Developments in PIL Domestic legislation Normative texts Views and comments

Online Symposium on the Law Governing Arbitration Agreements: A View from the Strand

The post below was written by Manuel Penades, who is a Reader in International Commercial Law at King’s College London. It is the second contribution to the EAPIL online symposium on the English Law Commission’s proposed reform of the law governing arbitration agreements. The other posts are written by Alex Mills, George Bermann, Sylvain Bollée, Matthias Lehmann and Giuditta Cordero-Moss.

Readers are encouraged to participate in the discussion by commenting on the posts. 

This post examines the changes proposed by the Law Commission of England and Wales to the choice of law rules for arbitration agreements. Previous contributions to this Symposium have transcribed the text of the draft legislation, which can be found here. The Law Commission introduces three significant amendments that impact the three steps of the common law doctrine of the proper law of the contract. First, the proposal limits the types of choice of law clauses that can demonstrate an express selection of the law applicable to arbitration agreements. Second, it eliminates the possibility to choose the governing law impliedly. Third, it replaces the closest and most real connection test with a hard-and-fast rule in favour of the law of the seat.

Each of these changes requires analysis, followed by a reflection on the New York Convention.

Express Choice of Law

The new rule continues to respect the parties’ freedom to choose the law governing their arbitration agreement. Party autonomy, however, is tempered by proposed section 6A(2) of the Arbitration Act, which provides that an ‘agreement between the parties that a particular law applies to the contract to which the arbitration agreement forms part does not, of itself, constitute an express agreement that that law also applies to the arbitration agreement’. The rule is apparently simple and excludes the possibility to rely on a generic choice of law clause applicable to the contract that includes the arbitration agreement. Section 6A(2) AA, however, does not capture other scenarios, which might become a source of controversy.

The first refers to cases in which the only choice of law in the whole contract is found in the arbitration agreement itself (e.g., ‘the arbitrators shall decide the dispute in accordance with the law of X’). While these cases do not refer to the arbitration agreement specifically, they are express references to the governing law of the whole contract and are contained in the arbitration agreement itself. It is unclear whether these choices will be express enough to satisfy section 6A(2) AA.

The second scenario refers to cases in which the matrix contract not only includes an express choice of law clause applicable to the whole ‘Agreement’, but also a clause in the contract that defines ‘Agreement’ as encompassing all the clauses incorporated in the contractual document, including the arbitration agreement. The UKSC ruled in Kabab-Ji v Kout Food [2021] UKSC 48 that ‘the effect of these clauses is absolutely clear’ [39] and amounts to an express choice also for the arbitration agreement. The Law Commission’s proposal does not mention whether section 6A(2) AA intends to overrule Kabab-Ji. In these scenarios it is not the generic choice of law clause ‘of itself’ that supports the finding of an express choice of law but the combined reading of that clause alongside the definition of the term ‘Agreement’ expressly agreed by the parties in another clause of the contract.

Neither of these uncertainties would exist in the current regime under Enka v Chubb [2020] UKSC 38, as the same law would apply under express or implied choice of law.

The Elimination of Implied Choice of Law

Enka clarified that the designation of a seat does not amount to an implied choice of the law governing the arbitration agreement. This reduced, yet did not eliminate, the uncertainty surrounding implied choice. The proposal of the Law Commission goes much further; it eliminates implied choice altogether from the choice of law rules applicable to arbitration agreements. This is quite revolutionary and might come as a surprise.

Notwithstanding the complexities caused by its application, the courts of England have never questioned the acceptance of implied choice and the UKSC confirmed in Enka that ‘an implied choice is still a choice which is just as effective as a choice made expressly’ [35]. An implied choice is a manifestation of party autonomy, a principle which is at the root of English contract and private international laws.

The proposed new rule also runs contrary to the acceptance of implied choice of law in the vast majority of instruments governing international business transactions (see article 3.1 Rome I, article 14.1 Rome II or article 4 Hague Principles on Choice of Law).

Against this background, disregarding an implied choice of law might seem a step backwards in the common law tradition and global trends. The truth, however, is that decades of arbitration-related litigation in England demonstrate that the inquiry around implied choice is a source of significant uncertainty, expense and tactical litigation. The Law Commission is willing to adopt a regime that disregards cases of real (yet implied) choice of law in exchange for the certainty and savings produced by the elimination of implied choice of law. This less litigious regime makes for better arbitration regulation and strengthens the position of England as efficient arbitration destination.

The proposed solution does not necessarily curtail party autonomy. In fact, the rule after Enka that an implied choice of law for the matrix contract automatically amounts to an implied choice of law for the arbitration agreement, while apparently straightforward, might not always be reflective of the real intent of the parties. The proposed rule eliminates such risk of artificiality.

Further, case law shows that in most disputes where the issue of implied choice arises, English law offers the most arbitration-friendly outcome among the various alternative laws. Under the proposed reform, those cases will be resolved frequently in favour of English law pursuant to the default rule. This will generally protect the parties’ agreement to arbitrate more than under the current regime.

From a normative point of view, the Law Commission’s proposal also eliminates the somewhat artificial cases of double implication, where an implied choice of law for the matrix contract is used as evidence to find an implied choice of the law governing the arbitration agreement (see the conclusion of the minority in Enka [207, 228]).

Finally, the proposal eliminates the confusion sometimes perceived in English judgments between the test applicable to imply a choice of law and the (stricter) requirements to imply an ordinary contractual term [Enka [35] or Kabab-Ji [53]].

The Law of the Seat and Role of the Validation Principle

Under the proposed regime, the absence of an express choice results in the application of the law of the seat. Hard-and-fast rules are alien to the common law doctrine, where the reference to the closest and most real connection permits certain room for manoeuvre in the determination of the applicable law. Other choice of law regimes that provide hard-and-fast rules incorporate escape clauses that allow for the exceptional disapplication of the identified law (e.g., article 4.3 Rome I). In contrast, the proposed rule lacks any reference to the possibility to escape from the law of the seat.

One could wonder whether this could be a residual role for the validation principle. This principle was used in Enka to support the application of the law of the seat when an implied choice in favour of the law of the matrix contract led to a serious risk that the arbitration agreement would be invalid or ineffective. The expulsion of implied choice from the proposed regime would eliminate the raison d’etre of the validation principle. Still, the Law Commission does not exclude the principle in absolute terms, and rather states that ‘we do not need the validation principle for that purpose’ [Para. 12.56]. The question then arises whether other purposes exist.

One option would be to retain the application of the validation principle to correct express choices of law that render the arbitration agreement invalid or ineffective. The answer should be negative. The role of courts is not to improve the contract (Arnold v Britton [2015] UKSC 35, [20]). The validation principle allows the court to resort to the more favourable interpretation when the contract allows for various possible interpretations. When the choice is express, however, there is only one undisputable choice, even if it renders the arbitration agreement invalid or ineffective. In those cases, party autonomy (and the pathologies derived from it) must prevail. Any deviation from the principle of party autonomy would have required an express rule in the Law Commission’s proposal.

The other possible application of the validation principle would be in the context of the default rule, when the law of the seat renders the arbitration agreement invalid or ineffective. Indeed, the majority of the UKSC in Enka suggested (but did not confirm) that the closest connection test might itself be subject to the validation principle [146]. As noted by the Law Commission [para. 12.58], my response to the Second Consultation said that it would be odd to apply the validation principle to escape from an invalidity provided by English law itself under the default rule. However, the proposed default rule is not just in favour of English law, but in favour of the law of any seat. This approach could open the door to the application of the validation principle when, unlike the law of the seat, English law rendered the arbitration agreement valid and effective. While the Final Report of the Law Commission does not explore this option, such extended reach of the validation principle would deviate from the finality and simplicity with which the Law Commission views the default rule. Also, it might not be an appropriate and efficient policy to use English law to enforce a foreign arbitration agreement when the parties have not selected the governing law and the law of the seat would render it invalid or ineffective.

The Conflict with the NYC

Article V(1)(a) NYC provides that arbitration agreements shall be governed by the law to which the parties subjected it or, failing any indication thereon, by the law of the country where the award was made. The default rule in the Law Commission’ proposal aligns English law with the NYC, which is a welcome result.

Section 103(2)(b) AA incorporates article V(1)(a) NYC and therefore allows ‘any indication’ of choice of law made by the parties. The UKSC concluded unanimously in Kabab-Ji that ‘the word “indication” signifies that something less than an express and specific agreement will suffice’ [33]. It is unclear whether the Law Commission intends the new choice of law rule to apply in the context of section 103 AA. The UKSC said in Kabab-Ji that the common law rules on choice of law for arbitration agreements were not ‘directly applicable’ in the context of NYC enforcement actions [35]. Also, awards caught by section 103 AA have a foreign seat by definition and are not English arbitrations. Still, the proposal makes it clear that ‘the new rule would apply whether the arbitration was seated in England and Wales, or elsewhere’ [12.75]. An option would be to interpret this statement as referring to every scenario in which English courts examine an arbitration agreement (whether seated in England and Wales or elsewhere) with the exception of cases caught by section 103 AA. That is, two different choice of law treatments would co-exist within the Act. This internal dealignment would be undesirable and could lead to serious inconsistencies. The same arbitration agreement in favour of an arbitration seated abroad could be subject to different laws in pre-award disputes (e.g., section 9 AA) and post-award litigation (e.g., section 103). The UKSC said in Enka [136] and in Kabab Ji [35] that this divide would be ‘ilogical’.

The better interpretation is that the Law Commission’s proposal also extends to section 103 AA cases. Nothing in the proposal expressly excludes this reading. In fact, the Report argues that the NYC allows, but does not require, the recognition of implied choices [12.47] and concludes that the proposal is compatible with the NYC [12.52]. Ultimately, the new rule replaces the common law doctrine with a statutory provision, which becomes part of the of the regulatory fabric of English arbitration law and should not be limited, unless otherwise provided, to areas originally governed by the common law. Section 100(2) AA shows that critical parts of the notion of arbitration agreement in Part III (where section 103 AA belongs) ‘have the same meaning as in Part I’ (where the new section 6A AA would be placed). Such internal coherence of English arbitration law supports the application of the proposed rule across the board. Still, should Parliament adopt of the Law Commission’s proposal, they would need to be aware of two undesirable (yet tolerable) dealignments.

The first is that English law would move away from the prevailing interpretation of article V(1)(a) NYC as regards the acceptance of implied choice. The UKSC in Kabab-Ji objected to this departure and held that ‘it is desirable that the rules set out in article V(1)(a) for determining whether there is a valid arbitration agreement should not only be given a uniform meaning but should be applied by the courts of the contracting states in a uniform way’ [32]. Still, England would not be alone in this travel. For instance, France has also departed from the choice of law rule in the NYC. Moreover, the generally pro-arbitration results usually achieved by the proposed rule could well place the reform within the favourable gateway of article VII NYC.

The second consequence is that the same arbitration agreement (and award) might be treated differently between English and foreign courts if an existing implied choice of law disregarded in England is effective in other jurisdictions. It should be noted, however, that retaining the possibility of implied choice does not guarantee the uniformity of outcome. For instance, the same dealignment of outcome could occur between two legal systems that accepted the possibility of implied choice of law if one favoured the law of the matrix contract whereas the other veered toward the law of the seat.


The reasons above support the view that the potential disregard of real (yet implied) choice in some exceptional cases and the risk of some disfunctions derived from the described dealignments would be compensated by the significant simplification and savings produced by the Law Commission’s proposal. The draft Bill is therefore well-founded, courageous and beneficial to reinforce English law’s position at the forefront of international arbitration globally.

0 comments on “Online Symposium on the Law Governing Arbitration Agreements: A View from the Strand

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: