Apropos of a recent decision in ConocoPhillips v. Venezuela (ICSID Case No ARB/07/30), this post discusses the potential underlying concerns an arbitral tribunal may consider when deciding whether it can revise earlier decisions within the context of fragmented proceedings.


The ICSID proceedings in ConocoPhillips v. Venezuela (ICSID Case No ARB/07/30) commenced in November 2007 under the Netherlands-Venezuela Bilateral Investment Treaty (the “BIT”) and Venezuela’s Foreign Investment law (the “Foreign Investment Law”). The arbitration concerns the Claimants’ interests in two extra-heavy oil projects located in the region in Venezuela known as the Orinoco Oil Belt (Faja Petrolífera del Orinoco) – the “Petrozuata Project” and the “Hamaca Project”, and in an offshore project for the extract of light to medium crude oil – the “Corocoro Project”. According to the Claimants, by passing a number of changes in law and adopting measures concerning these projects, the Respondent breached the BIT and the Foreign Investment Law.

On 12 June 2010, the arbitral tribunal in this case (Judge Kenneth Keith, President, L Yves Fortier, CC, QC and Professor Georges Abi-Saab) (the “Arbitral Tribunal”), issued Procedural Order No 3, in which it ordered that it would decide, at a first stage, issues of jurisdiction, liability and some issues relevant to the determination of compensation. The calculation of compensation was left for a later stage.

On 3 September 2013, the Arbitral Tribunal issued a decision on jurisdiction and merits (the “Decision”). Amongst other things, the Arbitral Tribunal found that it had jurisdiction under the BIT (but not under the Foreign Investment Law) and concluded that the Respondent “breached its obligation to negotiate in good faith for compensation for its taking of the ConocoPhillips assets […]”.

On 8 September 2013, the Respondent formally requested the Arbitral Tribunal to hold a hearing for submissions on the issue of good faith negotiations. The Respondent indicated that the Arbitral Tribunal would have been under “certain misapprehensions” with respect to the progress of the negotiations between the Respondent and ConocoPhillips. The Respondent also asked the Arbitral Tribunal to consider new evidence disclosed by Wikileaks, which would demonstrate that the Respondent had engaged in good faith negotiations with ConocoPhillips.

Grounding its request, the Respondent argued that there is no specific provision in the ICSID framework dealing with the power of an arbitral tribunal to revise a decision and that this presented a lacuna. However, the Respondent added that under Article 44 of the ICSID Convention, the Arbitral Tribunal had the power to fill this gap and therefore had the power to reconsider the Decision. The Respondent added that a tribunal that is “still in session” can always revise “interim” and “preliminary” decisions. On the basis that, under Article 38(2) of the ICSID Rules, an arbitral tribunal has the power to re-open a closed proceeding, the Respondent argued that, a fortiori, an arbitral tribunal had the power to reconsider a decision rendered before the closing of the proceedings.

The Claimants, in response, emphasised the principle of finality of arbitration pursuant to Article 53 of the ICSID Convention under which there would be a limited level of review only in respect of final awards (an “Award” in this context). They added that the findings of the Arbitral Tribunal in the Decision have res judicata effect.

The majority of the Arbitral Tribunal (Judge Kenneth Keith and L Yves Fortier, CC, QC) focussed on whether the Arbitral Tribunal had the power to reconsider a decision. It dismissed the Respondent’s request for reconsideration. It considered that the Decision would not be an Award within the meaning of Article 53 of the Convention. In the majority’s view, an Award is only a decision reached at the end of the whole proceeding. In the majority’s view, the only reason underpinning the Respondent’s contention is a temporal one: “a further stage in the proceedings, relating to quantum, remains” (paragraph 20 of the Decision). The majority, however, pointed out that the Arbitral Tribunal’s Decision was not “interim” or “preliminary” and would be incorporated in the Award. The majority indicated that the Decision resolved an issue in dispute between the parties and would have res judicata effect.

The majority dismissed the argument that Article 44 of the Convention would entitle it to revise the Decision. According to the majority, there would be no gap in the ICSID framework as under the ICSID Convention a tribunal’s decision can be reviewed once the Award has been rendered.

By contrast, Professor Georges Abi-Saab took a teleological view of the Arbitral Tribunal’s power to revisit its Decision. Mr Abi-Saab considered that the majority concluded that the Respondent had not negotiated in good faith with ConocoPhillips on the basis of “extraordinary speculative reasoning”. In Mr Abi-Saab’s view, the Wikileaks cables would provide irrefutable evidence that the majority’s finding on good faith negotiations was wrong.

Mr Abi-Saab harshly criticised what he viewed as the majority’s narrow and formal interpretation of the ICSID framework. Mr Abi-Saab pointed out that under Article 48(3) of the Convention, the Award shall deal with every question submitted to the Arbitral Tribunal. This would mean that earlier decisions only become final when they are incorporated into the Award. Before that, earlier decisions would be interlocutory and subject to review by an arbitral tribunal. Thus, only the Award would have res judicata effect.


The majority decision and the dissenting opinion are worlds apart. This is understandable for a number of reasons, but, for the purposes of this blog post, we will focus on one. As a starting point, the ICSID framework does not provide specific guidance as to whether an arbitral tribunal that retains jurisdiction would have the power to revise earlier decisions before an Award within the meaning of Article 53 of the Convention is made. Given this lack of specific guidance, in theory, a decision by an arbitral tribunal as to whether it has the power to revise earlier decisions (not even if it should exercise that power) would at the very least be influenced by relevant underlying principles and the ultimate consequences of any course it adopted (call it legal realism, if you wish). An arbitral tribunal deciding whether to revise earlier decisions on the basis of new evidence or supervening events would be faced with a number of competing principles and interests.

Principles supporting revision

It could be said that reasons of legal efficiency, the integrity of the process and certainty could influence an ICSID arbitral tribunal to conclude that it does not have the power to re-open earlier decisions.

The issue in hand ultimately arises from the common practice consisting of fragmenting the resolution of a dispute (often referred to as “bifurcation”). It is not uncommon for cases to be bifurcated in stages of jurisdiction and merits, or even fragmented into three stages, that is, jurisdiction, merits and quantum. Fragmenting a dispute into sequential stages can be cost-efficient.

It could be argued that fragmentation would not further efficiency if a decision made by an arbitral tribunal on, say, jurisdiction can be revisited until an Award is made. Obviously, allowing the re-opening of issues would delay the resolution of the dispute, increase costs and add complexity. In fact, if earlier decisions in a fragmented arbitration are open to review as long as the arbitral tribunal has retained a scintilla of jurisdiction, it would be more efficient to resolve a dispute in a non-fragmented setting.

If the conclusion is that an arbitral tribunal can revisit issues decided at an earlier stage in a fragmented proceeding, then it would be difficult to argue that the issues at stake are dissectible. This conclusion, by reference to the criteria laid down in cases such as Glamis Gold v. USA (NAFTA/UNCITRAL, Procedural Order No 2), would militate against the fragmentation of the conduct of arbitral proceedings.

Giving an arbitral tribunal the power to revisit earlier decisions could give rise to abuse of process issues. For example, with a view to disrupting the conduct of the proceedings, a party on opportunistic grounds might attempt to re-open issues that have already been decided. Laying down the foundations for such abusive tactics might be seen as a betrayal to the very purpose of the fragmentation of an arbitration.

In addition, allowing an arbitral tribunal to revise earlier decisions might give rise to uncertainty as to where the parties stand in an arbitration. A claimant that has established the jurisdiction of ICSID and the competence of an arbitral tribunal would unlikely be at ease knowing that the arbitral tribunal might change its mind right up until an Award is made.

Principles in favour of revision

A principle that could be advanced in favour of revision would be that of fairness. It might be perceived as being unfair for an arbitral tribunal to ignore new evidence or a significant change in circumstances when this has an impact on an earlier decision in the same proceedings.

Suppose, for example, that A, an investor, is suing State X under a bilateral investment treaty pursuant to which shares would amount to a protected investment. The arbitral tribunal in that case concluded that it had jurisdiction to deal with the claim. C, a third party, sues A before the courts of State Z and obtains a declaration that the contract under which A acquired the shares in issue is void ab initio; thus the shares have never belonged to A. This declaration would have erga omnes effect and therefore A would lack locus standi to sue State B. Should in this scenario an arbitral tribunal just close its eyes and say: “sorry, it is simply too late?”

The potential unfairness of this situation may be exacerbated by the fact that Article 51 of the ICSID Convention (dealing with the revision of an Award) does not sit well with fragmented proceedings. Under this provision, for revision to be possible, amongst other things, the supervening fact has to be unknown to the tribunal and the applicant at the time the Award is rendered. What we are discussing here precisely concerns situations that are known to the relevant party and possibly the arbitral tribunal before the Award is rendered. In other words, if a supervening fact becomes known to a party after a decision to which the fact is relevant has been made, but before an Award is made, in the view espoused by the majority, a party would have no remedy at all.


On balance, there is no easy answer as to what principles should prevail and what an arbitral tribunal should do (and this blog post does not venture a response). On the one hand, justice delayed, is justice denied. On the other, the ultimate purpose of any dispute resolution system is (or should be) the fair resolution of a dispute. Due to the complexity of the issue discussed in this post, it would have been commendable if the majority had analysed on an ex hypothesi basis whether it was willing to exercise the power to reconsider the Decision.

By Matthew Weiniger and Alejandro I Garcia, Herbert Smith Freehills LLP. The views expressed are those of the authors.


Leave a Reply

Your email address will not be published. Required fields are marked *