THE ARBITRATOR’S JURISDICTION TO DETERMINE JURISDICTION*
The boundaries of my language signify the borders of my world.
Legal phrases, maxims and rules often enhance efficient dispute resolution by providing intellectual hooks on which to hang analysis, as well as mental handles with which to arrange otherwise complex arguments. Like the questions we ask, the language of the law can shape the choices ultimately made by arbitrators and judges.
Words can beget misunderstanding as well as insight, however. Expressions which bear multiple meanings often find themselves employed with promiscuous disregard to context and function.
The disorienting effect of language finds illustration in the principle that arbitrators may rule on their own authority. Often expressed as Kompetenz-Kompetenz2 (literally “jurisdiction on jurisdiction”), the precept has been applied to questions such as who must arbitrate, what must be arbitrated, and which powers arbitrators may exercise.3
As we shall see, this much-vexed principle possesses a chameleon-like quality that changes color according to the national and institutional background of its application. The basic rule that arbitrators may decide on their own jurisdiction says nothing about who ultimately decides a particular case. Rather, the rule states only that the question of “who decides what” may itself be addressed by an arbitrator. At least until a competent court directs otherwise, arbitral proceedings need not stop just because one side challenges the arbitrator’s authority.
To say that arbitrators may make jurisdictional decisions tells only part of the story.4 Every jurisdictional ruling by an arbitrator begs two further questions, one relating to timing and the other to finality.
The timing question asks when judges should intervene in the arbitral process to monitor possible jurisdictional excess. If an unhappy respondent denies having agreed to arbitrate, a court might be requested to declare the arbitration clause invalid. Should a judge entertain a “mid-arbitration” request to stop the proceedings? Or should the respondent be required to wait until an award has been rendered, and only then seek vacatur for alleged jurisdictional excess?
Each alternative carries its own risks and opportunities for mischief. Delay in judicial scrutiny can subject respondents to the expense of unauthorized proceedings before overreaching arbitrators.5 However, early access to courts increases opportunities for dilatory tactics. In the business world, determining the scope of arbitration clauses may implicate time-consuming investigations into complex questions of fact and law related to matters such as agency relationships and the corporate veil.6
The second question relates to the effect that judges should give to arbitrators’ jurisdictional rulings. In what circumstances (if any) should an arbitrator’s decision on his or her authority be final?
Legal systems differ on whether and when an arbitrator’s decision on his or her authority should foreclose judicial determination on the matter. Some countries (notably the United States) implement the litigants’ agreement to have arbitral authority determined by the arbitrators themselves. Judges, of course, must still ask what (if anything) the parties actually expected the arbitrator to decide.7 Assuming such an agreement exists, however, it will be respected.
Other countries (notably Germany) seem to preclude such agreements to arbitrate arbitrability. This approach sacrifices liberty of contract in order to provide an extra measure of protection against inadvertent loss of the proverbial day in court.
This dual line of inquiry, looking at the timing of judicial intervention and the effect of arbitral determinations, can remove much of the mystification afflicting jurisdictional discourse in arbitration law. From a policy perspective, the correct answers will not always be self-evident. However, asking the right questions, rather than simply reciting a catch-phrase, permits attention to the costs and benefits of each alternative, enhancing the transactional security and economic cooperation that can be facilitated by arbitration.8
The principle in primitive form
In its most primitive form, the principle that arbitrators may rule on their jurisdiction serves as a measure to protect against having an arbitration derailed before it begins. The arbitral tribunal (and/or the relevant arbitral institution) need not halt the proceedings just because one side questions its authority. The principle reduces the prospect that proceedings will be derailed through a simple allegation that an arbitration clause is unenforceable, due to any number of contract law defenses. In most legal systems,9 arbitrators can get on with their work until ordered to stop by a judge with authority to do so.10
The rule is not foolproof, of course, given the eternal ingenuity with which fools often acquit themselves. Recalcitrant parties can still mount troublesome court challenges (even if not ultimately successful) designed to slow the train.11 However, the principle does avoid conceptual barriers to arbitration that would exist if legal systems considered jurisdictional powers of judges and of arbitrators to be mutually exclusive.12
On occasion, analogies have been made between arbitral jurisdiction and the power of courts to construe constitutional provisions related to their authority. Such comparisons should be resisted. Few non-circular options exist for interpreting judicial authority, at least in western legal systems. By contrast, in commercial arbitration the enforcement of arbitral authority (initially a matter of the litigants’ consent13) normally rests with national courts, which must undertake some investigation into the legitimacy of that authority as part of the enforcement process.
Diversity: the timing and impact of court intervention
Although most countries accept that a jurisdictional objection does not automatically stop an arbitration, little consensus exists on other aspects of an arbitrator’s ruling on his or her authority. National practice diverges in both (i) the timing of court examination of arbitral authority and (ii) the impact that an arbitrator’s jurisdictional ruling will have in a judicial proceeding.
Diversity results from the fact that an arbitrator’s jurisdictional power, at least in commercial arbitration,14 derives from national law and institutional rules,15 not from the treaty framework imposed by the New York Convention.16 Consequently, the expression Kompetenz-Kompetenz has thus taken on several lives, giving rise to a constellation of related but distinct notions, often subject to undue mystification.17 While commentators sometimes refer to “the internationally recognized doctrine” of Kompetenz-Kompetenz,18 it would be more accurate to speak of doctrines in the plural.19 Variations derive both from disparate implementations of the principle and from divergent views on what exactly is meant by a “jurisdictional question”.20
To illustrate, if German courts are asked to hear a matter which one side asserts is subject to arbitration, they would decide immediately on the validity and scope of the arbitration agreement.21 In neighboring France, challenge of arbitration clauses must normally wait until an award has been rendered.22
Across the Channel in England, litigants have a right to declaratory decisions on arbitral authority only if they take no part in the arbitration.23 In Germany the admissibility of such applications depends on whether the arbitral tribunal has already been constituted.24 In Sweden and in Finland,25 parties may apply to courts for jurisdictional declarations at any time.
In Switzerland, courts asked to appoint an arbitrator will normally apply a prima facie standard in deciding whether the arbitration clause is valid, but engage in full consideration of jurisdiction (at least as to law) in the context of award review.26 American courts, however, may order full examination of the validity of an arbitration clause at any stage of the arbitral process to determine whether, as a matter of fact and law, the parties have indeed agreed to arbitrate.27
The United States generally permits parties to give arbitrators the final word on some aspects of arbitral power.28 A similar result would seem to obtain in Finland.29 In other countries, however, the effect of such agreements remains far from clear.30
In light of this multiplicity of applications, the temptation exists to suggest that the term Kompetenz-Kompetenz be exiled from the arbitration lexicon, and that scholars abandon any hope of rationalizing the principle. Such radical change would be ill-advised, however. Only the most compelling reasons justify banishment of time-honored notions.31 The remedy for confusion will normally lie in a fuller appreciation of the contextual application of the term, a task to which we now turn.
The modest suggestions of this chapter are threefold. First, discourse about arbitral jurisdiction suffers considerable damage through loose jargon divorced from specific national practice. What matters is when courts examine the limits of arbitral authority and the effect (if any) that judges give the parties’ agreements about such authority. Second, agreements to submit jurisdictional questions to arbitration should be honored but not presumed. Finally, although arguments about the timing of judicial intervention remain finely balanced, the weightier considerations argue for postponing most jurisdictional inquiry until after the award has been rendered. In any event, more coherent and precise rules remain desirable.
The shadow of public power
On its face, Kompetenz-Kompetenz addresses the powers of arbitrators, in particular their right to make jurisdictional rulings. The flip side of the equation, however, reveals a rule about courts, and the limitations on judges’ ability to hear certain matters imposed when litigants decide (or allegedly decide) to submit controverted questions to private dispute resolution. This reverse perspective highlights the heart of understanding how the principle works in practice.
Although private, arbitration proceeds in the shadow of public coercion. Arbitrators have no marshals or sheriffs, and thus parties often ask judges to stay litigation, compel arbitral proceedings, seize assets or grant res judicata effect to an award so as to preclude competing court actions. The contours of arbitral power thus concern not only arbitrator and litigants, but also national legal systems which must establish guidelines for when and to what extent courts may intervene to review or to pre-empt the arbitrator’s jurisdictional ruling.
From the perspective of a national legal system, challenges to an arbitrator’s authority raise two distinct questions. The first relates to the point in the arbitral process when courts ought to examine arbitral authority to prevent or correct an excess of jurisdiction.32 The second addresses the matter of when (if ever) courts should defer to an arbitrator’s jurisdictional determination as final.
The first inquiry concerns the timing of judicial intervention. Paradigms range from the American approach (courts may intervene at any moment) to the French model (courts wait until after an award is rendered). The difference becomes significant when one side to the dispute makes application to a court with supervisory (curial) competence over the arbitration, asking that the proceedings be stopped or that a case be heard notwithstanding an alleged arbitration clause.33
Between these two extremes, many legal systems provide hybrid timing solutions that vary according to the specific posture in which arbitral jurisdiction has been challenged. One standard might apply when a legal action is brought in respect of matters purportedly referred to arbitration. Another standard might pertain to motion for declaratory judicial determination of preliminary jurisdictional questions. Distinctions might be made depending on whether the applicant has or has not taken part in the arbitration.34
Effect of an arbitrator’s determination
The other question relates to the effect of an arbitration agreement on jurisdictional questions. A legal system might take the position that all arbitral decisions on jurisdiction may be reviewed de novo by the appropriate court.35 However, such is not the only option, or even the most sensible one. An alternative would be for courts to ask what jurisdictional matters the parties agreed the arbitrator would decide, and to defer accordingly.
In systems where courts may defer to an arbitrator’s jurisdictional determination, judges must still examine arbitral authority. However, the analysis is placed at a different level, asking whether the parties intended an arbitrator to have the last word on a particular jurisdictional issue. The pertinent question is what the contract provides.36
With the obvious exception of challenges based on public policy (non-arbitrable subjects), analysis would normally focus on the parties’ pre-dispute intent. Courts must examine the facts of each case as they bear on the parties’ pre-dispute expectations. If (and only if) the litigants intended arbitration of a particular jurisdictional question, would the matter be given to the arbitrator for ultimate disposition, not just an expression of preliminary views. However, in all events courts would first look seriously at the parties’ expectations.
A cautionary tale about an (allegedly) lazy professor
The point about the binding nature of an agreement to arbitrate jurisdiction might be illustrated by the following scenario. Imagine a publication dispute between a law journal and a learned professor. Having written an article for the journal, the professor asserts that the editor agreed to pay a $1,000 royalty. On failure to pay, she files an arbitration claim based on what she believes to constitute a valid arbitration clause in the license to publish. “No way!” the editor replies. “On its face, the clause provides that an arbitrator has jurisdiction only over claims filed within thirty days after the dispute arises. This lazy professor missed that deadline, having waited to file her claim on the 4th of July, more than two months after our early May disagreement on the matter.”
The professor sees things quite differently. She replies that no differences arose until the middle of June. She recalls no discussion in May, and challenges the editor’s recollection. To settle the matter, the author and the journal sign a written agreement to have an arbitrator determine when the disagreement arose: the first week of May or the middle of June. They agree that the same proceeding will address entitlement to the $1,000, and that both questions will be addressed by an eminent French scholar whom they jointly appoint as sole arbitrator.
After hearing the witnesses and reading the parties’ submissions, the arbitrator finds no evidence of a disagreement before 15 June.37 This means the author’s claim was timely filed on 4 July. Further, the arbitrator finds that the editor did indeed agree to pay $1,000 for the article.
Does any sound policy give the editor a jurisdictional escape hatch? Having agreed to submit the time limits to arbitration, why should the Editor be allowed to renege on the bargain and ask courts to decide de novo when the disagreement occurred? Although a judge might be skilled at weighing the evidence, the parties submitted the question to arbitrators.
Any such a second bite at the jurisdictional apple would seem inconsistent with the whole thrust of modern arbitration law, which aims to give res judicata effect to arbitral awards based on valid agreements. Any arbitral award should, of course, be subject to challenge for jurisdictional infirmity, such as physical coercion or forgery in the arbitration clause. However, an agreement accepted with informed consent, followed by fair proceedings, should bind both sides. Having lost the arbitration, the editor should not be permitted to refuse the author her $1,000 fee by re-opening the dispute. Indeed, if the award is made abroad, refusal to grant recognition might well violate the New York Convention.
In this cautionary scenario, a particular issue (the date of the Editor/author disagreement) started out being characterized as “jurisdictional” but ended (by the parties’ consent) as a matter of the “substantive merits” in their dispute. There is nothing unusual about such transformation. Without going as far as the proverbial Humpty Dumpty, most thoughtful people accept that words have different meanings in different contexts, and that language would be misapplied if labels used against one background are transferred to another with no adjustment to take into account their function.38
A word on procedural context
The way courts evaluate arbitration clauses often depends on the procedural context in which the clauses present themselves. Some countries apply different standards to pre-award and post-award judicial scrutiny, distinguishing between prima facie and full review. On occasion, legal systems permit jurisdictional challenges brought in the course of court actions (“stop the lawsuit so we can arbitrate”) but deny requests for declarations about ongoing arbitrations (“stop the arbitration because we should be in court”).39 In some instances a court will address jurisdiction differently depending on whether or not the arbitration has actually begun.40 Different rules might also apply according to whether the arbitration is conducted locally or abroad.41
Several scenarios merit consideration. First, applications to review awards (partial or final) may be brought at the place of arbitration as motions to vacate or to confirm. Review might also be sought at the enforcement situs through motions to recognize or enforce the award. Second, a respondent in a court action might assert an arbitration clause as a bar to a lawsuit brought on the merits of claim, usually for breach of contract. Finally, a litigant considering that a dispute should be heard in court rather than in arbitration may petition for a judicial declaration (combined with an injunction in some countries) about the scope or validity of an action clause. Such actions might be brought either before or during the arbitration.
Differences relate not only to when and whether courts may address arbitral jurisdiction, but on the standards of review applied when they do examine the validity of the arbitration clause. The most significant dividing line relates to whether the judge will make a full inquiry into the parties’ intent, or simply a summary examination, applying what is sometimes called a prima facie standard.
For example, a seller might bring a judicial action to collect the price of an engine. In response, the buyer (who alleges the engine was defective) might move to stay litigation, asserting that the parties had agreed to arbitrate their dispute. The seller might reply with allegations that the arbitration clause was void.
In the alternative, the buyer might file an arbitration for product malfunction, alleging an engine explosion that caused personal injury and loss of profits. Here it would be the seller (preferring to be in court) who asks a judge to address the validity and scope of the arbitration agreement, perhaps arguing that the person who signed the clause lacked authority, or that the clause was not broad enough to cover the tort action for personal injury or the financial claim for lost profits.
German law illustrates how these procedural postures evoke different judicial responses. Courts in Germany would address challenges to the arbitration clause in the context of a lawsuit, with the buyer arguing that the claim should be heard by arbitrators. A judge could also hear the seller’s application (if brought before arbitration began) for a declaration that the arbitration clause was invalid. However, if the arbitration was in progress (and no lawsuit had been brought), the arbitrators would simply rule on their own jurisdiction and proceed with the case. Judicial pronouncement on the allegedly defective arbitration clause would await challenge to an award, whether partial or final.42
Matters get even more complicated in legal systems where different standards of review apply according to the procedural posture of the arbitration. French judges, for example, asked to hear a claim can address the validity of an arbitration clause only in the most superficial manner, and only in the event no arbitral tribunal has been constituted. At that point the court can ask whether the clause was clearly void (for example the document might lack any signature), but must put off until later any more complex questions (such as disputes about whether the scope of the arbitration clause covers the dispute).43 Once the arbitration has started, however, judges must sit on their hands until the award is made, when they provide a full examination of alleged defects in the arbitration clause.44
In some countries, courts distinguish between arbitration held at home or abroad. Swiss courts, for example, make a full and comprehensive review of the validity of the arbitration clause when the arbitration has its seat abroad. By contrast, when the arbitration is held in Switzerland, judges engage only in a summary examination of arbitral jurisdiction (examen sommaire). Full review must wait until the award stage.
In other nations (such as the United States) courts engage in full examination of arbitral power regardless of whether the arbitration has begun, and irrespective of whether they are being asked to hear the merits of the claims. The court might decide that the lawsuit should stop and the arbitration should proceed. Or vice versa. Or, the court might pass this jurisdictional question back to the arbitrators themselves for their determination.
As a general matter, pre-award requests motions for declarations and injunctions implicate a preventive role for courts. The jurisdictional foundation of an arbitral proceeding must be monitored before anyone knows what the arbitrator will decide. The arbitrator’s jurisdiction becomes an issue because judges are asked to make a respondent participate, or to tell a claimant that the arbitration lacks jurisdictional foundation.45
By contrast, when arbitral jurisdiction becomes an issue in the endgame, after an award is rendered, judges exercise a remedial function, correcting mistakes that allegedly occurred earlier in the arbitral process. The validity of an award might be subject to judicial scrutiny at the arbitral seat, through motions to vacate or to confirm under local law.46 Or the award might be subject to scrutiny when presented for recognition abroad, by a winning claimant seeking to attach assets or a prevailing respondent asserting the award’s res judicata effect to block competing litigation. Normally (but not always) the New York Arbitration Convention would be invoked.47 At this point, a different set of options present themselves. Courts then face the choice of either giving effect to the award (by confirmation, recognition or enforcement) or rejecting its validity (by vacatur or non-recognition).
When questions are raised about the validity or scope of a particular arbitration clause, one option would be for the arbitration to stop automatically, until matters have been clarified by a judge. It is against this extreme position, which denies arbitrators any right at all to rule on their own authority, that one must begin to explore the various meanings of Kompetenz-Kompetenz.48
If a legal system does allow the arbitration to proceed in the face of a jurisdictional challenge, the story could unfold in several ways. At least three different approaches might be envisaged.
First, the arbitrators might offer an opinion on the limits of their own authority, but without in any way restricting the court’s consideration of the same question. Although the arbitration does not necessarily stop, neither do related judicial actions. Courts proceed pursuant to whatever motions might be available under local law. Second, courts could refrain from entertaining any jurisdictional motions until after an award had been rendered. The arbitrators would then have the first word on jurisdiction.
The third meaning given to Kompetenz-Kompetenz requires that courts defer completely to an arbitrator’s decision about his or her own authority. The arbitrator gets the last word as well as the first. However, such a result requires that judges first determine that the parties did in fact agree to such finality.49
No automatic stop to the arbitration
Under the first hypothesis, the arbitrator’s right to make jurisdictional rulings operates in tandem with a rule allowing judges to examine an arbitrator’s jurisdiction before an award has been rendered. In some countries, courts may step in from day one, at any time in almost any circumstance.50 In others, courts might have full power to address arbitral jurisdiction in the context of lawsuits on the merits of the claim, but only limited margin to maneuver through declaratory judgments.51
The arbitrator’s right to rule on jurisdiction holds significant practical value (at least for the party wishing to arbitrate) notwithstanding the possibility of court intervention. A recalcitrant respondent cannot bring the proceedings to a halt just by challenging jurisdiction.52 Moreover, whether courts ultimately substitute their own views for those of the arbitrators depends on the facts of each case. In some instances a judge might order the proceedings suspended, either permanently or until the jurisdictional facts have been determined.53 In others, the arbitration clause may be found to be robust enough to cover the controverted dispute.54
As mentioned, even in countries that permit courts to address arbitral jurisdiction before an award is rendered, distinctions are often made between judicial actions on the merits of a dispute (where a defendant asserts that the action is preempted by an arbitration clause) and requests for declaratory judgments about potential or ongoing arbitrations (where a respondent asserts defects in the arbitration clause). With respect to court actions on the merits, judges usually possess full power to address jurisdictional questions, particularly in countries following the UNCITRAL Model Arbitration Law. For declaratory decisions, however, the law sometimes limits the circumstances in which such applications may be made.55
In this regard, it is important not to confuse the allocation of functions between arbitrators and the supervisory arbitral institution with the allocation of responsibility between arbitrators and national courts. For example, under the Arbitration Rules of the International Chamber of Commerce, if the ICC Court is “prima facie satisfied” that an arbitration agreement may exist, any jurisdictional challenge of a deeper nature goes to the arbitrators. This does not mean, however, that national courts will be deprived of power to make jurisdictional determinations when asked to stay litigation, enjoin arbitration or vacate an award.56
Giving arbitrators the first word
In other legal systems, recourse to courts must wait until the end of arbitration, after an award has been rendered. This version of Kompetenz-Kompetenz lays down rules about the stages in the arbitral process at which judges may intervene. The positive part of the principle addresses itself to arbitrators, permitting them to decide challenges to their own authority. The so-called “negative effect” of the principle speaks to courts,57 telling judges to wait until arbitration ends before inquiring about the validity or effect of an arbitration clause.58
Best exemplified by French law, this approach means that if an arbitrator has already begun to hear a matter, courts must decline to hear the case. The judge has a limited opportunity to hear a case only if the arbitration has not begun, and only if the alleged arbitration agreement is found to be clearly void (manifestement nulle). Given the importance of French doctrine in this field, the full text merits consideration:
When a dispute which has been brought before an arbitral tribunal pursuant to an arbitration agreement is brought before a governmental court, the court must declare itself without jurisdiction. If the dispute has not yet been brought before the arbitral tribunal, the court must also declare itself without jurisdiction unless the arbitration agreement is clearly void.59
At issue here is the timing, rather than the extent, of judicial review. Going to court at the beginning of the proceedings can save expense for a defendant improperly joined to the arbitration. On the other hand, judicial resources may be conserved by delaying review until the end of the process, by which time the parties might have settled.
Even in countries that allow judicial intervention before an award is rendered, a core element of modern arbitration law resides in recognition of separate spheres of responsibility for courts and arbitrators. Pale hints of the negative aspect of Kompetenz-Kompetenz can be found, for example, in a recent U.S. Supreme Court decision which adopted a “wait and see” approach with respect to public policy questions related to arbitration of treble damages claims.60
This is not to say, however, that the French timing mechanism itself has gained widespread acceptance. Such across-the-board deference to arbitrators (or alleged arbitrators) commands no wide international consensus. Outside the French hexagon, legal systems follow a more flexible and nuanced approach with respect to court intervention. Significant departures from French practice can be seen not only in the United States, but also in important arbitral venues such as England, Sweden and Switzerland, as well as nations such as Germany which follow the UNCITRAL Model Arbitration Law.61
The point is not trivial, since scholars sometimes cloak the Gallic perspective with wider acceptance than may actually be the case. Readers even encounter references to principles ‘generally recognized in comparative law“62 which, on closer examination, describe nothing more than the parochial French approach (however commendable). Whatever the optimum policy might be on timing judicial intervention, the temptation to blur lines between the “is” and the “ought” of legal doctrine must be resisted.63
The arbitrator’s decision is final
Jurisdiction as a question of substantive merits Regardless of when judges entertain motions on arbitral jurisdiction, the parties might agree, expressly or impliedly, to subject the jurisdictional question to arbitration. In legal systems following this third approach, jurisdictional questions themselves are considered capable of settlement by arbitration, pursuant to agreement by the parties.64 Under these circumstances, an arbitrator’s determination on his or her own authority will be final. The parties’ agreement transforms the jurisdictional difference into a disputed question of fact or law, whose substantive merits the litigants submit to final determination by an arbitrator.65
The application of this line of reasoning will in all events depend on the facts of each case. In some instances, the parties may indeed have agreed to submit a jurisdictional question to final and binding arbitration. In other instances, an assertion that they have done so will be preposterous, unable to withstand analysis except by ignoring reality in favor of fiction. The parties’ agreement, determined on a case-by-case basis, will determine whether this brand of Kompetenz-Kompetenz makes sense. In each instance, the question for judges will be: what did the parties intend to submit to arbitration?
Legal systems disagree on whether judges should ever be permitted to accord finality to an arbitrator’s decision on his or her authority, even on a finding of the parties’ prior consent. In practice, giving arbitrators the last word on jurisdictional questions means that some litigants may well lose their access to court. The peril derives not so much from isolated mistakes, whether by arbitrators or by courts, but from the risk that an overburdened judiciary might fall into a systematic proclivity toward granting jurisdictional authority to arbitrators, even when contracts are ambiguous on the matter.
Long gone are the days when judges exhibited blanket hostility to arbitration. Today, courts often perceive arbitration as a way to clear crowded dockets. Even the best of judges may be tempted to exchange rigorous reasoning for the convenience of a finding that the parties really did want jurisdictional questions addressed by arbitrators. Such a tendency seems to have been a factor in recent changes in German law, which now reduces the prospect of never-bargained-for arbitrations by requiring that all questions of arbitral authority go to judges.66
This approach, however, may create even more problems than it resolves. A judicial monopoly on final resolution of jurisdictional questions imposes serious restrictions on party autonomy, particularly among sophisticated business managers. The result is a serious limit on the liberty of contract that has long bolstered healthy commercial transactions in free economies.
German doctrine: then and now Prior to Germany’s adoption of the UNCITRAL Model Law in 1998, court decisions had recognized that an arbitral tribunal might be granted the power to rule on its own jurisdiction pursuant to a specific clause, accepted by both parties, that implicitly dispensed with subsequent judicial review. In a landmark decision, Germany’s highest court, the Bundesgerichthof, had decided that the parties to a commercial contract could submit the question of arbitral authority to final and binding arbitration.67 What the court called a Kompetenz-Kompetenz-Klausel, or “jurisdiction on jurisdiction clause”, was deemed sufficient to insulate the arbitrator’s decision on the matter from judicial scrutiny.
Currently, the prevailing opinion in Germany (both scholarly and judicial) seems to hold that such Kompetenz-Kompetenz clauses are invalid. German case law has held that the parties may not restrict judges from examining arbitral jurisdiction in the context of challenges to either interim or final awards.68 Whether such a position constitutes sound doctrine remains open to debate, as discussed later.
French and Swedish perspectives French law seems to include a more nuanced position with respect to the finality of an arbitrator’s jurisdictional determinations.69 Unlike Germany, France has known no marked break with prior case law, which may account for the fact that the Gallic position seems to be elaborated through scholarly comment.70
As a starting point, it is clear that the legal framework for judicial review of awards bears a mandatory character. The grounds on which courts set awards aside (enumerated in Nouveau code de procedure civile) may not be abrogated by contract.71
Less evident, however, is the proposition that a ban on waiver of statutory annulment standards necessarily means limitation of the questions which by contract may be submitted for final determination by arbitrators. What might be a jurisdictional question of jurisdiction in some contexts could become a matter of the merits in an arbitration where both sides clearly submitted the issue to arbitration.
For example, one party to a bill of lading might contest that it incorporated by reference the arbitration clause in a related charter-party. Posed in those terms, the matter would normally be a jurisdictional question ultimately to be determined by courts. However, nothing in French law suggests that the two parties cannot, in a clear and distinct agreement, agree to be bound by an arbitrator’s determination on that question.
Sweden seems to take a similar position, albeit in a more explicit fashion. After providing that arbitrators may rule on their own jurisdiction, Section 2 of the Swedish Arbitration Act adds that this principle “shall not prevent a court from determining such a question” and that the decision of arbitrators on their jurisdiction “is not binding” but rather subject to the full panoply of grounds for challenging awards.72 Again, it is not certain that the mandatory nature of judicial review necessarily prohibits parties from submitting jurisdictional questions to arbitration if they so wish.
The American “arbitrability question” In the United States, a clear line of judicial pronouncements holds that in some situations arbitrators may rule on their own powers without subsequent de novo review by courts. In the sense used by American courts, such grants of jurisdictional power are not legal fictions, but require evidence of the parties’ real intent expressed in concrete language either in the main contract and or in a separate agreement.
Jurists from outside the United States may find the terminology unfamiliar. Court decisions speak of the “arbitrability question” in the same way that the rest of the world refers to a jurisdictional issue. If an “arbitrability question” has been submitted to arbitration, then courts defer to the arbitrator on the matter.
Admittedly, the words “Kompetenz-Kompetenz clause” do not figure in American cases. However, after the “arbitrability question” decisions have been broken down and decorticated, one finds that judges in the United States have been using the same conceptual framework as the pre-1998 German cases. In this context, one might recall how the middle-aged cloth merchant in Molière’s Bourgeois Gentilhomme learned, much to his delight, that he had actually been speaking prose all along, without ever being aware of this rhetorical skill.73 The difference, however, is that courts in the United States seem happily oblivious to the link between American legal notions and the doctrines elaborated in the rest of the world to meet similar juridical problems.
In addressing jurisdiction, American courts sometimes say the issue is not only “who decides what,” but also “who decides who decides”. This formulation provides another way of asking when arbitrators may determine the contours of their own decision-making authority.74
The American approach often involves the transformation of a jurisdictional matter (normally for courts) into the substantive merits in the arbitration itself (for the arbitrators). Jurisdictional challenges usually relate to the arbitrator’s authority to decide an issue or to exercise a particular procedural power. Once it has been determined that the parties agreed to entrust to the arbitrator the adjudication of disputes on such questions, then almost by definition the question is no longer one of jurisdiction. Arbitrators receive their power from the parties’ consent. If a court decides that the parties asked the arbitrator to decide a matter (for example time eligibility requirements for arbitration), then in essence this constitutes the court’s jurisdictional determination.
When the very existence of an agreement to arbitrate is disputed, however, courts will generally refuse to compel arbitration until they resolve whether the arbitration clause exists at all.75 In some events, the existence and content of the parties’ agreement may have to be determined by a jury.76
The recent decision in Alliance Bernstein Investment v. Schaffran77 illustrates the various ramifications of the American approach. A former employee of a New York hedge fund alleged wrongful termination, claiming that he had been fired for cooperating with government investigations into wrongdoings by his employer.78
The employment relationship was subject to rules of the National Association of Securities Dealers (NASD), which provided for mandatory arbitration with one important exception: claims of employment discrimination.79 Normally, the arbitration clause would have been invoked by the employer, on the assumption that juries tend to possess a more sympathetic predisposition toward employees.
In this case roles were reversed. For reasons that are not entirely clear from the face of the decision, the employer (not the employee) moved for a declaratory judgment that the “whistle blower” action (alleging retaliation for cooperation with a government investigator) constituted an “employment discrimination” claim by the employee, and thus was not subject to arbitration. The arbitrator’s authority thus depended on whether the employee’s claim could be characterized as an “allegation of discrimination” within the meaning of the NASD Rules.
The court did not see its role as deciding whether or not the arbitrator possessed jurisdiction to hear the claim.80 Rather, the question was who (judge or an arbitrator) would decide whether the allegations of termination for “whistle blowing” were subject to arbitration, or instead amounted to the type of discrimination claim that was carved out of the scope of the arbitration clause.81
The starting point for analysis lay in the relevant NASD Rules, under which the arbitrators were expressly “empowered to interpret and determine the applicability of all provisions under the [NASD] Code” in a way that was “final and binding” on the parties.82 The court’s job, therefore, was to ascertain what the parties meant by that language.83
According to the Second Circuit, a presumption exists that the parties would normally intend an “arbitrability question” to be determined by a judge.84 The presumption might, however, be overcome.85 To do so would require “clear and unmistakable” evidence that the parties wished the question to be decided by arbitrators.86 This intent could be found, for example, in a separate agreement providing for arbitration of “any and all controversies” including interpretation of the provisions of the relevant arbitration rules.87
In Alliance Bernstein the court was careful to keep close to the language of the arbitration provisions. The language of the relevant rule did not provide for “any and all” matters to be arbitrated, but only for power to “interpret and determine the applicability of” provisions under the NASD code, which would include the scope of the exclusion for discrimination claims.88 Since there was no disagreement that both sides had accepted the NASD rules,89 the parties’ intent to arbitrate their differing interpretations of the rules could be ascertained from the four corners of the documentation before the court. The arbitrator’s decision on this matter would not be subject to later judicial second guessing. Rather, the “whistle blower” claim would be subject to arbitration if, and only if, the arbitrator so determined.
The Court found the question of whether “whistle blower” claims were arbitrable was for the arbitrators, and thus insulated that finding from review for “excess of powers” under the Federal Arbitration Act.90 The award might well be attacked on other jurisdictional grounds, however. For example, the arbitrator would still lack power if an irregularity could be found in signature of the agreement containing the reference to arbitration. Perhaps the person who signed was not authorized to do so. Or, the signature might have been compelled by a gun at the head. Or, maybe the signature was a forgery. But the decision on jurisdiction over the “whistle blower” claim could not be disregarded because a judge later disagreed with the arbitrator’s interpretation of the rules.
Under this approach, once a precise question has clearly been delegated to an arbitrator, it ceases to be “jurisdictional” in the context of the case to which it is relevant. Since the arbitrator has been empowered to hear the matter, any further inquiry must be limited to “What did the arbitrator decide?” That decision might relate to a matter which, in the abstract, would be characterized as jurisdictional. However, the parties’ intent prevails, and the arbitrator will determine the matter in a final way.
Such allocation of functions between judges and arbitrators explains itself principally by reference to contract principles. Absent an express or implied waiver of the right to go to court, a litigant will not normally be denied recourse to otherwise competent tribunals. But once such a waiver has been given in the form of an arbitration clause, it is hard to see why a litigant should be permitted to renege on this bargain to arbitrate.
Reasonable people, of course, might argue about what the parties had in mind when they made their bargain. One judge might think that another judge got it wrong or was misguided in her reading of how the relevant arbitration rules or contract affected the questions that would be submitted to arbitration. But these debatable matters of fact, do not call into question the jurisdictional principle that the parties to a dispute may empower arbitrators to decide controversies about the pre-conditions to arbitration.
American judges who review questions of jurisdiction must look beyond labels, and instead fix their scrutiny on the parties’ real deal. If two litigants intended to submit a question to final and binding arbitration, then the arbitral determination holds, regardless of whether the question would initially have fallen within the arbitrator’s mission. In this sense, the Kompetenz-Kompetenz clause remains alive and well in the United States.
In defining “arbitrable questions”, courts are in the business of drawing lines between jurisdiction and merits, often in a manner that enlarges the arbitrator’s authority. Although issues of substance (merits of the dispute) and jurisdiction (arbitrator’s right to hear the case) should be treated differently by courts, the two categories are not fixed immutably in the real world. A particular question might be characterized as “substantive merits” in one dispute and “jurisdiction” in another. If indeed the parties to an arbitration agreement clearly intend for a matter to be decided by arbitration, then a “jurisdictional” label would be inappropriate if it were to lead courts to usurp the arbitral function. To ignore this possibility might, in some circumstances, put a country in breach of its New York Convention obligations. The wrinkles on this topic are sizeable, and thus it has been addressed in greater detail below in the discussion of the new German arbitration law.
Fixing the point in time for court intervention involves a relatively clear (albeit difficult) choice between costs and benefits related to the expenditure of either public or private resources. Under one model, a party unhappy with having to arbitrate may go to court at any moment for the purpose of contesting arbitral power. Another paradigm, however, provides for court challenge of arbitral authority only after an award is rendered.
Court challenge to jurisdiction at the beginning of the process can save time and expense for the litigants. If a judge finds the alleged arbitral clause to be void, or too narrow in scope to cover the dispute, then neither side need waste time or money in arbitration. The parties are free to pursue their litigation in the appropriate judicial forum.
By contrast, government funds can be preserved by delaying judicial review until after the award has been rendered. If questions of authority are left to the end game, perhaps there will not even be a jurisdictional challenge in court. The case might settle, or the party resisting arbitration might prevail. And if the matter does go to court, the arbitrator may have done much of the intellectual heavy lifting, sorting facts and law to provide the reviewing judge a helpful analytic road-map.
Extremes: France and the United States
United States American arbitration law traditionally has given parties a right to raise a matter of arbitral authority at any time, whether before or after the award. Such determinations would usually be made pursuant to litigation under Sections 3 and 4 of the Federal Arbitration Act, providing for stay of court litigation and orders to compel arbitration.91 This approach means that a party who never agreed to arbitrate will not need to waste time and money in a proceeding that lacks an authoritative foundation. Moreover, either side can request clarification of the scope of the arbitrator’s power before substantial sums are spent needlessly. The prospect of award vacatur on jurisdictional grounds cannot be excluded, but it may be less likely to hang as a Sword of Damocles in cases of obvious jurisdictional defect.
France By contrast, the French model delays court consideration of jurisdictional matters until the award review stage.92 This approach reduces the prospect of dilatory tactics designed to derail an arbitration. A bad-faith respondent will be less able to add the cost of a court challenge at the same time as the arbitration is going forward.93
Another benefit from the French paradigm lies in its potential for higher quality jurisdictional review by judges, who will be able to benefit from the arbitrators’ earlier consideration of the matter. And government resources may be conserved for the simple reason that a settlement might obviate the need for judicial review.
A cynic, of course, might note that the French rule can have practical advantages for arbitrators themselves, who will not be declared incompetent until after collecting their fees. But as Rudyard Kipling might have written, that is another story.
Hybrids: England, Switzerland and the UNCITRAL Model
Countries that delay judicial intervention until the award stage aim to preserve government resources. By contrast, legal systems that permit court rulings on arbitral jurisdiction at any moment allow litigants to avoid the expense of an invalid proceeding.
Attempts to find a middle way in the timing of jurisdictional challenge have not always proved easy. Like the man who hoped to get his girlfriend drunk without emptying the wine bottle,94 efforts at meeting both goals have often served disappointment.
Nevertheless, some legal systems do explore hybrid solutions. England, Switzerland and the UNCITRAL Model provide examples.
England95 The position in England seems once to have been roughly analogous to that in the United States, in that arbitrators addressed jurisdiction subject to general control by the competent court.96 Such remains the case with respect to final awards, where dissatisfied litigants may challenge arbitrators’ mistakes on substantive and procedural jurisdiction.97 Things have become a bit more complex since 1996.98 Today, the English Arbitration Act gives an arbitral tribunal the right to rule on its own substantive jurisdiction.99 The right to challenge arbitral jurisdiction by declaration or injunction is open only to a person “who takes no part in the proceedings.”100 This power can be particularly useful in connection with what are sometimes called “unilateral” arbitration clauses, which permit one side the option to litigate in court rather than to arbitrate. If the other side begins an arbitration before the option has been exercised, the power to request a declaration provides the machinery for vindicating the right to litigate.101
Most challenges to substantive competence must wait until an award has been rendered.102 At that time courts will have an opportunity to review excess of authority as well as the arbitrators’ improper arrogation of powers.103 On occasion, a jurisdictional ruling may also give rise to allegations of procedural irregularity.104
The Act does permit application for judicial determination on a “preliminary point of jurisdiction.” In this latter context, courts may consider the matter only on the agreement of all parties, or if the arbitral tribunal grants permission and a court finds that addressing the question is likely to produce substantial savings in costs.105
Otherwise, a party seeking annulment of an arbitrator’s decision for excess of jurisdiction may do so only after attempting to remedy the problem through the appropriate arbitral procedures. In the interest of arbitral efficiency, court challenges to awards can only be brought after any available institutional review.106 And a “use it or lose it” principle requires that challenges for excess of authority must be made “forthwith” or within the time provided by the arbitration agreement.107 To rebut the presumption that the right to object has been waived, the challenging party must show that it did not know, and could not with reasonable diligence have discovered, the grounds for objection.108 In this respect, the Act leads to a result different from the one obtaining in Switzerland, where defendants may lose their right to challenge an award’s jurisdictional underpinnings by boycotting the proceedings.109
Arbitral jurisdiction might also be tested in court if one party brings a court action for a claim which the other party says is covered by the arbitration agreement. Arbitral authority is put at issue in a motion to stay legal proceedings, and the point is decided then and there. Like analogous provisions in Article 8 of the UNCITRAL Model Law, English arbitration law contemplates that in some instances there might be simultaneous proceedings by courts and arbitrators regarding the competence of the arbitral tribunal.110
Switzerland111 Although not free from scholarly debate,112 Swiss case law seems to distinguish between arbitration held inside and outside of the country. Federal statute provides that arbitral tribunals shall rule on their authority, normally through interlocutory decisions,113 and that objections to jurisdiction must be raised before the tribunal prior to any defense on the merits.114 Moreover, state courts must decline jurisdiction unless they find the arbitration clause void, inoperative or incapable of being applied.115
If the seat of the arbitration is in Switzerland, courts engage in only summary examination of arbitral authority.116 When the arbitral seat lies outside of Switzerland, however, the Tribunal fédéral has called for a fuller and more comprehensive examination of the validity of the arbitration agreement.117 This inquiry would generally occur at the time the clause is invoked in a Swiss court action on the merits of the dispute, allegedly brought in disregard of the agreement to arbitrate. In applying Article II of the New York Convention (requiring reference to arbitration unless the clause is void, inoperative or incapable of being performed), courts would not limit themselves to a summary (prima facie) examination of the validity of the agreement to arbitrate.
The logic of this distinction (which has not gone unquestioned118) seems to be that when arbitration occurs abroad, Swiss courts might not get a chance at a later time to correct an arbitrator’s erroneous decision about jurisdiction under the questionable agreement. By contrast, most arbitration conducted inside Switzerland will be subject to judicial review on the grounds enumerated in the federal conflicts of law statute, which include excess of jurisdiction.119
Comparisons are sometimes made between jurisdictional review in France and in Switzerland. Notwithstanding some inferences to the contrary,120 one can see as many (or more) differences as similarities. For arbitrations inside the forum state, both countries delay full judicial review of arbitral authority until the award stage. There the similarity ends, however. Swiss law contains nothing equivalent to the extreme French position that requires courts, while the arbitration is ongoing, to refrain from addressing even the clearest indications of an arbitration clause’s invalidity.121 On the contrary, Swiss courts verify the validity of arbitration clauses in a summary fashion (prima facie) when asked either to appoint an arbitrator or to hear disputes allegedly subject to arbitration.122