Law Applicable to Substantive Validity of Arbitration Agreements: Reflections under the Model Law

Arbitration is based on consent and generally on an agreement between the parties in dispute. Such agreement is considered separable from the main agreement, mainly for preserving its validity.

The arbitration agreement’s formation, interpretation, termination, scope and effects are subject to the law applicable to it, which is generally referred to as the law applicable to the substantive validity of the arbitration agreement.

Such law is largely considered to include aspects such as its interpretation, as well as issues of consent, error, fraud, and the scope of the arbitration agreement.

Other aspects of an arbitration agreement may be governed by different laws, such as:

• its formal validity (mostly governed by the law of the seat),
• the subjective capacity of the parties (normally governed by the law of their nationality in case of a natural person and its seat in case of a corporation),
• the authority of representatives to enter into an arbitration agreement (governed by the law applicable to the power of attorney),
• matters of subrogation and transfer, such as non-signatories (however, these are most of the times considered to be also governed by the law applicable to the substantive validity of the arbitration agreement, see e.g. Kabab-Ji which concerned non-signatories, and where it was “undisputed that the law governing the validity of the arbitration agreement governs the question of whether KFG became a party to the arbitration agreement”),
• the question of objective arbitrability (governed normally by lex fori), or consumer protection aspects, etc.

Determining the law applicable to the substantive validity of an international arbitration agreement might be relevant during all stages of the arbitral process, for example when jurisdiction of the tribunal is contested, to determine whether there is a valid arbitration agreement, or when anti-suit injunctions are sought, and of course at the set-aside and enforcement stages.

According to the doctrine, few options exist as to which might be the law applicable to the substantive validity of the arbitration agreement: (a) the law of the forum where judicial enforcement of the agreement is sought; (b) the law expressly or impliedly chosen by the parties to govern the arbitration agreement itself; (c) the law of the arbitral seat; (d) a “validation” principle; and (e) international law. (see Chapter 2: International Arbitration Agreements: Legal Framework, in Gary B. Born, International Arbitration: Law and Practice (Second Edition), 2nd edition (© Kluwer Law International; Kluwer Law International 2015) pp. 47 -72).

Practical Implications: Dallah and Kabab-Ji sagas

The importance of getting the applicable law right (and uniformly) cannot be stressed enough. The so-called Dallah or Kabab-Ji sagas are illustrative in this respect.

In both the above cases, courts in England and France had divergent views as to the applicable law to the arbitration agreements. Applying (or interpreting divergently) separate laws to the arbitration agreements, English courts deemed the tribunal had no jurisdiction over a non-signatory, whereas under the French law at the seat, it had. The result was that the French court – the seat court – upheld the award, and dismissed the annulment action, but the English court refused enforcement of the same.

The courts reached different conclusions because they applied different laws to the arbitration agreement, according to their choice-of-law systems.

On one hand, according to French law jurisprudence, the arbitration agreement is deemed to be independent from any domestic law. Its validity depends only on the common intention of the parties, without it being necessary to make reference to a national law (Court of cassation, 20 December 1993, Dalico v. Khoms et El Mergeb).

On the other hand, the English courts followed the Sulamerica standard applying the English conflict of laws rules, which require to consider (i) whether the parties expressly chose a law governing the arbitration agreement; (ii) if they did not, whether the parties made an implied choice of such law; and (iii) in the absence of either choice, which system of law has the “closest and most real connection” with the arbitration agreement.

Neither France, nor England are parties to the Model Law, but, in addition to both Paris and London being important seats of arbitration, the above sagas underscore the practical implications of determining the applicable law to arbitration agreements.

The Applicable Law to Arbitration Agreements under the Model Law

The UNCITRAL Model Law on International Commercial Arbitration (1985), with amendments as adopted in 2006 (the “Model Law”), does not contain an express choice-of-law rule for the law applicable to the substantive validity of the arbitration agreement.

It only contains provisions dealing with the setting aside and enforcement, namely articles 34(2)(a)(i), respectively 36(1)(a)(i).

These provisions basically say, in a similar fashion to article V1(a) of the NYC (United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards (New York, 10 June 1958), that an award might be set aside or enforcement refused if the arbitration agreement was not valid under the law to which the parties subjected it, or in the absence of any indication thereon, under the law of the seat.

The above provisions are largely considered to apply to questions of substantive validity of the arbitration agreement as well.

The conclusions that seem to be clear from the above provisions are that:

• the parties’ agreement is the starting point to determine the applicable law, reflecting the key principle of party autonomy in arbitral proceedings,
• second, parties’ choice can be express or implied; this conclusion steams from the wording used by the Model Law: “failing any indication thereon”, which should be construed to allow not only express, but also implied choice, and lastly,
• in the absence of parties’ agreement, the fallback is on the law at the seat.

Looking at few Model Law jurisdictions, there is rarely a reference to the Model Law by courts when determining the law applicable to the substantive validity of arbitration agreements. However, courts often rely on the NYC provisions, which basically are the same as those of the Model Law, as mentioned above.

I have identified four main categories in few analyzed Model Law jurisdictions – either statutory or case law-based conclusions, specifically in cases where parties agreed on the law governing their main contract and selected the seat of the arbitration in a different jurisdiction:

1. Countries applying the law of the main contract, mainly as implied choice (e.g. Austria – with fallback on the law of the seat applying NYC, Singapore).
2. Countries applying the law of the seat, either as provided by statutory provisions or by case law interpretation, but mainly as the fallback (Scotland, Belgium, Bulgaria, Cyprus, Greece, Germany).
3. In between the two systems: Hungary, whose statutory provisions stipulate that if no law is chosen by the parties, the law governing the main contract applies. However, the law of the seat may be used if it has a closer connection to the agreement.
4. So-called validation principle, where courts apply any of the prescribed laws that validate the arbitration agreement (e.g. Spain, Peru), the laws mainly considered being the law governing the substance of the dispute, the law of the main contract, and the law of the seat. Similar provisions in some non-Model Law countries, including chiefly Switzerland, and also Portugal and Romania.

This might be seen as an inconsistent interpretation of the Model Law. However, it might be just a reporting issue, as jurisdictions applying the law of the seat to the substantive validity of the arbitration agreements, might already have found there was no express or implied choice of the parties.

Proposed Interpretation of the Model Law

My take on the interpretation of articles 34(2)(a)(i), respectively 36(1)(a)(i) of the Model Law (likewise on article V(1)(a) NYC), which, as mentioned, are generally considered to apply to the substantive validity of the arbitration agreements as well, is that the likely intention of the legislator was to consider parties’ choice for the law of the main contract as an implied choice for the law applicable to the arbitration agreement.  

The starting point for my analysis is that there are arguably two main systems that might govern the substantive validity of an arbitration agreement: the law of the main contract and the law of the seat.

Further, it should not be debatable that the Model Law allows for implicit choice, given the use of the wording “failing any indication thereon” in articles 34(2)(a)(i), respectively 36(1)(a)(i).

In addition, in the absence of choice (express or implied), the fallback, as an objective connector, is on the law of the seat.

Therefore, under both the NYC and the Model Law, the implicit choice was most likely intended for the law of the main contract, as the fallback is on the law of the seat in both mentioned instruments, as an objective connector that comes into play in the absence of parties’ choice (express or implied), which in turn, is of subjective nature.

This makes sense because it gives prevalence to the intention and choice of the parties in a process of contractual interpretation, whereas determining that the objective factor/closest connection leading to the law of the seat, is an objective process applied irrespective of the parties’ intentions.

The interpretation seems to be consistent with the approach in most jurisdictions, apart from those that only provide a two-stage test: express choice of the parties and fallback on the law of the seat (e.g. English Arbitration Act 2025). But apart from those situations, most of the analysed jurisdictions apply a three-stage test: express choice, implied choice for the law of the main contract, with fallback on the law of the seat.

This might mean that in practice less situations might trigger the applicable law as being the law of the seat, because in most cases there might not be sufficient elements pointing to another solution than that the parties implicitly opted for the same law to apply to all their clauses in the main contract, including to the arbitration agreement.

The issue remains on what the principles in assessing whether there is an implied choice or not are, as the more expansive the interpretation, the less cases falling on the law of the seat would be.

It is worth mentioning in this context the evolution we have seen in UK courts on the issue of whether an implied choice of the parties should be solved in favor of the seat or of the main contract:

• implied choice for the main contract (Sonatrach [2001] EWHC 481, Svenska EWHC [2005] 2437),
• implied choice for the seat (C v D, [2007] EWCA Civ 1282, Abuja [2012] EWHC 87),
• implied choice for the main contract (SULAMÉRICA [2012] EWCA Civ 638),
• implied choice for the seat (Enka [2020] EWCA Civ 574),
• implied choice for the main contract (Enka [2020] UKSC 38),
• no implied choice, with fallback on the seat English Arbitration Act 2025.

Areas of interpretation also remain in the absence of parties’ agreement for the law of the main contract and/or for the seat.

Takeaways

There are pros and cons for either approach, but the decisive factor should be creating uniformity and certainty. Parties choosing arbitration should be able to rely on a simple system without the need to litigate over the law applicable to their arbitration agreements.

It might be for this reason that the UK legislator settled the matter in favor of the seat, despite its Supreme Court favoring the law of the main contract in Enka and subsequently in Unicredit. Moreover, given the extended scope of the express agreement in Kabab-Ji, the English Arbitration Act 2025 clarifies that such clause does not represent express consent: “[..] agreement between the parties that a particular law applies to an agreement of which the arbitration agreement forms a part does not constitute express agreement that that law also applies to the arbitration agreement.” Therefore, if there is no express agreement recorded by the parties for the law governing the arbitration agreement, the fallback is the law at the seat.

We also see a trend that more institutions recommend parties to expressly agree on the applicable law to their arbitration agreements. Notable examples in this respect could be found in the SIAC UNCITRAL Rules Model Clause, revised as of 9 December 2024, or the Arbitration under the HKIAC Administered Arbitration Rules.

In case parties do not agree, legislators have started expressly providing for the fallback if parties do not (expressly) agree: the law of the seat, as legislated in the English Arbitration Act 2025, or the law of the main contract, as proposed as an amendment to the Singapore Arbitration Act.

Furthermore, arbitration rules provide for the fallback if parties do not agree. For example the LCIA Arbitration Rules 2020 provide that “[..]  the law applicable to the Arbitration Agreement and the arbitration shall be the law applicable at the seat of the arbitration, unless and to the extent that the parties have agreed in writing on the application of other laws or rules of law and such agreement is not prohibited by the law applicable at the arbitral seat.”

It remains to be seen if all these changes will bring more clarity and consistency, and most importantly, predictability to users.