Despite the common use of bifurcation in arbitration proceedings, there are many arguments against the bifurcation of most arbitral proceedings.
Bifurcation refers to the separation of issues, typically dividing the proceedings into jurisdictional or procedural phases and the merits phase, or merits and quantum phases, allowing the arbitral tribunal to address and decide on specific issues sequentially rather than concurrently. Arbitration proceedings may also be “trifurcated”, such as being separated into separate jurisdictional, merits and quantum phases.
Although bifurcation was conceived to promote cost-efficiency and time effectiveness and, hence, to ensure procedural economy by separating issues for preliminary or independent determination into a separate phase of proceedings, many scholars have advocated against it. Bifurcation rarely promotes efficiency in arbitration proceedings. In fact, it implies a number of risks that will be analysed below.
The primary arguments against bifurcation include:
- The increased time and costs of the arbitration;
- Difficulties to strictly separate issues between different phases of the proceedings;
- A lack of objectivity – or the appearance thereof – of the arbitral tribunal, should jurisdiction be found;
- Questions regarding the nature of the decision allowing for bifurcation and possible challenges arising from it.
Increased Time and Costs in the Event of Bifurcation
The primary argument against bifurcation concerns the fact that it only enhances procedural economy to the extent that the objection raised is successful and disposes of the entirety of the case.[1] In other situations – the lack of a truly dispositive issue – bifurcation will generally have the opposite effect and result in the increased time and costs of the proceedings.[2]
Furthermore, abusive bifurcation requests are often made whose only purpose is precisely to delay the course of the arbitral proceedings.[3] If the arbitration continues after bifurcation, this will necessarily increase the overall time and costs of the proceedings.
In 2011, Lucy Greenwood published an analysis of available empirical data on the bifurcation of ICSID arbitrations in the Journal of International Arbitration. The study unsurprisingly demonstrated that, on average, bifurcated proceedings took longer to conclude than non-bifurcated ones:[4]
- bifurcated cases took an average of 3.62 years to conclude with a final award;
- non-bifurcated cases reached a final award and required, on average, 3.04 years to conclude.
Between 2016 and 2018, Lucy Greenwood renewed her analysis and similar findings were reached:[5]
- 38 ICSID cases resulting in a final award were reviewed;
- The cases that were bifurcated took an average of 4 years, 3 months, to conclude;
- Non-bifurcated cases took, on average, 3 years, 2 months, to reach a final award.
This data was also corroborated by the ICSID Working Group in relation to the proposed amendments of the ICSID Rules published in August 2018.[6]
Moreover, little time was saved even in the event of a jurisdictional objection being upheld:[7]
- The cases where the jurisdictional objection was upheld required an average of 2 years, 4 months to reach an award;
- Bifurcated cases where the jurisdictional challenge failed required an average of 5 years, 2 months, to conclude.
These analyses demonstrate the considerable impact that an “unsuccessful” jurisdictional challenge can have on the overall length of arbitral proceedings. Furthermore, the efficiency of the bifurcation process is, by definition, only confirmed at the end of its exercise.[8]
Holding separate hearings for each phase can lead to the duplication of efforts and additional preparation and legal fees. Parties often end up paying for two or more sets of hearings, additional witness testimonies, and additional expert reports. Thus, it is not only time that is lost but also costs.
Complexity of the Dispute
Another of the arguments against bifurcation resides in the fact that assessing the chances of success of bifurcation at the very beginning of the case can prove to be an extremely burdensome exercise for the arbitral tribunal, as the case may be very different after twelve months of proceedings.[9]
With regard to requests for bifurcation based on jurisdiction, the tribunal may struggle to separate preliminary request from the merits, as they may be inextricably linked (see, e.g., Glamis Gold v. USA or Gavrilovic v. Croatia, P.O. No. 2, 31 May 2005, para. 25).
As Professor Schreuer notes, “[S]ome jurisdictional questions are so intimately linked to the merits of the case that it is impossible to dispose of them in preliminary form.”[10]
Bifurcating in terms of liability and damages is also problematic: the issues of liability and damages are often so intertwined that separating them may not provide any clarity. Instead, it may simply make the proceedings more complex. A single, comprehensive hearing often allows the arbitrator(s) to have a full understanding of the case, which may lead to a better decision.
In any event, the arbitral tribunal should by all means avoid prejudging the case or stifling a genuine claim by the early determination of a disputed matter in breach of due process.[11]
CBI and Kentz (CJKV) v. Chevron Australia is an example of bifurcated proceedings where the tribunal separated the liability phase from the damages phase. It clearly illustrates the difficulties of bifurcation, as the tribunal, after having considered the issue of liability by a first partial award, calculated damages while reconsidering the basis on which payment was to be made to the claimant.[12] This issue would have been readily avoided by not bifurcating.
Perceived Lack of Objectivity of the Arbitral Tribunal
Doctrine raises awareness regarding the issue of having one and the same arbitral tribunal decide on jurisdiction and the merits phase of the proceedings, as this can create doubt as to the objectivity of the tribunal. If arbitrators decide to render a negative decision on jurisdiction, this necessarily would adversely affect the arbitrators’ overall fees, which are not insignificant. The temptation may thus be for the tribunal to allow a case to proceed on the merits when bifurcated on jurisdictional grounds.[13]
Although this is only an apparent lack of objectivity, it does not contribute to enhancing trust in the arbitral process. Although scholars have suggested the systematic change of the members of the arbitral tribunal once the jurisdictional phase has been decided[14], this does not decrease time or costs, but quite the opposite. Again, this militates in favour of not bifurcating unless there is a very high probability of the case being disposed of at the jurisdictional phase.
Uncertain Nature of the Decision on Bifurcation
Among the arguments against bifurcation is also the uncertain nature of the decision rendered by an arbitral tribunal on preliminary matters. Separate phases can result in inconsistent or incompatible findings. This implies determining whether the preliminary award has res judicata effects or not. A positive answer would prohibit the reconsideration of such decisions prior to the rendering of a final award. However, the arbitral tribunal may be tempted to revise its findings in light of new evidence that was not previously available.
Moreover, ICSID proceedings and commercial arbitration obey different regimes:
- In principle, partial awards in commercial arbitration can be immediately challenged/annulled (based on the applicable law);[15]
- In ICSID proceedings, there is no express basis for reconsidering a partial award.[16]
Applicants have attempted to obtain the reconsideration of ICSID partial awards on several grounds under the ICSID Convention, including:
- Article 44: the power of the tribunal to decide upon any question of procedure not covered by the Convention, the ICSID Rules or any other rules agreed by the parties;
- Article 49: omitted questions, the rectification of clerical and mathematical errors;
- Article 51: revision of the award subject to the existence of a new decisive fact unknown by the parties and the tribunal prior to the final award without negligence on their part.
The effects of a partial award, however, may still be uncertain, given arbitral case law.
In ConocoPhillips v. Venezuela, the tribunal rejected the application of Article 44 of the ICSID Convention and held: “Those decisions [partial awards] in accordance with practice are to be incorporated in the Award. It is established as a matter of principle and practice that such decisions that resolve points in dispute between the Parties have res judicata effect. ‘They are intended to be final and not to be revisited by the Parties or the Tribunal in any later phase of their arbitration proceedings.’”
Perenco v. Ecuador concurred and rejected all grounds invoked by the respondent for the reconsideration of an interlocutory decision.
More recently, however, the arbitral tribunal in Burlington v. Ecuador considered that interlocutory decisions do not have res judicata effects and allowed reconsideration of a partial award under Article 51 of the ICSID Convention.[17]
Conclusion
Bifurcation is in fact a double-edged sword that is used far too commonly: if tribunals get it absolutely right, proceedings will be more efficient; if they get it wrong, procedural economy will be sacrificed.
A decision to bifurcate the proceedings when this was not necessary can nearly double the amount of time needed until the rendering of the final award. Yet, arbitral tribunals typically have little knowledge regarding a case when the decision on bifurcation must be made.
Moreover, the regime of the partial award, at least in ICSID proceedings, remains uncertain. Potential challenges thereon can further delay the issue of a final award and increase costs.
Hence, there should be a strong presumption against bifurcation, absent a high probability that bifurcation will result in the early termination of the proceedings.
[1] L. Greenwood, “Revisiting Bifurcation and Efficiency in International Arbitration Proceedings”, in Maxi Scherer (ed), Journal of International Arbitration, Kluwer Law International 2019, Volume 36 Issue 4, p. 422.
[2] L. Greenwood, “Revisiting Bifurcation and Efficiency in International Arbitration Proceedings”, in Maxi Scherer (ed), Journal of International Arbitration, Kluwer Law International 2019, Volume 36 Issue 4, p. 425.
[3] O. Caprasse, M. Papadatou, « De la bifurcation en arbitrage commercial et en arbitrage CIRDI », Revue de l’Arbitrage, Comité Français de l’Arbitrage 2022, Volume 2022 Issue 2, p. 550; L. Greenwood, “Does Bifurcation Really Promote Efficiency?”, Journal of International Arbitration, Kluwer Law International 2011, Volume 28 Issue 2, p. 108.
[4] L. Greenwood, “Does Bifurcation Really Promote Efficiency?”, Journal of International Arbitration, Kluwer Law International 2011, Volume 28 Issue 2, p. 107.
[5] L. Greenwood, “Revisiting Bifurcation and Efficiency in International Arbitration Proceedings”, in Maxi Scherer (ed), Journal of International Arbitration, Kluwer Law International 2019, Volume 36 Issue 4, p. 424.
[6] L. Greenwood, “Revisiting Bifurcation and Efficiency in International Arbitration Proceedings”, in Maxi Scherer (ed), Journal of International Arbitration, Kluwer Law International 2019, Volume 36 Issue 4, p. 424.
[7] L. Greenwood, “Revisiting Bifurcation and Efficiency in International Arbitration Proceedings”, in Maxi Scherer (ed), Journal of International Arbitration, Kluwer Law International 2019, Volume 36 Issue 4, p. 424.
[8] O. Caprasse, M. Papadatou, « De la bifurcation en arbitrage commercial et en arbitrage CIRDI », Revue de l’Arbitrage, Comité Français de l’Arbitrage 2022, Volume 2022 Issue 2, p. 550
[9] L. Greenwood, “Does Bifurcation Really Promote Efficiency?”, Journal of International Arbitration, Kluwer Law International 2011, Volume 28 Issue 2, p. 110.
[10] C. Schreuer, “The ICSID Convention: A Commentary” (2d ed.), Cambridge University Press 2009, p. 537.
[11] M. Benedettelli, “To Bifurcate or Not To Bifurcate? That is the (Ambiguous) Question”, in William W. Park (ed), Arbitration International, Oxford University Press 2013, Volume 29 Issue 3, p. 504.
[12] J. Browne, “Perils of bifurcated hearings”, Commercial Litigation Blog, 3 February 2023.
[13] G. Meijer, “Dynamic Bifurcation and the Objectivity of Arbitrators”, in S. Brekoulakis (ed), Arbitration: The International Journal of Arbitration, Mediation and Dispute Management, Chartered Institute of Arbitrators (CIArb), Sweet & Maxwell 2019, Volume 85 Issue 1, p. 79.
[14] G. Meijer, “Dynamic Bifurcation and the Objectivity of Arbitrators”, in S. Brekoulakis (ed), Arbitration: The International Journal of Arbitration, Mediation and Dispute Management, Chartered Institute of Arbitrators (CIArb), Sweet & Maxwell 2019, Volume 85 Issue 1, p. 81.
[15] O. Caprasse, M. Papadatou, « De la bifurcation en arbitrage commercial et en arbitrage CIRDI », Revue de l’Arbitrage, Comité Français de l’Arbitrage 2022, Volume 2022 Issue 2, p. 575.
[16] O. Caprasse, M. Papadatou, « De la bifurcation en arbitrage commercial et en arbitrage CIRDI », Revue de l’Arbitrage, Comité Français de l’Arbitrage 2022, Volume 2022 Issue 2, p. 576.
[17] See also Standard Chartered Bank v. Tanzania, Award, ICSID Case No. ARB/10/20, 12 September 2016; paras. 312-314 and 319-320 (for reconsideration of the partial award on a different ground).
