England hosts roughly 5,000 arbitrations per year, contributing an estimated GBP 2.5 billion to the national economy. Yet the governing framework for arbitration and English commercial law just underwent the most significant reform in three decades. The Arbitration Act 2025, which received Royal Assent on 24 February 2025 and came into force on 1 August 2025, amends the Arbitration Act 1996 by codifying Supreme Court rulings and reversing controversial precedents.
The central question is whether this framework delivers justice for commercial parties through certainty, autonomy, and efficiency. The statutory architecture, foundational doctrines, key case law, the AA 2025 reforms, enforcement mechanisms, and London's global competitive position all bear on that question. The answer is yes, but with tension points that every LLB and LLM student should understand and be prepared to critique.
Finality conflicts with legal development. Party autonomy conflicts with public interest safeguards. The implied common law duty of confidentiality leaves gaps that statute does not fill. Section 69 appeals on points of law make England an outlier among arbitration-friendly jurisdictions. These tensions run through every section that follows, and they are precisely the points that distinguish a strong exam answer from a descriptive one.
The Statutory Framework: Arbitration Act 1996 and Its Guiding Principles
The Arbitration Act 1996 opens not with definitions but with a statement of philosophy. Section 1 sets out three principles that function as the interpretive lens for every provision of arbitration and English commercial law that follows.
Section 1(a) provides that the object of arbitration is the fair resolution of disputes by an impartial tribunal without unnecessary delay or expense. Section 1(b) enshrines party autonomy: parties should be free to agree how their disputes are resolved, subject only to such safeguards as are necessary in the public interest. Section 1(c) limits court intervention. A court should not intervene in arbitral matters except as provided by the Act.
These are not decorative. They are operative principles that courts apply when interpreting ambiguous provisions. Party autonomy, in particular, means that most provisions of the Act are default rules. Parties can contract out of them.
The Act distinguishes between mandatory and non-mandatory provisions, listed in Schedule 1. Mandatory provisions include the requirements for a valid arbitration agreement, the duty of the tribunal to act fairly and impartially under s.33, and the court's power to remove an arbitrator under s.24. Non-mandatory provisions can be modified by agreement, giving parties direct control over procedural rules, evidential standards, and remedial powers. The practical significance of this boundary was illustrated in Lesotho Highlands Development Authority v Impregilo SpA [2005] UKHL 43, where the House of Lords held that s.48(4), which confers a power to award compound interest, is non-mandatory and that a tribunal purporting to exercise a broader remedies power than the parties had agreed would exceed its jurisdiction under s.68.
A further mandatory provision of immediate practical importance is s.9, which requires a court to grant a stay of legal proceedings brought in breach of an arbitration agreement. In Halki Shipping Corporation v Sopex Oils Ltd [1998] 1 WLR 726, the Court of Appeal confirmed that the stay is mandatory once an arbitration agreement covering the dispute exists, regardless of whether the claim is genuinely disputed. Parties cannot litigate around an arbitration clause by alleging that the claim is unanswerable.
England consciously rejected the UNCITRAL Model Law. The Departmental Advisory Committee, chaired by Lord Justice Mustill, published its 1989 report identifying gaps in the Model Law, untested provisions, and bifurcation risks. Lord Justice Steyn defended the decision on the basis that English arbitration law had a sufficiently developed common law foundation to justify an independent statute. The result is a framework broadly compatible with the Model Law in substance but distinctly English in structure. This choice explains why English arbitration law sometimes diverges from the approach in Singapore, Hong Kong, or Germany. Students should understand the reasoning: the DAC concluded that adopting the Model Law wholesale would have required abandoning settled English principles without a clear gain in international compatibility.
Separability, Kompetenz-Kompetenz, and Jurisdictional Doctrines
Can an arbitration clause survive if the contract containing it was procured by bribery? The House of Lords said yes. This answer rests on two doctrines that underpin the jurisdictional architecture of arbitration and English commercial law.
Section 7 of the AA 1996 codifies the separability doctrine. An arbitration agreement that forms part of another agreement is to be treated as a distinct agreement. It is not regarded as invalid, non-existent, or ineffective simply because the main contract is invalid, did not come into existence, or has become ineffective. The arbitration clause lives its own legal life. The doctrine has common law roots predating the Act. In Harbour Assurance Co (UK) Ltd v Kansa General International Insurance Co Ltd [1993] QB 701, the Court of Appeal held that an arbitration clause survived an allegation that the insurance contract containing it was illegal, on the basis that the clause was separable from the main agreement. Section 7 gave statutory force to that reasoning.
The leading authority is Fiona Trust & Holding Corp v Privalov [2007] UKHL 40. The dispute involved allegations of bribery in the chartering of vessels. The defendants argued that if the underlying charters were voidable for bribery, the arbitration clauses within them were equally tainted. Lord Hoffmann rejected this. He applied a "rational businessperson" test: rational parties intending to refer disputes to arbitration would not have intended the scope of that reference to depend on whether their dispute was characterised as one "under" or "arising out of" the contract. The phrase "disputes under this contract" was held to encompass disputes about the contract's validity, including claims for rescission. The effect is to prevent parties from escaping arbitration by attacking the main contract.
Section 30 codifies the Kompetenz-Kompetenz principle. The tribunal may rule on its own substantive jurisdiction, including whether there is a valid arbitration agreement, whether the tribunal is properly constituted, and what matters have been submitted to arbitration. This is subject to challenge before the courts under s.67.
Previously, under the approach in Dallah Real Estate and Tourism Holding Company v Ministry of Religious Affairs of the Government of Pakistan [2010] UKSC 46, the court conducted a full rehearing of the jurisdictional question, with fresh evidence permitted. The AA 2025 reformed s.67 to remove this full rehearing, bringing English law closer to other major arbitration jurisdictions where courts give weight to the tribunal's own findings on jurisdiction.
Appeals on Points of Law: Why Section 69 Makes English Arbitration Distinctive
Most arbitration-friendly jurisdictions treat awards as final. England allows appeals on points of law. This single feature most clearly distinguishes arbitration and English commercial law from the international mainstream.
Section 69 of the AA 1996 permits a party to appeal to the High Court on a question of English law arising out of an award. The test operates on two tracks. First, where the tribunal is "obviously wrong" on a question of law, the court will grant permission to appeal. Second, where the question is one of "general public importance" and the tribunal's decision is "open to serious doubt," permission may also be granted.
The test under s.69 has antecedents in the pre-1996 regime. In Pioneer Shipping Ltd v BTP Tioxide Ltd (The Nema) [1982] AC 724, Lord Diplock set out guidelines for granting leave to appeal under the Arbitration Act 1979 (the predecessor provision). The courts applied those guidelines strictly, granting leave only in exceptional cases involving one-off contracts. The AA 1996 codified and refined that approach, but the underlying judicial reluctance to interfere with awards persists.
The provision is opt-outable. Many institutional rules exclude it by default. Article 26.8 of the LCIA Rules constitutes an agreement to exclude s.69 unless the parties expressly opt back in. The appeal must be brought within 28 days and requires permission from the court. The opt-out mechanism means parties retain control over whether this right exists in their arbitration.
The tension is real. Section 69 serves an important function: it develops English commercial law. Arbitral awards that remain unchallenged cannot generate precedent. Without s.69, the body of English commercial law in areas like shipping, insurance, and commodities would stagnate. This is a genuine public interest argument.
Against this stands the principle of finality. Parties choose arbitration partly because they want a binding resolution without the delays of appellate litigation. The counter-evidence is instructive: between 2015 and 2017, only 1 of 112 challenges under s.68 (serious irregularity) succeeded. English courts are highly reluctant to interfere with awards. The theoretical availability of appeal does not translate into routine disruption.
No equivalent to s.69 exists in Model Law jurisdictions. Whether England's approach is a strength or a weakness depends on whether one prioritises legal certainty and the development of commercial law, or absolute finality. The opt-out mechanism strikes an acceptable balance by giving parties the choice. Students should be prepared to defend either position in an exam setting.
The Arbitration Act 2025: Key Reforms and Their Impact
The Arbitration Act 2025 received Royal Assent on 24 February 2025 and came into force on 1 August 2025. It contains 18 clauses. Paul Hastings described it as "Refinement Not Revolution." The reforms are targeted legislative responses to specific judicial decisions that created uncertainty or inefficiency in arbitration and English commercial law.
Section 6A: Governing law of the arbitration agreement. This is the most doctrinally significant reform. In Enka Insaat ve Sanayi AS v OOO Insurance Company Chubb [2020] UKSC 38, the Supreme Court held that the governing law of the arbitration agreement, absent express choice, followed the law of the main contract rather than the law of the seat. This created complexity where the seat was in England but the main contract was governed by foreign law. Section 6A reverses this: the governing law of the arbitration agreement now defaults to the law of the seat. For London-seated arbitrations, this means English law governs the arbitration agreement unless the parties agree otherwise.
Section 39A: Summary disposal. The tribunal now has an express statutory power to dispose of issues on a summary basis where a party has "no real prospect of success." This mirrors the standard under CPR Part 24 for summary judgment in litigation. Before this reform, tribunals were reluctant to dispose of claims summarily because no express power existed. Section 39A removes the doubt and promotes efficiency.
Section 23A: Arbitrator disclosure duty. This codifies the Supreme Court's decision in Halliburton Company v Chubb Bermuda Insurance Ltd [2020] UKSC 48, which held that an arbitrator has a duty to disclose circumstances that might reasonably give rise to justifiable doubts about their impartiality. The particular concern is repeat appointments by the same party or the same law firm in related arbitrations. Codification provides certainty and reinforces the integrity of the arbitral process.
Section 67 reform. The removal of full rehearing on jurisdiction challenges strengthens the tribunal's authority and reduces the scope for tactical disruption.
Strengthened arbitrator immunity. The Act extends the existing immunity of arbitrators from civil liability, providing greater protection for arbitrators acting in good faith.
These reforms address certainty (s.6A), efficiency (s.39A), and integrity (s.23A). As Ashurst Perkins Coie observed: "After 1 August 2025, London will have one of the most up-to-date procedural laws globally."
Confidentiality in English Arbitration: The Implied Duty and Its Limits
English law imposes no statutory duty of confidentiality on arbitration. Neither the AA 1996 nor the AA 2025 contains one. The duty that exists under arbitration and English commercial law is implied at common law.
In Dolling-Baker v Merrett [1990] 1 WLR 1205, the Court of Appeal first recognised an implied duty of confidentiality arising from "the nature of arbitration itself." The reasoning was that the private character of arbitral proceedings implied an obligation not to disclose documents or information generated in the arbitration.
The Court of Appeal confirmed and developed this in Ali Shipping Corporation v Shipyard Trogir [1998] EWCA Civ 3054. The implied duty prevents parties from using arbitration materials in other proceedings without consent or court permission. The court identified four exceptions: consent of all parties, order of the court, reasonable necessity for the protection of a party's legitimate interests, and the interests of justice. The scope of the "reasonable necessity" exception was examined in Hasseneh Insurance Co of Israel v Stuart J Mew [1993] 2 Lloyd's Rep 243, where Colman J held that a party may disclose an arbitral award to the extent necessary to establish or protect its legal rights in subsequent court proceedings. The award, unlike the documents produced in the arbitration, carries a stronger claim to disclosure because it is the definitive resolution of the parties' dispute. These exceptions create real gaps. A party involved in related litigation or regulatory proceedings may find that the implied duty does not prevent disclosure to third parties such as regulators, insurers, or co-defendants.
The Law Commission considered codification in its 2023 Final Report and declined to recommend it. The common law duty was considered sufficiently developed and flexible. This stands in contrast to Singapore, where confidentiality in arbitration is codified in statute.
The practical recommendation is direct. Parties who want robust confidentiality protections should include express confidentiality clauses in their arbitration agreements, specifying the scope of protected information, permitted disclosures, and remedies for breach. The implied duty provides a baseline, not comprehensive protection.
Enforcement of Awards: The New York Convention and English Law
Enforcement is the practical payoff of choosing London as a seat for arbitration and English commercial law disputes. An award that cannot be enforced across borders is of limited commercial value.
The Convention on the Recognition and Enforcement of Foreign Arbitral Awards 1958 (the New York Convention) is implemented in English law via Part III of the AA 1996, sections 100 to 104. A party holding a New York Convention award may apply to the High Court under s.101 for leave to enforce the award as if it were a court judgment.
The grounds for refusing enforcement under s.103 mirror Article V of the New York Convention: incapacity of a party, invalidity of the arbitration agreement, lack of proper notice, the award dealing with matters beyond the scope of the submission, procedural irregularity in the composition of the tribunal or the conduct of proceedings, the award not yet being binding or having been set aside, and public policy. The burden of proof falls on the party resisting enforcement.
English courts adopt a pro-enforcement bias. Refusal of enforcement is the exception. The High Court requires the resisting party to establish one of the s.103 grounds to a high standard of proof. This judicial attitude is well established and reinforces the attractiveness of London as a seat. In Yukos Capital SARL v OJSC Rosneft Oil Company [2012] EWCA Civ 855, the Court of Appeal upheld enforcement of a Russian arbitral award despite the fact that Russian courts had annulled it. The Court of Appeal held that the Russian annulment decisions were the product of a partial and dependent judiciary and therefore could not be recognised in England as a matter of public policy. The case illustrates both the strength of the pro-enforcement presumption and the breadth of the public policy exception in s.103(3).
Domestic awards (those made in England) are enforced under s.66 of the AA 1996, which permits enforcement with the leave of the court in the same manner as a judgment. The applicant files the original award or a certified copy, together with the original arbitration agreement, and the court grants leave unless the respondent can show a ground for refusal.
The combination of a London seat, English governing law, and the New York Convention gives parties access to enforcement in 172 contracting states. This is a powerful practical advantage that underpins London's position as a global arbitration seat.
London as a Global Seat: Competing with Singapore and Paris
The 2025 Queen Mary University of London / White & Case International Arbitration Survey found London and Singapore tied at 31 to 34 percent as the most preferred arbitral seats globally. London led in four of six surveyed regions. The competition is real, and the AA 2025 is partly a response to it. Arbitration and English commercial law must evolve to retain London's position.
The LCIA reported 365 referrals in 2025, with 87 percent of cases seated in London, 79 percent governed by English law, and parties drawn from 103 jurisdictions. Median costs were US$117,653. These figures reflect sustained demand.
London's advantages are structural. The Commercial Court and its specialist judges have unmatched depth of experience in arbitration matters. The pool of specialist arbitrators is large and diverse. The body of established case law provides predictability. English is the working language. The GMT timezone permits communication with both Asian and American markets within a single working day.
Singapore is London's closest competitor and has launched its own arbitration law consultation in 2025. Paris, home to the ICC, remains a strong third option, particularly for civil law parties. The competition is healthy and pushes all three seats toward modernisation.
The AA 2025 reforms are designed to maintain London's edge. Summary disposal (s.39A) addresses a complaint that arbitration is too slow. The seat-law default (s.6A) removes a source of uncertainty that could have driven parties toward Singapore or Paris. The reforms signal that English arbitration law is actively adapting, not resting on reputation. The GBP 2.5 billion annual economic contribution provides ample incentive to keep the framework current. For students, the competitive dynamics between London, Singapore, and Paris offer a rich source of comparative analysis in any exam on arbitration and English commercial law.