The post below was written by George A Bermann, who is Walter Gellhorn Professor of Law and Jean Monnet Professor of European Union Law at Columbia Law School. It is the third 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, Manuel Penades, Sylvain Bollée, Matthias Lehmann and Giuditta Cordero-Moss.
Readers are encouraged to participate in the discussion by commenting on the posts.
The relationship between the law (if any) chosen in the arbitration clause and the law of the seat is unsettled in the US.
It was taken up in the Restatement of the US Law of International Commercial and Investor-State Arbitration. The gist of the Restatement is that, while the law of the seat governs the conduct of the arbitration, it does not govern the interpretation of the arbitration agreement. Interpretation of the arbitration agreement should of course be governed by the law, if any, chosen in the arbitration clause itself. (I note that the court in Enka v. Chubb cited the Restatement in support.)
There was debate over whether, in the absence of a choice of applicable law in the arbitration clause, the arbitration agreement should be governed by the law, if any, chosen in the main contract. The view that ultimately prevailed is that more respect on matters of choice of law should be given to any expression of preference as to choice of law in the contract (even if not in the arbitration clause) over the law of the seat. That was a choice of seat, not a choice of law (other than the law of arbitration of the seat).
Unfortunately, the Restatement drew no distinction between issues of the interpretation and the validity of the arbitration agreement. More on that below.
The Restatement did not go much further, but the thinking behind it can be amplified and extrapolated. I attempt to do so below. I hasten to add that what follows happens also to be what I think the law should be.
It is this framework that I would use in assessing the differences between US law and the law advanced by the Law Commission.
As a general matter, I believe that Report in some cases fails to make an important distinction and in other cases, acknowledges the distinction, but makes the wrong choice.
I set out below what I consider to be these important distinctions:
- Distinction between the purposes underlying a choice of law in the arbitration clause (absent which in the law of the main contract) and the purposes underlying a choice of law function of the arbitration law of the seat
When parties choose a seat, they are choosing a seat, full stop. We should not suppose they are choosing an applicable law of any kind other than the arbitration law of the seat (lex arbitri).
By contrast, when parties indicate an applicable law in their arbitration agreement they are making a choice of applicable law. But, absent an indication of an applicable law in the arbitration clause, where else did the parties express a choice of law preference? They expressed it in the choice of law clause in the main contract. There too they are making a choice of applicable law, and their choice of an applicable law should be respected.
- Distinction between the law of the arbitration agreement and the law of the main contract
This result should be unaffected by the principle of separability. The principle of separability exists for one reason: to ensure that the demise of the main contract (as invalid) does not entail the demise of its arbitration clause. That is why we have the separability principle. It should not be extended to functions (such as choice of law) for which it was not intended.
- Distinction between “arbitration law of the seat” and “law of the seat”
It is vital to distinguish between the arbitration law of a jurisdiction (lex arbitri) and the whole body of law at the seat, and we too often fail to do so by referring sloppily to “the law of the seat”. An arbitration statute should make clear what it is talking about when it refers to “the law of the seat.”
When the parties chose a seat they certainly chose the lex arbitri of the seat. But, notwithstanding, it seems to be assumed that when the law of the seat is referred to, it includes at least some parts of the law of the seat outside the lex arbitri. For example, if the formation or validity of an arbitration agreement is called into question, the law of the seat may include the law of contract of the seat. If contract law at the seat treats coerced contracts as invalid, then that would apply to a claim that the arbitration agreement was coerced.
- Distinction between the issues of interpretation and issues of validity
As I mentioned, the Restatement fails to distinguish between issues of interpretation and validity, but it should have.
The law chosen in an arbitration agreement most fundamentally determines the interpretation of that agreement (such as its scope). There is absolutely no reason why the law of the seat should have anything to say about the meaning and scope of the arbitration agreement. If there is no choice of law in the arbitration agreement, then interpretation of the arbitration agreement should be governed by the law chosen in the main contract (on the reasoning set out above).
The question of the validity of the arbitration agreement is slightly more subtle.
Suppose the arbitration agreement is invalid under the law, if any, designated in the arbitration agreement, failing which the law governing the main contract, then it is invalid. It should not matter that it happens to be valid under the law of the seat.
On the other hand, conversely, if the arbitration agreement is invalid under the law of the seat, it is invalid, even if it would be valid under the law, if any, designated in the arbitration agreement, failing which the law governing the main contract. Why? Because the seat has a legitimate interest in the validity of the arbitration agreement giving rise to an arbitration on its territory.
More generally, one should not assume that if the law of the arbitration agreement is not the law of the seat, the seat’s policies risk being impaired. But that is not the case. Under no circumstance can the law of the arbitration agreement or the law of the main contract override the mandatory norms of the arbitration law of the seat (or the public policy of the seat).
The approach set out here is of course contrary to the so-called “validation principle,” and deliberately so. The impetus is a belief that the law chosen by the parties (even that in the main contract) deserves a measure of respect, as does the law of the seat. More delineation should be given to the matter than is generally given. I believe it is sometimes assumed that, unless you give as much weight as you possibly can to the law of the seat, you are not pro-arbitration, which is not the case.
- Distinction between the mandatory and default law provisions of the lex arbitri
Focusing now on the lex arbitri, it contains both mandatory and default rules. Its mandatory law provisions (and principles of public policy at the seat more generally) must be respected. But its default rules can be contracted around by the parties.
How can parties contract around the arbitration law at the seat? Obviously parties can contract around default rules of the seat by a term of their arbitration agreement. But they should also be allowed to contract around the default rules of the seat via the law designated in the arbitration clause.
Whether they can contract around the default rules of the seat via the law governing the main contract will be more controversial, but, for the reasons set out above, they should be able to do so.
Thoughts on Specific Provisions of the Report and Recommendation
- 12.17: I do not share the view that subjecting an arbitration agreement to the law of the main contract is a threat to the UK as a seat. It is no more a threat than application of a law chosen in the arbitration clause itself; yet the Report allows the latter to apply in lieu of the law of the seat (sec. 12.17).
- 12.18, 12:47: The Report treats a choice of law clause in the main contract as only an “implied” choice of law for the arbitration agreement. Driving a wedge between the law designated in the arbitration agreement and the law designated in the main contract is an unwarranted extrapolation of the separability principle.
- 12.19: I see nothing wrong with the law designated in either the arbitration clause or the main contract with displacing the non-mandatory law of the seat.
- 12.22: the rule in Enka v. Chubb is not “too complex and unpredictable”.
- 12.25: as may be expected, I, like those commenters referred to here, do not believe the placement of the choice of law clause in a contract should be determinative.
- 12.35: This is just another assertion of separability where it doesn’t belong.
- 12.40: This view is correct. When parties choose a seat, they do not think they are choosing anything more than the seat. Maybe they should be bound by the lex arbitri, but why by the law of the seat writ large?
- 12.53: What is said here makes no sense. To have the law chosen in the main contract govern the arbitration agreement in no way compromises the parties’ decision to arbitrate. The parties will still arbitrate, won’t they? The arbitration clause is 100% intact. What the Report is in effect doing is to convert the notion of “the decision to arbitrate” into the notion of “the decision to arbitrate under the law of the seat.” In other words, the remark already assumes what the Report wants to establish, namely necessarily subject the arbitration agreement to the law of the seat. Moreover, if giving effect to a choice of law (other than the law of the seat) in the arbitration clause itself – which the Report clearly allows – does not undermine the decision to arbitrate, then giving effect instead to the applicable law clause in the main contract doesn’t undermine that decision either. Here, the Report is “question-begging.”
- 12.66: I do not understand the Report’s aversion to using the law designated in the main contract in the rare situation that no seat was yet chosen.
- 12.73: Here and elsewhere it is said that we can’t allow a choice of law in the main contract to override the parties’ intent to arbitrate. But it doesn’t override. We can easily give effect to the parties’ intent to arbitrate without subjecting the arbitration agreement in all respects to the law of the seat.
This post hits the nails on their head and says pretty much all that there is to say. When it comes to the elucidation of what the parties intended and agreed to settle by arbitration it is more or less ludicrous to suggest that the law of the place of arbitration should answer that question. Which other question about the meaning of the parties’ contractual agreement is to be answered by saying ‘let’s ask the law of the place where they will sort out their differences and see what it would have to say’ ?
It is probably true that if the law of the place of the seat says that a matter is not arbitrable, that has to be respected, but arbitrability is in this sense an issue which is separate and distinct. The question was addressed by the Singapore Court of Appeal in a jolly thoughtful judgment: Anupam Mittal v Westbridge Ventures [2023] SGCA 1. The existence of this judgment, which pre-dated the second consultation paper, was drawn to the attention of the Commission in a comment on the consultation paper, in the hope that it might find it helpful. The dismal reference to it in the Report as a footnote to the discussion of the validation principle (p 144 n 366) cannot be said to display much in the way of intellectual engagement.
As for Enka v Chubb, there would have been a fair point to be made about the uncertainty of the law before that decision and that in Kabab-ji. But we now have our answers, with only a small number of gaps to be filled. The idea that we pretend that the law is as uncertain as it might have appeared to be in 2013, and proceed from there, is not one based on coherent reflections.