Argued March 22, 1995 Decided May 22, 1995
This case arose out of disputes centered on a "workout" agreement, embodied in four documents, which governs the "working out" of debts owed by respondents — Manuel Kaplan, his wife, and his wholly owned investment company, MK Investments, Inc. (MKI) — to petitioner First Options of Chicago, Inc., a firm that clears stock trades on the Philadelphia Stock Exchange. When First Options' demands for payment went unsatisfied, it sought arbitration by a stock exchange panel. MKI, which had signed the only workout document containing an arbitration agreement, submitted to arbitration, but the Kaplans, who had not signed that document, filed objections with the panel, denying that their disagreement with First Options was arbitrable. The arbitrators decided that they had the power to rule on the dispute's merits and ruled in First Options' favor. The District Court confirmed the award, but the Court of Appeals reversed. In finding that the dispute was not arbitrable, the Court of Appeals said that courts should independently decide whether an arbitration panel has jurisdiction over a dispute, and that it would apply ordinary standards of review when considering the District Court's denial of respondents' motion to vacate the arbitration award.
1. The arbitrability of the Kaplan/First Options dispute was subject to independent review by the courts. Pp. 942-947.
(a) The answer to the narrow question whether the arbitrators or the courts have the primary power to decide whether the parties agreed to arbitrate a dispute's merits is fairly simple. Just as the arbitrability of the merits of a dispute depends upon whether the parties agreed to arbitrate that dispute, see, e.g., Mastrobuono v. Shearson Lehman Hutton, Inc., ante, at 52, so the question "who has the primary power to decide arbitrability" turns upon whether the parties agreed to submit that question to arbitration. If so, then the court should defer to the arbitrator's arbitrability decision. If not, then the court should decide the question independently. These two answers flow inexorably from the fact that arbitration is simply a matter of contract between the parties. Pp. 942-943.
(b) The Kaplans did not agree to arbitrate arbitrability. Courts generally should apply ordinary state-law principles governing contract formation in deciding whether such an agreement exists. However, courts should not assume that the parties agreed to arbitrate arbitrability unless there is "clea[r] and unmistakabl[e]" evidence that they did so. See, e.g., ATT Technologies, Inc. v. Communications Workers, 475 U.S. 643, 649. First Options cannot show a clear agreement on the part of the Kaplans. The Kaplans' objections to the arbitrators' jurisdiction indicate that they did not want the arbitrators to have binding authority over them. This conclusion is supported by (1) an obvious explanation for their presence before the arbitrators ( i.e., Mr. Kaplan's wholly owned firm was arbitrating workout agreement matters); and (2) Third Circuit law, which suggested that they might argue arbitrability to the arbitrators without losing their right to independent court review. First Options' counterarguments are unpersuasive. Pp. 943-947.
2. Courts of appeals should apply ordinary standards when reviewing district court decisions upholding arbitration awards, i.e., accepting findings of fact that are not "clearly erroneous" but deciding questions of law de novo; they should not, in those circumstances, apply a special "abuse of discretion" standard. It is undesirable to make the law more complicated by proliferating special review standards without good reason. More importantly, a court of appeals' reviewing attitude toward a district court decision should depend upon the respective institutional advantages of trial and appellate courts, not upon what standard of review will more likely produce a particular substantive result. Nothing in the Arbitration Act supports First Options' claim that a court of appeals should use a different standard when conducting review of certain district court decisions. Pp. 947-949.
3. The factbound question whether the Court of Appeals erred in its ultimate conclusion that the dispute was not arbitrable is beyond the scope of the questions this Court agreed to review. P. 949.
19 F.3d 1503, affirmed.
BREYER, J., delivered the opinion for a unanimous Court.
James D. Holzhauer argued the cause for petitioner. With him on the briefs were Timothy S. Bishop, Stephen P. Bedell, Timothy G. McDermott, and Kenneth E. Wile.
John G. Roberts, Jr., argued the cause for respondents. With him on the brief for respondent Manuel Kaplan were Donald L. Perelman, Richard A. Koffman, and David G. Leitch. Gary A. Rosen filed a brief for respondent Carol Kaplan.
Briefs of amici curiae urging reversal were filed for the National Futures Association et al. by Daniel J. Roth; and for the Philadelphia Stock Exchange, Inc., et al. by Lydia Gavalis.
Gerald F. Rath, Steven W. Hansen, and Stuart J. Kaswell filed a brief for the Securities Industry Association as amicus curiae.
In this case we consider two questions about how courts should review certain matters under the federal Arbitration Act, 9 U.S.C. § 1 et seq. (1988 ed. and Supp. V): (1) how a district court should review an arbitrator's decision that the parties agreed to arbitrate a dispute, and (2) how a court of appeals should review a district court's decision confirming, or refusing to vacate, an arbitration award.
The case concerns several related disputes between, on one side, First Options of Chicago, Inc., a firm that clears stock trades on the Philadelphia Stock Exchange, and, on the other side, three parties: Manuel Kaplan; his wife, Carol Kaplan; and his wholly owned investment company, MK Investments, Inc. (MKI), whose trading account First Options cleared. The disputes center on a "workout" agreement, embodied in four separate documents, which governs the "working out" of debts to First Options that MKI and the Kaplans incurred as a result of the October 1987 stock market crash. In 1989, after entering into the agreement, MKI lost an additional $1.5 million. First Options then took control of, and liquidated, certain MKI assets; demanded immediate payment of the entire MKI debt; and insisted that the Kaplans personally pay any deficiency. When its demands went unsatisfied, First Options sought arbitration by a panel of the Philadelphia Stock Exchange.
MKI, having signed the only workout document (out of four) that contained an arbitration clause, accepted arbitration. The Kaplans, however, who had not personally signed that document, denied that their disagreement with First Options was arbitrable and filed written objections to that effect with the arbitration panel. The arbitrators decided that they had the power to rule on the merits of the parties' dispute, and did so in favor of First Options. The Kaplans then asked the Federal District Court to vacate the arbitration award, see 9 U.S.C. § 10 (1988 ed., Supp. V), and First Options requested its confirmation, see § 9. The court confirmed the award. Nonetheless, on appeal the Court of Appeals for the Third Circuit agreed with the Kaplans that their dispute was not arbitrable; and it reversed the District Court's confirmation of the award against them. 19 F.3d 1503(1994).
We granted certiorari to consider two questions regarding the standards that the Court of Appeals used to review the determination that the Kaplans' dispute with First Options was arbitrable. 513 U.S. 1040 (1994). First, the Court of Appeals said that courts "should independently decide whether an arbitration panel has jurisdiction over the merits of any particular dispute." 19 F.3d, at 1509 (emphasis added). First Options asked us to decide whether this is so ( i.e., whether courts, in "reviewing the arbitrators' decision on arbitrability," should "apply a de novo standard of review or the more deferential standard applied to arbitrators' decisions on the merits") when the objecting party "submitted the issue to the arbitrators for decision." Pet. for Cert. i. Second, the Court of Appeals stated that it would review a district court's denial of a motion to vacate a commercial arbitration award (and the correlative grant of a motion to confirm it) " de novo." 19 F.3d, at 1509. First Options argues that the Court of Appeals instead should have applied an "abuse of discretion" standard. See Robbins v. Day, 954 F.2d 679, 681-682 (CA11 1992).
The first question — the standard of review applied to an arbitrator's decision about arbitrability — is a narrow one. To understand just how narrow, consider three types of disagreement present in this case. First, the Kaplans and First Options disagree about whether the Kaplans are personally liable for MKI's debt to First Options. That disagreement makes up the merits of the dispute. Second, they disagree about whether they agreed to arbitrate the merits. That disagreement is about the arbitrability of the dispute. Third, they disagree about who should have the primary power to decide the second matter. Does that power belong primarily to the arbitrators (because the court reviews their arbitrability decision deferentially) or to the court (because the court makes up its mind about arbitrability independently)? We consider here only this third question.
Although the question is a narrow one, it has a certain practical importance. That is because a party who has not agreed to arbitrate will normally have a right to a court's decision about the merits of its dispute (say, as here, its obligation under a contract). But, where the party has agreed to arbitrate, he or she, in effect, has relinquished much of that right's practical value. The party still can ask a court to review the arbitrator's decision, but the court will set that decision aside only in very unusual circumstances. See, e.g., 9 U.S.C. § 10 (award procured by corruption, fraud, or undue means; arbitrator exceeded his powers); Wilko v. Swan, 346 U.S. 427, 436-437 (1953) (parties bound by arbitrator's decision not in "manifest disregard" of the law), overruled on other grounds, Rodriguez de Quijas v. Shearson/American Express, Inc., 490 U.S. 477 (1989). Hence, who — court or arbitrator — has the primary authority to decide whether a party has agreed to arbitrate can make a critical difference to a party resisting arbitration.
We believe the answer to the "who" question ( i.e., the standard-of-review question) is fairly simple. Just as the arbitrability of the merits of a dispute depends upon whether the parties agreed to arbitrate that dispute, see, e.g., Mastrobuono v. Shearson Lehman Hutton, Inc., ante, at 57.; Mitsubishi Motors Corp. v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 626 (1985), so the question "who has the primary power to decide arbitrability" turns upon what the parties agreed about that matter. Did the parties agree to submit the arbitrability question itself to arbitration? If so, then the court's standard for reviewing the arbitrator's decision about that matter should not differ from the standard courts apply when they review any other matter that parties have agreed to arbitrate. See ATT Technologies, Inc. v. Communications Workers, 475 U.S. 643, 649 (1986) (parties may agree to arbitrate arbitrability); Steelworkers v. Warrior Gulf Nav. Co., 363 U.S. 574, 583, n. 7 (1960) (same). That is to say, the court should give considerable leeway to the arbitrator, setting aside his or her decision only in certain narrow circumstances. See, e.g., 9 U.S.C. § 10. If, on the other hand, the parties did not agree to submit the arbitrability question itself to arbitration, then the court should decide that question just as it would decide any other question that the parties did not submit to arbitration, namely independently. These two answers flow inexorably from the fact that arbitration is simply a matter of contract between the parties; it is a way to resolve those disputes — but only those disputes — that the parties have agreed to submit to arbitration. See, e.g., ATT Technologies, supra, at 649; Mastrobuono, supra, at 57-58, and n. 9; Allied-Bruce Terminix Cos. v. Dobson, 513 U.S. 265, 271 (1995); Mitsubishi Motors Corp., supra, at 625-626.
We agree with First Options, therefore, that a court must defer to an arbitrator's arbitrability decision when the parties submitted that matter to arbitration. Nevertheless, that conclusion does not help First Options win this case. That is because a fair and complete answer to the standard-of-review question requires a word about how a court should decide whether the parties have agreed to submit the arbitrability issue to arbitration. And, that word makes clear that the Kaplans did not agree to arbitrate arbitrability here.
When deciding whether the parties agreed to arbitrate a certain matter (including arbitrability), courts generally (though with a qualification we discuss below) should apply ordinary state-law principles that govern the formation of contracts. See, e.g., Mastrobuono, ante, at 62-63, and n. 9; Volt Information Sciences, Inc. v. Board of Trustees of Leland Stanford Junior Univ., 489 U.S. 468, 475-476 (1989); Perry v. Thomas, 482 U.S. 483, 492-493, n. 9 (1987); G. Wilner, 1 Domke on Commercial Arbitration § 4:04, p. 15 (rev. ed. Supp. 1993) (hereinafter Domke). The relevant state law here, for example, would require the court to see whether the parties objectively revealed an intent to submit the arbitrability issue to arbitration. See, e.g., Estate of Jesmer v. Rohlev, 241 Ill. App.3d 798, 803, 609 N.E.2d 816, 820 (1993) (law of the State whose law governs the workout agreement); Burkett v. Allstate Ins. Co., 368 Pa. 600, 608, 534 A.2d 819, 823-824 (1987) (law of the State where the Kaplans objected to arbitrability). See generally Mitsubishi Motors, supra, at 626.
This Court, however, has (as we just said) added an important qualification, applicable when courts decide whether a party has agreed that arbitrators should decide arbitrability: Courts should not assume that the parties agreed to arbitrate arbitrability unless there is "clea[r] and unmistakabl[e]" evidence that they did so. ATT Technologies, supra, at 649; see Warrior Gulf, supra, at 583, n. 7. In this manner the law treats silence or ambiguity about the question " who (primarily) should decide arbitrability" differently from the way it treats silence or ambiguity about the question " whether a particular merits-related dispute is arbitrable because it is within the scope of a valid arbitration agreement" — for in respect to this latter question the law reverses the presumption. See Mitsubishi Motors, supra, at 626 ("`[A]ny doubts concerning the scope of arbitrable issues should be resolved in favor of arbitration'") (quoting Moses H. Cone Memorial Hospital v. Mercury Constr. Corp., 460 U.S. 1, 24-25 (1983)); Warrior Gulf, supra, at 582-583.
But, this difference in treatment is understandable. The latter question arises when the parties have a contract that provides for arbitration of some issues. In such circumstances, the parties likely gave at least some thought to the scope of arbitration. And, given the law's permissive policies in respect to arbitration, see, e.g., Mitsubishi Motors, supra, at 626, one can understand why the law would insist upon clarity before concluding that the parties did not want to arbitrate a related matter. See Domke § 12.02, p. 156 (issues will be deemed arbitrable unless "it is clear that the arbitration clause has not included" them). On the other hand, the former question — the "who (primarily) should decide arbitrability" question — is rather arcane. A party often might not focus upon that question or upon the significance of having arbitrators decide the scope of their own powers. Cf. Cox, Reflections Upon Labor Arbitration, 72 Harv. L. Rev. 1482, 1508-1509 (1959), cited in Warrior Gulf, 363 U.S., at 583, n. 7. And, given the principle that a party can be forced to arbitrate only those issues it specifically has agreed to submit to arbitration, one can understand why courts might hesitate to interpret silence or ambiguity on the "who should decide arbitrability" point as giving the arbitrators that power, for doing so might too often force unwilling parties to arbitrate a matter they reasonably would have thought a judge, not an arbitrator, would decide. Ibid. See generally Dean Witter Reynolds Inc. v. Byrd, 470 U.S. 213, 219-220 (1985) (Arbitration Act's basic purpose is to "ensure judicial enforcement of privately made agreements to arbitrate").
On the record before us, First Options cannot show that the Kaplans clearly agreed to have the arbitrators decide ( i.e., to arbitrate) the question of arbitrability. First Options relies on the Kaplans' filing with the arbitrators a written memorandum objecting to the arbitrators'jurisdiction. But merely arguing the arbitrability issue to an arbitrator 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. To the contrary, insofar as the Kaplans were forcefully objecting to the arbitrators deciding their dispute with First Options, one naturally would think that they did not want the arbitrators to have binding authority over them. This conclusion draws added support from (1) an obvious explanation for the Kaplans' presence before the arbitrators ( i.e., that MKI, Mr. Kaplan's wholly owned firm, was arbitrating workout agreement matters); and (2) Third Circuit law that suggested that the Kaplans might argue arbitrability to the arbitrators without losing their right to independent court review, Teamsters v. Western Pennsylvania Motor Carriers Assn., 574 F.2d 783, 786-788 (1978); see 19 F.3d, at 1512, n. 13.
First Options makes several counterarguments: (1) that the Kaplans had other ways to get an independent court decision on the question of arbitrability without arguing the issue to the arbitrators ( e.g., by trying to enjoin the arbitration, or by refusing to participate in the arbitration and then defending against a court petition First Options would have brought to compel arbitration, see 9 U.S.C. § 4); (2) that permitting parties to argue arbitrability to an arbitrator without being bound by the result would cause delay and waste in the resolution of disputes; and (3) that the Arbitration Act therefore requires a presumption that the Kaplans agreed to be bound by the arbitrators' decision, not the contrary. The first of these points, however, while true, simply does not say anything about whether the Kaplans intended to be bound by the arbitrators'decision. The second point, too, is inconclusive, for factual circumstances vary too greatly to permit a confident conclusion about whether allowing the arbitrator to make an initial (but independently reviewable) arbitrability determination would, in general, slow down the dispute resolution process. And, the third point is legally erroneous, for there is no strong arbitration-related policy favoring First Options in respect to its particular argument here. After all, the basic objective in this area is not to resolve disputes in the quickest manner possible, no matter what the parties' wishes, Dean Witter Reynolds, supra, at 219-220, but to ensure that commercial arbitration agreements, like other contracts, "`are enforced according to their terms,'" Mastrobuono, ante, at 54 (quoting Volt Information Sciences, 489 U.S., at 479), and according to the intentions of the parties, Mitsubishi Motors, 473 U.S., at 626. See Allied-Bruce, 513 U.S., at 271. That policy favors the Kaplans, not First Options.
We conclude that, because the Kaplans did not clearly agree to submit the question of arbitrability to arbitration, the Court of Appeals was correct in finding that the arbitrability of the Kaplan/First Options dispute was subject to independent review by the courts.
We turn next to the standard a court of appeals should apply when reviewing a district court decision that refuses to vacate, see 9 U.S.C. § 10 (1988 ed., Supp. V), or confirms, see § 9, an arbitration award. Although the Third Circuit sometimes used the words " de novo" to describe this standard, its opinion makes clear that it simply believes (as do all Circuits but one) that there is no special standard governing its review of a district court's decision in these circumstances. Rather, review of, for example, a district court decision confirming an arbitration award on the ground that the parties agreed to submit their dispute to arbitration, should proceed like review of any other district court decision finding an agreement between parties, i.e., accepting findings of fact that are not "clearly erroneous" but deciding questions of law de novo. See 19 F.3d, at 1509.
One Court of Appeals, the Eleventh Circuit, has said something different. Because of federal policy favoring arbitration, that court says that it applies a specially lenient "abuse of discretion" standard (even as to questions of law) when reviewing district court decisions that confirm (but not those that set aside) arbitration awards. See, e.g., Robbins v. Day, 954 F.2d, at 681-682. First Options asks us to hold that the Eleventh Circuit's view is correct.
We believe, however, that the majority of Circuits is right in saying that courts of appeals should apply ordinary, not special, standards when reviewing district court decisions upholding arbitration awards. For one thing, it is undesirable to make the law more complicated by proliferating review standards without good reasons. More importantly, the reviewing attitude that a court of appeals takes toward a district court decision should depend upon "the respective institutional advantages of trial and appellate courts," not upon what standard of review will more likely produce a particular substantive result. Salve Regina College v. Russell, 499 U.S. 225, 231-233 (1991). The law, for example, tells all courts (trial and appellate) to give administrative agencies a degree of legal leeway when they review certain interpretations of the law that those agencies have made. See, e.g., Chevron U.S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 843-844 (1984). But no one, to our knowledge, has suggested that this policy of giving leeway to agencies means that a court of appeals should give extra leeway to a district court decision that upholds an agency. Similarly, courts grant arbitrators considerable leeway when reviewing most arbitration decisions; but that fact does not mean that appellate courts should give extra leeway to district courts that uphold arbitrators. First Options argues that the Arbitration Act is special because the Act, in one section, allows courts of appeals to conduct interlocutory review of certain antiarbitration district court rulings ( e.g., orders enjoining arbitrations), but not those upholding arbitration ( e.g., orders refusing to enjoin arbitrations). 9 U.S.C. § 16 (1988 ed., Supp. V). But that portion of the Act governs the timing of review; it is therefore too weak a support for the distinct claim that the court of appeals should use a different standard when reviewing certain district court decisions. The Act says nothing about standards of review.
We conclude that the Court of Appeals used the proper standards for reviewing the District Court's arbitrability determinations.
Finally, First Options argues that, even if we rule against it on the standard-of-review questions, we nonetheless should hold that the Court of Appeals erred in its ultimate conclusion that the merits of the Kaplan/First Options dispute were not arbitrable. This factbound issue is beyond the scope of the questions we agreed to review.
The judgment of the Court of Appeals is affirmed.