New York is the hub of international commerce in the United States. It is therefore also a key jurisdiction for lawyers drafting and invoking international arbitration agreements. Recent developments in the United States and overseas have complicated this area of practice by adopting an undesirable default rule for the law that governs the validity or effectiveness of an international arbitration agreement. Fortunately, many of the problems can be avoided by making a modest change to the language of standard arbitration clauses.
In the past, lawyers drafting arbitration clauses have not typically focused on this issue. Although most lawyers appreciate the need for a general choice of law clause in a contract, few pause to consider the law that governs the arbitration clause specifically. Until recently, moreover, the need to address that narrower issue was far less acute. Recent developments, however, significantly alter the consequences of overlooking this issue and warrant much closer attention when drafting an arbitration clause.
Arbitration Clauses and Choice of Law Provisions
Parties to international commercial contracts often insist on an arbitration clause to ensure a neutral forum for dispute resolution. Parties may legitimately fear that, if a dispute arises, the other party will seek “hometown justice” by trying to litigate in its own country’s home courts. Those concerns are particularly acute for investor-state agreements, where the counterparty is often the foreign government itself.
International commercial contracts also typically include choice of law clauses that specify what law governs the contract. Sometimes that clause may refer to the same jurisdiction as the arbitral seat – i.e., the place of arbitration specified in the arbitration clause. Often, however, it does not. For example, where a foreign investor agrees with a developing country’s government to build infrastructure in that country, the choice of law clause often points to the host country’s law. It is sensible and convenient for the parties’ substantive obligations to be interpreted under the law of the place where the contract will be performed, even if the parties agree to arbitrate disputes in a neutral third country.
Where the arbitral seat and the choice of law clause point in different directions, serious complications can arise if a party later challenges the validity of the arbitration clause. No one doubts that the choice of law clause governs the parties’ substantive obligations under the contract. It is also generally accepted that the law of the seat governs the procedures for any arbitration. But it is far less clear what law governs the construction and validity of the arbitration clause. Where the contractual relationship breaks down, one party may seek to avoid arbitration or challenge a resulting award by contesting the validity or existence of the arbitration agreement. What law applies to those challenges is a critical issue.
According to one school of thought, the contract’s general choice law of law clause should apply to the arbitration clause, just as it applies to any other provision in the contract. Another commonly held view is that the law of the arbitral seat should govern. Which law applies can have major implications for whether the arbitration proceeds and whether any resulting award is enforceable. Courts and commentators have debated that question for decades. Recent developments in the United States and England, however, mark a significant change in the legal landscape.
The New York Convention
Where a party seeks to enforce an arbitral award in a country different from the one where it was rendered, the New York Convention ordinarily governs whether courts of that country must recognize and enforce the award. That Convention has its own built-in choice of law clause for disputes over the validity of the arbitration agreement that formed the basis for the award. Article V.1(a) permits denial of recognition where the arbitration agreement “is not valid under the law to which the parties have subjected it or, failing any indication thereon, under the law of the country where the award was made.”
Unfortunately, that provision is ambiguous regarding the relevance of a general choice of law clause in the broader contract. Is a general choice of law clause sufficient to “subject” the arbitration clause to that law? Or does the Convention require a choice of law clause specific to the arbitration provision? Might the selection of the arbitral seat itself be an implied choice of law for the arbitration clause, even in the face of a general choice of law clause that points somewhere else? The Convention does not say.
Commentators have taken diverging views. Albert Jan van den Berg, a leading authority on the New York Convention, opined that the “better view” is that “the choice of law clause for the contract in general is not sufficient as choice of law for the arbitral clause.” He relied on the fundamentally distinct purpose of an arbitration clause. “The main contract and the arbitral clause have different objects: the main contract concerns the relationship between the parties as to the substance; the arbitral clause is concerned with the procedure for settling disputes arising out of the main contract.” In his view, the selection of the arbitral seat also amounted to an implied choice of law: “[I]f a contract contains a general choice of law clause and provides in the arbitral clause that arbitration is to be held in a country with a different law, the latter indication must be deemed to prevail over the former.”
Other authorities weighed in on both sides. Gary Born agreed that a “general choice-of-law clause in [the] underlying contract” should not determine the law for the arbitration clause, and that the “default rule . . . is the law of the arbitral seat, not the law governing the underlying contract.” He relied on the “presumptively separable” status of arbitration clauses: Arbitration clauses are severable and thus enforceable even if a party contests the validity of the broader contract. By contrast, Redfern & Hunter, a leading English treatise, took the opposite view: “Since the arbitration clause is only one of many clauses in a contract, it would seem reasonable to assume that the law chosen by the parties to govern the contract will also govern the arbitration clause.”
United States Developments
Recent case law in the United States illustrates how the law governing an arbitration clause can be a significant focus of litigation.
In Balkan Energy Ltd. v. Republic of Ghana, the Government of Ghana sought to address a severe power shortage by contracting with the local subsidiary of an English company, Balkan Energy, to refurbish an unused power barge. The power purchase agreement contained a general choice of law clause stating that the contract was governed by Ghanaian law, but it also included an arbitration clause requiring settlement of any disputes by the Permanent Court of Arbitration in The Hague, Netherlands. After the parties’ relationship broke down, Balkan Energy brought an arbitration. The Government responded by obtaining an injunction from its local Ghanaian courts purporting to restrain the arbitration. The arbitration clause, the Government claimed, was invalid because the power purchase agreement was not ratified by parliament. The arbitral tribunal proceeded with the arbitration anyway and rendered an award for Balkan Energy.
Balkan Energy then sought to enforce its award in the U.S. District Court for the District of Columbia. The Government of Ghana opposed recognition on the ground that the lack of parliamentary ratification rendered the arbitration agreement invalid under Ghanaian law. The district court disagreed. Citing van den Berg, the court held that the arbitration clause’s validity should be determined by the law of the arbitral seat. Because Ghana did not claim that the arbitration clause was invalid under Dutch law, the court enforced the award.
Meanwhile, in OJSC Ukrnafta v. Carpatsky Petroleum Corp., Carpatsky, a Texas oil and gas company, entered into a joint venture to develop gas reserves in Ukraine with Ukrnafta, an oil and gas company indirectly majority owned by the Ukrainian government. The joint activity agreement included a Ukrainian choice of law clause but also an arbitration clause providing for arbitration at the Stockholm Chamber of Commerce, a popular venue for contracts with parties in former Soviet republics. After a dispute arose over Ukrnafta’s refusal to allow Carpatsky to participate fully in the venture, Carpatsky brought an arbitration. Ukrnafta responded by obtaining rulings from its local Ukrainian courts that the agreement was ineffective under Ukrainian law because Carpatsky had changed from a Texas corporation to a Delaware corporation before signing the contract. The Stockholm tribunal went ahead regardless and awarded Carpatsky around $150 million.
After Ukrnafta preemptively sued Carpatsky in Texas, Carpatsky moved to confirm the award. The district court entered judgment for Carpatsky. On appeal, a key issue was what law governed Ukrnafta’s challenge to the arbitration clause. Carpatsky argued that Swedish law applied because the agreement designated Stockholm as the seat of arbitration. The Swedish courts, it noted, had upheld the agreement under Swedish law. Ukrnafta responded that Ukrainian law governed due to the contract’s choice of law clause, and Ukrainian courts had found that no agreement was formed. The Fifth Circuit took a third approach: It held that the dispute was fundamentally a question of Carpatsky’s capacity to enter into the arbitration agreement, and as such, was governed by the law of Delaware, Carpatsky’s domicile. Under Delaware law, Carpatsky’s change of domicile had no impact on the arbitration clause.
While those cases were unfolding, the American Law Institute was deliberating over the same question in the Restatement of the U.S. Law of International Commercial and Investor-State Arbitration, a project spearheaded by Reporter George Bermann of Columbia Law School. The Restatement adopts a position favoring the general choice of law clause over the arbitral seat: “In determining the validity of [an] arbitration agreement . . . a court applies the law to which the parties have subjected the arbitration agreement or, if no such law has been selected, . . . the law identified in the general choice-of-law clause in the underlying contract or, in the absence of such a clause, . . . the law of the seat of arbitration.”
The Reporter’s note acknowledges that “[c]ommentators are divided on this approach.” It opines that the Restatement’s approach “facilitates a consistent interpretive framework for the entire contract, and best reflects the parties’ intent.” The American Law Institute approved the proposed final draft on May 20, 2019, and expects to publish the Restatement in late 2021.
United Kingdom Developments
Similar disputes have unfolded in other jurisdictions. England, in particular, has seen major legal developments in recent years.
In two cases, XL Insurance Ltd. v. Owens Corning and C v. D, English courts considered disputes involving so-called “Bermuda Form” insurance contracts. Dissatisfied with how American judges and juries were deciding insurance disputes, certain foreign insurers began insisting on London arbitration clauses when writing policies for American companies, even while retaining New York choice of law clauses. When disputes arose over the validity or effect of those arbitration clauses, courts had to decide whether New York or English law governed. In both XL Insurance and C v. D, the courts chose English law on the basis of the arbitral seat.
A third case, Sulamérica Cia Nacional de Seguros S.A. v. Enesa Engenharia S.A., took a different approach. That case was an insurance dispute arising out of a hydroelectric project in Brazil. The insurance contracts contained Brazilian choice of law clauses but also London arbitration clauses. The Brazilian insured parties tried to avoid arbitration on the ground that, under Brazilian law, the insurers could not invoke the arbitration clauses without their consent. The court opined that “an express choice of law governing the substantive contract is a strong indication of the parties’ intention in relation to the agreement to arbitrate” that should ordinarily govern absent factors pointing the other way. In that particular case, however, the court did find countervailing factors – Brazilian law would have rendered the arbitration clauses totally ineffectual. The court thus applied English law despite its general rule.
That disarray led the U.K. Supreme Court to take up the issue in Enka Insaat Ve Sanayi AS v. OOO Insurance Co. Chubb, a case decided in October 2020. The central issue in Enka was “which system of national law governs the validity and scope of the arbitration agreement when the law applicable to the contract containing it differs from the law of the seat of the arbitration.” A Russian insurer had sued a Turkish construction company in Russia after a fire caused severe damage at a Russian power plant. The construction company sought an anti-suit injunction in England on the basis of a London arbitration clause in the construction contract. The contract did not contain a general choice of law clause, but the parties agreed that Russian law applied to the broader contract based on the close connection to that country.
The court began by addressing what rule should apply when the parties do include a general choice of law clause in the contract. The court adopted a rule favoring the general choice of law clause over the arbitral seat: “[I]t is natural to interpret . . . a governing law clause, in the absence of good reason to the contrary, as applying to the arbitration clause for the simple reason that the arbitration clause is part of the contract which the parties have agreed is to be governed by the specified system of law.” The court perceived several advantages to that approach, including the complexity of applying different laws to different parts of the same contract. Although an arbitration clause is severable, the court reasoned, it is still part of the broader contract. The general choice of law clause should therefore presumptively apply.
The court then turned to the fact pattern before it in which the contract did not contain a general choice of law clause. In that circumstance, the court held, the law of the seat – in that case, England – should govern.
Where Things Stand
The foregoing developments represent a significant shift in momentum. Previously, van den Berg and Born probably could have claimed to represent the dominant view on whether the law governing an arbitration clause is the law of the arbitral seat or the law of the general choice of law clause. The Restatement and Enka cast serious doubt on that status.
Regardless of which side has the better legal argument, the Restatement/Enka position is undesirable from the perspective of arbitration policy. The law governing the arbitration agreement matters principally because it is the law that applies when one party wants to avoid arbitration by challenging the validity or effectiveness of the arbitration clause. The same considerations that justify a neutral arbitral forum to resolve disputes also favor a neutral legal regime to determine whether the arbitration will proceed.
Parties that agree to arbitrate disputes in a neutral third country typically do so because one or both of the parties does not trust the home courts of the other party to render impartial justice. Rules that invite intervention by the parties’ home courts impede that goal. The Restatement/Enka rule has precisely that effect. General choice of law clauses often point to one of the parties’ home states, particularly in foreign investment contracts. In those circumstances, a rule that determines the validity or effectiveness of the arbitration clause by looking to the general choice of law clause is a recipe for mischief. It invites one of the parties to resort to its home courts to challenge the arbitration agreement or the resulting award, and it invites those courts to intervene by attempting to enjoin the arbitration or by refusing to recognize the award.
National court rulings are not necessarily binding on an international arbitral tribunal even where the court purports to apply its own country’s laws. Nor are those rulings necessarily binding in proceedings to enforce an award in other jurisdictions. At a minimum, however, home court rulings will have a much stronger claim to deference when they purport to interpret and apply their own country’s laws. If an arbitration clause is governed by Egyptian law, and Egyptian courts hold that the arbitration agreement is invalid under Egyptian law, a court in another jurisdiction deciding whether to recognize the award is more likely to be receptive to a challenge to the agreement’s validity. At the very least, the Egyptian court rulings will seriously complicate and protract any enforcement proceedings.
That sort of regime undermines the goals of international arbitration. Where parties provide for arbitration in a neutral third country, they do so because they want disputes adjudicated impartially. That includes disputes over whether the arbitration should proceed at all. A general choice of law clause in the broader contract does not detract from that objective in the least. In investor-state contracts, for example, parties often include general choice of law clauses pointing to the law of the host state because the investment contract will be performed in that country, and it is convenient for that country’s laws to govern the parties’ substantive obligations. That does not mean the host country’s laws should also determine whether that state can avoid an agreement to resolve disputes in a neutral forum.
A default rule that favors the law of the arbitral seat helps avoid home court resistance or intervention. If an American company and an Italian company agree to arbitrate in London, and the London tribunal holds that the arbitration agreement is valid and binding under English law, courts in other countries are unlikely to think twice about enforcing an award merely because Italian or American courts assert that the English tribunal misinterpreted English law. The law of the seat rule helps avoid unwarranted home court intervention by tying the governing law to the neutral forum specified in the contract. The Restatement/Enka rule has the opposite effect.
The Drafting Solution
Fortunately, all is not lost. The rule that the Restatement and Enka adopt is only a default rule that applies where the parties have not specifically selected the law that governs the arbitration clause. Where a contract includes a choice of law clause that specifically applies to the arbitration clause, there is no need to debate whether the contract’s general choice of law clause or the arbitral seat is the better indication of the parties’ intent. The parties have made their intent clear by explicitly selecting the law for the arbitration clause. The Restatement and Enka both acknowledge that parties may contract around their default rule in that manner.
To be sure, it is not common practice to include a specific choice of law clause that applies only to the arbitration clause. As Professor Born observes: “[P]arties in practice virtually never expressly choose the law governing their arbitration agreement. Instead, commercial parties ordinarily include only a general choice-of-law clause in their underlying contract.” It is hard to blame them. Although most contracting parties appreciate the benefits of a general choice of law clause, few are prescient enough to consider whether to exclude particular contractual provisions, like an arbitration clause, from that general choice of law. Moreover, so long as the arbitral seat supplied the governing law for the arbitration clause, there was no obvious reason to specify an exception to the general choice of law for the arbitration clause – that law would not apply to the arbitration clause anyway. The Restatement and Enka now put the consequences of failing to address this issue explicitly into much starker relief.
The Restatement and Enka provide good reason to reevaluate the traditional drafting practice. Where parties have gone to the trouble of providing for arbitration in a neutral third country, it is highly undesirable for the enforceability of that arbitration clause to depend on one of the contracting parties’ own legal regimes as interpreted by that party’s own home courts, rather than on the legal regime of the neutral forum selected as the arbitral seat. Parties should therefore contract around the Restatement/Enka default rule by explicitly selecting the law of the arbitral seat for the arbitration clause.
This proposal is not hard to implement. It amounts simply to this: In any international business agreement that includes both a general choice of law clause pointing to Country A and an arbitration clause providing for arbitration in Country B, the parties should include within the arbitration clause a provision – which need only be a single sentence long – stating that all questions relating to the arbitration clause, including any disputes over the validity, existence, or effectiveness of that clause, are governed by the laws of Country B. After the Restatement and Enka, adding this one sentence could greatly improve the enforceability of the arbitration clause. At a minimum, this sentence could help avoid time-consuming and costly litigation if a party later attempts to contest the arbitration clause.
Neither the American Law Institute nor the U.K. Supreme Court issues decisions that are binding in New York or any other U.S. jurisdiction. Nonetheless, the Restatement and Enka will surely be cited as persuasive authority. Moreover, arbitration enforcement disputes often wind up in multiple countries, so it is important to take a global view. Careful lawyers – in New York and elsewhere – would do well to pay attention to these developments when drafting or litigating international arbitration agreements. In this instance, the law of international arbitration has taken a small step backward. But the problem is easy to avoid with only a minor change in how arbitration agreements are typically drafted.
Robert K. Kry is a partner at MoloLamken. He represents both claimants and respondents in arbitration-related disputes in New York, Washington, D.C., and other federal courts around the country. He represented the claimants in two of the cases discussed in this article, OJSC Ukrnafta v. Carpatsky Petroleum Corp., 957 F.3d 487 (5th Cir. 2020), and Balkan Energy Ltd. v. Republic of Ghana, 302 F. Supp. 3d 144 (D.D.C. 2018).
. See, e.g., Karaha Bodas Co. v. Perusahaan Pertambangan Minyak Dan Gas Bumi Negara, 364 F.3d 274, 291 (5th Cir. 2004).
. Convention on the Recognition and Enforcement of Foreign Arbitral Awards, June 10, 1958, 21 U.S.T. 2517.
. Id. art. V.1(a), 21 U.S.T. at 2520.
. Albert Jan van den Berg, The New York Arbitration Convention of 1958: Towards a Uniform Judicial Interpretation 293 (1981).
. 3 Gary B. Born, International Commercial Arbitration § 26.05[C][e][i]–, at 3463–69 (2d ed. 2014).
. 3 id. § 26.05[C][e][i], at 3464–65; see also Buckeye Check Cashing, Inc. v. Cardegna, 546 U.S. 440, 445 (2006).
. Alan Redfern & Martin Hunter, Law and Practice of International Commercial Arbitration § 2-86, at 125 (4th ed. 2004).
. 302 F. Supp. 3d 144 (D.D.C. 2018), appeal dismissed by stipulation, No. 18-7061, 2018 WL 5115572 (D.C. Cir. Oct. 12, 2018).
. Id. at 147.
. Id. at 147–49.
. Id. at 152–53, 157, 159. The court also held in the alternative that the award should be enforced even if Ghanaian law applied. Id. at 153–54.
. 957 F.3d 487 (5th Cir. 2020).
. Id. at 493–94.
. Id. at 494–95.
. Id. at 497–98.
. Restatement of the U.S. Law of International Commercial and Investor-State Arbitration (proposed final draft Apr. 24, 2019).
. Id. § 4.10(c); see also id. cmt. c.
. Id. reporter’s note c.
.  1 All E.R. (Comm.) 530.
.  EWCA Civ. 1282.
. See XL Insurance,  1 All E.R. (Comm.) at 540–43; C v. D,  EWCA Civ. 1282, ¶¶ 21–29.
.  EWCA Civ. 638.
. Id. ¶ 26.
. Id. ¶¶ 30, 32.
.  UKSC 38.
. Id. ¶ 2.
. Id. ¶¶ 10, 13–19, 147–148.
. Id. ¶ 43.
. Id. ¶¶ 53, 61, 170(iv)–(v).
. Id. ¶¶ 120, 145, 170(viii), 171.
. See, e.g., Carpatsky Petroleum Corp. v. OJSC Ukrnafta, No. SCC V 124/2007, Final Award ¶ 208 (Arb. Inst. of Stockholm Chamber of Commerce Sept. 24, 2010) (arbitral tribunal “does not consider itself bound by the decisions of a national court”).
. See, e.g., Karaha Bodas Co. v. Perusahaan Pertambangan Minyak Dan Gas Bumi Negara, 500 F.3d 111, 124 (2d Cir. 2007); cf. Animal Sci. Prods., Inc. v. Hebei Welcome Pharm. Co., 138 S. Ct. 1865, 1873 (2018).
. Enka mitigates that result to some extent by suggesting that a “validation principle” could weigh against applying the general choice of law clause where “there is ‘at least a serious risk’ that a choice of that law would ‘significantly undermine’ [the arbitration] agreement.”  UKSC 38, ¶¶ 95–109. Nonetheless, a rule favoring the law of the seat has advantages over that validation approach. The law of the seat rule ensures that a neutral legal regime applies to questions relating to the arbitration clause even when they fall short of rendering that clause completely ineffective. In addition, some courts may hesitate to adopt what they perceive to be a result-oriented pro-claimant rule. The law of the seat rule is a neutral choice of law rule that helps ensure that disputes over the arbitration clause are resolved impartially. It does not attempt to stack the deck in favor of a preferred result.
. See Restatement of the U.S. Law of International Commercial and Investor-State Arbitration, supra note 19, § 4.10(c); Enka,  UKSC 38, ¶ 170(iv).
. 3 Born, supra note 7, § 26.05[C][e][i], at 3464.