolution recommends that breach of a mandatory rule not forming part of a State’s fundamental public policy should not bar enforcement, even where the rule belongs to the law of the forum. Indian courts have not adopted the label “international public policy” — Cruz City records their scepticism about any “workable definition” — so the Resolution is invoked here only as external corroboration of a distinction Renusagar and Vijay Karia already draw in domestic terms, not as a standard India follows. No Indian court has cited the Resolution or a test of the Article 9(1) kind in a Section 48 case; the mapping onto Renusagar and Vijay Karia is this article’s proposed synthesis, not reported doctrine.
Three fact patterns test the filter.
Undiscoverable arbitrator non-disclosure. In MSA Global LLC (Oman) v Engineering Projects (India) Limited (Delhi High Court, 2026), the enforcement court refused enforcement under Section 48(2)(b) because the arbitrator had failed to disclose a prior arbitral association with the claimant’s Chairman. The objection is forum-connected — it concerns the integrity of the arbitral process itself, not a fact the tribunal or seat court decided — and it engages a value courts across jurisdictions treat as basic (impartiality of the adjudicator). MSA Global should be cited only for that refusal; it does not establish, as sometimes reported, that non-disclosure “survives” issue estoppel as a general proposition.
Fraud to the root, undiscoverable within seat limitation. The same structural point applies: fraud discovered only after the seat proceedings concluded was never available to the seat court, so no forum-neutral finding exists for it to be recoloured against. Stage 1 is satisfied because there is nothing to be estopped from; Stage 2 is satisfied because fraud vitiating the award sits squarely within Explanation 1(a)’s express fraud ground.
A genuine overriding mandatory rule not determined at the seat. The case an award-debtor will reach for is NAFED v Alimenta SA (Supreme Court, three-judge bench, 2020), which refused enforcement of a FOSFA award on the ground that supply in breach of the Government’s export control would offend the fundamental public policy of India relating to export. NAFED has not been overruled; it remains a three-judge-bench decision on the books. But its route, not its label, is the difficulty. Whether the export restriction was genuine or self-induced was a question of fact, and — on the tribunal’s findings as recorded in the enforcement proceedings — the FOSFA tribunal had decided it against NAFED. In the author’s analysis, the Supreme Court could reach its Section 32 conclusion (contingency triggered, contract void, no damages) only by displacing that finding. Read against Nagaraj, that displacement is instructive: a factual finding of this kind is now the paradigm of a forum-neutral issue closed at the seat, and an objection that depends on reopening it is, after Nagaraj, materially less likely to be sustained — whatever public-policy label is attached to it. NAFED thus survives as an illustration of the category the filter is meant to police, and as a caution that the label can be misapplied even at the highest level.
[Verification note — remove before publication: the tribunal’s self-inducement finding and the precise scope of the Supreme Court’s treatment of it are drawn from the reported proceedings and secondary analysis; both should be confirmed against the primary award and judgment before the point is pressed in any filing.]
A related question — untouched by Nagaraj — is whether a party who could have raised an objection at the seat but stayed silent may raise it for the first time at enforcement. The judgment’s ratio rests on issues actually decided; whether “ought to have been raised” independently bars a party is, per contemporaneous commentary, unresolved. The stronger route to the same result avoids resting on issue estoppel at all: waiver, abuse of process, and the residual Section 48 discretion each supply a basis for treating undue silence as costly, reinforced — by analogy, not direct authority — by the settled principle that what cannot be done directly cannot be done indirectly.
- The Practitioner’s Protocol
| For award-holders: plead the seat determination first, run the filter in the petition, and look behind every public-policy label. For award-debtors: only three grounds credibly survive; anything else invites exemplary costs. For drafters: make every argument at the seat — after Nagaraj, silence is priced. |
The two contributions of this Part — the substantive filter, and the estoppel gate it sits behind — are worth only what a practitioner can do with them. This section converts them into instructions for each side of the enforcement contest, and for the drafter who, by the time enforcement is live, acted years earlier.
For the award-holder’s counsel.
Plead the seat determination first. Before arguing that an objection fails the fundamental-policy standard, argue that the debtor cannot raise it at all. Set out what the seat court or tribunal actually decided, and invoke transnational issue estoppel to close every forum-neutral issue — contractual construction, factual findings, commercial characterisation — at the threshold. Stage 1 is cheaper than Stage 2; win there first.
Run the filter in the petition, not the reply. Anticipate each Section 48(2)(b) objection the debtor is likely to raise, and pre-empt it in the enforcement petition itself, characterised through the two-stage filter. Leaving it to the reply cedes the debtor the opening frame.
Isolate ordinary mandatory rules from overriding ones. Most statutory-breach objections — exchange control (Vijay Karia), stamping (the Shriram EPC obiter), and their like — invoke ordinary domestic mandatory rules that do not, without more, engage fundamental policy. Say so expressly, and reserve the fundamental-policy standard for the narrow class of genuinely overriding rules the framework in Sections 3 and 4 identifies.
Look behind the label. A public-policy objection is frequently a factual or contractual dispute the debtor lost at the seat, re-dressed in the language of Section 48(2)(b). Before conceding that an objection is forum-connected, ask what the tribunal actually decided and whether the objection can succeed without reopening it. If it cannot, it is a forum-neutral point (Stage 1) and should be met there, not on the merits. NAFED illustrates how readily the two are conflated: in the author’s analysis, the objection there turned on a factual finding of self-inducement the tribunal had already made, yet it was sustained under a public-policy heading. The lesson is not that the surviving grounds are illusory, but that they must be isolated from the merits dispute they are so often made to carry.
For the award-debtor’s counsel.
Only three grounds credibly survive the gate: undiscoverable arbitrator non-disclosure; fraud going to the root, undiscoverable within the seat’s limitation period; and a genuine overriding mandatory rule the seat never determined. If the objection is not one of these, it should not be run. After Nagaraj, a recoloured merits point does not merely fail — it draws exemplary costs. Speculative resistance is now a priced risk, not a free option.
For the drafter.
This is where Parts I and II close the loop. Every argument the party might later want to make in India must be made at the seat: silence is now costly, and the estoppel gate will not reopen for a point that could have been taken and was not. The clause itself does work here too — Part I’s informed-choice provision (Element 7), by recording the parties’ deliberate selection of seat and governing law, forecloses at source the opportunistic objection that a party never understood what it agreed to.
Clause, court, doctrine. Across three Parts, the seat has carried more weight than it is usually credited with. In Part I it was the foundation — the clause fixing where the award is made and which law governs it. In Part II it set the enforcement route into India. Here it acquires a third function: it fixes the weight of estoppel, because what the seat decides is what the Indian court will not reopen. The clause is the foundation, the protocol the structure, the doctrine the roof. A foreign-seated award drafted and argued with all three in view arrives at the Indian court’s door with very little left standing against it.
- Services We Provide
Disclaimer: This article is for general informational purposes only and does not constitute legal advice or create an advocate–client relationship. The analysis reflects the law as at the date of publication and includes the author’s own proposed synthesis, expressly flagged as such where it goes beyond decided authority. Readers should obtain specific professional advice before acting on any matter discussed. R & D Law Chambers LLP accepts no liability for actions taken in reliance on this article.