One of the great advantages of arbitration is that it is a “one-shot” dispute resolution mechanism that does not allow for a series of appeals. Indeed, many users stress the finality of awards and the lack of an appeals mechanism as a valuable characteristic of arbitration. However, there may be situations where the “one shot” misses the mark, leading to a fundamentally unjust award. This may, in particular, be the case if a party is found to have acted fraudulently during the proceedings and to have obtained a favourable award as a result (e.g. when criminal courts subsequently find that a key piece of evidence that influenced the award was forged). More generally, the post-award discovery of a fact which would have impacted on the award always raises the question whether the award should be allowed to stand.
In litigation, the solution to such situations is revision, which is a remedy that has existed in many jurisdictions for a long time. Revision is an extraordinary remedy based on the alleged discovery of new facts and aimed at the amendment of a final and binding decision in light of these new facts, usually by the original adjudicator. In arbitration, the remedy of revision is less developed. With this post, we intend to give an overview of the existing legal framework in national arbitration laws and institutional rules, and to analyse whether it is time for the arbitration community to undertake a renewed effort to regulate the revision of arbitral awards.
The existing framework in national arbitration laws
National arbitration laws take different approaches to the revision of arbitral awards. Jurisdictions that have adopted the UNCITRAL Model Law on International Commercial Arbitration of 1985 (“Model Law“), such as e.g. Germany, Singapore, and Denmark, generally do not have rules on revision. This is due to the fact that the Model Law itself only provides for the setting aside of arbitral awards and does not expressly contemplate revision as an alternative remedy.
Incidentally, the travaux préparatoires of the Model Law show that the lack of a revision mechanism was a deliberate decision. The drafters considered establishing a separate set-aside ground with longer deadlines “for such cases as fraud or false evidence which had materially affected the award” (see here, p. 36, para. 299). However, the drafters ultimately decided against such a regime because of the “need for speedy and final settlement of disputes in international commercial relationships” (see here, p. 36, para. 300).
Given the absence of an express rule under this regime, the revision of an award by the arbitral tribunal is arguably not possible because the tribunal’s mandate is exhausted once the final award is rendered. Moreover, the discovery of a new fact does not necessarily give rise to a ground for setting aside the award either. Nonetheless, this does not necessarily mean that an award will always be enforceable no matter what information is uncovered after the award was rendered. For example, in Germany, if it can be proven that a party obtained an award in a fraudulent way, the award debtor may be entitled to an order prohibiting the award creditor from enforcing the award, or requiring the award creditor to repay what it has already received in satisfaction of the award. In other words, under German law, the award debtor has a substantive right to not be subjected to the enforcement of an award that was obtained in a fraudulent manner.
Conversely, in jurisdictions that expressly provide for the revision of arbitral awards (like Switzerland, the Netherlands, Spain, and France) some broad similarities in the design of the remedy can be observed. Thus, in general, only previously unknown facts which could not have been discovered during the arbitral proceedings can provide grounds for revision (cf. Article 190a(a) Swiss PILA, Article 1068(c) Dutch Code of Civil Procedure, Article 510(i) Spanish Civil Procedure Act (in conjunction with Article 43 of the Spanish Arbitration Act), Sections 1502, 595(2) French Code of Civil Procedure). Moreover, a request for revision can be filed if an award was influenced by criminal acts or by false evidence (cf. Article 190a(b) Swiss PILA, Article 1068(a)-(b) Dutch Code of Civil Procedure, Article 510(ii)-(iv) Spanish Civil Procedure Act (in conjunction with Article 43 of the Spanish Arbitration Act), Sections 1502, 595(1), (3)-(4) French Code of Civil Procedure), a situation which was present e.g. in a 2009 Swiss case that was previously discussed on this blog.
At the same time, however, there are also important differences. Some jurisdictions provide for revision proceedings to be conducted before the original tribunal and only refer the matter to the courts if the original tribunal cannot be reconstituted (cf. Article 1502 of the French Code of Civil Procedure). Other regimes provide for an application to the courts and then allow the courts to remit the award to the original tribunal for reconsideration (or to set the award aside itself) (cf. Section 68(2)(g) and (3)(a) of the English Arbitration Act, in relation to fraud). Conversely, in other jurisdictions, applications are only heard by domestic courts which have the power to revoke the award if the requirements are fulfilled (cf. Article 1068 of the Dutch Code of Civil Procedure, Article 516 Spanish Civil Procedure Act (in conjunction with Article 43 of the Spanish Arbitration Act)).
Also, there are considerable differences insofar as time limits are concerned. While Spanish and Dutch law require an application to be made within three months from the discovery of the ground for revision, French and Swiss law provide for 2‑month and 90‑day time limits, respectively. Moreover, there are also varying absolute time limits. Thus, in Switzerland, a request for revision is generally excluded after 10 years from the date when the award became legally binding, whereas Spain imposes an absolute time limit of 5 years. In yet other jurisdictions, laws are silent on this matter.
The existing framework in arbitration rules
In contrast to domestic law, revision is generally not addressed in the rules of arbitral institutions. Indeed, none of the major institutions for international commercial arbitration (such as the ICC, SIAC, HKIAC, LCIA, SCC or SAC) have included a provision on revision in their rules. That said, one major institution providing for revision in its rules is ICSID. Under Article 51 of the ICSID Convention, an application for the revision of an ICSID award may be brought, “on the ground of discovery of some fact of such a nature as decisively to affect the award, provided that when the award was rendered that fact was unknown to the Tribunal and to the applicant and that the applicant’s ignorance of that fact was not due to negligence“. The applicable time limits are relatively strict, requiring the application to be made within 90 days after the discovery of the fact, and in any event within three years from the date on which the award was rendered. If possible, the decision on revision shall be rendered by the tribunal that rendered the original award.
Should the arbitration community revisit the issue of revision?
The above analysis shows that approaches to revision differ widely. Some jurisdictions do not provide for any set of rules for revision, while others regulate revision in a detailed manner. Where there are no such rules, courts may be required to fill the gap, which can lead to uncertainty (especially if there is not a lot of case law). Furthermore, among those jurisdictions with detailed rules on revision, there are marked differences in approach, such as with regard to the question whether the revision proceedings are to be brought before a court or the original tribunal.
Against this background, one may wonder whether the issue of revision should be revisited in future discussions concerning potential amendments to the Model Law. The drafters’ initial decision to not include a rule on revision so as to favour the finality of arbitral awards certainly made sense at the time. However, experience has shown that there are cases in which new relevant facts are uncovered belatedly. Such cases are exceptional, and cases in which the new evidence could not have been uncovered during the proceedings is even rarer. Nevertheless, rather than leaving it to domestic courts to find a solution for these extraordinary situations, it would ensure greater legal certainty to have a clear set of rules from the outset. Moreover, purposefully formulating these rules in a very narrow fashion would ultimately contribute to the finality of arbitral awards.
Further, one may well wonder whether arbitral institutions should not step in to “fill the gap” by adding express provisions on revision in their rules. However, the general reluctance of (major) arbitral institutions to address revision is understandable if one considers that there is not always a gap to fill because the applicable lex arbitri may indeed provide for a revision regime. And if it does, the domestic rules will often be mandatory, meaning that institutional rules cannot opt out of or modify such rules. Accordingly, the safest option for institutions may indeed be to not address the issue at all.
At the same time, not addressing the issue in institutional rules may leave a lacunae if the lex arbitri provides for revision proceedings before the original tribunal. In particular, this may mean that there are no clear rules on the payment of arbitrator fees for the revision proceedings (e.g. how the fees are calculated, which party pays the advance on fees, etc.). Against this background, institutions should at least consider adding rules that regulate these specific issues in case the lex arbitri allows for the revision of awards by the original tribunal.