(delivered the opinion of the Court)
Article 8 of an investment treaty between the United Kingdom and Argentina contains a dispute-resolution provision, applicable to disputes between one of those nations and an investor from the other. See Agreement for the Promotion and Protection of Investments, Art. 8(2), Dec. 11, 1990, 1765 U. N. T. S. 38 (hereinafter Treaty). The provision authorizes either party to submit a dispute “to the decision of the competent tribunal of the Contracting Party in whose territory the investment was made,” i.e., a local court. Art. 8(1). And it provides for arbitration [Art. 8(2)(a)]
The Treaty also entitles the parties to agree to proceed directly to arbitration. Art. 8(2)(b).
This case concerns the Treaty’s arbitration clause, and specifically the local court litigation requirement set forth in Article 8(2)(a). The question before us is whether a court of the United States, in reviewing an arbitration award made under the Treaty, should interpret and apply the local litigation requirement de novo, or with the deference that courts ordinarily owe arbitration decisions. That is to say, who – court or arbitrator – bears primary responsibility for interpreting and applying the local litigation requirement to an underlying controversy? In our view, the matter is for the arbitrators, and courts must review their determinations with deference.
In the early 1990’s, the petitioner, BG Group plc, a British firm, belonged to a consortium that bought a majority interest in an Argentine entity called MetroGAS. MetroGAS was a gas distribution company created by Argentine law in 1992, as a result of the government’s privatization of its state-owned gas utility. Argentina distributed the utility’s assets to new, private companies, one of which was MetroGAS. It awarded MetroGAS a 35-year exclusive license to distribute natural gas in Buenos Aires, and it submitted a controlling interest in the company to international public tender. BG Group’s consortium was the successful bidder.
At about the same time, Argentina enacted statutes providing that its regulators would calculate gas “tariffs” in U. S. dollars, and that those tariffs would be set at levels sufficient to assure gas distribution firms, such as MetroGAS, a reasonable return.
In 2001 and 2002, Argentina, faced with an economic crisis, enacted new laws. Those laws changed the basis for calculating gas tariffs from dollars to pesos, at a rate of one peso per dollar. The exchange rate at the time was roughly three pesos to the dollar. The result was that MetroGAS’ profits were quickly transformed into losses. BG Group believed that these changes (and several others) violated the Treaty; Argentina believed the contrary.
In 2003, BG Group, invoking Article 8 of the Treaty, sought arbitration. The parties appointed arbitrators; they agreed to site the arbitration in Washington, D. C.; and between 2004 and 2006, the arbitrators decided motions, received evidence, and conducted hearings. BG Group essentially claimed that Argentina’s new laws and regulatory practices violated provisions in the Treaty forbidding the “expropriation” of investments and requiring that each nation give “fair and equitable treatment” to investors from the other. Argentina denied these claims, while also arguing that the arbitration tribunal lacked “jurisdiction” to hear the dispute. App. to Pet. for Cert. 143a–144a, 214a–218a, 224a–232a. According to Argentina, the arbitrators lacked jurisdiction because:
BG Group was not a Treaty-protected “investor”;
BG Group’s interest in MetroGAS was not a Treaty-protected “investment”; and
BG Group initiated arbitration without first litigating its claims in Argentina’s courts, despite Article 8’s requirement. Id., at 143a–171a.
In Argentina’s view, “failure by BG to bring its grievance to Argentine courts for 18 months renders its claims in this arbitration inadmissible.” Id., at 162a.
In late December 2007, the arbitration panel reached a final decision. It began by determining that it had “jurisdiction” to consider the merits of the dispute. In support of that determination, the tribunal concluded that BG Group was an “investor,” that its interest in MetroGAS amounted to a Treaty-protected “investment,” and that Argentina’s own conduct had waived, or excused, BG Group’s failure to comply with Article 8’s local litigation requirement. Id., at 99a, 145a, 161a, 171a. The panel pointed out that in 2002, the President of Argentina had issued a decree staying for 180 days the execution of its courts’ final judgments (and injunctions) in suits claiming harm as a result of the new economic measures. Id., at 166a–167a. In addition, Argentina had established a “renegotiation process” for public service contracts, such as its contract with MetroGAS, to alleviate the negative impact of the new economic measures. Id., at 129a, 131a. But Argentina had simultaneously barred from participation in that “process” firms that were litigating against Argentina in court or in arbitration. Id., at 168a–171a. These measures, while not making litigation in Argentina’s courts literally impossible, nonetheless “hindered” recourse “to the domestic judiciary” to the point where the Treaty implicitly excused compliance with the local litigation requirement. Id., at 165. Requiring a private party in such circumstances to seek relief in Argentina’s courts for 18 months, the panel concluded, would lead to “absurd and unreasonable result[s].” Id., at 166a.
On the merits, the arbitration panel agreed with Argentina that it had not “expropriate[d]” BG Group’s investment, but also found that Argentina had denied BG Group “fair and equitable treatment.” Id., at 222a–223a, 240a–242a. It awarded BG Group $185 million in damages. Id., at 297a.
In March 2008, both sides filed petitions for review in the District Court for the District of Columbia. BG Group sought to confirm the award under the New York Convention and the Federal Arbitration Act. See Convention on the Recognition and Enforcement of Foreign Arbitral Awards, Art. IV, June 10, 1958, 21 U. S. T. 2519, T. I. A. S. No. 6997 (New York Convention) (providing that a party may apply “for recognition and enforcement” of an arbitral award subject to the Convention); 9 U. S. C. §§204, 207 (providing that a party may move “for an order confirming [an arbitral] award” in a federal court of the “place designated in the agreement as the place of arbitration if such place is within the United States”). Argentina sought to vacate the award in part on the ground that the arbitrators lacked jurisdiction. See §10(a)(4) (a federal court may vacate an arbitral award “where the arbitrators exceeded their powers”).
The District Court denied Argentina’s claims and confirmed the award. 764 F. Supp. 2d 21 (DC 2011); 715 F. Supp. 2d 108 (DC 2010). But the Court of Appeals for the District of Columbia Circuit reversed. 665 F. 3d 1363 (2012). In the appeals court’s view, the interpretation and application of Article 8’s local litigation requirement was a matter for courts to decide de novo, i.e., without deference to the views of the arbitrators. The Court of Appeals then went on to hold that the circumstances did not excuse BG Group’s failure to comply with the requirement. Rather, BG Group must “commence a lawsuit in Argentina’s courts and wait eighteen months before filing for arbitration.” Id., at 1373. Because BG Group had not done so, the arbitrators lacked authority to decide the dispute. And the appeals court ordered the award vacated. Ibid.
BG Group filed a petition for certiorari. Given the importance of the matter for international commercial arbitration, we granted the petition. See, e.g., K. Vandevelde, Bilateral Investment Treaties: History, Policy& Interpretation 430–432 (2010) (explaining that dispute-resolution mechanisms allowing for arbitration are a “critical element” of modern day bilateral investment treaties); C. Dugan, D. Wallace, N. Rubins, & B. Sabahi, Investor-State Arbitration 51–52, 117–120 (2008) (referring to the large number of investment treaties that provide for arbitration, and explaining that some also impose pre-arbitration requirements such as waiting periods, amicable negotiations, or exhaustion of local remedies).
As we have said, the question before us is who – court or arbitrator – bears primary responsibility for interpreting and applying Article 8’s local court litigation provision. Put in terms of standards of judicial review, should a United States court review the arbitrators’ interpretation and application of the provision de novo, or with the deference that courts ordinarily show arbitral decisions on matters the parties have committed to arbitration? Compare, e.g., First Options of Chicago, Inc v Kaplan, 514 U. S. 938, 942 (1995) (example where a “court makes up its mind about [an issue] independently” because the parties did not agree it should be arbitrated), with Oxford Health Plans LLC v Sutter, 569 U. S. ___, ___ (2013) (slip op., at 4) (example where a court defers to arbitrators because the parties “ ‘bargained for’ ” arbitral resolution of the question (quoting Eastern Associated Coal Corp v Mine Workers, 531 U. S. 57, 62 (2000) )). See also Hall Street Associates, L. L. C v Mattel, Inc., 552 U. S. 576, 588 (2008) (on matters committed to arbitration, the Federal Arbitration Act provides for “just the limited review needed to maintain arbitration’s essential virtue of resolving disputes straightaway” and to prevent it from becoming “merely a prelude to a more cumbersome and time-consuming judicial review process” (internal quotation marks omitted)); Eastern Associated Coal Corp., supra, at 62 (where parties send a matter to arbitration, a court will set aside the “arbitrator’s interpretation of what their agreement means only in rare instances”).
In answering the question, we shall initially treat the document before us as if it were an ordinary contract between private parties. Were that so, we conclude, the matter would be for the arbitrators. We then ask whether the fact that the document in question is a treaty makes a critical difference. We conclude that it does not.
Where ordinary contracts are at issue, it is up to the parties to determine whether a particular matter is primarily for arbitrators or for courts to decide. See, e.g., Steelworkers v Warrior & Gulf Nav. Co., 363 U. S. 574, 582 (1960) (“[A]rbitration is a matter of contract and a party cannot be required to submit to arbitration any dispute which he has not agreed so to submit”). If the contract is silent on the matter of who primarily is to decide “threshold” questions about arbitration, courts determine the parties’ intent with the help of presumptions.
On the one hand, courts presume that the parties intend courts, not arbitrators, to decide what we have called disputes about “arbitrability.” These include questions such as “whether the parties are bound by a given arbitration clause,” or “whether an arbitration clause in a concededly binding contract applies to a particular type of controversy.” Howsam v Dean Witter Reynolds, Inc., 537 U. S. 79, 84 (2002) ; accord, Granite Rock Co. v Teamsters, 561 U. S. 287–300 (2010) (disputes over “formation of the parties’ arbitration agreement” and “its enforceability or applicability to the dispute” at issue are “matters .... the court must resolve” (internal quotation marks omitted)). See First Options, supra, at 941, 943–947 (court should decide whether an arbitration clause applied to a party who “had not personally signed” the document containing it); AT&T Technologies, Inc. v Communications Workers, 475 U. S. 643, 651 (1986) (court should decide whether a particular labor-management layoff dispute fell within the arbitration clause of a collective-bargaining contract); John Wiley & Sons, Inc v Livingston, 376 U. S. 543–548 (1964) (court should decide whether an arbitration provision survived a corporate merger). See generally AT&T Technologies, supra, at 649 (“Unless the parties clearly and unmistakably provide otherwise, the question of whether the parties agreed to arbitrate is to be decided by the court, not the arbitrator”).
On the other hand, courts presume that the parties intend arbitrators, not courts, to decide disputes about the meaning and application of particular procedural preconditions for the use of arbitration. See Howsam, supra, at 86 (courts assume parties “normally expect a forum-based decision maker to decide forum-specific procedural gateway matters” (emphasis added)). These procedural matters include claims of “waiver, delay, or a like defense to arbitrability.” Moses H. Cone Memorial Hospital v Mercury Constr. Corp., 460 U. S. 1, 25 (1983) . And they include the satisfaction of “‘prerequisites such as time limits, notice, laches, estoppel, and other conditions precedent to an obligation to arbitrate.’” Howsam, supra, at 85 (quoting the Revised Uniform Arbitration Act of 2000 §6, Comment 2, 7 U. L. A. 13 (Supp. 2002); emphasis deleted). See also §6(c) (“An arbitrator shall decide whether a condition precedent to arbitrability has been fulfilled”); §6, Comment 2 (explaining that this rule reflects “the holdings of the vast majority of state courts” and collecting cases).
The provision before us is of the latter, procedural, variety. The text and structure of the provision make clear that it operates as a procedural condition precedent to arbitration. It says that a dispute “shall be submitted to international arbitration” if “one of the Parties so requests,” as long as “a period of eighteen months has elapsed” since the dispute was “submitted” to a local tribunal and the tribunal “has not given its final decision.” Art. 8(2). It determines when the contractual duty to arbitrate arises, not whether there is a contractual duty to arbitrate at all. Cf. 13 R. Lord, Williston on Contracts §38:7, pp. 435, 437; §38:4, p. 422 (4th ed. 2013) (a “condition precedent” determines what must happen before “a contractual duty arises” but does not “make the validity of the contract depend on its happening” (emphasis added)). Neither does this language or other language in Article 8 give substantive weight to the local court’s determinations on the matters at issue between the parties. To the contrary, Article 8 provides that only the “arbitration decision shall be final and binding on both Parties.” Art. 8(4). The litigation provision is consequently a purely procedural requirement – a claims-processing rule that governs when the arbitration may begin, but not whether it may occur or what its substantive outcome will be on the issues in dispute.
Moreover, the local litigation requirement is highly analogous to procedural provisions that both this Court and others have found are for arbitrators, not courts, primarily to interpret and to apply. See Howsam, supra, at 85 (whether a party filed a notice of arbitration within the time limit provided by the rules of the chosen arbitral forum “is a matter presumptively for the arbitrator, not for the judge”); John Wiley, supra, at 555–557 (same, in respect to a mandatory pre-arbitration grievance procedure that involved holding two conferences). See also Dialysis Access Center, LLC v. RMS Lifeline, Inc., 638 F. 3d 367, 383 (CA1 2011) (same, in respect to a pre-arbitration “good faith negotiations” requirement); Lumbermens Mut. Cas. Co v Broadspire Management Servs., Inc., 623 F. 3d 476, 481 (CA7 2010) (same, in respect to a pre-arbitration filing of a “Disagreement Notice”).
Finally, as we later discuss in more detail, see infra, at 13–14, we can find nothing in Article 8 or elsewhere in the Treaty that might overcome the ordinary assumption. It nowhere demonstrates a contrary intent as to the delegation of decisional authority between judges and arbitrators. Thus, were the document an ordinary contract, it would call for arbitrators primarily to interpret and to apply the local litigation provision.
We now relax our ordinary contract assumption and ask whether the fact that the document before us is a treaty makes a critical difference to our analysis. The Solicitor General argues that it should. He says that the local litigation provision may be “a condition on the State’s consent to enter into an arbitration agreement.” Brief for United States as Amicus Curiae 25. He adds that courts should “review de novo the arbitral tribunal’s resolution of objections based on an investor’s non-compliance” with such a condition. Ibid. And he recommends that we remand this case to the Court of Appeals to determine whether the court-exhaustion provision is such a condition. Id., at 31–33.
We do not accept the Solicitor General’s view as applied to the treaty before us. As a general matter, a treaty is a contract, though between nations. Its interpretation normally is, like a contract’s interpretation, a matter of determining the parties’ intent. Air France v Saks, 470 U. S. 392, 399 (1985) (courts must give “the specific words of the treaty a meaning consistent with the shared expectations of the contracting parties”); Sullivan v Kidd, 254 U. S. 433, 439 (1921) (“[T]reaties are to be interpreted upon the principles which govern the interpretation of contracts in writing between individuals, and are to be executed in the utmost good faith, with a view to making effective the purposes of the high contracting parties”); Wright v Henkel, 190 U. S. 40, 57 (1903) (“Treaties must receive a fair interpretation, according to the intention of the contracting parties”). And where, as here, a federal court is asked to interpret that intent pursuant to a motion to vacate or confirm an award made in the United States under the Federal Arbitration Act, it should normally apply the presumptions supplied by American law. See New York Convention, Art. V(1)(e) (award may be “set aside or suspended by a competent authority of the country in which, or under the law of which, that award was made”); Vandevelde, Bilateral Investment Treaties, at 446 (arbitral awards pursuant to treaties are “subject to review under the arbitration law of the state where the arbitration takes place”); Dugan, Investor-State Arbitration, at 636 (“[T]he national courts and the law of the legal situs of arbitration control a losing party’s attempt to set aside [an] award”).
The Solicitor General does not deny that the presumption discussed in Part III, supra (namely, the presumption that parties intend procedural preconditions to arbitration to be resolved primarily by arbitrators), applies both to ordinary contracts and to similar provisions in treaties when those provisions are not also “conditions of consent.” Brief for United States as Amicus Curiae 25–27. And, while we respect the Government’s views about the proper interpretation of treaties, e.g., Abbott v Abbott, 560 U. S. 1, 15 (2010) , we have been unable to find any other authority or precedent suggesting that the use of the “consent” label in a treaty should make a critical difference in discerning the parties’ intent about whether courts or arbitrators should interpret and apply the relevant provision.
We are willing to assume with the Solicitor General that the appearance of this label in a treaty can show that the parties, or one of them, thought the designated matter quite important. But that is unlikely to be conclusive. For parties often submit important matters to arbitration. And the word “consent” could be attached to a highly procedural precondition to arbitration, such as a waiting period of several months, which the parties are unlikely to have intended that courts apply without saying so. See, e.g., Agreement on Encouragement and Reciprocal Protection of Investments, Art. 9, Netherlands-Slovenia, Sept. 24, 1996, Netherlands T. S. No. 296 (“Each Contracting Party hereby consents to submit any dispute .... which they can not [sic] solve amicably within three months .... to the International Center for Settlement of Disputes for settlement by conciliation or arbitration”), online at www.rijksoverheid.nl/documenten-en-publicaties/besluiten/2006/10/17/slovenia.html (all Internet materials as visited on Feb. 28, 2014, and available in Clerk of Court’s case file); Agreement for the Promotion and Protection of Investments, Art. 8(1), United Kingdom-Egypt, June 11, 1975, 14 I. L. M. 1472 (“Each Contracting Party hereby consents to submit” a dispute to arbitration if “agreement cannot be reached within three months between the parties”). While we leave the matter open for future argument, we do not now see why the presence of the term “consent” in a treaty warrants abandoning, or increasing the complexity of, our ordinary intent-determining framework. See Howsam, 537 U. S., at 83–85; First Options, 514 U. S., at 942–945; John Wiley, 376 U. S., at 546–549, 555–559.
In any event, the treaty before us does not state that the local litigation requirement is a “condition of consent” to arbitration. Thus, we need not, and do not, go beyond holding that, in the absence of explicit language in a treaty demonstrating that the parties intended a different delegation of authority, our ordinary interpretive framework applies. We leave for another day the question of interpreting treaties that refer to “conditions of consent” explicitly. See, e.g., United States-Korea Free Trade Agreement, Art. 11.18, Feb. 10, 2011 (provision entitled “Conditions and Limitations on Consent of Each Party” and providing that “[n]o claim may be submitted to arbitration under this Section” unless the claimant waives in writing “any right” to press his claim before an “administrative tribunal or court”), online at www.ustr.gov/trade-agreements/free-trade-agreements/korus-fta/final-text; North American Free Trade Agreement, Arts. 1121–1122, Dec. 17, 1992, 32 I. L. M. 643–644 (providing that each party’s “[c]onsent to [a]rbitration” is conditioned on fulfillment of certain “procedures,” one of which is a waiver by an investor of his right to litigate the claim being arbitrated). See also 2012 U. S. Model Bilateral Investment Treaty, Art. 26 (entitled “Conditions and limitations on Consent of Each Party”), online at www.ustr.gov / sites / default / files / BIT text for ACIEP Meeting.pdf. And we apply our ordinary presumption that the interpretation and application of procedural provisions such as the provision before us are primarily for the arbitrators.
A treaty may contain evidence that shows the parties had an intent contrary to our ordinary presumptions about who should decide threshold issues related to arbitration. But the treaty before us does not show any such contrary intention. We concede that the local litigation requirement appears in ¶(1) of Article 8, while the Article does not mention arbitration until the subsequent paragraph, ¶(2). Moreover, a requirement that a party exhaust its remedies in a country’s domestic courts before seeking to arbitrate may seem particularly important to a country offering protections to foreign investors. And the placing of an important matter prior to any mention of arbitration at least arguably suggests an intent by Argentina, the United Kingdom, or both, to have courts rather than arbitrators apply the litigation requirement.
These considerations, however, are outweighed by others. As discussed supra, at 8–9, the text and structure of the litigation requirement set forth in Article 8 make clear that it is a procedural condition precedent to arbitration – a sequential step that a party must follow before giving notice of arbitration. The Treaty nowhere says that the provision is to operate as a substantive condition on the formation of the arbitration contract, or that it is a matter of such elevated importance that it is to be decided by courts. International arbitrators are likely more familiar than are judges with the expectations of foreign investors and recipient nations regarding the operation of the provision. See Howsam, supra, at 85 (comparative institutional expertise a factor in determining parties’ likely intent). And the Treaty itself authorizes the use of international arbitration associations, the rules of which provide that arbitrators shall have the authority to interpret provisions of this kind. Art. 8(3) (providing that the parties may refer a dispute to the International Centre for the Settlement of Investment Disputes (ICSID) or to arbitrators appointed pursuant to the arbitration rules of the United Nations Commission on International Trade Law (UNCITRAL)); accord, UNCITRAL Arbitration Rules, Art. 23(1) (rev. 2010 ed.) (“[A]rbitral tribunal shall have the power to rule on its own jurisdiction”); ICSID Convention, Regulations and Rules, Art. 41(1) (2006 ed.) (“Tribunal shall be the judge of its own competence”). Cf. Howsam, supra, at 85 (giving weight to the parties’ incorporation of the National Association of Securities Dealers’ Code of Arbitration into their contract, which provided for similar arbitral authority, as evidence that they intended arbitrators to “interpret and apply the NASD time limit rule”).
The upshot is that our ordinary presumption applies and it is not overcome. The interpretation and application of the local litigation provision is primarily for the arbitrators. Reviewing courts cannot review their decision de novo. Rather, they must do so with considerable deference.
The dissent interprets Article 8’s local litigation provision differently. In its view, the provision sets forth not a condition precedent to arbitration in an already-binding arbitration contract (normally a matter for arbitrators to interpret), but a substantive condition on Argentina’s consent to arbitration and thus on the contract’s formation in the first place (normally something for courts to interpret). It reads the whole of Article 8 as a “unilateral standing offer” to arbitrate that Argentina and the United Kingdom each extends to investors of the other country. Post, at 9 (opinion of Roberts, C. J.). And it says that the local litigation requirement is one of the essential “‘terms in which the offer was made.’” Post, at 6 (quoting Eliason v. Henshaw, 4 Wheat. 225, 228 (1819); emphasis deleted).
While it is possible to read the provision in this way, doing so is not consistent with our case law interpreting similar provisions appearing in ordinary arbitration contracts. See Part III, supra. Consequently, interpreting the provision in such a manner would require us to treat treaties as warranting a different kind of analysis. And the dissent does so without supplying any different set of general principles that might guide that analysis. That is a matter of some concern in a world where foreign investment and related arbitration treaties increasingly matter.
Even were we to ignore our ordinary contract principles, however, we would not take the dissent’s view. As we have explained, the local litigation provision on its face concerns arbitration’s timing, not the Treaty’s effective date; or whom its arbitration clause binds; or whether that arbitration clause covers a certain kind of dispute. Cf. Granite Rock, 561 U. S., at 296–303 (ratification date); First Options, 514 U. S., at 941, 943–947 (parties); AT&T Technologies, 475 U. S., at 651 (kind of dispute). The dissent points out that Article 8(2)(a) “does not simply require the parties to wait for 18 months before proceeding to arbitration,” but instructs them to do something – to “submit their claims for adjudication.” Post, at 8. That is correct. But the something they must do has no direct impact on the resolution of their dispute, for as we previously pointed out, Article 8 provides that only the decision of the arbitrators (who need not give weight to the local court’s decision) will be “final and binding.” Art. 8(4). The provision, at base, is a claims-processing rule. And the dissent’s efforts to imbue it with greater significance fall short.
The treatises to which the dissent refers also fail to support its position. Post, at 3, 6. Those authorities primarily describe how an offer to arbitrate in an investment treaty can be accepted, such as through an investor’s filing of a notice of arbitration. See J. Salacuse, The Law of Investment Treaties 381 (2010); Schreuer, Consent to Arbitration, in The Oxford Handbook of International Investment Law 830, 836–837 (P. Muchlinski, F. Ortino, & C. Schreuer eds. 2008); Dugan, Investor-State Arbitration, at 221–222. They do not endorse the dissent’s reading of the local litigation provision or of provisions like it.
To the contrary, the bulk of international authority supports our view that the provision functions as a purely procedural precondition to arbitrate. See 1 G. Born, International Commercial Arbitration 842 (2009) (“A substantial body of arbitral authority from investor-state disputes concludes that compliance with procedural mechanisms in an arbitration agreement (or bilateral investment treaty) is not ordinarily a jurisdictional prerequisite”); Brief for Professors and Practitioners of Arbitration Law as Amici Curiae 12–16 (to assume the parties intended de novo review of the provision by a court “is likely to set United States courts on a collision course with the international regime embodied in thousands of [bilateral investment treaties]”). See also Schreuer, Consent to Arbitration, supra, at 846–848 (“clauses of this kind .... creat[e] a considerable burden to the party seeking arbitration with little chance of advancing the settlement of the dispute,” and “the most likely effect of a clause of this kind is delay and additional cost”).
In sum, we agree with the dissent that a sovereign’s consent to arbitration is important. We also agree that sovereigns can condition their consent to arbitrate by writing various terms into their bilateral investment treaties. Post, at 9–10. But that is not the issue. The question is whether the parties intended to give courts or arbitrators primary authority to interpret and apply a threshold provision in an arbitration contract – when the contract is silent as to the delegation of authority. We have already explained why we believe that where, as here, the provision resembles a claims-processing requirement and is not a requirement that affects the arbitration contract’s validity or scope, we presume that the parties (even if they are sovereigns) intended to give that authority to the arbitrators. See Parts III, IV–A and IV–B, supra.
Argentina correctly argues that it is nonetheless entitled to court review of the arbitrators’ decision to excuse BG Group’s noncompliance with the litigation requirement, and to take jurisdiction over the dispute. It asks us to provide that review, and it argues that even if the proper standard is “a [h]ighly [d]eferential” one, it should still prevail. Brief for Respondent 50. Having the relevant materials before us, we shall provide that review. But we cannot agree with Argentina that the arbitrators “‘exceeded their powers’” in concluding they had jurisdiction. Ibid. (quoting 9 U. S. C. §10(a)(4)).
The arbitration panel made three relevant determinations:
“As a matter of treaty interpretation,” the local litigation provision “cannot be construed as an absolute impediment to arbitration,” App. to Pet. for Cert. 165a;
Argentina enacted laws that “hindered” “recourse to the domestic judiciary” by those “whose rights were allegedly affected by the emergency measures,” id., at 165a–166a; that sought “to prevent any judicial interference with the emergency legislation,” id., at 169a; and that “excluded from the renegotiation process” for public service contracts “any licensee seeking judicial redress,” ibid.;
under these circumstances, it would be “absurd and unreasonable” to read Article 8 as requiring an investor to bring its grievance to a domestic court before arbitrating. Id., at 166a.
The first determination lies well within the arbitrators’ interpretive authority. Construing the local litigation provision as an “absolute” requirement would mean Argentina could avoid arbitration by, say, passing a law that closed down its court system indefinitely or that prohibited investors from using its courts. Such an interpretation runs contrary to a basic objective of the investment treaty. Nor does Argentina argue for an absolute interpretation.
As to the second determination, Argentina does not argue that the facts set forth by the arbitrators are incorrect. Thus, we accept them as valid.
The third determination is more controversial. Argentina argues that neither the 180-day suspension of courts’ issuances of final judgments nor its refusal to allow litigants (and those in arbitration) to use its contract renegotiation process, taken separately or together, warrants suspending or waiving the local litigation requirement. We would not necessarily characterize these actions as rendering a domestic court-exhaustion requirement “absurd and unreasonable,” but at the same time we cannot say that the arbitrators’ conclusions are barred by the Treaty. The arbitrators did not “‘stra[y] from interpretation and application of the agreement’” or otherwise “‘effectively “dispens[e]”’” their “‘own brand of .... justice.’” Stolt-Nielsen S. A. v AnimalFeeds Int’l Corp., 559 U. S. 662, 671 (2010) (providing that it is only when an arbitrator engages in such activity that “‘his decision may be unenforceable’” (quoting Major League Baseball Players Assn. v Garvey, 532 U. S. 504, 509 (2001) (per curiam)).
Consequently, we conclude that the arbitrators’ jurisdictional determinations are lawful. The judgment of the Court of Appeals to the contrary is reversed.
It is so ordered.
(concurring in part)
I agree with the Court that the local litigation requirement at issue in this case is a procedural precondition to arbitration (which the arbitrators are to interpret), not a condition on Argentina’s consent to arbitrate (which a court would review de novo). Ante, at 8, 14. Importantly, in reaching this conclusion, the Court acknowledges that “the treaty before us does not state that the local litigation requirement is a ‘condition of consent’ to arbitration.” Ante, at 12. The Court thus wisely “leave[s] for another day the question of interpreting treaties that refer to ‘conditions of consent’ explicitly.” Ibid. I join the Court’s opinion on the understanding that it does not, in fact, decide this issue.
I write separately because, in the absence of this express reservation, the opinion might be construed otherwise. The Court appears to suggest in dictum that a decision by treaty parties to describe a condition as one on their consent to arbitrate “is unlikely to be conclusive” in deciding whether the parties intended for the condition to be resolved by a court. Ante, at 11. Because this suggestion is unnecessary to decide the case and is in tension with the Court’s explicit reservation of the issue, I join the opinion of the Court with the exception of Part IV–A–1.
The Court’s dictum on this point is not only unnecessary; it may also be incorrect. It is far from clear that a treaty’s express use of the term “consent” to describe a precondition to arbitration should not be conclusive in the analysis. We have held, for instance, that “a gateway dispute about whether the parties are bound by a given arbitration clause raises a ‘question of arbitrability’ for a court to decide.” Howsam v Dean Witter Reynolds, Inc., 537 U. S. 79, 84 (2002) . And a party plainly cannot be bound by an arbitration clause to which it does not consent. See Granite Rock Co v Teamsters, 561 U. S. 287, 299 (2010) (“Arbitration is strictly ‘a matter of consent’ ” (quoting Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 479 (1989) ).
Consent is especially salient in the context of a bilateral investment treaty, where the treaty is not an already agreed-upon arbitration provision between known parties, but rather a nation state’s standing offer to arbitrate with an amorphous class of private investors. In this setting, a nation-state might reasonably wish to condition its consent to arbitrate with a previously unspecified investor counterparty on the investor’s compliance with a requirement that might be deemed “purely procedural” in the ordinary commercial context, ante, at 9. Moreover, as The Chief Justice notes, “[i]t is no trifling matter” for a sovereign nation to “subject itself to international arbitration” proceedings, so we should “not presume that any country .... takes that step lightly.” Post, at 9 (dissenting opinion).
Consider, for example, the United States-Korea Free Trade Agreement, which as the Court recognizes, ante, at 12–13, includes a provision explicitly entitled “Conditions and Limitations on Consent of Each Party.” Art. 11.18, Feb. 10, 2011. That provision declares that “[n]o claim may be submitted to arbitration” unless a claimant first waives its “right to initiate or continue before any administrative tribunal or court .... any proceeding with respect to any measure alleged to constitute a breach” under another provision of the treaty. Ibid. If this waiver condition were to appear without the “consent” label in a binding arbitration agreement between two commercial parties, one might characterize it as the kind of procedural “‘condition precedent to arbitrability’” that we presume parties intend for arbitrators to decide. Howsam, 537 U. S., at 85. But where the waiver requirement is expressly denominated a “condition on consent” in an international investment treaty, the label could well be critical in determining whether the states party to the treaty intended the condition to be reviewed by a court. After all, a dispute as to consent is “the starkest form of the question whether the parties have agreed to arbitrate.” Post, at 13. And we ordinarily presume that parties intend for courts to decide such questions because otherwise arbitrators might “force unwilling parties to arbitrate a matter they reasonably would have thought a judge .... would decide.” First Options of Chicago, Inc. v Kaplan, 514 U. S. 938, 945 (1995) .
Accordingly, if the local litigation requirement at issue here were labeled a condition on the treaty parties’ “consent” to arbitrate, that would in my view change the analysis as to whether the parties intended the requirement to be interpreted by a court or an arbitrator. As it is, however, all parties agree that the local litigation requirement is not so denominated. See Agreement for the Promotion and Protection of Investments, Art. 8(2), Dec. 11, 1990, 1765 U. N. T. S. 38. Nor is there compelling reason to suppose the parties silently intended to make it a condition on their consent to arbitrate, given that a local court’s decision is of no legal significance under the treaty, ante, at 8–9, and given that the entire purpose of bilateral investment agreements is to “reliev[e] investors of any concern that the courts of host countries will be unable or unwilling to provide justice in a dispute between a foreigner and their own government,” Brief for Professors and Practitioners of Arbitration Law as Amici Curiae 6. Moreover, Argentina’s conduct confirms that the local litigation requirement is not a condition on consent, for rather than objecting to arbitration on the ground that there was no binding arbitration agreement to begin with, Argentina actively participated in the constitution of the arbitral panel and in the proceedings that followed. See Eastern Airlines, Inc. v Floyd, 499 U. S. 530, 546 (1991) (treaty interpretation can be informed by parties’ post-enactment conduct).
In light of these many indicators that Argentina and the United Kingdom did not intend the local litigation requirement to be a condition on their consent to arbitrate, and on the understanding that the Court does not pass on the weight courts should attach to a treaty’s use of the term “consent,” I concur in the Court’s opinion.
Chief Justice Roberts
(with whom Justice Kennedy joins, dissenting)
The Court begins by deciding a different case, “initially treat[ing] the document before us as if it were an ordinary contract between private parties.” Ante, at 6. The “document before us,” of course, is nothing of the sort. It is instead a treaty between two sovereign nations: the United Kingdom and Argentina. No investor is a party to the agreement. Having elided this rather important fact for much of its analysis, the majority finally “relax[es] [its] ordinary contract assumption and ask[s] whether the fact that the document before us is a treaty makes a critical difference to [its] analysis.” Ante, at 10. It should come as no surprise that, after starting down the wrong road, the majority ends up at the wrong place.
I would start with the document that is before us and take it on its own terms. That document is a bilateral investment treaty between the United Kingdom and Argentina, in which Argentina agreed to take steps to encourage U. K. investors to invest within its borders (and the United Kingdom agreed to do the same with respect to Argentine investors). Agreement for the Promotion and Protection of Investments, Dec. 11, 1990, 1765 U. N. T. S. 33 (Treaty). The Treaty does indeed contain a completed agreement for arbitration – between the signatory countries. Art. 9. The Treaty also includes, in Article 8, certain provisions for resolving any disputes that might arise between a signatory country and an investor, who is not a party to the agreement.
One such provision – completely ignored by the Court in its analysis – specifies that disputes may be resolved by arbitration when the host country and an investor “have so agreed.” Art. 8(2)(b), 1765 U. N. T. S. 38. No one doubts that, as is the normal rule, whether there was such an agreement is for a court, not an arbitrator, to decide. See First Options of Chicago, Inc. v Kaplan, 514 U. S. 938–945 (1995).
When there is no express agreement between the host country and an investor, they must form an agreement in another way, before an obligation to arbitrate arises. The Treaty by itself cannot constitute an agreement to arbitrate with an investor. How could it? No investor is a party to that Treaty. Something else must happen to create an agreement where there was none before. Article 8(2)(a) makes clear what that something is: An investor must submit his dispute to the courts of the host country. After 18 months, or an unsatisfactory decision, the investor may then request arbitration.
Submitting the dispute to the courts is thus a condition to the formation of an agreement, not simply a matter of performing an existing agreement. Article 8(2)(a) constitutes in effect a unilateral offer to arbitrate, which an investor may accept by complying with its terms. To be sure, the local litigation requirement might not be absolute. In particular, an investor might argue that it was an implicit aspect of the unilateral offer that he be afforded a reasonable opportunity to submit his dispute to the local courts. Even then, however, the question would remain whether the investor has managed to form an arbitration agreement with the host country pursuant to Article 8(2)(a). That question under Article 8(2)(a) is – like the same question under Article 8(2)(b) – for a court, not an arbitrator, to decide. I respectfully dissent from the Court’s contrary conclusion.
The majority acknowledges – but fails to heed – “the first principle that underscores all of our arbitration decisions: Arbitration is strictly ‘a matter of consent.’” Granite Rock Co v Teamsters, 561 U. S. 287, 299 (2010) (quoting Volt Information Sciences, Inc v Board of Trustees of Leland Stanford Junior Univ., 489 U. S. 468, 479 (1989) ); see ante, at 7. We have accordingly held that arbitration “is a way to resolve those disputes – but only those disputes – that the parties have agreed to submit to arbitration.” First Options of Chicago, Inc., supra, at 943. The same “first principle” underlies arbitration pursuant to bilateral investment treaties. See C. Dugan, D. Wallace, N. Rubins, & B. Sabahi, Investor-State Arbitration 219 (2008) (Dugan); J. Salacuse, The Law of Investment Treaties 385 (2010); K. Vandevelde, Bilateral Investment Treaties: History, Policy, and Interpretation 433 (2010). So only if Argentina agreed with BG Group to have an arbitrator resolve their dispute did the arbitrator in this case have any authority over the parties.
The majority opinion nowhere explains when and how Argentina agreed with BG Group to submit to arbitration. Instead, the majority seems to assume that, in agreeing with the United Kingdom to adopt Article 8 along with the rest of the Treaty, Argentina thereby formed an agreement with all potential U. K. investors (including BG Group) to submit all investment-related disputes to arbitration. That misunderstands Article 8 and trivializes the significance to a sovereign nation of subjecting itself to arbitration anywhere in the world, solely at the option of private parties.
The majority focuses throughout its opinion on what it calls the Treaty’s “arbitration clause,” ante, at 1, but that provision does not stand alone. Rather, it is only part – and a subordinate part at that – of a broader dispute resolution provision. Article 8 is thus entitled “Settlement of Disputes Between an Investor and the Host State,” and it opens without so much as mentioning arbitration. 1765 U. N. T. S. 37. Instead it initially directs any disputing investor and signatory country (what the Treaty calls a “Contracting Party”) to court. When “an investor of one Contracting Party and the other Contracting Party” have an investment-related dispute that has “not been amicably settled,” the Treaty commands that the dispute “shall be submitted, at the request of one of the Parties to the dispute, to the decision of the competent tribunal of the Contracting Party in whose territory the investment was made.” Art. 8(1), id., at 37–38. (emphasis added). This provision could not be clearer: Before taking any other steps, an aggrieved investor must submit its dispute with a Contracting Party to that Contracting Party’s own courts.
There are two routes to arbitration in Article 8(2)(a), and each passes through a Contracting Party’s domestic courts. That is, the Treaty’s arbitration provisions in Article 8(2)(a) presuppose that the parties have complied with the local litigation provision in Article 8(1). Specifically, a party may request arbitration only (1) “after a period of eighteen months has elapsed from the moment when the dispute was submitted to the competent tribunal of the Contracting Party in whose territory the investment was made” and “the said tribunal has not given its final decision,” Art. 8(2)(a)(i), id., at 38, or (2) “where the final decision of the aforementioned tribunal has been made but the Parties are still in dispute,” Art. 8(2)(a)(ii), ibid. Either way, the obligation to arbitrate does not arise until the Contracting Party’s courts have had a first crack at the dispute.
Article 8 provides a third route to arbitration in paragraph 8(2)(b) – namely, “where the Contracting Party and the investor of the other Contracting Party have so agreed.” Ibid. In contrast to the two routes in Article 8(2)(a), this one does not refer to the local litigation provision. That omission is significant. It makes clear that an investor can bypass local litigation only by obtaining the Contracting Party’s explicit agreement to proceed directly to arbitration. Short of that, an investor has no choice but to litigate in the Contracting Party’s courts for at least some period.
The structure of Article 8 confirms that the routes to arbitration in paragraph (2)(a) are just as much about eliciting a Contracting Party’s consent to arbitrate as the route in paragraph 8(2)(b). Under Article 8(2)(b), the requisite consent is demonstrated by a specific agreement. Under Article 8(2)(a), the requisite consent is demonstrated by compliance with the requirement to resort to a country’s local courts.
Whereas Article 8(2)(a) is part of a completed agreement between Argentina and the United Kingdom, it constitutes only a unilateral standing offer by Argentina with respect to U. K. investors – an offer to submit to arbitration where certain conditions are met. That is how scholars understand arbitration provisions in bilateral investment treaties in general. See Dugan 221; Salacuse 381; Brief for Practitioners and Professors of International Arbitration Law as Amici Curiae 4. And it is how BG Group itself describes this investment treaty in particular. See Brief for Petitioner 43 (the Treaty is a “standing offer” by Argentina “to arbitrate”); Reply Brief 9 (same).
An offer must be accepted for a legally binding contract to be formed. And it is an “undeniable principle of the law of contracts, that an offer .... by one person to another, imposes no obligation upon the former, until it is accepted by the latter, according to the terms in which the offer was made. Any qualification of, or departure from, those terms, invalidates the offer.” Eliason v Henshaw, 4 Wheat. 225, 228 (1819) (emphasis added). This principle applies to international arbitration agreements just as it does to domestic commercial contracts. See Dugan 221–222; Salacuse 381; Schreuer, Consent to Arbitration, in The Oxford Handbook of International Investment Law 830, 836–837 (P. Muchlinski, F. Ortino, & C. Schreuer eds. 2008).
By incorporating the local litigation provision in Article 8(1), paragraph 8(2)(a) establishes that provision as a term of Argentina’s unilateral offer to arbitrate. To accept Argentina’s offer, an investor must therefore first litigate its dispute in Argentina’s courts – either to a “final decision” or for 18 months, whichever comes first. Unless the investor does so (or, perhaps, establishes a valid excuse for failing to do so, as discussed below, see infra, at 17), it has not accepted the terms of Argentina’s offer to arbitrate, and thus has not formed an arbitration agreement with Argentina.
Although the majority suggests that the local litigation requirement would not be a “condition of consent” even if the Treaty explicitly called it one, the Court’s holding is limited to treaties that contain no such clear statement. See ante, at 11–13. But there is no reason to think that such a clear statement should be required, for we generally do not require “talismanic words” in treaties. Medellín v Texas, 552 U. S. 491, 521 (2008) . Indeed, another arbitral tribunal concluded that the local litigation requirement was a condition on Argentina’s consent to arbitrate despite the absence of the sort of clear statement apparently contemplated by the majority. See ICS Inspection & Control Servs. Ltd. v Argentine Republic, PCA Case No. 2010–9, Award on Jurisdiction, ¶262 (Feb. 10, 2012). Still other tribunals have reached the same conclusion with regard to similar litigation requirements in other Argentine bilateral investment treaties. See Daimler Financial Servs. AG v Argentine Republic, ICSID Case No. ARB/05/1, Award, ¶¶193, 194 (Aug. 22, 2012); Wintershall Aktiengesellschaft v Argentine Republic, ICSID Case No. ARB/04/14, Award, ¶116 (Dec. 8, 2008).
In the face of this authority, the majority quotes a treatise for the proposition that “ ‘[a] substantial body of arbitral authority from investor-state disputes concludes that compliance with procedural mechanisms in an arbitration agreement (or bilateral investment treaty) is not ordinarily a jurisdictional prerequisite.’” Ante, at 16 (quoting 1 G. Born, International Commercial Arbitration 842 (2009)). But that simply restates the question. The whole issue is whether the local litigation requirement is a mere “procedural mechanism” or instead a condition on Argentina’s consent to arbitrate.
BG Group concedes that other terms of Article 8(1) constitute conditions on Argentina’s consent to arbitrate, even though they are not expressly labeled as such. See Tr. of Oral Arg. 57 (“You have to be a U. K. investor, you have to have a treaty claim, you have to be suing another party to the treaty. And if those aren’t true, then there is no arbitration agreement” (emphasis added)). The Court does not explain why the only other term – the litigation requirement – should be viewed differently.
Nor does the majority’s reading accord with ordinary contract law, which treats language such as the word “after” in Article 8(2)(a)(i) as creating conditions, even though such language may not constitute a “clear statement.” See 13 R. Lord, Williston on Contracts §38:16 (4th ed. 2013). The majority seems to regard the local litigation requirement as a condition precedent to performance of the contract, rather than a condition precedent to formation of the contract. Ante, at 8–9; see 13 Lord §§38:4, 38:7. But that cannot be. Prior to the fulfillment of the local litigation requirement, there was no contract between Argentina and BG Group to be performed. The Treaty is not such an agreement, since BG Group is of course not a party to the Treaty. Neither the majority nor BG Group contends that the agreement is under Article 8(2)(b), the provision that applies “where the Contracting Party and the investor of the other Contracting Party have so agreed.” An arbitration agreement must be formed, and Article 8(2)(a) spells out how an investor may do that: by submitting the dispute to local courts for 18 months or until a decision is rendered.
Moreover, the Treaty’s local litigation requirement certainly does not resemble “time limits, notice, laches, estoppel,” or the other kinds of provisions that are typically treated as conditions on the performance of an arbitration agreement, rather than prerequisites to formation. Revised Uniform Arbitration Act of 2000 §6(c), Comment 2, 7 U. L. A. 26 (2009). Unlike a time limit for submitting a claim to arbitration, see Howsam v Dean Witter Reynolds, Inc., 537 U. S. 79, 85 (2002) , the litigation requirement does not simply regulate the timing of arbitration. As the majority recognizes, ante, at 15–16, the provision does not simply require the parties to wait for 18 months before proceeding to arbitration, but instead requires them to submit their claims for adjudication during that period. And unlike a mandatory pre-arbitration grievance procedure, see John Wiley & Sons, Inc v Livingston, 376 U. S. 543–559 (1964), the litigation requirement sends the parties to court – and not just any court, but a court of the host country.
The law of international arbitration and domestic contract law lead to the same conclusion: Because paragraph (2)(a) of Article 8 constitutes only a unilateral standing offer by the Contracting Parties to each other’s investors to submit to arbitration under certain conditions, an investor cannot form an arbitration agreement with a Contracting Party under the Treaty until the investor accepts the actual terms of the Contracting Party’s offer. Absent a valid excuse, that means litigating its dispute in the Contracting Party’s courts to a “final decision” or, barring that, for at least 18 months.
The nature of the obligations a sovereign incurs in agreeing to arbitrate with a private party confirms that the local litigation requirement is a condition on a signatory country’s consent to arbitrate, and not merely a condition on performance of a pre-existing arbitration agreement. There are good reasons for any sovereign to condition its consent to arbitrate disputes on investors’ first litigating their claims in the country’s own courts for a specified period. It is no trifling matter for a sovereign nation to subject itself to suit by private parties; we do not presume that any country – including our own – takes that step lightly. Cf. United States v Bormes, 568 U. S. ___, ___ (2012) (slip op., at 4) (Congress must “unequivocally express[ ]” its intent to waive the sovereign immunity of the United States (quoting United States v Nordic Village, Inc., 503 U. S. 30, 33 (1992) ; internal quotation marks omitted)). But even where a sovereign nation has subjected itself to suit in its own courts, it is quite another thing for it to subject itself to international arbitration. Indeed, “[g]ranting a private party the right to bring an action against a sovereign state in an international tribunal regarding an investment dispute is a revolutionary innovation” whose “uniqueness and power should not be over-looked.” Salacuse 137. That is so because of both the procedure and substance of investor-state arbitration.
Procedurally, paragraph (3) of Article 8 designates the Arbitration Rules of the United Nations Commission on International Trade Law (UNCITRAL) as the default rules governing the arbitration. Those rules authorize the Secretary-General of the Permanent Court of Arbitration at The Hague to designate an “appointing authority” who – absent agreement by the parties – can select the sole arbitrator (or, in the case of a three-member tribunal, the presiding arbitrator, where the arbitrators nominated by each of the parties cannot agree on a presiding arbitrator). UNCITRAL Arbitration Rules, Arts. 6, 8–9 (rev. 2010 ed.). The arbitrators, in turn, select the site of the arbitration (again, absent an agreement by the parties) and enjoy broad discretion in conducting the proceedings. Arts. 18, 17(1).
Substantively, by acquiescing to arbitration, a state permits private adjudicators to review its public policies and effectively annul the authoritative acts of its legislature, executive, and judiciary. See Salacuse 355; G. Van Harten, Investment Treaty Arbitration and Public Law 65–67 (2007). Consider the dispute that gave rise to this case: Before the arbitral tribunal, BG Group challenged multiple sovereign acts of the Argentine Government taken after the Argentine economy collapsed in 2001 – in particular, Emergency Law 25,561, which converted dollar-denominated tariffs into peso-denominated tariffs at a rate of one Argentine peso to one U. S. dollar; Resolution 308/02 and Decree 1090/02, which established a renegotiation process for public service contracts; and Decree 214/02, which stayed for 180 days injunctions and the execution of final judgments in lawsuits challenging the effects of the Emergency Law. Indeed, in awarding damages to BG Group, the tribunal held that the first three of these enactments violated Article 2 of the Treaty. See App. to Pet. for Cert. 241a–242a, 305a.
Perhaps they did, but that is not the issue. Under Article 8, a Contracting Party grants to private adjudicators not necessarily of its own choosing, who can meet literally anywhere in the world, a power it typically reserves to its own courts, if it grants it at all: the power to sit in judgment on its sovereign acts. Given these stakes, one would expect the United Kingdom and Argentina to have taken particular care in specifying the limited circumstances in which foreign investors can trigger the Treaty’s arbitration process. And that is precisely what they did in Article 8(2)(a), requiring investors to afford a country’s own courts an initial opportunity to review the country’s enactments and assess the country’s compliance with its international obligations. Contrast this with Article 9, which provides for arbitration between the signatory countries of disputes under the Treaty without any preconditions. Argentina and the United Kingdom considered arbitration with particular foreign investors to be different in kind and to require special limitations on its use.
The majority regards the local litigation requirement as toothless simply because the Treaty does not require an arbitrator to “give substantive weight to the local court’s determinations on the matters at issue between the parties,” ante, at 9; see also ante, at 15–16, but instead provides that “[t]he arbitration decision shall be final and binding on both Parties,” Art. 8(4), 1765 U. N. T. S. 38. While it is true that an arbitrator need not defer to an Argentine court’s judgment in an investor dispute, that does not deprive the litigation requirement of practical import. Most significant, the Treaty provides that an “arbitral tribunal shall decide the dispute in accordance with .... the laws of the Contracting Party involved in the dispute.” Art. 8(4), ibid. I doubt that a tribunal would give no weight to an Argentine court’s authoritative construction of Argentine law, rendered in the same dispute, just because it might not be formally bound to adopt that interpretation.
The local litigation requirement can also help to narrow the range of issues that remain in controversy by the time a dispute reaches arbitration. It might even induce the parties to settle along the way. And of course the investor might prevail, which could likewise obviate the need for arbitration. Cf. McKart v. United States, 395 U. S. 185, 195 (1969) .
None of this should be interpreted as defending Argentina’s history when it comes to international investment. That history may prompt doubt that requiring an investor to resort to that country’s courts in the first instance will be of any use. But that is not the question. Argentina and the United Kingdom reached agreement on the term at issue. The question can therefore be rephrased as whether it makes sense for either Contracting Party to insist on resort to its courts before being compelled to arbitrate anywhere in the world before arbitrators not of its choosing. The foregoing reasons may seem more compelling when viewed apart from the particular episode before us.
Given that the Treaty’s local litigation requirement is a condition on consent to arbitrate, it follows that whether an investor has complied with that requirement is a question a court must decide de novo, rather than an issue for the arbitrator to decide subject only to the most deferential judicial review. See, e.g., Adams v Suozzi, 433 F. 3d 220, 226–228 (CA2 2005) (holding that compliance with a condition on formation of an arbitration agreement is for a court, rather than an arbitrator, to determine). The logic is simple: Because an arbitrator’s authority depends on the consent of the parties, the arbitrator should not as a rule be able to decide for himself whether the parties have in fact consented. Where the consent of the parties is in question, “reference of the gateway dispute to the court avoids the risk of forcing parties to arbitrate a matter that they may well not have agreed to arbitrate.” Howsam, 537 U. S., at 83–84.
This principle is at the core of our arbitration precedents. See Granite Rock Co., 561 U. S., at 299 (questions concerning “the formation of the parties’ arbitration agreement” are for a court to decide de novo). The same principle is also embedded in the law of international commercial arbitration. 2 Born 2792 (“[W]here one party denies ever having made an arbitration agreement or challenges the validity of any such agreement, .... the possibility of de novo judicial review of any jurisdictional award in an annulment action is logically necessary”). See also Restatement (Third) of U. S. Law of International Commercial Arbitration §4–12(d)(1) (Tent. Draft No. 2, Apr. 16, 2012) (“a court determines de novo .... the existence of the arbitration agreement”).
Indeed, the question in this case – whether BG Group accepted the terms of Argentina’s offer to arbitrate – presents an issue of contract formation, which is the starkest form of the question whether the parties have agreed to arbitrate. In Howsam v Dean Witter Reynolds, Inc., we gave two examples of questions going to consent, which are for courts to decide: “whether the parties are bound by a given arbitration clause” and “whether an arbitration clause in a concededly binding contract applies to a particular type of controversy.” 537 U. S., at 84. In both examples, there is at least a putative arbitration agreement between the parties to the dispute. The only question is whether the agreement is truly binding or whether it covers the specific dispute. Here, by contrast, the question is whether the arbitration clause in the Treaty between the United Kingdom and Argentina gives rise to an arbitration agreement between Argentina and BG Group at all. Cf. ante, at 2 (Sotomayor, J., concurring in part) (“Consent is especially salient in the context of a bilateral investment treaty, where the treaty is not an already agreed-upon arbitration provision between known parties”).
The majority never even starts down this path. Instead, it preempts the whole inquiry by concluding that the local litigation requirement is the kind of “procedural precondition” that parties typically expect an arbitrator to enforce. Ante, at 8–9. But as explained, the local litigation requirement does not resemble the requirements we have previously deemed presumptively procedural. See supra, at 8. It does not merely regulate the timing of arbitration. Nor does it send the parties to non-judicial forms of dispute resolution.
More importantly, all of the cases cited by the majority as examples of procedural provisions involve commercial contracts between two private parties. See ante, at 9. None of them – not a single one – involves an agreement between sovereigns or an agreement to which the person seeking to compel arbitration is not even a party. The Treaty, of course, is both of those things.
The majority suggests that I am applying “a different kind of analysis” from that governing private commercial contracts, just because what is at issue is a treaty. Ante, at 15. That is not so: The key point, which the majority never addresses, is that there is no completed agreement whatsoever between Argentina and BG Group. An agreement must be formed, and whether that has happened is – as it is in the private commercial contract context – an issue for a court to decide. See supra, at 12–13.
The distinction between questions concerning consent to arbitrate and mere procedural requirements under an existing arbitration agreement can at times seem elusive. Even the most mundane procedural requirement can be recast as a condition on consent as a matter of technical logic. But it should be clear by now that the Treaty’s local litigation requirement is not a mere formality – not in Buenos Aires, not in London. And while it is true that “parties often submit important matters to arbitration,” ante, at 11, our precedents presume that parties do not submit to arbitration the most important matter of all: whether they are subject to an agreement to arbitrate in the first place.
Nor has the majority pointed to evidence that would rebut this presumption by showing that Argentina “ ‘clearly and unmistakably’ ” intended to have an arbitrator enforce the litigation requirement. Howsam, supra, at 83 (quoting AT&T Technologies, Inc. v Communications Workers, 475 U. S. 643, 649 (1986) ). As the majority notes, ante, at 14, the Treaty incorporates certain arbitration rules that, in turn, authorize arbitrators to determine their own jurisdiction over a dispute. See Art. 8(3). But those rules do not operate until a dispute is properly before an arbitral tribunal, and of course the whole question in this case is whether the dispute between BG Group and Argentina was before the arbitrators, given BG Group’s failure to comply with the 18-month local litigation requirement. As a leading treatise has explained, “[i]f the parties have not validly agreed to any arbitration agreement at all, then they also have necessarily not agreed to institutional arbitration rules.” 1 Born 870. “In these circumstances, provisions in institutional rules cannot confer any [such] authority upon an arbitral tribunal.” Ibid.
I also see no reason to think that arbitrators enjoy comparative expertise in construing the local litigation requirement. Ante, at 14. It would be one thing if that provision involved the application of the arbitrators’ own rules, cf. Howsam, supra, at 85, or if it were “intertwined” with the merits of the underlying dispute, John Wiley & Sons, 376 U. S., at 557. Neither is true of the litigation requirement. A court can assess compliance with the requirement at least as well as an arbitrator can. Given the structure of Article 8 and the important interests that the litigation requirement protects, it seems clear that the United Kingdom and Argentina thought the same.
Although the Court of Appeals got there by a slightly different route, it correctly concluded that a court must decide questions concerning the interpretation and application of the local litigation requirement de novo. 665 F. 3d 1363, 1371–1373 (CADC 2012). At the same time, however, the court seems to have simply taken it for granted that, because BG Group did not submit its dispute to the local courts, the arbitral award in BG Group’s favor was invalid. Indeed, the court addressed the issue in a perfunctory paragraph at the end of its opinion and saw “ ‘only one possible outcome’”: “that BG Group was required to commence a lawsuit in Argentina’s courts and wait eighteen months before filing for arbitration.” Id., at 1373 (quoting Stolt-Nielsen S. A v Animal Feeds Int’l Corp., 559 U. S. 662, 677 (2010) ).
That conclusion is not obvious. A leading treatise has indicated that “[i]t is a necessary implication from [a unilateral] offer that the offeror, in addition, makes a subsidiary offer by which he or she promises to accept a tender of performance.” 1 Lord §5:14, at 1005. On this understanding, an offeree’s failure to comply with an essential condition of the unilateral offer “will not bar an action, if failure to comply with the condition is due to the offeror’s own fault.” Id., at 1005–1006.
It would be open to BG Group to argue before the Court of Appeals that this principle was incorporated into Article 8(2)(a) as an implicit aspect of Argentina’s unilateral offer to arbitrate. Such an argument would find some support in the background principle of customary international law that a foreign individual injured by a host country must ordinarily exhaust local remedies – unless doing so would be “futile.” See Dugan 347–357. In any event, the issue would be analyzed as one of contract formation, and therefore would be for the court to decide. I would accordingly vacate the decision of the Court of Appeals and remand the case for such an inquiry.
I respectfully dissent.
 The dissent discounts the significance of Argentina’s conduct on the ground that Argentina “object[ed] to the [arbitral] tribunal’s jurisdiction to hear the dispute.” Post, at 16, n. 2. But there is a difference between arguing that a party has failed to comply with a procedural condition in a binding arbitration agreement and arguing that noncompliance with the condition negates the existence of consent to arbitrate in the first place. Argentina points to no evidence that its objection was of the consent variety. This omission is notable because Argentina knew how to phrase its arguments before the arbitrators in terms of consent; it argued separately that it had not consented to arbitration with BG Group on the ground that BG was not a party to the license underlying the dispute. See App. to Pet. for Cert. 182a–186a. First Options of Chicago, Inc v Kaplan, 514 U. S. 938 (1995) , is not to the contrary, as that case held that “arguing the arbitrability issue to an arbitrator” did not constitute “clea[r] and unmistakabl[e]” evidence sufficient to override an indisputably applicable presumption that a court was to decide whether the parties had agreed to arbitration. Id., at 944, 946. The question here, by contrast, is whether that presumption attaches to begin with – that is, whether the local litigation requirement was a condition on Argentina’s consent to arbitrate (which would trigger the presumption) or a procedural condition in an already binding arbitration agreement (which would not). That Argentina apparently took the latter position in arbitration is surely relevant evidence that the condition was, in fact, not one on its consent.
 To be clear, the only question is whether BG Group formed an arbitration agreement with Argentina. To say that BG Group never formed such an agreement is not to call into question the validity of its various commercial agreements with Argentina.
 Justice Sotomayor contends that “Argentina’s conduct confirms that the local litigation requirement is not a condition on consent, for rather than objecting to arbitration on the ground that there was no binding arbitration agreement to begin with, Argentina actively participated in the constitution of the arbitral panel and in the proceedings that followed.” Ante, at 4 (opinion concurring in part). But as the arbitral tribunal itself recognized, Argentina did object to the tribunal’s jurisdiction to hear the dispute. App. to Pet. for Cert. 99a, 134a, 143a, 161a–163a. And we have held that “merely arguing the arbitrability issue to an arbitrator” – say, by “filing with the arbitrators a written memorandum objecting to the arbitrators’ jurisdiction” – “does not indicate a clear willingness to arbitrate that issue, i.e., a willingness to be effectively bound by the arbitrator’s decision on that point.” First Options of Chicago, Inc. v Kaplan, 514 U. S. 938, 946 (1995) . The concurrence contends that Argentina “apparently” argued its jurisdictional objection in terms of procedure rather than consent, ante, at 4, n., but the one piece of evidence cited – a negative inference from the arbitrator’s characterization of Argentina’s argument on a subsidiary issue – hardly suffices to distinguish First Options.
all rights reserved