The underlying idea

The idea behind bifurcation is to split the arguments in a case into two phases when the decision on one aspect would terminate the proceeding and thus save further time and costs. This, of course, is key to any consideration on bifurcation. If the decision on the first phase does not terminate the proceeding, then the cost-saving effect turns into its opposite, meaning that such a decision should not be taken lightly and only after having assessed the pros and cons carefully.

Availability of bifurcation on the applicable legal regime

The starting point is «whether bifurcation is available under the applicable legal regime». Bifurcation is provided for in some, but not all the major arbitral rules used by commercial and state parties. Mostly, however, arbitration rules do not explicitly address bifurcation, but rather set out a general concept to be used by the tribunal which allows the bifurcation of the proceeding. For example, The Swiss Rules of International Arbitration 2021 provide in their article 34 that in addition to making a final award, the arbitral tribunal may make «interim or partial awards». The International Chamber of Commerce (ICC) Rules of Arbitration 2021 simply provide in their Appendix IV a) that the tribunal «may bifurcate» the proceedings, or issue a partial award, when doing so may genuinely be expected to result in a more efficient resolution of the case. Similarly, the tribunal has wide discretion with respect to bifurcation under the Rules of the London Court of International Arbitration (LCIA), which provide that the tribunal may determine the stage of the arbitration at which any issue or issues shall be determined, and in what order (Article 14.6 (iv)). For arbitrations having their seat in Switzerland, art. 188 of the Swiss Private International Act (SPILA) stipulates that unless the parties have agreed otherwise, the arbitral tribunal may render «partial awards», allowing it to bifurcate the proceeding.

Typical situations where bifurcation might be considered

In international arbitrations, parties typically choose to bifurcate and thus split the proceeding either into (1) a decision on «jurisdiction and one on the subject-matter»: Only if jurisdiction is affirmed, the arbitral tribunal decides on liability and quantum». (2) Furthermore, bifurcation often happens between «liability and quantum in damages claim situations». Although typically the tribunal would first decide on liability and then hear the parties on the quantum, it is also possible to hear the parties on damages first. The option might be appropriate if the decision as to whether a party had suffered loss or was otherwise entitled to damages is not intrinsically intertwined with the question of liability. Or (3) the arbitral tribunal first decides whether the principal «owes the agent a commission», and in the affirmative, it decides on the amount of the commission. (4) Another setting where we see bifurcation is when the arbitral tribunal first decides on the question of «whether there is a contract» (or: the claimed deliverables fall within the scope of the contract), and if yes, on the question of defects and damages. And of course, (5) if «statute of limitation defences» are raised, the parties and the arbitral tribunal may choose to argue and decide on these first.

Types of awards rendered in bifurcated proceedings

Under Swiss law, we generally distinguish between two kind of awards in bifurcated proceedings: According to the dominant doctrine, an interim award is limited to the clarification of a substantive prejudicial position (e.g. affirmation or denial of the liability of a party on the merits «without simultaneously deciding on the quantitative elements of the concrete liability», as expressly stated in the below quoted commentary), whereas the preliminary decision deals with a procedural preliminary issue (e.g. rejection of the plea of lack of jurisdiction). The arbitral tribunal may issue partial or interim awards, but is in principle not obliged to do so.

What considerations should be made when deciding on applying for, and deciding on, bifurcation?

As we have seen above, bifurcation will not always be advisable and will need to be determined on a case-by-case basis. The litmus test there is whether bifurcation in the case at hand may genuinely be expected to result in a more cost- and time-efficient resolution of the case. This may be so, e.g., when the arguments pertaining to, and the quantification of, «damages are particularly complex» and require substantive submissions and expert reports. By dealing with the question of liability first, a party may not be required to incur costs on extensive submissions and quantum expert reports if a claim (or claims) are dismissed on liability. Bifurcation may also work the other way around, namely when the question whether claimant is entitled to damages is a very limited one. This was the case in one of our recent arbitrations where claimant based its claim on the third-party liquidation theory («Drittschadenstheorie»). After having bifurcated the proceeding, the arbitral tribunal decided that in the case at hand, the purchaser of transhipping cranes was not entitled to claim damages suffered by the leasee of the transhipping barge, and the case stopped there. Potentially high expert report and expert witness costs as well as disclosure issues («why should we substantiate the damages in a damage claim or open our books for inspection in an agency compensation claim case when the claim maybe will be dismissed on liability grounds?») are usually drivers for bifurcation. Bifurcation may also be appropriate if damages may be easily segregated from the liability issues in dispute. However, bifurcation may be less advisable if damages are closely linked to liability and both pertain to intrinsically entwined factual elements, or the same witnesses would need to testify equally on liability and quantum. 

The same considerations appliy for document production: For time efficiency reasons, if document production in the first (bifurcated) stage is an issue, one might think of including documents which would be relevant in a second stage, at least if these are of limited extent.

Bifurcation is regularly considered if the decision on the first aspect (such as liability, the entitlement to a commission payment, or the validity of the contract) could considerably increase the likelihood of an early settlement of all or at least a substantial part of the claims.  

Fundamentally, it will be for the parties and the tribunal to consider whether bifurcation is appropriate for any particular case but the risk of slowing down proceedings rather than creating expediency (and thus saving costs) should be carefully considered prior to making any application. Let us for this reason look at some typical «bifurcation traps» we meet in practice.

Typical «bifurcation traps»

For efficiency purposes, when «purely» procedural and legal issues are concerned, bifurcation may be an easy choice and promising. In practice though, such easy choices are rare. They may apply if in the first phase, «solely contract interpretation issues» would be involved, such as the legal construction of the (accepted) arbitration clause, whether a certain legal concept (such as in the third-party liquidation, case mentioned above) arise, or when statute of limitation defences are raised.

However, even in situations where bifurcation may seem an obvious choice, a thorough analysis of the case at hand is advisable: A party may later argue that the contract was amended or that at the time of the conclusion of the contract, the parties had a certain common understanding of the contract clause in dispute. Then, witnesses might have to be heard on the entire factual setting of the case anyway, and invariably the question arises «whether the potentially saved costs outweigh the risk of extra costs in case the bifurcated proceeding continues and a second hearing needs to be held». Such elements need to be taken into account before deciding on bifurcation.

This said, it doesn’t come as a surprise that in practice, the segregation of the two phases is often tricky and entail considerable risks: First, to have the benefit of a potential reduction of cost, bifurcation applications should be filed «early in the proceeding». However, at an early stage of the proceeding, it might not yet be clear whether bifurcating certain aspects of the case is appropriate in the first place. Second, for the same reason, the delineation of the issues to be decided in the first phase might not be sufficiently clear and in the second phase, the question may arise whether a certain aspect of the case was not already part of the first phase. It may not always be obvious, e.g., what elements relevant to the quantum discussion do not already need to be argued, substantiated or even proven in the liability phase of the arbitration. But there is more: If a certain element has been argued in the liability stage and liability is admitted, if needed, may this argument, or other aspects of the same factual element, be argued more extensively or even anew in the quantum stage? This is relevant because the award which the arbitral tribunal renders at the end of the first phase has «res iudicata effect» and the question easily raises «right to be heard concerns» and entails setting-aside risks (see, e.g., art. 190(2)(d) SPILA). In this context, it is to be noted that under the Swiss «lex fori», the «res judicata» effect only applies to the operative part of the award and not its reasoning.  

Furthermore, we have seen cases where the bifurcation is requested by Respondent only when Claimant has already filed a full-fletched statement of claim (the need for bifurcation is often not clear at the time of the case management conference, in particular if the notice for arbitration and the answer thereto are of very limited extent and in particular Respondent’s counsel did not have the time to thoroughly examine the case yet). Respondent then had to file a full-fletched statement of defence containing the application for bifurcation. In such cases, the «liquidation of the costs» in case the claim is dismissed can be tricky. The situation becomes even more unsatisfactory when the arbitral tribunal holds that there is liability, but then Claimant predominantly or altogether loses in the quantum stage. Respondent then might have to pay high legal costs for losing phase 1 although (partly, or even predominantly) winning the case in the second phase.

Another consideration when deciding on bifurcation concerns the «availability of witnesses» for the proceeding for the second stage; a party representative (witness) is all of the sudden an unrelated witness not under our control, and our contact may be limited.

Somme recommendations

Because bifurcation is a two-edged sword and entails complex questions, addressing it early in the proceeding is key to ensuring its time and cost saving effects.

This is all the more true in damages claim cases because of «res judicata» issues.

For «clarity of delineation» purposes, some authors advocate to draw up a damages (or other) protocol where issues such as the type of damages claimed; whether the parties will be addressing mitigation of loss; any application of contractual limitation clauses (or limitation under law); the format of any expert submissions, including any methodology and underlying calculations to explain the ultimate quantum claimed (or disputed); a good faith estimate of any damages claimed; and the guidelines for presentation of expert evidence at the hearing (see, e.g., Louise Woods, Ciara Ros, Elena Guillet and Edward Wilson, «Procedural Issues», GAR online publication, 9 June 2021). Although not questioning the wisdom of such protocols in general, in this author’s view, the complexity of drawing them up and the uncertainty regarding the further development of the case poses multiple problems and risks, and at the end of the day, the parties end up struggling more with procedural (delineation and «res iudicata») issues than with the facts themselves. It may therefore be advisable to agree in the terms of reference that in stage two (if any) the parties are allowed to bring up «not the identical, but only related» issues (if needed) as in stage one. Absent such rule, the above legal considerations apply.   

Furthermore, because of the limitations of the «res iudicata» effect (which, as we have seen, only pertains to the operative part and not the grounds of the award), parties may consider covering some aspects of the case by requesting the arbitral tribunal to issue a «declaratory or negative declaratory decision», i.e. a decision that asserts the existence or the non-existence or exclusion of certain facts («Feststellungsklage» and «negative Feststellungsklage»).

Finally, the parties and the arbitral tribunal should give due consideration to the cost issue and carefully evaluate and regulate in the terms of reference whether the cost of the arbitration will be determined by stage or exclusively at the end of the arbitration.

Conclusion

Yes, bifurcation may have cost- and time-saving effects. Its benefits become often more uncertain, though, when a detailed analysis of the pros and cons in the case at hand is made. Unless it may «genuinely be expected» to result in a more cost- and time-efficient resolution of the case, in particular because the issues for the two stages can be delineated with high certainty and the chances of terminating the arbitration after the first stage are high, bifurcation might not be the first choice. If bifurcation is agreed, it is imperative to carefully and clearly determine what issues must be argued and substantiated in what phase. Finally, if the parties and the arbitral tribunal decide in favour of bifurcation, it should be discussed how and when the costs of stage one will be decided and distributed if the arbitration continues.