The influence of the principle of ‘severability’ of arbitration agreements, some may think, has extended far beyond the core problem it was intended to deal with: allowing an arbitral Tribunal to determine the existence or validity of the contract that contains the arbitration clause. Nor is it necessarily confined to arbitration agreements: the underlying principle is that a defendant who denies that a contract ever came into existence, or that it is valid, must first run that argument before the forum chosen (arbitral Tribunal or court) in the putatively valid contract. This principle is illustrated in the classic speech of Lord Hoffmann in Fiona Trust v Privalov. But it is not—as Lord Hoffmann himself recognised—without exceptions and it has also given rise to difficult questions about the effect it has on other well-established principles of arbitration law or the conflict of laws. Hamblen J considered two such issues in his recent judgment in Habas v VSC. The first, the law governing the arbitration agreement, is a well-known controversy in Indian law as well. It is a question of great practical importance in India, with the rapid growth of neutral arbitration clauses especially where the law governing the main contract is not the law of the country of the seat (for example an Indian and a Japanese company enter into a contract governed by Indian law but with a London arbitration clause). The second issue in Habas is a novel and interesting point that highlights the importance of a close analysis of the Fiona Trust principle—suppose a defendant enters into a contract through an agent who, it is said, lacked actual authority to agree to a particular seat of arbitration (but not to arbitration itself as a mode of dispute resolution), can the seat be treated as a relevant factor in the choice of law analysis?
The first question has produced somewhat contradictory answers in the English courts in recent years. The starting point is uncontroversial enough: as with any contract, the applicable law is the law expressly or impliedly chosen by the parties or, in the absence of choice, the system of law with which the transaction has its ‘closest and most real connection’. However, the principle of severability has led some to suggest (notably in C v D) that the arbitration agreement may be more closely connected to the seat of arbitration even if the matrix contract is not, on the footing that it is a (separate) agreement to resolve disputes. As Shantanu has pointed out, the most recent Court of Appeal authority on this point—Sulamerica v Enesa—is not easy to interpret. Indeed, signalling that the courts may be retreating from the obiter views expressed by Longmore LJ in C v D, Andrew Smith J recently held, after considering Sulamerica, that the choice of a law to govern the matrix contract is an ‘implied’ (and perhaps even ‘express’) choice of the law governing the arbitration agreement. However, the seat retains its importance in cases where the parties do not choose a law to govern the matrix agreement: it is likely in such cases that the law of the country of the seat will govern the matrix contract as well as the arbitration agreement, unless there are strong indications to the contrary. In Habas, Hamblen J accepted this (at ) but was faced with a more difficult question: what if the defendant alleges that the agent lacked actual authority to agree to that particular seat of arbitration? Can the seat then be used as an indication of closest connection?
This question arose in a dispute between a Turkish manufacturer of steel [“Habas”] and a Hong Kong engineering company [“VSC”] over whether a contract for the sale (by Habas) of steel had indeed been concluded. In the course of negotiations for this contract, Habas and VSC never communicated with each other directly: Habas acted through its agent, Steel Park, which communicated with VSC or another agent, Charter Alpha. In cases like this, it is more usual to find the defendant alleging that there was no contract because some essential term was not agreed or that its agent was bribed and so on and therefore that the chosen forum lacks jurisdiction. In this case, however, the disagreement between the parties during negotiations was not about the commercial terms but about dispute resolution: VSC had initially proposed Hong Kong arbitration and Hong Kong governing law while Habas had proposed Turkish arbitration and Turkish governing law. The last draft of the contract that Habas directly signed contained an arbitration clause with a Turkish seat. However, VSC later proposed to amend the seat to London and Steel Park accepted that amendment. A London arbitral Tribunal found that it had jurisdiction and awarded VSC damages for Habas’ failure to supply the agreed steel. Habas challenged the jurisdiction of the Tribunal on the ground that, under Turkish law, Steel Park would not be treated as having had actual or ostensible authority to bind it to London arbitration. This was notwithstanding the fact that Habas, in October 2009, had issued an ‘Agency Letter’ to Steel Park which had been forwarded to Charter Alpha and VSC—this letter would, under English (but not Turkish) law, have given Steel Park ostensible authority to conclude an agreement to sell steel with a London arbitration clause. So to resolve the jurisdiction question, it was necessary to ask the first question set out above: what law governs the arbitration agreement? Since the seat of arbitration under the putatively valid contract was London, and since the contract contained no choice of law clause for the matrix agreement, Hamblen J held that English law prima facie governed the arbitration agreement.
It is important to notice that this case differs from many cases on this subject in this way: here the defendant’s attack on the seat of arbitration was not parasitic on an attack on the main agreement (bribery, failure to agree essential terms, contract invalid under proper law etc): the attack was specifically on the seat of arbitration since Steel Park was said to have lacked actual or ostensible authority to agree to London arbitration (although not to arbitration itself). So Habas argued that the fact that London was the seat of arbitration had to be disregarded in the choice of law analysis in which event Turkish law would become the proper law, under which Steel Park would have lacked ostensible authority. Virtually the only authority for this proposition is the following passage from Dicey, Morris and Collins in relation to a choice of law clause:
… it may be thought unlikely that P [ie principal] could be bound and entitled by virtue of a law which governed the contract with the third party only because A [ie agent], in excess of his actual authority, agreed to its selection as the applicable law. The problem is similar to that raised by the question of capacity and can be resolved in a similar way. Where the agent exceeds his authority in choosing the law to govern his contract with the third party, P should only be regarded as entitled or bound if he would be so under the law applicable in the absence of choice.
Hamblen J declined to accept the argument made in this passage for a number of reasons, the most important of which (apart from prior Court of Appeal authority) were that there is no “logical or principled link between the issue of authority and the issue of the law with which the contract has its closest connection”  and that “it involves English law according special treatment to actual authority for conflicts purposes”. In other words, since the (putatively valid) contract contained a London seat, the arbitration agreement was governed by English law; since it was governed by English law, Steel Park did not lack ostensible authority to agree to a London seat; therefore the Tribunal did not lack jurisdiction. It is respectfully submitted that this conclusion (although perhaps inevitable at first instance given prior authority) may need to be reconsidered in appeal because it does not appear to sufficiently distinguish between a direct attack on the seat or the choice of law and an attack on those matters through an attack on the main agreement. If Habas had argued that the entire contract was void and therefore there was no arbitration clause or London seat, Hamblen J’s decision would have been perfectly right; but it is not inevitable that the same answer must be given when the argument is that the arbitration clause is void or (as in this case) the chosen seat could not have been chosen by the agent.
The parallel to Fiona Trust is clear, with the important difference that Habas’ argument did not deny authority to conclude an arbitration agreement (which Lord Hoffmann accepts is an exception to his general rule, at ) but rather said that although the agent could have agreed to arbitration, it could not have agreed to a London seat. Hamblen J is correct that there is no ‘logical link’ between this lack of authority and the applicable law, since the seat is only taken as an indication of with which system the contract is most closely connected, but it is not clear why there should be, as long as the choice of this seat is specifically attacked (whatever the reason for the lack of authority). Of course, Hamblen J rightly points out that Habas’ argument went further than this: it was not confined (as Dicey, Morris and Collins is) to the choice of law clause itself but extends to any clause that is relevant in the choice of law analysis. One may indeed ask: if the DMC argument is extended (as Habas argued it should be) to an attack on the seat, why should it not be extended to an attack on any other clause that is relevant to the choice of law question, such as place of performance? Ultimately that may be the strongest argument in favour of Hamblen J’s view and it seems likely that this question will be considered by the appellate courts in the future.