Authors: Kshitij Saruparia & Apeksha Kachhawaha*
| Jurisdictions:
|
Topics: |
I. Introduction
On 29 May 2026, the Supreme Court of India ordered a state-owned road corporation to pay an arbitral award rendered in August 2014, under a concession agreement terminated in July 2007. Madhya Pradesh Road Development Corporation Ltd. v. Jabalpur Corridor Pvt. Ltd., 2026 INSC 590, has circulated almost entirely on the strength of its opening paragraph, where Justice J.K. Maheshwari writes that arbitration in India has not failed but that courts have sometimes failed arbitration, and that judicial interference in alternative dispute resolution has often been “a cure without a disease”. The aphorism has already been quoted, reposted, and set beside a decade of similar exhortations.
The passage that deserves the attention of an international audience sits at paragraphs 21 to 23, and it has attracted almost none. There the Court records that the award creditor, a special purpose vehicle incorporated by a Malaysian construction company, is a protected indirect investment under Article 2 of the India-Malaysia bilateral investment treaty of 1995; that although India terminated that treaty in 2017, Article 14(4) preserves protection for existing investments for a further ten years; that the investor has standing to seek compensation under it; and that the Malaysian High Commission had taken up the delayed adjudication with India’s Ministry of External Affairs through diplomatic channels. The Court then cites SAIPEM S.p.A. v. Bangladesh, ICSID Case No. ARB/05/07, in which a tribunal found that interference by the Bangladeshi judiciary with an ICC award amounted to expropriation through a denial of justice. The bench is explicit that it is not deciding a treaty claim. It is equally explicit that the submissions are “not a mere red herring” but an unfortunate reality check on the judicial handling of arbitration in India.
This comment makes three claims. First, those paragraphs are the first instance of the Supreme Court of India reasoning about its own supervisory jurisdiction under Sections 34 and 37 of the Arbitration and Conciliation Act, 1996 in the register of state responsibility: the set-aside court as an organ of the State whose conduct is attributable to India on the international plane. Second, the precedent the Court chose is not the closest one on the facts, and the closer authority, White Industries v. India, goes uncited; both the choice and the omission repay attention, though not in the way an initial reading suggests. Third, the internalisation does real work, of an interpretive rather than doctrinal kind, and its most durable audience sits outside India’s courtrooms altogether, where it will be read in two directions at once.
II. What the Court Decided
The dispute has a long tail. MPRDC, wholly owned by the State of Madhya Pradesh, awarded a build-operate-transfer road concession in 2002 to a Malaysian bidder, which incorporated Jabalpur Corridor Pvt. Ltd. as its project vehicle and financed construction through a loan of Rs 80.85 crore from EXIM Bank Malaysia. Delays in handing over encumbrance-free land stalled the project; MPRDC terminated the concession in July 2007; arbitration commenced in 2011; and in August 2014 a majority award granted the concessionaire a termination payment of approximately Rs 49.47 crore, with pre-award interest at the contractual rate of 14.75 percent and post-award interest at the then statutory 18 percent. The Section 34 application failed in February 2016, the Section 37 appeal failed in December 2016, and the corporation’s special leave petition then sat before the Supreme Court for nine years.
The judgment disposes of three questions. On jurisdiction, MPRDC’s objection that the concession was a works contract falling exclusively under the Madhya Pradesh Madhyastham Adhikaran Adhiniyam, 1983 had been rejected by the arbitral tribunal under Section 16, relitigated through Section 14 proceedings, decided against the corporation by the High Court in 2013, and closed when the Supreme Court dismissed the special leave petition against that decision and, in March 2016, the review. MPRDC revived the point in June 2017 in a rejoinder affidavit, after a Full Bench of the High Court overruled the 2013 judgment in Viva Highways. The Court held the issue final between the parties: a later change in the law does not reopen adjudications that have attained conclusiveness, and raising the plea in a rejoinder at the fourth tier was an abuse of process. Following M.P. Rural Road Development Authority v. L.G. Chaudhary (2018) 10 SCC 826 and Gayatri Project Ltd. v. M.P. Road Development Corpn. Ltd. (2025) 10 SCC 750, an award predating the change of position could not in any event be annulled on that ground alone.
On the merits, the Court treated the tribunal’s construction of the termination payment clauses as at the least a plausible view, protected by three concurrent findings below and by what the judgment calls a “narrowing pyramid”: with each appellate layer, the threshold for disturbing an award rises. This is orthodox doctrine, restated from Associate Builders v. Delhi Development Authority (2015) 3 SCC 49 and Ssangyong Engineering v. NHAI (2019) 15 SCC 131 through the Section 37 line of cases.
On interest lies the judgment’s quiet doctrinal news. MPRDC invoked the Constitution Bench in Gayatri Balasamy v. ISG Novasoft Technologies Ltd., 2025 INSC 605, which held by a four-to-one majority that courts possess a limited power under Sections 34 and 37 to modify awards, extending in some circumstances to post-award interest, and asked the Court to reduce the rates. The Court accepted that the power exists and refused to exercise it: the pre-award rate was the parties’ own bargain, claimed by MPRDC itself at the identical figure in its counterclaim, and the corporation’s conduct across nineteen years of resistance disentitled it to the discretion. Balasamy’s first prominent appearance in the restraint jurisprudence is thus as a discretion withheld from a dilatory state litigant, not as a new instrument for award debtors.
III. The International Law Turn
What distinguishes MPRDC from the Court’s earlier restraint decisions is where it locates the reason for restraint. Paragraphs 21 to 23 are structured as a justification: because the award creditor is a treaty-protected foreign investment, because a supervisory judiciary can itself commit an internationally wrongful act, and because the Court’s recent decision in State of U.P. v. Reliance Industries Ltd., 2026 SCC OnLine SC 864, authored by the same judge, recognises that transactions involving foreign investment carry an inherent expectation of stability in the rule of law, the Court must take care to apply the 1996 Act in a non-discriminatory manner. The international material is not decoration. It supplies the premise for the strictness with which the judgment then patrols the boundaries of Sections 5, 34 and 37, and for the diagnosis at paragraph 74 that a hesitancy in some courts, rooted in suspicion of the arbitral process, produces re-examination of evidence and re-interpretation of contracts at the set-aside and appellate stages.
Extra-judicial statements in this register are not new. Chief Justice Surya Kant told the Indian Council of Arbitration’s international conference in April 2026 that excessive judicial intervention risks weakening faith in arbitration as a reliable method of dispute resolution. Judgments are different instruments. A reportable decision of the Supreme Court citing an investor-state award, in the course of explaining its own standard of review, becomes part of the material every Section 34 court is expected to reckon with.
IV. The Precedent Not Cited
The authority the Court chose is instructive, and so is the one it did not. SAIPEM was an annulment case. The Bangladeshi courts revoked the authority of an ICC tribunal mid-proceeding and later treated its award as a nullity, and the ICSID tribunal found an expropriation because judicial measures had destroyed the award’s value. Nothing comparable happened here. The Indian courts did not destroy the award; they upheld it at every stage of the challenge, from the Section 34 court through to the Supreme Court. The pathology in MPRDC was not judicial destruction of an award but time: nineteen years from termination to an enforceable direction to pay, twelve of them after the award.
For delay, the precedent squarely in point is White Industries Australia Ltd. v. India, the UNCITRAL final award of 30 November 2011 and still the only publicly known investment-treaty ruling against India arising from its courts’ handling of a commercial award. White Industries complicates the Court’s gesture. The tribunal there held that nine years of Indian proceedings over an ICC award, including a jurisdictional appeal the Supreme Court had not managed to hear, did not amount to a denial of justice: that standard is exacting, and the conduct of the individual courts disclosed no bad faith. India was held liable instead under the less demanding obligation to provide “effective means of asserting claims and enforcing rights”, imported into the Australia-India treaty through its most-favoured-nation clause from the India-Kuwait BIT.
India then closed that route. The 2016 Model BIT, drafted in substantial reaction to White Industries, abandoned the MFN clause and contains no free-standing effective-means guarantee; what survives in its closed-list fair and equitable treatment provision is denial of justice, the very standard the delay claim in White Industries failed to satisfy. Stated honestly, the doctrinal position is this: a claim by the Malaysian investor on the MPRDC facts would have had to proceed under the 1995 treaty’s older standards within the sunset window, and would have confronted the same stringent denial-of-justice threshold that defeated a claim built on comparable delay; for future investors under India’s current architecture, delay alone is harder still to litigate. The Court’s own judgment is, moreover, the strongest answer to any prospective claim, since the system, however slowly, ultimately delivered.
Why, then, cite SAIPEM rather than White Industries? The uncharitable reading is that the Court reached for the more alarming authority and would have found the more analogous one inconvenient, since citing White Industries means acknowledging that India has already been held liable for something adjacent to this pathology and rewrote its treaty programme in response. That reading is available. It is not, on reflection, the best one. Read against what the Court was actually doing in paragraphs 21 to 23, the choice is defensible on grounds that are strategic without being evasive, and the distinction matters.
The two authorities operate at different levels of abstraction, and only one of them is pitched at the level the Court needed. White Industries holds that delay of that magnitude did not cross the denial-of-justice threshold, and that India’s liability arose instead under an effective-means guarantee reached through a most-favoured-nation clause, a route India has since closed. Its operative proposition is therefore doubly ill-suited to a judgment addressed to Section 34 courts: it says that judicial delay was not, on those facts, an internationally wrongful act, and the standard that did produce liability no longer exists in India’s treaty architecture. A bench seeking to impress on subordinate courts that their supervisory conduct carries consequences under international law cannot build that on an authority whose holding, read plainly, is reassuring. SAIPEM supplies what is needed: a supervisory court can itself commit an internationally wrongful act. That is a proposition about the character of judicial conduct rather than about India’s quantified exposure, and it is the proposition a Section 34 judge has to internalise. On this reading the citation is not a substitution of alarm for accuracy but a selection of the authority that carries the relevant doctrinal content, and the Court’s own express disclaimer that no treaty claim is before it is consistent with a bench that knows the difference between the two registers.
The omission of White Industries remains an omission, and it is not costless. A judgment that wanted to make the strictest possible case for restraint could have cited both, using White Industries to establish that Indian courts have already generated an adverse award and SAIPEM to establish what supervisory conduct can amount to at its worst. That the Court took only the second limb is a choice about how much of the history to put on the record, and it is fair to notice it. But noticing it does not carry the further inference that the Court was concealing anything. Both readings survive the text; the strategic reading survives it better.
The selectivity of the internalisation shows more clearly in a second omission: Devas v. Antrix. Within the past four years, the same Court held a USD 562.5 million ICC award in favour of foreign-backed investors to be the fruit of fraud, the Delhi High Court set the award aside on the strength of those findings, and the Supreme Court declined to interfere in October 2023, all while treaty tribunals under the India-Mauritius and India-Germany BITs had already ruled for the investors and a Dutch appellate court would later refuse to recognise the annulment. The claim here is about framing, and it should be stated narrowly, because the alternative claim is not one the record supports. Fraud is a legitimate ground of annulment in every major system, including under Article V(2)(b) of the New York Convention, and the treaty tribunals that ruled for the Devas investors did so on substantive treaty standards rather than by holding the annulment itself internationally wrongful. Nothing in this comment suggests that the state-responsibility register should have produced a different outcome in Devas, or that the Court there was obliged to reason as though the set-aside were itself an internationally wrongful act. The point is narrower and, for that reason, harder to answer. The vocabulary of state responsibility is available to the Court whenever it chooses to reach for it, and the record shows it reaching for that vocabulary when the Court is enforcing an award in favour of a foreign investor and leaving it untouched when the Court is upholding an annulment against one. The asymmetry is one of register, not of result. It tells us something about when the Court finds it useful to describe itself as an organ of the State.
The judgment’s first paragraph contains a subtler attribution point. Even the Government’s role, the Court says, cannot be ignored. That sentence has attracted no commentary, and it may be the most technically accurate line in the passage, because it locates the attribution question where international law actually locates it. Much of the delay in MPRDC was not the judiciary’s doing but the state litigant’s. A wholly state-owned corporation raised serial jurisdictional objections, pursued a Section 14 application when the Section 16 route failed, sought review after losing in the Supreme Court, and produced its final jurisdictional plea for the first time in a rejoinder affidavit at the fourth tier. In attribution terms this matters, because the litigation conduct of a state entity resisting payment is more straightforwardly chargeable to the State than the pace of independent courts. Whether the bench arrived at that sentence by doctrinal route or by rhetorical instinct is unknowable from the judgment and, for present purposes, immaterial. Its precision is a property of the sentence, not of the intention behind it, and a future tribunal reading the passage will take it as it finds it: an apex court recording that the State’s own litigant, not only its courts, produced the delay. Read against paragraphs 21 to 23, the sentence acknowledges that the appellant before the Court was, for treaty purposes, India.
V. What the Internalisation Does, and Does Not Do
The judgment creates no new ground of challenge and no new standard of review. Investment stability is not a criterion under Section 34, foreign-invested awards do not enjoy a different test, and any argument built on differential treatment would collide with the Court’s own insistence that the 1996 Act be applied in a non-discriminatory manner. Readers should also resist treating paragraphs 21 to 23 as a holding; the Court decided the appeal on issue estoppel, plausible-view deference, and party autonomy in the matter of interest, and said in terms that no treaty claim was before it.
What the passage supplies is an interpretive premise and an argumentative instrument. Section 5 confines intervention to what Part I of the Act provides, and the seven-judge bench in In re: Interplay Between Arbitration Agreements and the Stamp Act (2024) 6 SCC 1 directs that every provision be construed in light of that principle. MPRDC adds an account of why the principle carries the weight it does: the supervisory court acts for the State, and its conduct is ultimately measurable against standards the State does not author. Counsel resisting a merits-flavoured Section 34 challenge, or a Section 37 attempt to reopen concurrent findings, can now put before the court a Supreme Court judgment that connects the statutory limits to India’s international responsibility, delivered while enforcing an award in favour of a treaty-protected investor whose government had escalated the matter diplomatically. Associate Builders states the rule; MPRDC states the stakes.
Whether the courts that generate the conduct described in paragraph 74 will be moved is an empirical question, and a decade of restatements counsels modesty. The more interesting question is what the passage does when it is read abroad, and here the piece’s own account of the doctrine cuts against a simple answer. If a Malaysian claimant brought the MPRDC facts to a tribunal within the sunset window, India’s first submission would be this judgment. The award was upheld at every stage. The system, slowly, delivered. The apex court identified the pathology, named it, and corrected for it in the very case complained of. That is close to what the White Industries tribunal found dispositive in declining to find a denial of justice: the courts functioned, and the exacting standard was not met. Paragraphs 21 to 23 are, on that reading, exculpatory. They show a judiciary that monitors this precise failure.
The defensive reading is real and this comment does not resist it. It is also incomplete, because it takes the paragraphs as evidence of correction and ignores what they concede in order to perform the correction. A denial-of-justice claim is defeated by showing that the courts worked. It is not the only claim a tribunal will hear, and it is not the register in which a treaty argument about systemic delay is likely to be pitched. What paragraphs 21 to 23 record, over the signature of the Supreme Court of India, is that the delay is real, that it is systemic rather than incidental to one docket, that courts and government have both contributed to it, and that the Court is aware this pathology is measurable against standards India does not author. An admission of that kind survives the case it was made in. It is available to the next claimant regardless of what happened to this award creditor, and it forecloses the argument that Indian delay is anecdotal or that the criticism is a foreign construction. India can no longer say that. Its own apex court has said otherwise, in a reportable judgment, while citing an investment-treaty award.
The judgment therefore travels outward in two directions, and the tension between them is not a flaw in the reading but a feature of the text. On the standard, it is a defence: the corrective mechanism worked, and worked in this case. On the facts, it is an admission: the pathology exists, and the State’s highest court has certified it. Which limb a future tribunal reaches for will depend on the claim before it, and India will not always get to choose.
VI. Conclusion
MPRDC will be remembered for its aphorism, and the aphorism earns its circulation; a court rarely says of its own hierarchy that its interventions have supplied a cure without a disease. The judgment’s contribution to the field lies elsewhere. For the first time, the Supreme Court of India has examined its own supervisory function through the eyes of an investment tribunal and written down what it saw, choosing the authority pitched at the level of abstraction its domestic audience required and leaving the liability analysis deliberately unfinished. The discipline the judgment imposes is addressed inward, to the Section 34 and Section 37 courts, and it is the more credible for resting on a proposition about what supervisory courts are rather than on a warning about what India once paid. What travels outward is less obedient. The Court has produced, in the same paragraphs, the best evidence India will have that its judiciary corrects this pathology and the best evidence a claimant will have that the pathology is real and known. It cannot be read for one without being available for the other. Judgments of this kind do not settle what they will be used for.
* Kshitij Saruparia is an Advocate practising in India and a graduate of NALSAR University of Law, Hyderabad. Apeksha Kachhawaha is an Associate at ASV Legal, Delhi and a graduate of Maharashtra National Law University, Nagpur.