I have long argued – in articles, blog posts, and amicus briefs – that it violates due process to invoke a forum selection clause to obtain personal jurisdiction over a defendant who was not a party to the agreement in which the clause appears. This position has not yet achieved universal acceptance. The state courts in New York, in particular, have repeatedly held that forum selection clauses can be used to assert personal jurisdiction over non-party defendants who are “closely related” to the parties or the transaction. In this blog post, I use a recent case—Bandari v. QED Connect Inc.—decided by Magistrate Judge Gary Stein (SDNY) to highlight some of the problems with the “closely related” test.

The dispute in Bandari grew out of a stock purchase agreement. The plaintiff, Jalandher Bandari, was a resident of Texas. He agreed to purchase shares in QED Connect, Inc., a New York holding company, from David Rumbold, a resident of Illinois. The sale was orchestrated by Nanny Katharina Bahnsen, the chief executive officer of QED and a resident of Colombia. There were three parties to the stock purchase agreement: Bandari, Rumbold, and QED. (Bahnsen signed the contract on behalf of QED.) The agreement contained an exclusive forum selection clause choosing the state and federal courts sitting in New York City.

Although Bandari tendered the purchase price (approximately $150,000), he never received the shares he was promised. When Bandari asked for his money back, Bahnsen made excuses and eventually stopped responding to his emails. Bandari subsequently brought a lawsuit in federal court in New York against QED, Rumbold, and Bahnsen. After none of the defendants appeared to defend the suit, Bandari moved for a default judgment.

The federal courts in New York will not grant a default judgment until they determine that personal jurisdiction exists. The court quickly concluded that it had personal jurisdiction over Rumbold and QED because they had signed the contract containing the New York forum selection clause. The court then went on to conclude—wrongly, in my view—that Bahnsen was also subject to personal jurisdiction in New York because she had negotiated the sale and signed the contract on behalf of QED:

A party to a contract with a forum-selection clause may invoke that clause to establish personal jurisdiction over a defendant that is not party to the contract but that is “closely aligned” with a party, or “closely related” to the contract dispute itself, such as corporate executive officers. As the CEO of QED and the individual who negotiated the transaction with Bandari and signed the Agreement on behalf of QED, Bahnsen is “closely related” to both a party to the Agreement and to the dispute. Thus, she is also bound by the forum selection clause.

This conclusion is inconsistent with basic principles of agency law; an agent is not a party to a contract that the agent signs on behalf of a disclosed principal. It is inconsistent with basic principles of contract law; a person may not be bound by an agreement without their express consent. And it is inconsistent with basic principles of personal jurisdiction; a person who lacks minimum contacts with the forum is not subject to personal jurisdiction unless she consents. Nevertheless, the court concluded that Bahnsen was subject to personal jurisdiction in New York because she was “closely related” to the parties and the transaction.

This conclusion is made all the more jarring by that fact that the court also held that Bandari had failed to state a valid claim for breach of contract against Bahnsen because she was not a party to the agreement. In the court’s words:

[A]lthough Bandari’s breach of contract claim is asserted against all three Defendants, there is no basis for a finding of contract liability as to Bahnsen. Bahnsen is not a party to the Agreement and she signed the Agreement solely on behalf of QED. It is well established that a corporate officer who signs a contract on behalf of the corporation cannot be held personally liable for the corporation’s breach, absent a showing that the officer was the alter ego of the corporation. The Complaint does not adequately plead an alter ego theory of liability against Bahnsen and hence it does not state a viable breach of contract claim against her.

The court held, in other words, that Bahnsen (1) was subject to personal jurisdiction in New York by operation of the forum selection clause, but (2) could not be held liable for breach of contract because she was not a party to the agreement containing the forum selection clause. The hand that authored the personal jurisdiction section of the opinion was seemingly unaware of what the hand that authored the breach of contract section of the opinion was doing.

One can, of course, reconcile these conflicting statements by taking the position that forum selection clauses are not subject to the usual rules of agency law, contract law, and personal jurisdiction. There are, however, constitutional problems with such an approach. Under this line of reasoning, a person residing in a foreign country (Colombia) is subject to personal jurisdiction in New York when she negotiates and signs a contract that contains a New York forum selection clause on behalf of the entity that employs her even though she is not the alter ego of the company and is not herself a party to the agreement. These actions are, in my view, insufficient to subject her to personal jurisdiction in New York.

Although the court declined to enter a default judgment against Bahnsen on the claim for breach of contract, it did enter a default judgment against her on the plaintiff’s claims for securities fraud and common law fraud. A contract to which she was not a party, therefore, paved the way for the assertion of jurisdiction and the imposition of liability. New York has long sought to attract litigation business from around the world. It has been largely successful in those efforts. If that state continues to assert personal jurisdiction over foreign executives merely because they negotiate and sign contracts in their corporate capacity, however, one wonders whether these executives may start directing the company’s attorneys to choose another jurisdiction.

[This post is cross-posted at Transnational Litigation Blog.]

This guest post is posted on behalf of Kamakshi Puri, Senior Associate at Cyril Amarchand Mangaldas, Delhi, India, and dual-qualified lawyer (India and England and Wales).

The Singapore High Court recently clarified the scope of the court’s jurisdiction over foreign non-parties to the arbitration. In an application to set aside two interim injunctions, in Alphard Maritime Ltd. v Samson Maritime Ltd. & Ors. (2025) SGHC 154,[1] the court held that the the seat per se did not confer jurisdiction against non-parties to an arbitration, and that jurisdiction would first have to be established through regular service-out procedures before the seat court could grant an injunction against a non-party.

Factual Background

Briefly, the applicant, Alphard Maritime (“Alphard”), initiated SCMA arbitration[2] against its debtor, Samson Maritime (“Samson”), and Samson’s wholly owned subsidiary, Underwater Services (“Underwater”), for alleged breach of a settlement agreement for the sale of approx. nine vessels and Samson’s shareholding in Underwater to Alphard (“Subject Assets”). Alphard initiated arbitration upon receiving information of the pledge/mortgage of the Subject Assets to J M Baxi Marine Services (“Baxi”) in breach of the Settlement Agreement. In addition to the ex-parte freezing order against Samson and Underwater, Alphard had received from the seat court, acting in support of the arbitration, an ex-parte prohibitory injunction restraining Baxi and other creditors of Samson from assisting in or facilitating the dissipation of, or dealing with, any of Samson and Underwater’s assets worldwide. Baxi was not a party to the Settlement Agreement. While one of the defendants was based out of Singapore, Samson and Underwater were bound by the jurisdiction conferred to the seat court; however, Baxi was a foreign non-party to the arbitration.

While the interim freezing injunction against Samson and Underwater was vacated on the finding that there was no evidence of dissipation or risk of dissipation of assets, and the court observed that there was no basis for the injunction which in effect prohibited Baxi and/or the lenders from asserting their own contractual rights or enforcing proprietary rights against Samson which pre-dated the Settlement Agreement, the injunction was vacated primarily on the finding that the Singapore court, as the seat court, had no jurisdiction over Baxi or the foreign lenders.

Seat Court’s Jurisdiction over Foreign Defendants

A court must have in personam jurisdiction to grant an injunction against a party. Under Singapore law, which follows the English law on jurisdiction, jurisdiction is based on service of proceedings, and the court assumes jurisdiction over a foreign party (not having a presence in Singapore and not having submitted to the proceedings) through permission for service out of the claims. [3] The court allows permission for service out where “the Singapore Court is the appropriate forum for hearing the proceedings”.[4] For the assessment of whether permission for service out should be granted, i.e., that Singapore Court is the appropriate forum, the claimant is required to meet the following three-prong assessment: [5]

1. A good arguable case that there is sufficient nexus with the Singapore court;
2. Singapore is the forum conveniens; and
3. There is a serious question to be tried on the merits of the claim.

The “sufficient nexus” refers to the connection between the court and the defendant and follows the logic that a party may only be called to a foreign court where they have a sufficiently strong connection to the state. Practice Directions 63(3)(a) to (t) set out “Factors” that guide as to the possible connection that the foreign defendant may have with the Singapore court. [6]

Alphard relied on 2 factors – first, PD 63(3)(d), a claim to obtain relief in respect of the breach of a contract governed by the laws of Singapore. This was held to be inapplicable, as Baxi was neither a party to the contract, nor committed any breach. Second, PD 63(3)(n) claims made under any other written law of Singapore. In this regard, it was contended that the claim against Baxi was under Section 12A of the International Arbitration Act, i.e., an exercise of the Singapore court’s power to grant an injunction against non-parties in support of Singapore-seated arbitration, which wide power ensured that non-parties did not collude with the defendants to frustrate the fruits of a claim. The court accepted PD 63(3)(n) as a relevant factor.

However, since sufficient nexus with the court is not enough for permission to service out, the court proceeded to the next equity, i.e., whether Singapore was the ‘forum conveniens’. Forum conveniens is an exercise in determining the most appropriate court for deciding the lis. It is the assessment of the connection of the dispute with the Singapore court. The ‘dispute’ here was the prohibitory injunction against Baxi. The court held that to be the ‘appropriate court’ for interim relief against a specific party, it required more than the arbitration being seated in Singapore. The seat court would be the appropriate court if the dispute with the specific party could be traced to the arbitration, or assets/obligations were substantially that of party to the arbitration, i.e.,

1. Was the non-party bound by the arbitration agreement even if it was not a party to the arbitration?
2. did the non-party hold assets in Singapore, which arguably belonged beneficially to a party to the arbitration (non-party was a trustee / pass-through for the assets)
3. was the non-party a corporate entity held/owned by the party to the arbitration, and therefore, did the dissipation of assets of the party amount to the dissipation of value of the party (merger of identity between the party and non-party)?

The Court held that in the absence of any of the above, the seat court would not be the de facto appropriate forum for injunctions against all non-parties even when the injunction is in aid of Singapore-seated arbitration. The court did not find any reason for Baxi, an entity pursuing its independent remedy against the Alphard, to be brought before the Singapore court.

Notably, Alphard had already pursued interim relief under Section 9 of the (Indian) Arbitration and Conciliation Act, 1996, against the Defendants, including Baxi, before the High Court of Bombay. [7] The Bombay High Court, acting further to its power for making interim orders for protection of the subject matter in arbitration, including in international commercial arbitration where the place of arbitration is outside India [8], granted a status quo injunction, including on Baxi, on further dealing in or creating any further third-party interests in the shares held by Samson in Underwater and a disclosure order in respect to the transaction for pledge created in favour of Baxi.

Concluding Thoughts

For the known benefits of enforcement and limited grounds of challenge of awards under Singapore law and before Singapore courts, foreign parties regularly opt for Singapore as the neutral seat of arbitration. In such cases, the only nexus of the dispute with the court is its designation as the seat court. Separately, arbitral tribunals do not have jurisdiction over non-parties to an arbitration; thus, courts assume adjudication for interim relief applications against non-parties to the arbitration. With this decision, the Singapore court has confirmed the non-seat court’s interference for interim reliefs where parties require protective orders vis-a-vis non-parties to the arbitration.

[1] Available here.

[2] Arbitration under the Singapore Chamber of Maritime Arbitration (“SCMA”) Rules.

[3] S. 16(1)(a)(ii) of the Supreme Court of Judicature Act 1969: “16.—(1)  The General Division has jurisdiction to hear and try any action in personam where — (a) the defendant is served with an originating claim or any other originating process — …(ii) outside Singapore in the circumstances authorised by and in the manner prescribed by Rules of Court or Family Justice Rules.”

[4] Rules of Court 2021, Rule 1(1) of Order 8 of ROC 2021 “1.—(1)  An originating process or other court document may be served out of Singapore with the Court’s approval if it can be shown that the Court has the jurisdiction or is the appropriate court to hear the action” .

[5] Supreme Court Practice Directions 2021, Para 63(2).

[6] Prior to 2021, this condition was similar to English law, i.e., the “Good and arguable case that a gateway applies”. While “gateways” have been done away with, the Practice Directives have set out a non-exhaustive list of factors (PD 63(3)(a)–(t)) which a claimant “should refer to” in order to meet the requirement under PD 63(2)(a). These factors mirror the gateways with were earlier found in the Rules of Court 2014. See Ardavan Arzandeh, The New Rules of Court and the Service-Out Jurisdiction in Singapore, (2022) Singapore Journal of Legal Studies 191–201.

[7] Alphard Maritime Ltd. v Samson Maritime Limited & Ors. Commercial Arbitration Petition (L) No.7499 of 2025, Order dated 02.04.2025, available here.

[8] Section 9 read with Section 2(2) of the Arbitration Act, 1996.

This guest post was written by Ezzatollah Pabakhsh, Master’s Student at the University of Antwerp

The European Union has taken an unprecedented step by regulating artificial intelligence (AI) through the EU AI Act, which is the world’s first comprehensive legal framework for AI governance. According to Recital 61, Article 6(2) and Annex III, 8(a), AI tools used in legal or administrative decision-making processes—including alternative dispute resolution (ADR), when used similarly to courts and producing legal effects—are considered high risk. These tools must comply with the strict requirements outlined in Articles 8 through 27.

These provisions are designed to ensure transparency, accountability, and respect for fundamental rights. This obligation will take effect on August 2, 2026, according to Article 113. Notably, the Act’s extraterritorial scope, as outlined in Articles 2(1)(c) and (g), applies to any AI system that affects individuals within the European Union. This applies regardless of where the system is developed or used. It also applies to providers and deployers outside the EU whose output is used within the Union. This raises a critical question: can non?compliance with the EU AI Act serve as a basis for courts in EU Member States to refuse recognition or enforcement of an arbitral award on procedural or public?policy grounds?[1]

Consider the following scenario: Two EU-based technology companies, one Belgian and one German, agree to resolve their disputes through US-seated arbitration. Suppose the ADR center uses AI-powered tools that do not comply with the EU AI Act‘s high-risk system requirements. How would enforcement of the resulting award play out before national courts in the EU?

This scenario presents a direct legal conflict. If the winning party seeks to enforce the award in a national court of an EU Member State, two well-established legal grounds for refusing enforcement may arise.[2] First, the losing party may invoke Article V(1)(d) of the 1958 New York Convention, together with the applicable national arbitration law. They could argue that reliance on AI systems that do not comply with the EU AI Act constitutes a procedural irregularity, as it departs from the parties’ agreed arbitration procedure and undermines the integrity of the arbitral process.[3] Second, under Article V(2)(b) of the Convention, the enforcing court may refuse recognition on its own motion if it finds that using non-compliant AI violates the forum’s public policy, especially when fundamental rights or procedural fairness are at stake.[4] The following section will examine these two scenarios in more detail.

Scenario 1: Procedural Irregularity under Article V(1)

Imagine that the ADR center uses an AI tool to assist the tribunal in drafting the award during the proceedings. This AI system uses complex algorithms that cannot produce transparent, human-readable explanations of how key conclusions were reached. The final award relies on these outputs, yet it offers no meaningful reasoning or justification for several significant findings. Furthermore, the tribunal does not disclose the extent to which it relies on the AI system, nor is there any clear evidence of human oversight in the deliberation process.

When the losing party in Belgium contests enforcement of the award, they invoke Article V(1)(d) of the New York Convention, arguing that the arbitral procedure did not align with the parties’ expectations or the applicable law. This objection is also found in Article 1721 of the Belgian Judicial Code (BJC), inspired by Article 36 of the UNCITRAL Model Law and, to a large extent, mirroring the grounds of Article V of the New York Convention. Among these, two are especially relevant to the use of AI in the arbitral process and are central to the objection in this case.

First, under Article 1721(1)(d), a party may argue that the award lacks proper reasoning[5], which violates a core procedural guarantee under Belgian law.[6]  This requirement ensures that parties can understand the legal and factual basis for the tribunal’s decision and respond accordingly.[7] In this case, however, the award’s reliance on opaque, AI-generated conclusions, particularly those produced by “black box” systems, renders the reasoning inaccessible and legally inadequate.[8] The EU AI Act further reinforces this objection. Articles 13, 16, and 17 require transparency, traceability, and documentation for high-risk AI systems. Meanwhile, Article 86 grants limited right to explanation for affected persons where a deployer’s decision is based on Annex III systems and produces legal effects. If an award fails to meet these standards, it may not align with Belgian procedural norms.

Second, under Article 1721(1)(e), a party may argue that the tribunal’s composition or procedure deviated from the parties’ agreement or the law of the seat. For example, if the arbitration agreement contemplated adjudication by human arbitrators and the tribunal instead relied on AI tools that materially influenced its reasoning without disclosure or consent, this could constitute a procedural irregularity. According to Article 14 of the EU AI Act, there must be effective human oversight of high-risk AI systems. Where such oversight is lacking or merely formal and AI outputs are adopted without critical human assessment, the legitimacy of the proceedings may be seriously undermined. Belgian courts have consistently held that procedural deviations capable of affecting the outcome may justify refusal of recognition and enforcement.[9]

Scenario 2: Public Policy under Article V(2)(b)

In this scenario, the court may refuse to enforce the award on its own initiative if it is found to be contrary to public policy[10] under Article V(2)(b) of the New York Convention, Article 34(2)(b)(ii) of the UNCITRAL Model Law, or Article 1721(3) of the Belgian Judicial Code (BJC). These provisions allow courts to deny recognition and enforcement if the underlying procedure or outcome conflicts with fundamental principles of justice in national and European legal systems.[11]

In comparative international practice, public policy has both substantive and procedural dimensions. When a breach of fundamental and widely recognized procedural principles renders an arbitral decision incompatible with the core values and legal order of a state governed by the rule of law, procedural public policy is engaged. Examples include violations of due process, lack of tribunal independence, breach of equality of arms, and other essential guarantees of fair adjudication.[12]

In this case, the use of non-transparent AI systems may fall within this category.[13] If a tribunal relies on these tools without disclosing their use or without providing understandable justifications, the process could violate Article 47 of the Charter of Fundamental Rights of the European Union. This article guarantees the right to a fair and public hearing before an independent and impartial tribunal. This issue, along with case law, could provide a reasonable basis for refusal based on public policy.[14] When applying EU-relevant norms, Belgian courts are bound to interpret procedural guarantees in accordance with the Charter. [15]

Comparative case law provides additional support. In Dutco, for example, the French Cour de cassation annulled an arbitral award for violating the equality of arms in the tribunal’s constitution, which is an archetypal breach of procedural public policy.[16] Similarly, in a 2016 decision under § 611(2)(5) ZPO, the Austrian Supreme Court annulled an award where the arbitral procedure was found to be incompatible[17] with Austria’s fundamental legal values.[18] These rulings confirm that courts may deny enforcement when arbitral mechanisms, especially those that affect the outcome, compromise procedural integrity.

Belgian courts have consistently held that recognition and enforcement must be refused where the underlying proceedings are incompatible with ordre public international belge, particularly where fundamental principles such as transparency, reasoned decision-making, and party equality are undermined.[19] In this context, reliance on non-transparent AI—without adequate procedural safeguards—may constitute a violation of procedural public policy. As a result, enforcement may lawfully be denied ex officio under Article V(2)(b) of the New York Convention and Article 1721(3) of the Belgian Judicial Code, thereby preserving the integrity of both the Belgian and broader EU legal frameworks. Ultimately, courts retain wide discretion under public policy grounds to decide with real control whether or not to enforce AI-assisted awards.[20]

These potential refusals of enforcement within the EU highlight a broader trend, as domestic procedural safeguards are increasingly influenced by global regulatory developments, prompting questions about whether the EU’s approach to AI in arbitration will remain a regional standard or evolve into an international benchmark.

The EU AI Act as a Global Regulatory Model?

The EU has a proven history of establishing global legal benchmarks—rules that, while originating in Europe, shape laws and practices far beyond its borders.[21] The GDPR is the clearest example of this. Its extraterritorial scope, strict compliance obligations, and enforcement mechanisms have prompted countries ranging from Brazil to Japan to adopt similar data protection frameworks.[22]

In arbitration, a comparable pattern could emerge. If EU courts apply the EU AI Act’s high-risk requirements when deciding on the recognition and enforcement of arbitral awards, other jurisdictions may adopt comparable standards, encouraging convergence in AI governance across dispute resolution systems. Conversely, inconsistent enforcement approaches could foster fragmentation rather than harmonisation. In any case, the Act’s influence is already being felt beyond Europe, prompting arbitration stakeholders to address new questions regarding procedural legitimacy, technological oversight, and cross-border enforceability.

Conclusion

The interplay between the EU AI Act and the enforcement of arbitral awards highlights how technological regulation is shaping the concept of procedural fairness in cross-border dispute resolution. Whether the Act becomes a catalyst for global standards or a source of jurisdictional friction, parties and institutions cannot ignore its requirements. As AI tools move deeper into arbitral practice, compliance will become not just a regulatory obligation but a strategic necessity for ensuring the enforceability of awards in key jurisdictions.

[1] Tariq K Alhasan, ‘Integrating AI Into Arbitration: Balancing Efficiency With Fairness and Legal Compliance’ (2025) 42 Conflict Resolution Quarterly 523, 524.

[2] ibid 525.

[3] Jordan Bakst and others, ‘Artificial Intelligence and Arbitration: A US Perspective’ (2022) 16 Dispute Resolution International 7, 23; Sanjana Reddy Jeeri and Vinita Singh, ‘Soft Law, Hard Justice: Regulating Artificial Intelligence in Arbitration’ (2024) 17 Contemporary Asia Arbitration Journal 191, 222.

[4] Sean Shih and Eric Chin-Ru Chang, ‘The Application of AI in Arbitration: How Far Away Are We from AI Arbitrators?’ (2024) 17 Contemporary Asia Arbitration Journal 69, 81.

[5] Horst Eidenmuller and Faidon Varesis, ‘What Is an Arbitration? Artificial Intelligence and the Vanishing Human Arbitrator’ (2020) 17 New York University Journal of Law and Business 49, 72.

[6] Dilyara Nigmatullina and Beatrix Vanlerberghe, ‘Arbitration Related Lessons: Insights from the Supreme Courts around the World’ (2020) 2020 b-Arbitra | Belgian Review of Arbitration 307, 354.

[7] Gizem Kasap, ‘Can Artificial Intelligence (“AI”) Replace Human Arbitrators? Technological Concerns and Legal Implications’ (2021) 2021 Journal of Dispute Resolution 209, 230, 249.

[8] Shih and Chang (n 4) 79.

[9] Koen De Winter and Michaël De Vroey, ‘Belgium’ in Baker McKenzie International Arbitration Yearbook: 10th Anniversary Edition 2016–2017 (Baker McKenzie 2017), 81, 82, 85.

[10] Eidenmuller and Varesis (n 5) 80–81; Bernard Hanotiau, ‘Arbitrability; Due Process; and Public Policy Under Article V of the New York Convention Belgian and French Perspectives’ (2008) 25 Journal of International Arbitration 721, 729–730.

[11] Kasap (n 7) 252; Annabelle O Onyefulu, ‘Artificial Intelligence in International Arbitration: A Step Too Far?’ (2023) 89 Arbitration: The International Journal of Arbitration, Mediation and Dispute Management 56, 63.

[12] Nigmatullina and Vanlerberghe (n 6) 351–352.

[13] Shih and Chang (n 4) 86.

[14] Nigmatullina and Vanlerberghe (n 6) 353.

[15] A de Zitter, ‘The Impact of EU Public Policy on Annulment, Recognition and Enforcement of Arbitral Awards in International Commercial Arbitration’ (University of Oxford 2019) 5, 251–253.

[16] Stefan Kröll, ‘Siemens – Dutco Revisited? Balancing Party Autonomy and Equality of the Parties in the Appointment Process in Multiparty Cases | Kluwer Arbitration Blog’ <https://legalblogs.wolterskluwer.com/arbitration-blog/siemens-dutco-revisited-balancing-party-autonomy-and-equality-of-the-parties-in-the-appointment-process-in-multiparty-cases/> accessed 18 August 2025; Nigmatullina and Vanlerberghe (n 6) 351.

[17] Alexander Zollner, ‘Austrian Supreme Court Set aside an Arbitral Award Due to a Violation of the Procedural Ordre Public’ (Global Arbitration News, 21 June 2017) <https://www.globalarbitrationnews.com/2017/06/21/austrian-supreme-court-set-aside-arbitral-award-for-violation-of-public-policy/> accessed 18 August 2025. ; Franz Schwarz and Helmut Ortner, ‘Austria’ in Giacomo Rojas Elgueta, James Hosking and Yasmine Lahlou (eds), Does a Right to a Physical Hearing Exist in International Arbitration? (ICCA Reports, 2020) 26, https://www.arbitration-icca.org/right-to-a-physical-hearing-international-arbitration accessed 5 August 2025

[18] Nigmatullina and Vanlerberghe (n 6) 351.

[19] Alhasan (n 1) 5–6.

[20] Shih and Chang (n 4) 87; Hanotiau (n 10) 737.

[21] Arturo J Carrillo and Matías Jackson, ‘Follow the Leader? A Comparative Law Study of the EU’s General Data Protection Regulation’s Impact in Latin America’ (2022) 16 ICL Journal 177, 178; Michelle Goddard, ‘The EU General Data Protection Regulation (GDPR): European Regulation That Has a Global Impact’ (2017) 59 International Journal of Market Research 703, 703–704.

[22] Carrillo and Jackson (n 21) 242–245.

By Dr Sarah McKibbin, University of Southern Queensland

The Supreme Court of New South Wales’ decision in Isaacman v King [No 2]^[1] is the kind of case that tempts one to say ‘nothing to see here’, and yet it richly rewards a closer look. On a conventional application of Voth v Manildra Flour Mills^[2] — the leading Australian authority on forum non conveniens — Garling J stayed proceedings that attempted to litigate a New York relationship dispute in Sydney, being ‘well satisfied’ that the NSW Supreme Court was a clearly inappropriate forum.[3] The reasons, though brief by design,[4] illuminate the transaction costs of jurisdictional overreach,[5] show how the Voth framework handles an extreme set of facts, and offer a careful case study for empirical debates about Australian ‘parochialism’ in jurisdictional decision-making.

The Factual Background

The facts almost read like a hypothetical designed to test the outer limits of exorbitant, or long-arm, jurisdiction. A US biotech executive residing in New York sued his former partner, an Australian marketing consultant, in the NSW Supreme Court for alleged negligent transmission of herpes simplex virus during their relationship in New York. The relationship began and ended in New York; the alleged transmission occurred there; the plaintiff’s diagnosis and treatment took place there; and the defendant, though Australian, lived overseas and was only ordinarily resident in Victoria when in Australia. The plaintiff had a four-month period in 2022 split between Sydney, New South Wales, and Melbourne, Victoria, with visits to Queensland, while exploring business opportunities for skincare ventures. He pointed to social friendships in Sydney and his one-off membership of the North Bondi Returned Services League Club.[6]

None of this impressed Garling J as a meaningful link to New South Wales. As Garling J readily observed in the case’s earlier procedural judgment, there was ‘no connection whatsoever between either of the parties, and the pleaded cause of action and the State of New South Wales.’[7] The RSL membership did not establish ‘any connection at all with the forum’.[8] The pleading itself underscored the foreignness of the dispute: by notice under New South Wales’ court rules,[9] the plaintiff relied on New York law, in particular New York Public Health Law § 2307, alongside common law claims available under New York law.[10]

The decision

The stay analysis proceeded squarely under Voth. Garling J recited the familiar principles: the onus lies on the defendant; the question is whether the local court is a clearly inappropriate forum, not whether an alternative is more convenient; it is relevant that another forum can provide justice; and the need to determine foreign law is not conclusive but is a significant factor.[11] The only explicit nod to the English test in Spiliada Maritime Corporation v Cansulex Ltd[12] came through the High Court’s own endorsement in Voth of Lord Templeman’s aspiration for brevity in such applications. [13] Yet Garling J noted that an issue arising in oral submissions required further written submissions, precluding an ex tempore disposition, but nonetheless kept the reasons concise.[14]

On the facts, the connecting factors all pointed away from New South Wales. The conduct giving rise to the claim, the governing law, and the evidentiary base were in New York. Neither party had assets in NSW, so any judgment, whether for damages or for costs, would have to be enforced elsewhere, compounding expense.[15] Garling J accepted, and the parties did not dispute, that New York courts could exercise in personam jurisdiction over the defendant; that acceptance underpinned the conclusion that there was another forum where the plaintiff could ‘obtain justice’.[16] The upshot was decisive but orthodox: the Supreme Court of New South Wales was a clearly inappropriate forum, and the proceedings would be stayed.[17]

The conditional order deserves to be recorded with some precision. The stay was to take effect seven days after publication of the judgment. Within that same seven-day period, the defendant was to file and serve a written undertaking that, if the plaintiff brought civil proceedings in the State of New York concerning the subject matter of the NSW suit, she would not plead any New York limitations defence, provided the plaintiff commenced in New York within three months of the stay taking effect and provided the claims were not statute-barred when the NSW proceeding was commenced.[18] Framed this way, the undertaking did not expand the analysis beyond Voth. It neutralised limitation prejudice, as long as the plaintiff did not delay commencing proceedings, and ensured practical access to the natural forum. Garling J also ordered the plaintiff to pay the costs of the forum non conveniens application.[19]

Two ancillary applications were left untouched. A motion seeking transfer to the Supreme Court of Victoria and a late-filed non-publication motion were not determined.[20] Given the stay, it was not appropriate to go on to decide further issues between the parties. Garling J added that ordering a transfer could impinge on the plaintiff’s own choices about where to proceed next; and with the matter stayed, non-publication orders served no useful purpose.[21]

Comments

Situating Isaacman v King [No 2] in the post-Voth jurisprudence helps explain both the ease and the limits of the result. Voth’s ‘clearly inappropriate forum’ test was announced as only a slight departure from the English Spiliada test,[22] but, as Richard Garnett’s early survey of the doctrine shows,[23] its operation had been variegated.[24] In the years immediately after Voth, Australian courts often refused stays where there were meaningful Australian connections — even if the governing law or much of the evidence was foreign — and sometimes gave generous weight to local juridical advantages.[25] Mary Keyes’ analysis in the Australian family law context underscores why this felt unpredictable: a forum-centric test with broad judicial discretion risks certainty, predictability and cost.[26] Understandably then, Keyes argues for an explicitly comparative, Spiliada-style inquiry that focuses on effective, complete and efficient resolution, the parties’ ability to participate, costs and enforceability.[27]

At the same time, the High Court tempered Voth in specific contexts. In Henry v Henry,[28]  the majority effectively created a presumption in favour of a stay where truly parallel foreign proceedings between the same parties on the same controversy were already on foot, explicitly invoking comity and the risks of inconsistent outcomes.[29] In CSR Ltd v Cigna Insurance Australia Ltd,[30] the High Court went further. Even without identity of issues, the ‘controversy as a whole’ analysis could render local proceedings oppressive where their dominant purpose was to frustrate access to relief available only abroad.[31] These qualifications that, outside the special case of parallel litigation, Voth directs attention to the suitability of the local forum in its own terms; but where duplication looms in the form of parallel proceedings, the analysis necessarily broadens. That broader, comparative posture is also what Ardavan Arzandeh shows Australian courts actually do in practice, despite Voth’s formal language.[32]

Isaacman v King [No 2] belongs to a different, more straightforward strand in that story: the ‘little or no connection with Australia’ cases in which stays have been ordered because the action and the parties’ controversy are overwhelmingly foreign.[33] Unlike the contested margins Garnett identifies, there was no pleaded Australian statutory right of a kind sometimes relied on as a juridical advantage; no contest about the availability of a competent foreign forum; and no tactical race between parallel proceedings. Garling J canvassed the classic connecting factors, noted the New York law pleaded, recorded the practical burdens of proof and enforcement, and concluded that New South Wales was clearly an inappropriate forum. That emphasis on concrete, case-specific connections and on consequences for the conduct and enforcement of the litigation fits both Keyes’ call for structured, predictable decision-making and Arzandeh’s demonstration that Australian courts, in substance, weigh the same considerations as Spiliada.[34]

Two implications follow. First, the decision is a neat instance of Voth doing exactly what it was designed to do when the forum is only nominally engaged. It offers little purchase for testing the harder comparative question whether, at the margins, Voth’s rhetoric yields different outcomes from Spiliada’s ‘more appropriate forum’ inquiry. That is consistent with Arzandeh’s view that the supposed gap is, in practice, vanishingly small.[35] Secondly, it gives texture to the practical burdens that inappropriate forum choices impose. Expert evidence on New York law would have been required; witnesses and records are in the United States; neither party’s assets are in New South Wales; and the court itself, even in this ‘easy’ case, could not resolve the application wholly on the basis of oral submissions because an issue warranted further written argument. Those are precisely the private and public costs Keyes highlights as reasons to favour a clearer, more comparative framework ex ante, rather than leaving calibration to ex post discretion.[36]

There is, then, a narrow lesson and a broader one. Narrowly, Isaacman v King [No 2] confirms that Australian courts will not entertain a claim whose only local anchors are social relationships and what amounts to a meal-discount club card. Broadly, it supplies one more controlled observation for comparative and empirical work: an extreme outlier that aligns with ‘no connection’ line of authority.[37] It also leaves open — indeed, usefully highlights — the need for data drawn from genuinely contested cases, where juridical advantage and practical adequacy are engaged on the evidence, if we are to assess how far Voth diverges, in practice, from its common law counterparts.[38]

Conclusion

Isaacman v King [No 2] therefore earns its place not because it breaks doctrinal ground, but because it shows the doctrine working as intended. The plaintiff’s Sydney friendships and RSL membership could not anchor a transatlantic dispute in a NSW court; New York law, evidence and enforcement pointed inexorably elsewhere; and a conditional stay ensured that the plaintiff would not be procedurally disadvantaged by being sent to the forum where the dispute belongs. If some forum non conveniens applications can be resolved quickly,[39] this was not one of them. But it was, in the end, a straightforward exercise of judicial discipline about where litigation should be done.

[1] [2025] NSWSC 381.

[2] (1990) 171 CLR 538 (‘Voth’).

[3] Isaacman v King [No 2] (n 1) [50].

[4] Voth (n 2) 565.

[5] See Andrew Bell, Forum Shopping and Venue in Transnational Litigation (Oxford University Press, 2003; J J Spigelman, ‘Transaction Costs and International Litigation’ (2006) 80(7) Australian Law Journal 438, 441–3.

[6] Ibid [22].

[7] Isaacman v King [2024] NSWSC 1291, [85]. The earlier judgment dealt with preliminary procedural matters including the plaintiff’s failed attempt to proceed pseudonymously.

[8] Isaacman v King [No 2] (n 1) [41]–[42].

[9] Uniform Civil Procedure Rules 2005 (NSW).

[10] Isaacman v King [No 2] (n 1) [14], [45]–[46].

[11] Ibid [35]–[36].

[12] [1987] AC 460.

[13] Isaacman v King [No 2] (n 1) [37], quoting Voth (n 2) 565. One wonders how often Lord Templeman’s aspiration is realised.

[14] Isaacman v King [No 2] (n 1) [37]–[38].

[15] Ibid [43], [46]–[49].

[16] Ibid [47].

[17] Ibid [39]–[51].

[18] Ibid [4], [56].

[19] Ibid [56].

[20] Ibid [7], [52]–[53].

[21] Ibid [8], [52]–[55].

[22] Voth (n 2) 558.

[23] Richard Garnett, ‘Stay of Proceedings in Australia: A “Clearly Inappropriate” Test?’ (1999) 23(1) Melbourne University Law Review 30.

[24] Cf Ardavan Arzandeh, ‘Reconsidering the Australian Forum (Non) Conveniens Doctrine’ (2016) 65 International and Comparative Law Quarterly 475.

[25] Garnett (n 22) 39–48.

[26] Mary Keyes, ‘Jurisdiction in International Family Litigation: A Critical Analysis’ (2004) 27 UNSW Law Journal 42, 63–4.

[27] Ibid.

[28] (1995) 185 CLR 571.

[29] Ibid 590–1; Garnett (n 22) 52–4.

[30] (1997) 189 CLR 345.

[31] Ibid 400–1; Garnett (n 22) 57–9.

[32] Arzandeh (n 23) 485, 486.

[33] Garnett (n 22) 45–6.

[34] Keyes (n 26) 63–4; Arzandeh (n 23).

[35] Arzandeh (n 23) 491.

[36] Keyes (n 26) 59–60.

[37] Garnett (n 22) 45–6.

[38] On the need for empirical research in this area, see Christopher A Whytock, ‘Sticky Beliefs about Transnational Litigation’ (2022) 28(2) Southwestern Journal of International Law 284.

[39] Spiliada (n 12) 465.

THE INDONESIAN CONSTITUTIONAL COURT DECISION REAFFIRMED PARENTAL CHILD ABDUCTION IS A CRIMINAL OFFENCE

By: Priskila Pratita Penasthika[1]

INTRODUCTION

The Indonesian Constitutional Court Decision Number 140/PUU-XXI/2023, issued on 3 September 2024, confirms that parental child abduction is a criminal offence under Article 330(1) of the Indonesian Criminal Code. Prior to this Decision, Article 330(1) of the Criminal Code was understood as a provision that could not criminalise someone for child abduction if the abduction was committed by one of the biological parents.

After 3 September 2024, through this Constitutional Court Decision, the abduction of a child by one of the biological parents, when the parent does not have custody based on a final court decision, is reaffirmed as a criminal offence.

CONSTITUTIONAL COURT DECISION

Facts

On 15 November 2023, five single mothers (Petitioners) whose children have been abducted by their ex-husbands submitted a petition to the Constitutional Court on 11 October 2023, challenging Article 330 (1) of the Indonesian Criminal Code, which states, “Anyone who, with deliberate intent, removes a minor from the authority which in accordance with the laws is assigned to him, or from the supervision of a person authorised to do so, shall be punished by a maximum imprisonment of seven years.”

The Petitioners shared a common experience: after divorcing their husbands, they were granted custody of their children through a court ruling. However, they have been deprived of this right because their ex-husband abducted their child.

The Petitioners also asserted that they had reported the ex-husband’s actions to the police under Article 330 (1) of the Criminal Code. However, in practice, the report was either dismissed or considered invalid because the police were of the view that the person who abducted the child was the biological father himself and, therefore, could not be prosecuted.

Given this background, the Petitioners believe that the phrase “anyone” (“barang siapa” in Indonesian) in Article 330(1) of the Criminal Code could be interpreted to mean that the biological father or mother of a child cannot be held accountable for the accusation of abducting their own child. Therefore, they submitted a petition to the Constitutional Court requesting a judicial review of Article 330(1) of the Criminal Code.

The Petitioners argue that the phrase “anyone” in Article 330(1) of the Criminal Code should encompass all individuals, including the child’s biological father or mother, as a legal subject. There should be no exceptions that grant absolute authority to the father or mother and exclude him or her from any legal action if he or she violates the child’s rights. Protecting children’s rights is a fundamental aspect of human rights, and the state has a responsibility to provide protection, oversight, and law enforcement to promote children’s welfare. Consequently, the state has the authority to act against parents who violate children’s rights.

Furthermore, the Petitioners request the Constitutional Court to declare that the phrase “anyone” in Article 330(1) of the Criminal Code, which was derived from the Wetboek van Strafrecht voor Nederlandsch-Indië (Staatsblad 1915 Number 732), and later enacted under Law Number 1 of 1946 on the Criminal Code in conjunction with Law Number 73 of 1958 on the Entry into Force of Law No. 1 of 1946 on the Criminal Code for the Entire Territory of the Republic of Indonesia, is unconstitutional, insofar as it is not interpreted to mean “anyone, without exception the biological father or mother of the child.”

The Decision

The Decision of the Constitutional Court Number 140/PUU-XXI/2023, which consists of nine Constitutional Judges, rejected the Petitioners’ request in its entirety.

The Constitutional Court Judges believe that Article 330(1) of the Criminal Code is an explicit and well-defined provision (expressive verbis), so there is no need to interpret it or add any supplementary meaning to it. The Judges asserted that the phrase “anyone” encompasses every individual without exception, including the biological father or mother of the child. The Court also noted that adding a new meaning to Article 330(1) of the Criminal Code, as requested by the Petitioners, could potentially restrict the scope of the legal subjects covered by that provision and other provisions in the Criminal Code that use the phrase “anyone”. This could result in legal uncertainty, according to the Judges.

In its legal deliberation, the Constitutional Court Judges referred to the United Nations Convention on the Rights of the Child (UNCRC), to which Indonesia is a state party, and its provisions are incorporated into Law Number 23 of 2002 on Child Protection, as amended by Law Number 35 of 2014 (Law on Child Protection). Furthermore, the Law on Child Protection recognises that the best interests of the child, as stipulated in the UNCRC, are a fundamental principle for child protection. According to the Official Elucidation of Law on Child Protection, the best interests of the child mean that, in all actions concerning children undertaken by the government, society, legislative bodies, and judiciary, the child’s best interest must be the primary consideration.

In cases of parental child abduction, aside from the child being the victim, the Constitutional Court recognises that the parent, who is forcibly separated from their child by the other parent, can also become a victim, particularly on a psychological level. This indicates that the psychological bond between parents and their biological children should not be severed, emphasising that the child’s best interests must take precedence. In this context, the Constitutional Court Judges emphasise that criminalising one of the child’s biological parents who breaches the provisions of Article 330(1) of the Criminal Code should only be considered as a last resort (ultimum remedium).

In another part of its Decision, the Constitutional Court addressed the issue of the Petitioners whose reports were rejected by the police. The Constitutional Court Judges stated that they had no authority to assess this matter. However, they affirmed in the Decision that law enforcement officers, especially police investigators, should have no hesitation in accepting any report concerning the application of Article 330(1) of the Criminal Code, even if it involves the child’s biological parents. This is because the term “anyone” includes every individual without exception, including, in this case, the child’s biological father and mother.

The Constitutional Court concluded that Article 330(1) of the Criminal Code provides legal protection for children and ensures fair legal certainty as outlined in the 1945 Constitution of the Republic of Indonesia. Therefore, the Court states that the Petitioners’ request is rejected in its entirety.

Dissenting Opinion

The nine Constitutional Judges did not reach a unanimous decision. Judge Guntur Hamzah expressed his dissenting opinion, arguing that the Constitutional Court should have partially granted the Petitioners’ request.

Judge Hamzah views the Petitioners’ case as also involving the enforcement of a norm that breaches the principles of justice, the constitution, and human rights. Due to numerous cases of parental child abduction, often committed by biological fathers, Judge Hamzah believes it is appropriate for the Constitutional Court to act as the defender of citizens’ constitutional rights in this matter. This aims to safeguard the constitutional rights of biological mothers who hold custody, whether naturally or legally granted by the court, from acts of child abduction or forced removal by biological fathers. It not only ensures legal certainty but also offers reassurance to both the child and the parent who holds the legal custody rights.

Judge Hamzah is of the opinion that the Constitutional Court should have partially granted the Petitioners’ request by inserting the phrase “including the biological father/mother” into Article 330(1) of the Criminal Code. This would have made Article 330(1) of the Criminal Code to read, “Anyone who, with deliberate intent, removes a minor from the authority which in accordance with the laws is assigned to him, including his biological father/mother, or from the supervision of a person authorised to do so, shall be punished by a maximum imprisonment of seven years.”

REMARKS

It is worth noting that Law Number 1 of 2023 on the Criminal Code (New Criminal Code) was approved by the Indonesian House of Representatives on 2 January 2023. The New Criminal Code will come into effect on 2 January 2026. There are no significant changes regarding the concept of child abduction in the New Criminal Code. Article 452(1) of the New Criminal Code is equivalent to Article 330(1) of the current Criminal Code. Article 452(1) of the New Criminal Code states: “Every person who removes a Child from the authority which in accordance with the statutory regulations is assigned to him or from the supervision of a person authorised to do so, shall be punished by a maximum imprisonment of 6 (six) years or a maximum fine of category IV.”

It is quite unfortunate that there has been no shift in the perspective towards parental child abduction cases in Indonesia. In early 2023, Indonesian lawmakers, as indicated in Article 452(1) of the New Criminal Code, still regard parental child abduction cases primarily from a criminal perspective. This stance is later reaffirmed in 2024 by the Court, as stated in the Constitutional Court Decision Number 140/PUU-XXI/2023.

Although the Constitutional Court Judges, in their Decision, recognise the psychological bond between parents and the child as part of the child’s best interests and acknowledge that criminalising a parent over child abduction is a last resort, parental child abduction is still viewed from a criminal perspective. Consequently, this Constitutional Court Decision does not provide an effective solution. The five petitioners remain unable to access their abducted children because they do not know their children’s whereabouts or how to contact them.

The Constitutional Court Judges also hold conflicting views in their deliberations. On one hand, they acknowledge that the psychological bond between parents and a child must be prioritised as part of the child’s best interests. On the other hand, they affirm the provision of Article 330(1) of the Criminal Code, which permits the criminalisation and imprisonment of the parent who commits child abduction, albeit as a last resort. It seems that the judges overlooked the possibility that criminalising and imprisoning the parent involved in child abduction could also harm the child’s best interests, as it would deprive the child of access to that parent.

It is also regrettable that none of the Judges or the expert witnesses involved in the proceedings mentioned the HCCH 1980 Convention on the Civil Aspects of International Child Abduction (HCCH 1980 Child Abduction Convention), which provides a perspective on parental child abduction from its civil aspects. Consequently, the procedures for returning the wrongfully removed child to their habitual residence—while safeguarding access rights and prioritising the child’s best interests as stipulated by the Convention—remain unfamiliar and unexplored in Indonesia.

The Constitutional Court Decision Number 140/PUU-XXI/2023, which considers parental child abduction from its criminal aspect, reveals a legal gap in Indonesian law that can only be filled in by the HCCH 1980 Child Abduction Convention. The Convention could serve as an instrument providing civil measures in cases of parental child abduction in Indonesia and promote a more effective resolution by ensuring the child’s prompt return without depriving access to either parent. In other words, Indonesia’s accession to the Convention has become more urgent to ensure that the child’s best interests, as recognised by Indonesian Law on Child Protection, are adequately protected.

Recognising that many adjustments within Indonesian laws and regulations will still be necessary, the Author of this article has long hoped that Indonesia will eventually accede to the HCCH 1980 Child Abduction Convention, hopefully sooner rather than later.

[1] Assistant professor in private international law at the Faculty of Law, Universitas Indonesia.

“The 2025 International Arbitration Survey: The Path Forward”

Luke Nottage (University of Sydney)

The 14th Queen Mary University of London Survey, again in collaboration with international law firm White & Case, was dissected at an Australian launch seminar (expertly moderated by partner Lee Carroll) at their Melbourne office on 22 July 2025. Some “early insights” had been provided during Paris Arbitration Weeks, when the Survey report was not yet public. This analysis delves deeper into the report and key findings, drawing also on the discussion with our co-panellists, including some suggestions for future research.

Survey Methodology

This latest Survey shows how the responses have become more expansive and therefore reliable over time. Although not a random survey, 2402 responses were received for the written questionnaire (the response rate is unspecified). This is significantly greater than “more than 900” respondents for the 2022 Survey focused on energy disputes, 1218 for the general 2021 Survey, and just 103 for the inaugural Survey in 2006. This study was again mixed-method, adding qualitative research through 117 follow-up interviews.

This increase in Survey participation arguably indicates the growing awareness of the research and interest in its results, as well as the proliferation and diversification of international arbitration (IA) over the last two decades. Overall respondents in 2025 (Chart 26) primarily practiced or operated in the Asia-Pacific (47%), illustrating arbitration’s shift (along with economic activity) into Asia; separately in North America (a further 10% of respondents), Central and Latin America (7%); plus Europe (10%) and Africa (6%).

Respondents’ primary roles (Chart 23) were counsel (35%), arbitrators (17%), both (14%), arbitral institution staff (9%), academics (8%) and tribunal secretaries (2%). Surprisingly, there were few in-house counsel (3%), who historically and anecdotally tend to be more concerned eg about costs and delays. Few respondents were primarily experts (1%), which may reflect the declining professional diversity within IA.

Arbitration with or without ADR

The 2025 Survey asked again about respondents’ preferred method of resolving cross-border disputes (Chart 1). IA together with ADR was most popular (48%), compared to standalone IA (39%). The Survey contrasts this with 59% versus 31% in 2021 (p5). That shift could indicate that IA has been working effectively to address eg persistent complaints about its costs and delays.

However, more work needs to be done by IA stakeholders, as in the 2015 Survey only 34% of respondents had preferred IA with ADR, versus 56% preferring just IA. This indicates that the trend over the last decade remains towards combining IA with ADR. Additionally, future research could usefully ask what is meant by IA “together with ADR”. As co-panellist Leah Ratcliff remarked from her experience (now as in-house counsel in Australia), parties are more comfortable with clauses providing for (structured) negotiations rather than (potentially still quite expensive) mediation before IA. It would also be interesting to check respondents’ preferences regarding Arb-Med (arbitrators actively promoting settlement, or engaging an Arb-Med-Arb process as in Singapore – arguably showing up in the 2022 SIDRA Survey, Exhibit 8.1).

The 2025 Survey commentary also suggests that ADR preference may be partly “influenced by cultural factors” (p6), noting European respondents favoured more standalone IA (51%) compared Asia-Pacific respondents (37%). However, recall that overall 39% favoured IA anyway.

There also remains great diversity within Asia regarding legal culture – let alone general culture. For example, first there are common law jurisdictions (eg Singapore, Hong Kong, Australia) with strong traditions now of domestic mediation for commercial disputes, due to high costs and delays in litigation initially (and sometimes still). This carries over into more willingness to agree to multi-tiered clauses mandating even mediation before arbitration. Secondly, however, there are some common law jurisdictions in Asia (notably India, despite extensive court delays) with no such tradition of privately-supplied mediation services. Relatedly, their legal advisors and parties are more reluctant to propose Med-Arb clauses in international contracts (although they may agree to them if proposed, if obtaining other benefits through negotiations). Thirdly, civil law jurisdictions (like Japan, with more efficient courts plus some Court-annexed mediation, but also mainland China) also seem less amenable to Med-Arb clauses, although long comfortable with clauses providing for good faith negotiations prior to IA. In addition, there is even greater diversity across Asia regarding Arb-Med (basically only practiced intensively in China, partly in Japan).

Preferred Seats and Rules

Earlier surveys had started to identify Singapore, Hong Kong and mainland Chinese cities within top preferred seats, along with traditional venues like London and Paris. Yet it was unclear whether this reflected the growing proportion of Asia-Pacific (essentially Asian) respondents. The 2025 Survey helpfully helps to address this question. Globally, ie among all respondents (Chart 3), the most preferred seat is London (chosen, among up to five seats, by 34%), then Singapore and Hong Kong (31% each), then Beijing and Paris (19% each). However, London and Singapore were ranked in the top four for all regional respondents, and Paris too except for Asia-Pacific respondents (Chart 2). Otherwise, the European and Asia-Pacific respondents “show strong preferences for seats in their respective regions” (2025 Survey, p7).

Quite similarly, LCIA Rules (nominated globally by 25% of all respondents, again with up to five preferences) were preferred in all regions except the Asia-Pacific, while SIAC Rules (chosen by 25%) and UNCITRAL Rules (15%) were preferred for all regions except Central and Latin America (Charts 4 and 5). By contrast, HKIAC Rules (25%) were most preferred by Asia-Pacific respondents (36%), but not selected among top 5 preferences from respondents from other regions. As co-panelist (and experienced arbitrator) Michael Pryles noted at the launch seminar, Hong Kong and HKIAC Rules still benefit as a compromise for transactions and disputes involving mainland China. He also rightly suggested, as did an audience member, that asking about “preferences” may not give the full picture. This could be usefully compared with evolving actual practice, including arbitration case filings. Over 2024, for example, HKIAC handled 352 new arbitration cases (77% international) whereas SIAC handled 625 (91% international).

Co-panellist Diana Bowman, new Secretary-General of the ACICA, remarked that the ACICA Rules did not quite make Chart 5, despite the Australian Centre’s increased case filings in recent years. As a former Rules committee member (2004-2024), I added that arbitral institutions should not just be judged by case filing statistics. Those depend for example on geography, although there may be scope for Australia to focus on niches, such as the South America – Southeast Asia or South Asia trades, or (as Pryles also observed) specialist fields such as disputes over resources. In addition, improving Rules (and seats more generally) can allow local parties more credibly to propose them but then compromise in negotiations to obtain other contractual benefits.

Pryles also shared experiences and views about the growing impact on IA from  sanctions on parties or participants. Notably, 30% of respondents noted that sanctions led to a different seat being chosen (Chart 6).

The 2025 Survey also found that 39% thought awards set aside at the seat should be enforceable in other jurisdictions (Chart 8), whereas 61% thought not. The 39% proportion is surprisingly high, as only French courts uniformly adopt this approach. Courts elsewhere will usually not enforce, unless there is some particularly egregious flaw regarding the seat court (such as proven corruption) or seat jurisdiction (such as legislation retrospectively impacting arbitration agreements or awards). Perhaps the 39% of respondents agreed with enforcement but only in such exceptional circumstances, which might then be separated out as a third possibility in future research. Meanwhile, this trend (and growing deference towards decision of seat courts instead upholding challenged awards) should reinforce the importance of carefully choosing the seat.

IA Enforcement and Efficiency

Past Surveys (and other research) typically identified enforceability of IA awards (and agreements), neutrality and expertise of arbitrators, flexibility in procedures, then privacy and confidentiality, as major advantages over cross-border litigation. The 2025 Survey innovated by focusing on the growing awareness and engagement in various public interest elements (eg environmental) even in commercial IA, including its perceived advantages instead of litigation. Arbitrator expertise (47%), avoiding local courts and laws (42%) and (broader?) neutrality (28%) were often chosen from among three options (Chart 15). Confidentiality was selected by 34% of respondents, which seems understandable given these are still commercial disputes (not ISDS arbitrations involving greater public interests and so already associated with more transparency). Enforceability of awards was only chosen by 32%, but this may reflect greater actual or anticipated problems with public policy or arbitrability exceptions to enforcement.

Then 2025 Survey also usefully drilled down into another commonly posed question: voluntary compliance with IA awards (Chart 7). Interestingly respondents said this happened similarly, almost always or often, for non-ICSID awards against states (33%) as for ICSID awards (34%), despite most of the latter involving the more delocalised ICSID Convention enforcement regime. Also surprisingly, good compliance for non-ICSID private awards was only reported by 40% of respondents. This may also indicate persistent question around “formalisation” and over-lawyering in IA, discussed more broadly under “efficiency and effectiveness” in the 2025 Survey (pp15-19).

Notably, respondents were asked to chose up to three options for processes that would most improve efficiency in IA (Chart 10). The most popular were expedited arbitration (50%, generating further questions) and early determination of unmeritorious claims or defences (49%). But there was also interest in non-binding pre-arbitral assessments by an expert (13%), mandatory settlement discussions (12%) or mediation (11%) in procedural timetables, and even “baseball arbitration” (11%). Interestingly, as this remains a hot topic for multi-tiered clauses, 7% chose “limiting grounds to challenge pre-arbitration ADR outcomes in arbitration proceedings” (rather than in court). Less surprisingly, as these impact on fees earned by counsel (the largest respondent group) and are rarely mentioned in arbitral Rules, only 1% picked “sealed offers” as a mechanism to improve efficiency.

The survey found “perhaps most surprisingly, given the respondents’ generally favourable view of combining arbitration with ADR, the option of multi-tiered dispute resolution clauses with mandatory ADR processes was included by fewer than 1% of respondents as one of their three picks. To some interviewees, ADR adds an unnecessary procedural layer. Others question the utility …” (p16). However, this low response rate arguably is due to the question’s phrasing, which asked about measures to improve efficiency in arbitration (not the overall dispute resolution process).

A final hot topic canvassed in the 2025 Survey concerns AI in IA (pp27-33). Pryles was skeptical about arbitrators delegating too much to Artificial Intelligence for their reasoning. Surprisingly, however, although 71% of respondents had never used AI for “evaluating legal arguments” in the past 5 years, for the next 5 years this was expected to drop to 31% (Chart 18). Admittedly, some of this may be done by lawyers and so less problematic than for arbitrators.

Less controversial is the existing use of AI for “document review” (never used so far by only 41%, expected to drop to 10%). However, that raises the question of whether an even more efficient approach would be for arbitrators to more pro-actively help identify the issues to be determined, and hence relevant evidence. The 2012 Survey (Chart 9) had found that to be the best means experienced to expedite arbitral proceedings, even when phrased as arbitrators doing this “as soon as possible after constitution” of the tribunal (which is more controversial than as the arbitration progresses, eg under the JCAA Interactive Arbitration Rules).

Conclusion

The 2025 Survey, especially combined with the earlier ones, provides a rich resource to understand current practices and concerns in IA. It also helps identify future opportunities and challenges, as well as promising ongoing research into this always-evolving field.

by Patrick Ostendorf (HTW Berlin)

In the recent and interesting case of LLC Eurochem v Société Generale S.A. et al [2025] EWHC 1938 (Comm), the English High Court (Commercial Court) considered the extent to which economic sanctions enacted by a foreign jurisdiction (EU law in this instance) can impact the enforcement of contractual payment claims (governed by English law) in English courts. More broadly, the decision also highlights the somewhat diminishing role of the Rome I Regulation (and its interpretation by the European Court of Justice) in the English legal system, and probably that of conflict of laws rules in general.

The underlying facts

A Russian company, respectively its Swiss parent (the assignee of the claimed proceeds of the drawdown), both of which are ultimately controlled by a Russian oligarch, claimed €212 million from two banks (one French and one Dutch, the latter operating through its Italian branch) out of six on-demand bonds governed by English law, based on corresponding exclusive jurisdiction agreements in favour of English courts. The performance bonds had been issued by the defendant banks to secure the proper performance of a contract for the construction of a fertiliser plant in Russia, which was terminated as a consequence of Russia’s illegal invasion of Ukraine. When the Russian company called on the bonds to recover advance payments made under the construction contract, the banks refused to pay, arguing that doing so would violate applicable EU sanctions.

The Commercial Court agreed with the banks that payment under the bonds would indeed breach both Art. 2 of Council Regulation (EU) No 269/2014 and Art. 11 of Regulation (EU) No 833/2014. However, even though the ultimate owner of the claimant was also subject to UK sanctions, UK sanctions did not apply in this case, as payment under the bonds would not have involved any acts in the UK or by UK companies or persons.

The key question

The key question was therefore this: Could the banks rely on the EU sanctions as a defence against the payment claim in an English court, given that their contractual performance would be illegal under foreign law? According to the Ralli Brothers principle (as established by the English Court of Appeal in Ralli Brothers v Companie Naviera Sota y Aznar [1920] 2 KB 287 and also serving as a blueprint for Art. 9(3) of the Rome I Regulation), the answer would be yes if the contractual performance required an act to be carried out in a place where it would be unlawful to do so. However, was the place of performance in the EU in this case, despite the fact that, under English common law, the place of payment is generally where the creditor (here, the claimant, as the beneficiary) is located, unless otherwise agreed by the parties?

The court’s resolution

The resolution was straightforward in relation to the defendant Italian branch, as the corresponding bond incorporated the ICC Uniform Rules for Demand Guarantees (URDG) and Art. 20(c) of the URDG explicitly states that payment is to be made at the branch or office of the guarantor (para. 447). However, the Commercial Court also answered this question in the affirmative with regard to the payment claims against the French bank (the relevant five bonds had not incorporated the URDG). This was based on the general proposition that, in relation to on-demand instruments, the place of performance should generally be where the demand must be made — hence in this case in France rather than Russia or Switzerland (paras 449 ff.).

Public policy was the alternative reasoning offered by the Commercial Court

More interesting still is the alternative argument offered by the Commercial Court. The court explicitly agreed with the defendants that the bonds should not have been enforced, even if the place of performance were in Russia (in which case the Ralli Bros. principle could accordingly not apply). The court postulated that, even outside the Ralli Bros. rule, ‘a sufficiently serious breach of foreign law reflecting important policies of foreign states may be such that it would be contrary to public policy to enforce a contract’ (paras 466 et seq). According to the defendants (and as confirmed by the court), the principle of comity was engaged particularly strongly here, given that the defendants would have faced prosecution, significant fines and the risk of imprisonment for individuals acting on behalf of the banks in France and Italy if they had paid.

Comments

The alternative reasoning given by the Commercial Court for the unenforceability of the bonds based on public policy seems to have two flaws.

Firstly, the view that enforcing a contract may be contrary to public policy due to a sufficiently serious breach of foreign law even outside the Ralli Bros. rule cannot be based on a clear line of precedent. The Commercial Court only refers to two High Court decisions, the more recent of which is Haddad v Rostamani (2021) EWHC 1892, para. 88. These decisions are difficult to reconcile with the Court of Appeal’s finding in Celestial Aviation Services Limited v Unicredit Bank GmbH [2024] EWCA Civ 628, paras. 105 et seq and prior High Court precedents relied on in this judgment, in particular Banco San Juan Internacional Inc v Petróleos De Venezuela S.A. [2020] EWHC 2937 (Comm), para. 79, which states that, ‘the doctrine therefore offers a narrow gateway: the performance of the contract must necessarily involve the performance of an act illegal at the place of performance. Subject to the Foster v. Driscoll principle, […] it is no use if the illegal act has to be performed elsewhere’. In Banco San Juan, the High Court referred to the Foster v Driscoll principle as the only legitimate expansion of the Ralli Bros. rule. But this principle is not applicable in the present case: It is limited to contracts entered into by the parties with the intention of committing a criminal offence in a foreign state (Foster v Driscoll [1929] 1 KB 470, 519).

Secondly, it is somewhat ironic that, in order to give effect to EU sanctions law, the Commercial Court relies on English common law precedents that hardly align with Art. 9(3) of the Rome I Regulation. This is because the ECJ has expressly taken the view that Art. 9 contains an exhaustive list of situations in which a court may apply foreign overriding mandatory provisions not merely as a matter of fact (see ECJ, 18 Oct 2016, Case C-135/15, Nikiforidis: ‘Article 9 of the Rome I Regulation must therefore be interpreted as precluding the court of the forum from applying, as legal rules, overriding mandatory provisions other than those of the State of the forum or of the State where the obligations arising out of the contract have to be or have been performed’).

Although the Commercial Court does not mention the Rome I Regulation in this regard, it still forms part of English statutory law as ‘assimilated law’ (formerly ‘retained EU law’). The justification for ignoring the Regulation is probably the prevailing, though (against the background of the general function of private international law and the fact that Art. 9 of the Rome I Regulation explicitly and exhaustively deals with this very problem) unconvincing, view in England that the Ralli Bros principle, and consequently its potential expansion in the present case, is not a conflict of laws rule in the first place: Instead, it is considered a principle of domestic English contract law, therefore unaffected by the exhaustive nature of Art. 9(3) of the Rome I Regulation (in favour of this view, for example, Chitty on Contracts, Vol. I General Principles, 35th edition (2023), para. 34-290, Dicey, Morris & Collins, The Conflict of Laws, Vol. 2, 16^th edition (2022), para. 32-257 with further references. Contrary, A. Briggs, Private International Law in English Courts (2014) para. 7.251, who rightly notes that such a characterisation ‘was only possible by being deaf to the language and tone in which the judgments were expressed, and it is a happy thing that the Rome I Regulation puts this seemly principle on a statutory footing’ and characterises the Ralli Bros principle accordingly as a ‘rule of common law conflict of laws’ (A. Briggs, The Conflict of Laws, 4th edition, 2019, p. 239). For a full discussion of the history and characterisation of the Ralli Bros rule, see W. Day (2020) 79 CLJ 64 ff.)

The need to rely both on a questionable characterisation and expansion of the Ralli Bros principle in this case may be due to English contract law (at least in its substantive core) being ill-equipped to address factual impediments caused by foreign illegality for the parties. Unlike civil law jurisdictions, which can rely on the doctrine of (temporary) impossibility to address such cases — the recent decision of the Court of Arbitration in CAS 2023/A/9669, West Ham United Football Club v PFC CSKA & FIFA (applying Swiss law), is a case in point — the doctrine of frustration is apparently too limited in scope to recognise factual impediments triggered by foreign illegality. Furthermore, the doctrine of frustration does not offer the necessary flexibility as it results in the termination of the contract rather than merely suspending it temporarily.

I. Introduction

The question of the application of Islamic law in non-Muslim countries has triggered extensive discussions and debates regarding the consistency of Islamic law rules – whether codified in modern legislation or not – with the forum’s public policy. This issue has attracted particular attention in the field of family law, where various legal Islamic institutions (such as dower, polygamy, and early marriage) have sparked considerable controversy and posed significant challenges in both court practice and academic debate. This is particularly salient in the field of dissolution of marriage, as Islamic practices such as talaq and khul have often been the subject of intense discussions concerning their recognition and validity in non-Muslim jurisdictions.

The case presented here is another example of the complexity inherent in the reception of peculiar Islamic law institutions in private international law. Recently decided by the Nagoya High Court (second-instance court) in its ruling of 12 June 2025, it concerns a type of marital dissolution based on ila (an oath of sexual abstention). To the best of my knowledge, no comparable case involving ila has been decided before in any jurisdiction, which makes this ruling particularly important both in theory and in practice. This is especially so given that resorting to ila in this case appears to have been part of a litigation strategy, anticipating an unfavourable outcome if the case had been brought before the court as a talaq case (see infra V). As such, the case provides an opportunity to consider the nature of this unusual Islamic legal institution, its specific features, and the challenges it may raise when examined by foreign courts.

II. The Case:

The parties in this case are a Bangladeshi Muslim couple who married in accordance with Islamic law in Bangladesh and subsequently moved to Japan, where they had their children. All parties, including the children, are permanent residents of Japan.

The case concerns a divorce action filed by the husband (X) against his wife (Y), seeking dissolution of marriage primarily under Bangladeshi law, and alternatively under Japanese law. X argued that, in his complaint, he declared his intention “in the name of Allah” to abstain from sexual relations with his wife; and since four months had passed without any sexual relations with Y, a “talaq-divorce” had been effected and thereby completed in accordance with Bangladeshi law. The divorce action was filed as a result of continuous disagreement and disputes between the parties on various issues including property rights, management of the household finance, and alleged misbehaviour and even violence on the wife’s side. At the time the action was filed, X and Y had already been living separately for some time.

One of the main issues revolved around whether the application of Bangladeshi law, which provides for this form of marital dissolution (referred to in the judgment as “talaq-divorce”), should be excluded due to inconsistency with Japanese public policy under Article 42 of the Act on the General Rules of Application of Laws (AGRAL).

The court of first instance (Nagoya Family Court, judgment of 26 November 2024) held that the “talaq-divorce” (as referred to in the judgment) was valid under Bangladeshi law and that its recognition did not contravene Japanese public policy. Notably, the court emphasized that “any assessment of whether the legal rules applicable between spouses who share the same religious and cultural background violate Japanese public policy should be approached with a certain degree of restraint”, given the strong cultural and religious elements involved in the personal status of the parties, who are both originally Bangladeshi nationals and Muslims who were married in accordance with Islamic law, even if they had been living and residing in Japan for some time.

Dissatisfied with the judgment, Y appealed before the High Court.

Y challenged the first instance judgment on various grounds. She basically argued – inter alia – that, given the strong ties the parties and their children have with Japan and their established life there, the mere fact that the parties are Bangladeshi nationals and Muslims should not justify a restrained implication of public policy, especially considering that the effects and consequences of the divorce would take place in Japan.

III. The Ruling

The Nagoya High Court upheld the judgment of the court of first instance, stating as follows (only a summary is provided here, with modifications and adjustments):

Under Bangladeshi law, which governs the present divorce, a husband may dissolve the marriage either through talaq (a unilateral declaration of divorce by the husband) or through other modes. There are several forms of talaq-divorce available to the husband, including ila. The latter entails the husband taking an oath in the name of Allah to abstain from sexual relations with his wife. If no intercourse occurs within four months following the oath, the divorce is considered to have taken effect.

In the present case, considering that Bangladeshi law is the applicable law, the talaq-divorce would be deemed valid, and would be recognized, since a period of four months had passed without any sexual contact between the parties after X made his declaration in the complaint.

Generally, when determining the applicability of Article 42 of the AGRAL, it is not the foreign law’s provisions themselves that should be assessed in abstracto. Rather, the application of the foreign law as the governing law may be excluded [only] where (1)its concrete application would result in a consequence that is contrary to public policy, and (2) the case has a close connection with Japan.

Regarding (1), the marital relationship between the parties had deteriorated over time, and various elements, when taken together, indicate that the parties had already reached a serious state of discord that could reasonably be seen as leading to separation or divorce. Consequently, considering all these circumstances, and taking into account the background of the case, the nature of the parties’ interactions, and the duration of their separation, it cannot be said that applying Bangladeshi law and recognizing the talaq-divorce in this case would be contrary to public policy.

With respect to (2), Y argued that, due to the strong connection between the case and Japan, the exclusion of the application of Bangladeshi in application of article 42 of the AGRAL should be admitted. However, as previously noted, the application of Bangladeshi law in this case does not result in a violation of public policy. Therefore, even considering the strong connection of the case to Japan, the application of Article 42 of the AGRAL cannot be justified.

IV. Comments

(*) Unless otherwise indicated, all references to Islamic law here are about classical Islamic law as developed by the orthodox Sunni schools, and not Islamic law as codified and/or practiced in modern Muslim countries.

1. Islamic law before Japanese Court

There are several cases in which Japanese courts have addressed the application of foreign laws influenced by or based on Islamic law. These cases have involved matters such as the establishment of filiation, annulment of marriage, parental authority, adoption, and divorce (whether based on the unilateral will of the husband or not). While in few instances the courts have applied the relevant foreign law without particular difficulties (for example, allowing a Japanese woman married to a Pakistani Muslim man to seek and obtain a divorce under Pakistani law), in most cases, the courts have refused to apply such laws on the grounds that they were contrary to Japanese public policy. The rules found incompatible with public policy include, among others, the non-recognition of out-of-wedlock filiation, the prohibition of interfaith marriage, the prohibition of adoption, the automatic attribution of parental authority to the father, and talaq-based divorce (triple talaq). The foreign laws at issue in these cases originate either from Muslim-majority countries such as Iran, Pakistan, Indonesia, and Egypt, or from non-Muslim countries with Muslim minorities who are governed by their own personal status laws, such as Myanmar and the Philippines.

The case commented on here provides a new example of a Japanese court grappling with the application of foreign law grounded in Islamic legal principles.

2. Ila and dissolution of marriage

Like many other traditional – and in some views, “exotic” – Islamic legal institutions (such as zihar, li’an, khul, tamlik, tafwidh, mubara’a …… definitions are intentionally omitted), ila is often difficult to apprehend correctly, both in substance and in function.

a) What is ila?

Generally speaking, ila can be defined as “the swearing of an oath by a man that he will not have intercourse with his wife” for a period fixed in the Quran (chapter 2, verse 226) at four months (See Ibn Rushd (I. A. Khan Nyazee, trans.), The Distinguished Jurist’s Primer – Vol. II: Bidayat Al-Mujtahid wa Nihayat Al-Muqtasid (Garnet Publishing, 2000) 121).

It worth mentioning first that ila is not an Islamic invention but was practiced in pre-Islamic society. In that context, ila allowed the husband to place considerable pressure on his wife by placing her in a state of marital limbo, which can be for an indefinite period. This left the woman in a vulnerable and uncertain position, as she was neither fully married in practice, nor legally divorced.

Islamic Sharia addressed this practice and, while it did not abolish it – unlike some other pre-Islamic institutions and practices –, it attempted to alleviate its harmful effects, by introducing a period of four months, during which the husband is invited to reconsider his decision and either resume marital life (Quran chapter 2, verse 226)  or dissolve the marriage  (Quran chapter 2, verse 227).

b) Ila – Different Practices

However, regarding the actual operation of ila, the schools of Islamic religio-legal jurisprudence (fiqh) diverge significantly on several points (Ibn Rushd, op. cit.). Two issues are particularly relevant here:

i. The first concerns whether :

(i-a) the four-month period stated in the Quran represents a maximum period, at the end of which the marriage is dissolved; or

(i-b) the four-month period merely marks the threshold between an oath of abstention that does not lead to marital dissolution and one that does. According to this latter view, only an oath exceeding four months, or one made for an indefinite duration, qualifies as ila that may result in the dissolution of marriage.

ii. The second issue concerns whether

(ii-a) the marriage is automatically dissolved once the four-month period has elapsed, if the husband does take the necessary actions to resume the marital life, that is after performing an act of expiation (kaffara) in accordance with the Quranic prescriptions (notably Chapter 5, verse 89); or

(ii-b), upon expiry of the term, the wife may petition a qadhi (Muslim judge), requesting that her husband either end the marriage by pronouncing talaq, or resume marital relations after performing an act of expiation (Chapter 5, verse 89). In such a case, the qadhi would then grant the husband a specified period to decide. If the husband fails to take either course of action, the qadhi may pronounce the dissolution of the marriage on account of his inaction. Depending on the legal opinion, this dissolution may be categorized either as a talaq issued on behalf of the husband, or as a judicial annulment (faskh).

Traditionally, the Hanafi school, prevalent in Bangladesh, follows positions (1-a) and (2-a), while the other major schools adopt views (1-b) and (2-b).

3. Ila and talaq – what’s the difference?

It is not uncommon for ila to be described as “a form of talaq.” This appears to be the position of the High Court, seemingly based on the arguments presented by X’s representative during the trial. It is true that both ila and talaq are prerogatives reserved exclusively for men; women do not have equivalent right (except, in the case of talaq, where the husband may contractually delegate this right to his wife at the time of the marriage). It is also true that both ila and talaq may lead to the dissolution of marriage based on the unilateral intention of the husband. However, describing ila as a “form of talaq” is not – technically speaking – entirely accurate.

i. Under the majority of schools of fiqh – except for the Hanafi –, the distinction is quite clear. This is because unlike talaq, ila, by itself, does not lead to dissolution of marriage. A judicial intervention is required upon the wife’s request for the marriage to be dissolved (which is not required for talaq).

 ii. Under the Hanafi school, however, the distinction between ila and talaq may be blurred due to their substantial and functional similarities. In both cases, a qualified verbal formula places the marriage in a suspended state^(*) for a specified period (the waiting period (iddah) in the case of talaq, and the four-month period in the case of ila). If the husband fails to retract his declaration within this period, the marriage is dissolved.

^(*) However, this does not apply in the case of a talaq that immediately dissolves the marriage: that is, a talaq occurring for the third time after two previous ones (whether or not those resulted in the dissolution of the marriage), or in the case of the so-called triple talaq, where the husband pronounces three talaqs in a single formula with the intention of producing the effect of three successive talaqs.

Nevertheless, a number of important distinctions remain between the two, even within the Hanafi doctrine.

a. The first concerns the frequency with which talaq and ila may be resorted to. Similar to ila, talaq does not necessarily lead to the dissolution of the marriage if the husband retracts during the wife’s waiting period (iddah). However, its use – even if followed by retraction – is limited to two occurrences (Chapter 2, verse 229). A third pronouncement of talaq results in immediate and irrevocable dissolution of the marriage, and creates a temporary impediment to remarriage. This impediment can only be lifted if the woman marries another man and that subsequent marriage is irrevocably dissolved (Quran, Chapter 2, verse 230). By contrast, ila, does not have such limitation and can be repeated without restriction (in terms of frequency), provided that the husband retracts by performing the act of expiation each time.

 b. The second concerns the form of retraction. In the case of talaq, the husband can resume conjugal life at will. No particular formality is required; and retraction can be explicit or implied. In the case of ila, however, retraction must take the form of an act of expiation (kaffara) in accordance with the Quranic prescriptions (Chapter 5, verse 89) before marital relations may resume.

4. Ila and public policy

a) Ila – some inherent aspects

As previously noted, ila has traditionally been used as a means for a husband to exert pressure or express discontent within the marriage by vowing abstinence from sexual relations. Under Islamic Sharia, this practice is preserved: husbands – even without making any formal oath of abstinence (ila) – are allowed to “discipline their wives” in cases of marital discord by abstaining from sharing the marital bed (hajr) as a corrective measure (Quran, Chapter 4, verse 34). Indeed, it is not uncommon that Muslim scholars justify the “rationale” behind this practice by stating that “a man may resort to ila…when he sees no other option but to abstain from sharing the marital bed as a means of disciplining and correcting his wife (italic added)…. In this case, his abstention during this period serves as a warning to deter her from repeating such behavior” (O. A. Abd Al-Hamid Lillu, ‘Mirath al-mutallaqa bi-al-‘ila – Dirasa fiqhiyya muqarana ma’a ba’dh al-tashri’at al-‘arabiyya [The Inheritance Rights of a Woman Divorced by Ila’: A Comparative Jurisprudential Study with Selected Arab Legislations]’ (2020) 4(3) Journal of the Faculty of Islamic and Arabic Studies for Women 630). It is therefore not surprising that some would view ila as “troubling” due to its perceived “sexism” and the fact that wives may find themselves at their husbands’ “mercy” with little thing to do (Raj Bhala, Understanding Islamic Law (Shar’ia) (Carolina Academic Press, 2023) 803).

These aspects, in addition with inherent gender asymmetry in the rights involved, calls into question the compatibility of ila with the public policy of the forum.

b) The position of the Nagoya High Court

As the Nagoya High Court rightly indicated, the exclusion of foreign law under the public policy exception does not depend on the content of the foreign law itself, assessed in abstracto. On the contrary, as it is generally accepted in Japanese private international law, public policy may be invoked based on two elements: (1) the result of applying the foreign law in a concrete case is found unacceptable in the eyes of Japanese law, and (2) there is a strong connection between the case and the forum (see K. Nishioka & Y. Nishitani, Japanese Private International Law (Hart, 2019) 22).

The Nagoya High Court’s explicit adherence to this framework, notably by engaging in an in concreto examination of the foreign law and avoiding invoking public policy solely on the ground of its content as some earlier court decisions suggest (see e.g. Tokyo Family Court judgment of 17 January 2019; see my English translation in 63 (2020) Japanese Yearbook of International Law 373), is noteworthy and should be welcomed.

That said, the Court’s overall approach raises some questions. The impression conveyed by the Court’s reasoning is that it focused primarily on the irretrievable breakdown of the marital relationship and the period of separation to conclude that there was no violation of public policy. In other words, since the marital relationship had reached a dead end, dissolving the marriage on the basis of objective grounds or on the basis of ila does not alter the outcome.

Although this approach is understandable, it would have been more convincing if the Court had carefully considered the nature of ila and its specific implications in this case, and eventually explicitly state that such elements were not established. These aspects appear to have been largely overlooked by the High Court, seemingly due to its unfamiliarity with Islamic legal institutions. It would have been advisable for the Court to address these aspects, at least to demonstrate its concerns regarding the potential abusive use of ila.

V. Concluding Remarks: Ila as a litigation strategy?

One may wonder why the husband in this case chose to resort to ila to end his marriage. One possible explanation is that Japanese courts have previously ruled that a talaq divorce in the form of triple talaq is inconsistent with public policy (Tokyo Family Court judgment of 17 January 2019, op. cit.). It appears that, anticipating a similar outcome, the husband in this case was advised to take a “safer approach” by relying on ila rather than resorting to triple talaq (see the comment by the law firm representing the husband in this case, available here – in Japanese only).  To be sure, associating talaq solely with its most contested form (i.e., triple talaq) is not entirely accurate. That said, considering how the case under discussion was decided, it is now open to question whether it would have been simpler for the husband to perform a single talaq and then abstain from retracting during his wife’s waiting period (iddah). At least in this way, the aspect of “disciplining the wife” inherent in ila would not be an issue that the courts would need to address

In Sinclair v Venezia Turismo, 2025 SCC 27 (available here) the Supreme Court of Canada has, by 5-4 decision, held that the Ontario court does not have jurisdiction to hear claims by Ontario residents against three Italian defendants in respect of a tort in Italy.  The Sinclair family members were injured in a gondola collision in Venice that they alleged was caused by the Italian defendants.  But there were several connections to Ontario.  The trip to Italy had been booked by Mr Sinclair using a premium credit card’s concierge and travel agency service [4, 156] and the gondola ride had been arranged through that service [15, 160].  The card was with Amex Canada and one or more contracts connected to the gondola ride had been made in Ontario.  The Sinclairs were also suing Amex Canada and the travel service for carelessness in making the arrangements with the Italian defendants, and those defendants attorned in Ontario [167, 172].  A core overall issue, then, was whether the plaintiffs would be able to pursue all of their claims arising from the gondola collision, against various defendants, in one legal proceeding in Ontario.

For assumed jurisdiction, Canadian common law requires that the plaintiff establish a presumed connecting factor (PCF) in respect of each defendant.  Once established, the defendant can rebut the PCF by showing that it does not point to a real relationship, or only a weak relationship, with the plaintiff’s chosen forum [7, 49, 202, 216].  It is well established that damage sustained by the plaintiff abroad, and continuing to be suffered in the forum, is not a PCF.  While less clear, the better view of the law is that the defendant’s being a “proper party” to a proceeding advanced against a local defendant is not a PCF.  So neither of these routes to jurisdiction, familiar in some legal systems, was available despite their fitting the facts.

Canadian courts have held that the fact that a contract connected with a tort was made in the forum is a PCF.  This is controversial because many have questioned the strength of this connection, based as it is on the place of making a contract, but it has been repeatedly endorsed by the Supreme Court of Canada.  Sinclair turned on whether this PCF had been established and if so rebutted [1, 51, 146].  The majority (decision written by Justice Cote) found the defendants had rebutted the PCF; the dissent (decision written by Justice Jamal) found not.

The reasons are a challenging read.  The majority and dissent disagree on many discrete points (including the standard of review and the standard of proof).  Many of these are essentially factual.  Because they do not see the facts the same way, it is hard to compare the legal analysis.  A key example is on the issue of what contract(s) had been made in Ontario.  The majority is not overly satisfied that any contract had been, but is prepared to accept that Mr Sinclair’s cardmember agreement was made in Ontario [102-103].  That contract is in a loose sense connected with the tort in Italy, but it is easy to see how one might think this is at best a very weak link [9].  In contrast, the dissent has no issue with the cardmember agreement having been made in Ontario [253, 259] and finds an additional contract also made in Ontario in respect of arranging the specific gondola ride [268].  That second contract is more closely linked to the tort and so the rebuttal analysis would be expected to differ from that relating to the cardmember agreement.  The majority does not find any such second contract at all: it sees this as a reservation made to arrange that the gondola be available, which is not a separate contract but rather a part of the way Amex Canada performs its service obligations under the cardmember agreement [105-107].

The result of the appeal is highly fact-specific.  But some useful general points can be extracted from the reasons.  First, the decision may add to our understanding of the test for when a contract made in the forum is “connected” to the tort.  In Lapointe (available here) the court had said that this is satisfied if “a defendant’s conduct brings him or her within the scope of the contractual relationship” AND “the events that give rise to the claim flow from the relationship created by the contract” [58, 215].  I confess to having had trouble understanding what the former aspect means.  What is it to be brought within the scope of the contractual relationship?  Is this a factual or legal question?  In what way would the Italian defendants be brought within the scope of the cardmember agreement (this does not seem possible) or even the second contract between Amex Canada and Carey International to arrange a gondola?  Do they get brought within the scope just because they end up being the relevant gondola providers?  Anyway, in this case, both the majority and the dissent seem to focus all of their analysis of whether the contract is connected to the tort on the second aspect: whether the tort “flows” from the earlier contract (a pretty easy test to meet here for all contracts involved) [128, 246].

Second, the judges engage in a lively debate about the standard of establishing a PCF.  This is understandable given the extent to which they disagree about the facts.  But their debate ends up being inconclusive.  For the majority see [59] to [62] and the conclusion that this is not an appropriate case to develop the law on this point (so these paragraphs, then, are markers for arguments parties might make in future cases in which the law might be developed).  For the dissent see [224] to [236] and the conclusion that what it considers the status quo on the issue remains the law (yet this is in dissent).  There may be common ground, since in both discussions care is taken, at least in places, to refer specifically to the distinction between disputes about facts and disputes about the application of the law to those facts.  A standard of proof, whether a balance of probabilities or a good arguable case, must be about facts and not law.  It does not make sense to talk about the standard of proof for establishing a point of law or satisfying a legal test.

Third, few Canadian cases have provided a detailed analysis of how the rebuttal of a PCF works, so this case is most welcome on that specific issue.  The majority offers some general considerations that feed into the analysis [67-72].  It also rejects the contention that rebuttal is a “heavy” burden on the defendant [74].  It calls the rebuttal “a shift in burden and perspective, not a shift in difficulty” [74, quoting the intervener BC Chamber of Commerce].  This language is likely be repeated quoted in subsequent decisions.  The majority also says that the PCF and rebuttal stages work in tandem and are complementary [74-75].  This reflects the idea that if the PCF is broad, there should be more scope for rebuttal, and if the PCF is narrow, less so.  The dissent does not disagree with this stated approach to the rebuttal analysis [see 217].  However, the judges disagree about whether the defendant’s reasonable expectations of where it might be sued can be considered as part of the rebuttal analysis.  The dissent says no [218, 291].  The majority says yes [71-72].

Finally, on the broader question of how willing courts should be to take jurisdiction over a defendant on grounds of efficiency, access to justice and avoidance of multiple proceedings, most comments from the judges are indirect.  The majority stresses the importance of “fairness” to defendants [45].  It rejects “bootstrapping” and insists that a PCF must be shown for each defendant [63].  It cautions against a jurisdiction analysis that considers “the factual and legal situation writ large” [63].  In contrast, the dissent sees the proceeding as one that “claims inseparable damages for these integrally related torts” [281] and rejects focusing on the collision as something separate from other facts and claims [249].  More directly, it states “[i]n a case alleging multiple torts, as in this case, or a case raising claims under multiple heads of liability, focussing on the dispute as a whole ensures that a court does not inappropriately hear only part of the case in the forum while leaving related claims to be heard in the extra-provincial or foreign court” [244].  In doing so it quotes the notorious para 99 of Club Resorts (available here), language that continues to trouble courts more than a decade later.  After Sinclair, are we closer to a principled answer for cases with related claims against multiple defendants?  By focusing on the narrow and specific questions raised by the particular PCF at issue, including identifying whether and where certain contracts were made, the broader debate is being conducted covertly rather than in the open.

Written by Héloise Meur, Université Paris 8

In two rulings dated 28 May 2025, the French Cour de cassation (Supreme Court) ruled on the issue of the law applicable to a sub-purchaser’s direct action in a chain of contracts transferring ownership, under European private international law. The issue is sensitive. The contractual classification under French law —an outlier in comparative law— had not been upheld by the Court of Justice of the European Union (CJEU) to determine international jurisdiction under the Brussels system (CJEU, 17 June 1992, C-26/91, Jakob Handte). Despite CJEU’s position, the Cour de cassation had consistently refused to adopt a tort-based qualification to determine the applicable law (esp. Civ. 1^st, 18 dec. 1990, n° 89-12.177 ; 10 oct. 1995, n° 93-17.359 ; 6 feb. 1996, n° 94-11.143 ; Civ. 3^rd, 16 janv. 2019, n° 11-13.509. See also, Civ. 1^st, 16 jan. 2019, n° 17-21.477), until these two rulings rendered under the Rome II Regulation.

The proceedings

In the first case (No. 23-13.687), a Luxembourgian company made available to a Belgian company certain equipment it had obtained through two lease contracts. The lessor had acquired the equipment from a French intermediate seller, who had purchased it from a French distributor, who had sourced it from a Belgian manufacturer (whose rights were ultimately transferred to a Czech company).

Following a fire that destroyed the equipment, the Dutch insurer — subrogated in the rights of the Luxembourgian policyholder — brought proceedings against the French companies before the French courts on the basis of latent defects. The manufacturer’s general terms and conditions included a choice-of-law clause in favour of Belgian law. The Belgian and Luxembourg companies sought various sums based on latent defects, lack of conformity, and breach of the selller’s duty to advise. The manufacturer voluntarily joined the proceedings.

Applying French law, the Court of Appeal held the insurer’s subrogated claims admissible and dismissed the French intermediary seller’s claims. The Court ordered the Czech manufacturer and French companies jointly and severally liable to compensate the Luxembourg company for its uninsured losses and to reimburse the French intermediary seller for the insured equipment. The manufacturer appealed to the Cour de cassation, and the French distributor lodged a cross appeal.

In the second case (No. 23-20.341), a French company was in charge of designing and building a photovoltaic power plant in Portugal. The French company purchased the solar panels from a German company. The sales contract included a jurisdiction clause in favour of the courts of Leipzig and a choice-of-law clause in favour of German law. In 2018, the Portuguese company, as assignee of the original contract, brought proceedings against the French and German companies seeking avoidance of the successive sales and restitution of the purchase price. Alternatively, the Portuguese final purchaser invoked the contractual warranty granted by the German manufacturer and sought damages. The Court of Appeal dismissed the purchaser’s claim under German law, which was applicable to the original contract. The Court of Appeal also declined jurisdiction over the French company’s claims against the German company due to the jurisdiction clause. The purchaser appealed to the Cour de cassation.

The legal question

Both appeals raised the question of the determination of the law applicable to the sub-purchaser’s direct action in a chain of contracts transferring ownership under European private international law, especially where a choice-of-law clause is included in the original contract.

The rulings of 28 May 2025

The Cour de cassation adopted the reasoning of the Jacob Handte judgment. The Court held that, in conflict of laws, the sub-purchaser’s action against the manufacturer does not qualify as a “contractual matter” but must be classified as “non-contractual” and therefore be governed by the Rome II Regulation (§§ 16 seq n° 23-13.687 ; §§ 18 seq n° 23-20.341).

The Court concluded that: “A choice-of-law clause stipulated in the original contract between the manufacturer and the first purchaser, to which the sub-purchaser is not a party and to which they have not consented, does not constitute a choice of law applicable to the non-contractual obligation within the meaning of Article 14(1) of that Regulation.” (§ 20, n° 23-13.687 ; § 22, n° 23-20.341).

This solution should be also supported by the Refcomp ruling (§ 18, n° 23-13.687 ; § 16, n° 23-20.341), in which the Court held that a jurisdiction clause is not enforceable against the sub-purchaser, “insofar as the sub-purchaser and the manufacturer must be regarded, for the purposes of the Brussels I Regulation, as not being bound by a contractual relationship” (CJEU, 7 Feb. 2013, C-543/10, para. 33).

According to the Cour de cassation, the law applicable to sub-purchaser’s claims against the manufacturer is the law of the place where the damage occurred, pursuant to Article 4 of the Rome II Regulation.

Comments

Firstly, the rejection of the contractual classification does not necessarily entail a tortious classification. To do so, it must also be established that the action seeks the liability of the defendant, in accordance with the definition adopted in the Kalfelis judgment (ECJ, 27 Sept. 1988, Case 189/87). It was not the case here, where the claims were based on latent defects and avoidance of contract.

Secondly, the choice of a non-contractual classification appears contrary to the developments in CJEU’s recent case law (H. Meur, Les accords de distribution en droit international privé, Bruylant, 2024, pp. 325 seq.), For the CJEU, it is sufficient to establish that the action could not exist in the absence of a contractual link for it to qualify as a “contractual claim” under Brussels I Regulation (CJEU, 20 Apr. 2016, C-366/13, para. 55, Profit Investment). The European Court further held that the identity of the parties is irrelevant to determine whether the action falls within the scope of contractual matters ; only the cause of the action matters (CJEU, 7 Mar. 2018, Flightright, joined cases C-274/16, C-447/16, C-448/16; and CJEU, 4 Oct. 2018, Feniks, C-337/17). Thus, the Court has moved away from its Jacob Handte case law.

Thirdly, limiting the effect of the choice-of-law clause to the contracting parties alone is inappropriate, as it will lead to the applicable law to the contract to vary depending on who invokes it (H. Meur, Dalloz actualité, 16 June 2025). This solution is also contrary to the European regulations. It is in contradiction with Article 3.1 of the Rome I Regulation, which states that “a contract shall be governed by the law chosen by the parties.” It is also incompatible with Article 3.2 of the Regulation. This article provides that “any change in the law to be applied that is made after the conclusion of the contract shall not […] adversely affect the rights of third parties,” from which it must be inferred a contrario that the original choice-of-law clause is enforceable against third parties (see the report by Reporting Judge S. Corneloup, pp. 21 seq.; also see the Report on the Convention on the Law Applicable to Contractual Obligations, OJEC, C 282, 31 Oct. 1980, para. 7 under the commentary on Article 3). For the sake of consistency, this understanding of the principle of party autonomy should also apply to Article 14 of the Rome II Regulation. Finally, Article 12 of the Rome I Regulation confirms that it is for the law applicable to the contract to determine the persons entitled to invoke it and the conditions under which they may do so (by contrast, the Vienna Convention on the International Sale of Goods and the Hague Convention do not apply to the question of the effect of the contract on third parties – see in particular Hague Convention, 1955, Art. 5.4; Civ. 1st, 12 July 2023, No. 21-22.843).

Thus, the law applicable to the sub-purchaser’s direct action should be the one chosen by the parties to the original contract (regardless of the claiming party), provided that this choice is intended to govern the contract. In the absence of a chosen law, the law of the habitual residence of the seller, as the debtor of the characteristic performance, should apply. If the designated law recognises, in principle, that a third party may invoke the rights available to the original contracting purchaser, the Vienna and Hague Conventions, which are applicable before the French courts, may regain their relevance in determining the content of those rights (see V. Heuzé, RCDIP, 2019, p. 534; E. Farnoux, AJ Contrat, 2020, p. 521).

Unfortunately, this is not the path taken by the Cour de cassation in its rulings of 28 May 2025. In practice, the original seller may be bound in respect of certain sub-purchasers, particularly those established in France, even though it may have had no knowledge of the successive sales. Such a solution increases legal uncertainty.