The finality of International Arbitral Awards
Roma Tre – UNIDROIT 2nd Annual International Arbitration Lecture
23 September 2015
Prof. Piero Bernardini
One of the reasons that make international arbitration so widespread as a method for settling disputes at transnational level is the “finality” of the award. Finality does not mean that the award may not be subject to review and possibly be annulled. It means that any review of the award has a limited scope, the safeguarding against violation of fundamental principles of law by arbitral tribunal. Indeed, as stated by William Park, President of the LCIA:
“[e]fficient arbitration implicates a tension between the rival goals of finality and fairness. Freeing awards from judicial challenge promotes finality, while enhancing fairness calls for some measure of court supervision. An arbitration’s winner looks for finality, while the loser wants careful judicial scrutiny of doubtful decisions”
There are two potentially conflicting principles at work in the review process. One is the principle of finality; the other is the principle of fairness. Finality is designed to serve the purpose of efficiency in terms of expeditious and economical settlement of disputes. Fairness may be an elusive goal that takes time and effort and may involve several layers of control, a phenomenon that is well known from domestic court procedure. In international arbitration the principle of finality is often seen to take precedence over the principle of fairness. The desire to see a dispute settled is regarded as more important than the substantive correctness of the decision. Annulment is the preferred solution to balance these two objectives. It is designed to provide emergency relief for egregious violations of a few basic principles while preserving the finality of the decision in most respects.
As will be shown, each country develops its own balance between finality and review. The systems vary, but, at the end of the day, it appears that there is a global trend in favour of arbitration, limiting consequently the review of the awards. This can be illustrated by the limited legal grounds available for challenging an award and by the restrictive interpretation given to these grounds by national courts.
The favour arbitrandum and for finality of awards is reflected by the rules of arbitration adopted by the parties for the conduct of arbitral proceedings. The most frequently adopted rules record the parties’ common intend to hold that “the award is final and binding on the parties” and that “the parties undertake to carry out any award without delay” as provided by UNCITRAL Arbitration Rules Article 32(2). A similar provision is in ICC Rules of Arbitration (Article 34(2)), in the Rules of arbitration of the Stockholm Arbitration Institute (Article 45) and in the Rules of arbitration of ICSID Additional Facility. The 2014 LCIA Rules and the 2012 ICC Rules go even a step further by adding that the parties waive their right to any form of recourse to any state court insofar as such waiver can validly be made (respectively, Article 26(8) and 34(6))
The grounds that are available for annulling an international arbitral award in the place of arbitration are defined by the national law of this place (so-called forum law). Although there are in principle no limits on the annulment grounds that forum law may provide, it has been authoritatively held that an implied limit is prescribed by the New York Convention. Under Article II of that Convention, Contracting States (presently over 155 States) have accepted to “recognize” agreements to arbitrate and since it is in the very nature of such agreements to provide for the binding resolution of disputes by arbitration also awards setting forth such resolution are to be final and binding. Consistent with this, national courts in the United States have concluded that actions to annul international awards must be limited to the grounds specified in Article V of the New York Convention for refusing recognition and enforcement of awards. The link of annulment grounds with Article V of the New York Convention is established by UNCITRAL Model Law which provides grounds for annulment that parallel those set forth by said Article V. Whether or not Article II of the New York Convention provides for an implied limit to annulment grounds under national laws, clearly a de novo judicial review would ignore the essential character of arbitration and arbitral awards recognized by the generality of national arbitration laws, which is to finally resolve the parties’ dispute in a binding manner.
Most developed arbitration laws treat international arbitral awards as final and binding, conferring effects similar to those of court judgments, notably that of res judicata. The English Arbitration Act 1996 provides in Sect. 58 that
“1) Unless otherwise agreed by the parties, an award made by the tribunal pursuant to an arbitration agreement is final and binding both on the parties and on any persons claiming through or under them”.
According to the French NCPC, “Once it is made, the arbitral award is res judicata in relation to the disputes it resolves” (Article 1484). As with court judgments, the final and binding force of an award is in principle limited to the territory of the State where the award was made. The New York Convention provides for their recognition outside that territory so that arbitral awards may become part of national legal system of the State granting recognition. In most cases a party will request recognition of an arbitral award in order to raise a defense of res judicata and thus bar the re-litigation in court of issues that have already been resolved in a domestic or foreign arbitration. There are instances in which a party, contrary to the final and binding character of the award, rejects the outcome of an arbitration by commencing litigation (or a new arbitration) aimed at re-litigating the parties’ previous dispute. In these cases, developed legal systems contain principles of preclusion, generally formulated in the context of international court judgments and transposed to the arbitral setting, that give effect to the final character of previous arbitral awards involving the parties’ dispute. The existence and efficient application of principles of “preclusion” is essential to the efficacy of the arbitral system in achieving its objective of finally resolving the parties’ dispute.
Most modern arbitration legislations treat arbitral awards as presumptively valid in actions to annul and recognise awards while permitting annulment only on specified and limited grounds. The European Court of Justice has explained in the Eco Swiss v. Benetton decision of 1999 the policies underlying such provisions as follows:
“it is in the interest of efficient arbitration proceedings that review of arbitration awards should be limited in scope and that annulment of or refusal to recognize an award should be possible only in exceptional circumstances”.
In the last thirty years or so, many States have revised their laws on arbitration, specifically with the aim to facilitate arbitration taking place within their borders.
When an application for the setting aside of an award is made, a national court may have a different scope of review from one jurisdiction to another depending on a number of variables, such as traditional differences in legal systems. The prevailing degree of review is the minimalist review where only issues of procedure are examined and the very minimal review of substance.
I shall briefly review the law and jurisprudence of countries that have not adopted UNCITRAL Model Law but which are often the seat of international arbitration. I shall thus briefly examine the Model Law review system and then conclude by dealing with the ICSID Convention review system.
Reviews of annulment proceedings in Sweden, Switzerland, the United States, France and England, have concluded that annulment of international awards is an exceptional occurrence, with the overwhelming majority of awards being upheld in the face of annulment challenges, thus ensuring their finality.
In Sweden, challenge of an award may be either (i) through a request for a declaration of invalidity for the grounds mentioned in Section 33 of the Arbitration Act of 1999 (the award includes determination of issues that may not be decided by arbitrators or is incompatible with basic principles of Swedish law or does not fulfil the requirements of written form and signature) or (ii) through an application to set aside the award pursuant to Sect. 34 of the Act (award not covered by a valid arbitration agreement, made after expiration of the term or by arbitrators in excess of their mandate or by arbitrators appointed contrary to the parties’ agreement or if its outcome was influenced by irregularity during the proceeding).
The Stockholm (Svea) Court of Appeal, which has jurisdiction over international awards, in the case CME Lauder v. The Czech Republic in a judgement of 15 May 2003, rejecting the Republic’s challenge against the award, held as follows:
“In line with what might be deemed to be an expression of the legal situation in many other countries, by virtue of the Arbitration Act, the Swedish legislature has adopted a restrictive approach towards to the possibilities to successfully have an arbitration award declared invalid or set aside based on a challenge”.
Swedish law permits parties that have not their domicile or place of business in Sweden to waive any form of recourse against the award. According to a national report regarding Sweden, during the period 2002-2007, 114 challenges were brought before the Svea Court of Appeal. Out of these 114 cases, only 33 resulted in a final decision by the Court and out of these only two arbitral awards were set aside.
In England, under the Arbitration Act 1996, arbitration proceedings or an award may be reviewed on the ground of lack of jurisdiction, procedural impropriety or public policy. A party objecting to the substantive jurisdiction of a tribunal may choose to take no part in the relevant proceedings; in that case he can [under Section 72(1)] raise the question of substantive jurisdiction in court at once by seeking a declaration or injunction or can [under Section 72(2)] await any award and challenge it then in court [under Section 67]. In the latter case, the party must raise the objection with the tribunal at the earliest opportunity. Whatever course of events is taken, no decision by the arbitral tribunal as to its own jurisdiction binds the court. English law does not in this respect recognise a principle of Kompetenz-kompetenz and courts may determine for themselves factual and legal issues arising in connection with the jurisdiction of the arbitral tribunal.
Sec. 68(2) of the English 1996 Act provides a catalogue of the principle grounds for serious procedural irregularity justifying setting an award aside; these include failing to deal with all the issues presented to the tribunal, failing to conduct the arbitral proceedings in accordance with procedure agreed by the parties, failure to render an award free from ambiguity or to conduct the proceedings fairly and equitably giving each party a reasonable opportunity of putting his case and dealing with that of his opponent, or the tribunal exceeding its powers or the award being obtained by fraud or the award or the way in which it was procured being contrary to public policy. These procedural defects will only warrant setting an award aside if they caused “substantial injustice” to the party challenging the award. The threshold for an award to be annulled for serious irregularity is very high. Perhaps the most notable ground for review under the Arbitration Act is Section 69, which provides for an appeal on a point of English law in an award made during the proceedings, which is considered a very expansive court review. Section 69 is not mandatory and may be opted out of by the parties. In England, only a limited number of annulment proceedings have been successful. This is due also to a remarkable aspect of English judicial system for arbitration which permits to save awards by the remedy of remission whereby if there is something wrong with the award or the arbitral process it may be corrected by the arbitrators in the light of the court’s determination under the procedure of Sect. 69(7)(c) of the 1996 Act.
In summary, while English courts’ scope of review is expansive regarding jurisdiction of an arbitration tribunal they intervene only cautiously and rarely on grounds of procedural injustice, public policy or error in relation to the merits of any issue submitted to arbitration and they will give considerable weight to the determination by the arbitral tribunal of any issues of fact or foreign law relevant to these latter issues.
In the United States, the Federal Arbitration Act does not provide for a substantive review of the merits of an arbitral award. The Act goes so far as to prohibit parties from agreeing on a review of the merits of an award by the courts.
As one US court put it, using a language widely applicable in other jurisdictions,
“In reviewing an arbitration award courts do not sit to hear claims of factual or legal error by an arbitrator as an appellate court does in reviewing decisions of lower courts”.
The standard of review under the Federal Arbitration Act is extremely limited. The Supreme Court in its landmark 2008 Hall Street decision referred to the FAA
“as substantiating a national policy favouring arbitration with just the limited review needed to maintain arbitration’s essential virtue of resolving disputes straightaway. Any other reading opens the door to the full-bore legal and evidentiary appeals that “can render informal arbitration merely a prelude to a more cumbersome and time consuming judicial review process”.
In Switzerland, case law holds that the list of grounds to challenge an award under Article 190(2) PILA is exhaustive and that each ground must be interpreted narrowly. An application to challenge an international arbitral award must be brought directly before the Swiss Federal Tribunal, which is the highest court in Switzerland.
The Federal Tribunal will not review the facts of the case, nor the law applied by the arbitral tribunal. Broadly speaking, only serious procedural defects or ruling on substance that are contrary to international public policy will be considered. Specifically, it has been consistently held that an award will not be set aside on the grounds that it has been rendered on the basis of obviously wrong findings of fact – even if these findings are contradicted by clear evidence on record – or in clear violation of rules of law or equity. As held by an authoritative commentator, the reason for this very restrictive approach both by the Swiss legislator and by the Federal Tribunal is that parties, having agreed to arbitration, should be held to that agreement and should not be afforded a second opportunity to re-argue the merits of the case in court.
In France, awards are presumed to be valid. In order to rebut this presumption, a party seeking its setting aside has to prove that the award does not satisfy the requirements of Article 1520 of the Noveau Code de Procedure Civile as amended by the reform of 13 January 2011, which alone provides grounds upon which to set aside the award:
1. The arbitral tribunal has wrongly declared to be competent or not to be competent;
2. The arbitral tribunal has been irregularly constituted (which includes the requirements of independence and impartiality of the arbitrators);
3. The arbitral tribunal has decided without confirming to the mission that had been given to it;
4. The principle of contradiction has not been respected;
5. Recognition or enforcement of the award is contrary to international public order.
The party challenging the award must have objected during the arbitration procedure, as soon as it became aware of the irregularity which it is contesting before the national judge.
An international arbitral tribunal is not considered in France to be a court of first instance; the merits of the dispute are not to be examined again by the courts reviewing the award. There is therefore no substantive review, French courts being not supposed to determine whether the arbitrators’ decision was right or wrong.
As the Paris Court of Appeal has repeatedly stated:
“the scrutiny of the court, which excludes (...) any power [of the court] to conduct a substantive review of the arbitral decision, should not bear upon the arbitrators’ assessment of each party’s rights as the award can only be set aside if its enforcement is contrary to public policy”.
A major innovation has been introduced in France by the 2011 reform giving the parties the option under Article 1522 to exclude by “special agreement” all grounds of annulment of an award. The need of a special agreement (“par convention special”) seems to exclude that a waiver of general nature as provided by some arbitration rules may be sufficient. If the waiver is validly raised, the only court control of the award shall be if one of the parties requests the recognition or enforcement in France of the award.
All the mentioned jurisdictions provide for annulment of arbitral awards on public policy grounds, France, England, Switzerland by express law provisions and the United States by holding of the courts, although the scope of such ground for annulment under these jurisdictions is extremely narrow. In France, the Court of Appeal in 2004 in the well-known Thalès Air Défense v. GIE Euromissile case in the context of an action for annulment held as follows:
“the international public policy defense under Article 1502-5 NCPC can be accepted only when the execution of the award would violate in an unacceptable way the French legal order, such violation having to affect in a manifest manner an essential rule of law or a principle of fundamental importance”.
The Court of Appeal noted that, in principle, EU competition law express a fundamental public policy of the French legal system. In dismissing the recourse for annulment, the French court indicated the limits within which the public policy exception may be raised in an annulment context referring to the need for public policy violation to be “flagrante, effective et concrete”, (“blatant, effective and concrete”), conditions that are not reproduced by [Article 1520(5) of] French Code revision of 2011 when dealing with international public policy exception.
In England, in its July 3, 2008 judgement in R v. V the Commercial Court addressed the issue of the scope of review on grounds of public policy. The court held that in any case, it would:
“accord the award full faith and credit, even if it were appropriate to embark on any form of preliminary inquiry”
since in substance there was plenty of material before the arbitrators to demonstrate that the contract was not illegal.
Coming to the UNCITRAL Model Law, the latter set forth an influential approach to the annulment of international awards in the arbitral seat. The Model Law was adopted by the UN General Assembly Resolution of December 11, 1985. 67 States have so far adopted the Model Law. Article 34 of the Model Law provides a detailed list of grounds, divided into two categories, for “recourse to a court against an arbitral award”. As already mentioned, these grounds parallel those set forth by Article V of the New York Convention for refusing recognition and enforcement of foreign awards. Under Article 34(2)(a) an award may be annulled if the party making the application furnishes proof that (i) the arbitration agreement was invalid or a party thereto lacked capacity; or (ii) a party was unable to present its case, including for lack of due notice; or (iii) the award deals with matters outside the scope of the submission to arbitration; or (iv) the composition of the arbitral tribunal was not in accordance with the parties’ arbitration agreement. The award may be annulled under Article 34(2)(b) if the courts finds that (i) the dispute was non-arbitrable; or (ii) the award violates local public policy. If none of these specified grounds is present, then the award may not be annulled, Article 34(2) providing for the setting aside of the award “only if” one of the listed grounds is accepted. The burden of proving that one of the grounds under Article 34(2)(a) applies is on the party seeking to set aside the award, as confirmed by decisions of national courts. National courts have further indicated that Article 34 grounds for annulment are exhaustively enumerated in Article 34 and are to be narrowly construed.
I will limit my comments to the public policy grounds of annulment under the Model Law.
Article 34(2)(b)(ii) provides that an award may be annulled if the annulment court finds that: “the award is in conflict with the public policy of this State”. The public policy exception is frequently invoked in the Model Law jurisdictions as a basis for annulling arbitral awards and gives rise to substantial complexities similar to those arising in connection with the recognition of foreign arbitral awards under Article V(2)(b) of the New York Convention providing that this ground for refusing recognition or enforcement may be applied ex officio by the court. The standard of proof required to establish a public policy exception under this Article is a demanding one, courts having uniformly held that this ground is “exceptional” and “extremely narrow”, to be applied “sparingly” and with “extreme caution”, and to be interpreted “restrictively”, limiting annulment in exceptional cases of clear violations of fundamental, mandatory legal rules, not in cases of judicial disagreement with a tribunal’s substantive decisions or procedural rulings. The rationale is that only matters which are essential to the forum state’s legal system, and are considered mandatory even in transnational settings, will constitute international public policy.
Corruption and official bribery, breach of trade regulations or export and currency controls, penalties and punitive damages provided by the award or serious procedural irregularities amounting to violation of “procedural public policy” may be grounds for public policy exceptions.
Coming to the ICSID Convention, it is to be noted the unique feature of the ICSID system consisting in its autonomous nature. ICSID arbitration is known as self-contained, or delocalized, arbitration because local courts in any particular State, including those at the place when arbitration is held, have no role to play in the proceeding. ICSID awards are binding on the disputing parties, may not be appealed, and are not subject to any remedies except those provided for in the Convention. As a result, unlike other international arbitral awards, ICSID awards cannot be challenged before national courts but only within the framework of the Convention and pursuant to its provisions.
The drafting history of the ICSID Convention confirms the limited and exceptional nature of the annulment remedy, the finality of awards being a fundamental goal for the system. Annulment was designed purposefully to confer a limited scope of review which would safeguard against “violation of the fundamental principles of law governing the Tribunal’s proceedings”. The remedy has thus been characterised as one concerning the legitimacy of the process rather than an inquiry into the substance of the award.
Article 53 provides that “the award shall be binding on the parties and shall not be subject to any other remedy except those provided for in this Convention”.
The choice of remedies offered by the ICSID Convention reflects a deliberate election by the drafters of the Convention to ensure finality of awards. These remedies are:
1. rectification (Article 49) – the Tribunal can rectify any clerical, arithmetical, or similar error in its award;
2. supplementary decision (Article 49) – the Tribunal may decide any question it omitted to decide in its award;
3. interpretation (Article 50) – the Tribunal may interpret its award where there is a dispute between the parties as to the meaning or scope of the award rendered;
4. revision (Article 51) – the Tribunal may revise its award on the basis of a newly discovered fact of such a nature as to decisively affect the award; and
5. annulment (Article 52) – an ad hoc Committee may fully or partially annul an award on the basis of one of the grounds under Article 52(1).
Under Article 52(1) of the ICSID Convention, the grounds for annulment are as follows:
1. that the Tribunal was not properly constituted;
2. that the Tribunal has manifestly exceeded its powers;
3. that there was corruption on the part of a member of the Tribunal;
4. that there has been a serious departure from a fundamental rule of procedure; or
5. that the award has failed to state the reasons on which it is based.
The circumstance that the excess of powers must be “manifest” and that the departure from a rule of procedure must be “serious” and must concern a rule that is “fundamental” is indicative of the object and purpose of the Convention to provide for a very high threshold for annulment. Out of these different grounds, the Tribunal’s manifest excess of powers, a serious departure from a fundamental rule of procedure and the failure to state the reasons on which the award is based have been invoked in practically all requests for annulment by either the investors or States, with more than one of such grounds being relied upon in each case.
Decisions on annulment have established that:
1. the grounds listed in Article 52(1) of the Convention are the only grounds on which an award may be annulled;
2. annulment is an exceptional and narrowly circumscribed remedy and the role of annulment committees is limited;
3. these committees are not court of appeal, annulment is not a remedy against an incorrect decision, and annulment committees cannot substitute the tribunal’s determination on the merits for their own;
4. annulment committees should exercise their discretion not to defeat the object and purpose of the remedy or erode the binding force and finality of awards;
5. Article 52 should be interpreted in accordance with its object and purpose, neither narrowly nor broadly; and
6. committee’s authority to annul is circumscribed by the grounds specified in the request for annulment, but committees have discretion with respect to the extent of an annulment, which may be partial or full.
According to the more than forty annulment decisions rendered until now the following has been progressively held:
The excess of powers occurs where the tribunal has not respected the parties’ arbitration agreement by accepting jurisdiction where in fact jurisdiction is lacking, or in the inverse case by declining jurisdiction when it existed. The ICSID Convention prescribes certain mandatory requirements that must be fulfilled for a Tribunal to have jurisdiction. These jurisdictional requirements require: (i) a legal dispute; (ii) arising directly out of an investment (iii) between a Contracting State (or any constituent subdivision or agency of a Contracting State designated to the Centre by that State); (iv) and a national of another Contracting State; (v) which the parties to the dispute consent in writing to submit to the Centre. The parties cannot agree to derogate from these criteria and the Tribunal must decline jurisdiction where a mandatory is not met, even if neither party has raised any objection to jurisdiction. Excess of powers has been found in case of failure by the tribunal to exercise its mandate in accordance with the parties’ arbitration agreement. On that basis, it has been held that there is an excess of powers in case of failure by the tribunal to apply the proper law (as distinguished from errors in the application of the law which in principle is no ground for annulment) or if the tribunal acted ex aequo et bono without the required agreement of the parties. The excess of powers must in any case be “manifest” in order to found annulment, this being the case if it can be discerned without the need for an elaborate analysis of the award.
The serious departure from a fundamental rule of procedure requires that the rule in question be “fundamental” and the procedural violation be “serious”. As held by annulment committees, fundamental rules of procedure include the equal treatment of the parties, the right to be heard, an independent and impartial tribunal, the treatment of evidence and burden of proof and deliberations among the tribunal’s members. Not every departure from a fundamental rule of procedure justifies annulment: the departure must be “serious”, meaning that it must have had a material effect on the tribunal’s decision
“causing the tribunal to reach a result substantially different from what it would have awarded had such rule been observed”.
Failure to state reasons on which the award is based is a requirement ensuring that the parties can understand the facts and the law applied by the Tribunal in reaching its conclusion. The correctness of the reasoning or whether it is convincing is not relevant. Failure to state the reasons may occur in case of absence of reasons on a particular aspect of the award which is material to the tribunal’s decision; or of contradictory reasons. Failure to state reasons does not necessarily result in annulment in case the failure is remedied by the tribunal issuing a supplementary decision concerning the question not addressed upon the request of the dissatisfied party (Convention, Article 49(2)).
The more recent steady trend of rejection of annulment requests may favour users’ perception of the stability reached by ICSID review system although it is clear that the Convention’s goal of assuring the finality of awards does not mean that awards should be left annulled, infringements of ICSID Convention’s basic tenets being in no case to be tolerated. Prior annulment decisions, although offering useful guidance to subsequent committees, do not have precedential value, a stare decisis as meant for municipal jurisprudence. By contrast, annulment decisions of non-ICSID awards, entrusted as they are to a body of national judges of a higher level of jurisdiction, sometime with a specific knowledge and experience in the field of international arbitration (as in the case of the Paris Court of Appeal), are expected to follow the precedents established by other courts of the same judiciary, thus ensuring a higher level of consistency and predictability of the national jurisprudence with respect to ICSID awards review system.
I conclude by mentioning that the review made of national laws and an international Convention of great importance for arbitration in the field of investment, together with the relevant jurisprudence, short as time constraints and your patience has made possible, confirms that finality of awards is one of the main objectives of the generality of legal systems of arbitration.
I thank you for your attention.