Christoph Schreuer, a professor at the University of Vienna, has said that there are “three generations of ICSID annulment decisions”. In light of the quartet of decisions rendered this summer, Promod Nair and Claudia Ludwig of Herbert Smith in London consider whether there is an emerging “fourth generation” of ICSID annulment jurisprudence – and whether reforms are needed to the annulment regime.

The summer of 2010 will be remembered for significant developments in the evolution of ICSID annulment jurisprudence. In the space of a couple of months, four annulment decisions were issued – in Sempra Energy International v Argentina, Enron Corporation & Ponderosa Asset v Argentina, Helnan Hotels v Egypt and Vivendi v Argentina II. In three of these the awards were fully or partially annulled, undermining the hitherto well‑established position that annulment is an “extraordinary remedy for unusual and important cases”. In the light of these decisions, we consider whether the scope of ICSID annulment review is more expansive than previously thought and whether it is now necessary to reform the entire system altogether.

Annulment as an “extraordinary remedy”

One of the defining characteristics of the ICSID arbitration regime – and one of its key perceived strengths – is its internal mechanism for reviewing arbitral awards. While awards issued outside the framework of the ICSID arbitration regime are subject to review by domestic courts – which are often perceived as investor-unfriendly – applications to annul ICSID awards are heard by ad hoc committees specifically convened for this purpose by ICSID itself. (See Emmanuel Gaillard’s introduction to Annulment of ICSID Awards, a 2004 volume of articles edited by himself and Yas Banifatemi.)

The drafters of the Washington – or ICSID – Convention made a deliberate choice when they decided against having an appeals system. They took the view that such a system would increase the amount of time taken to reach a final decision and that costs would escalate owing to the increased length of proceedings. Consistent with this philosophy, Article 53 of the Washington Convention excludes any appeal or other remedy in respect of an ICSID award “except those provided for in [the] Convention”.

Article 52 of the Convention, providing for the annulment of awards, constitutes a limited and “extraordinary” exception to the finality principle enshrined in Article 53. It states that an award may be annulled only on one or more of the following grounds:

  1. the tribunal was not properly constituted;
  2. the tribunal manifestly exceeded its powers;
  3. there was corruption on the part of a member of the tribunal;
  4. there was a serious departure from a fundamental rule of procedure;
  5. the award does not state the reasons on which it is based.

As explained by arbitrator Hans Van Houtte, annulment by an ICSID ad hoc committee is, at the same time, a drastic and a limited remedy. It is a drastic remedy in that an annulment decision does not simply alter some elements of the award, but invalidates it completely. The ad hoc committee has a choice between leaving the original award intact or declaring it void – either fully or in parts (Article 52(3) of the convention). It may not substitute its own decision on the merits, and the only redress open to a party whose award has been annulled is to resubmit the dispute to another arbitral tribunal.

Annulment is also a limited remedy in that an ad hoc committee is not a court of appeal. It cannot rehear the substance of the dispute. It can only consider whether the award should be annulled – in whole or in part – on one of the grounds specified in Article 52.

As the ad hoc committee in MTD v Chile put it:

“[T]he role of an ad hoc committee in the ICSID system is a limited one. It cannot substitute its determination on the merits for that of the tribunal. Nor can it direct a tribunal on a resubmission how it should resolve substantive issues in dispute. All it can do is annul the decision of the tribunal: it can extinguish a res judicata but on a question of merits it cannot create a new one.”

Three generations of ICSID annulment decisions

In a much cited analysis, Christoph Schreuer classifies ICSID annulment jurisprudence into three generations:

The first generation (1985-1986)

The first generation of annulment decisions consists of the decisions in Klöckner v Cameroon I and Amco Asia v Indonesia I, in which ICSID awards were set aside on the ground that the arbitral tribunals had manifestly exceeded their powers. The ad hoc committees in those cases soon faced criticism for re-examining the merits of the cases and for failing to properly distinguish between annulment and appeal.

If this line of decisions were to continue, ICSID would “lose its appeal”, observers commented. They also accused the committees of undermining one of investment treaty arbitration’s fundamental objectives, the finality of awards, by interpreting their review function so expansively.

The second generation (1989-1992)

Against the backdrop of the overwhelming criticism of the first two annulment decisions, subsequent ad hoc committees were much more cautious.

In MINE v Guinea, the state applied for annulment on the basis that the tribunal had failed to state reasons. The ad hoc committee held that “[t]he adequacy of the reasoning is not an appropriate standard of review under paragraph 1(e), because it almost inevitably draws an ad hoc Committee into an examination of the substance of the tribunal’s decision, in disregard of the exclusion of the remedy of appeal by Article 53 of the Convention.”. Annulment committees also had to consider the award

s in the resubmitted first generation cases, Klöckner v Cameroon II and Amco Asia v Indonesia II. In both cases the applications for annulment were unsuccessful.

The third generation (2002-2009)

The third generation of annulment decisions – represented by the decisions of the ad hoc committees in Wena Hotels v Egypt, Vivendi v Argentina I and CMS v Argentina – were welcomed for their more balanced approach. They underlined that ad hoc committees would only intervene in “serious and important cases”.

In Wena, the ad hoc committee held that, even where an award showed a lack of reasons, it had a certain amount of judicial discretion and “the remedy need not be the annulment of the award”. It explained that it would be especially likely to exercise this discretion where “the reasons supporting the Tribunal’s conclusions can be explained by the ad hoc Committee itself” – a welcome deviation from the “hair trigger” approach of the committee in Klockner I, which held that annulment would be automatic if any of the possible grounds were established. The balanced approach continued in Vivendi v Argentina I. In that case, the ad hoc committee said: “it appears to be established that an ad hoc committee has a certain measure of discretion as to whether to annul an award, even if an annullable error is found. Article 52(3) provides that a committee shall have the authority to annul the award or any part thereof,” and this has been interpreted as giving committees some flexibility in determining whether annulment is appropriate in the circumstances. Among other things, it is necessary for an ad hoc committee to consider the significance of the error relative to the legal rights of the parties”.

The ad hoc committee in CMS Gas Transmission Company v Argentina, meanwhile, confirmed the Vivendi principle that, even if an award contains serious errors and defects, annulment does not follow as an automatic consequence:

“Throughout its consideration of the Award, the Committee has identified a series of errors and defects,” a passage from the decision reads. “The Award contained manifest errors of law. It suffered from lacunae and elisions. All this has been identified and underlined by the Committee. However, the Committee is conscious that it exercises its jurisdiction under a narrow and limited mandate conferred by Article 52 of the ICSID Convention. The scope of this mandate allows annulment as an option only when certain specific conditions exist…. the Committee cannot simply substitute its own view of the law and its own appreciation of the facts for those of the Tribunal.”

The third generation of ICSID annulment jurisprudence has come in for praise for successfully navigating a course between “the Scylla of complete fairness and the Charybdis of absolute finality”. On the one hand, the annulment process is not intended to give losing respondents a second bite at the cherry by providing a full fledged right of appeal. On the other, the process seeks to ensure a minimum level of oversight over ICSID tribunals – providing an essential safeguard that is fundamental in helping to preserve the legitimacy of the system.

After three generations of annulment decisions, the following principles regarding the scope of ad hoc committees’ role were well established:

  1. they have a narrow and limited mandate;
  2. they cannot correct errors of law;
  3. they enjoy a certain degree of discretion;
  4. they may reconstruct the reasons if the tribunal has failed to provide them; and
  5. they may not automatically annul an award, even if the tribunal makes a mistake in its search for the appropriate law.

The fourth generation: a lower threshold for ICSID annulment?

A quartet of annulment decisions this summer has cast a shadow of uncertainty over these principles. In Sempra v Argentina and Enron v Argentina, ad hoc committees annulled awards on the basis that the tribunals had not properly understood and applied the necessity defence advanced by Argentina. In Helnan v Egypt, the committee partially annulled the award on the ground that the tribunal had erred in its application of the exhaustion of local remedies rule. And in Vivendi v Argentina II, a committee criticised the conduct of both an arbitrator and the ICSID secretariat – generating much heat and sound but ultimately refusing to annul the award in question.

These decisions have raised fresh questions about the threshold of review in annulment proceedings, and the proper role to be discharged by ad hoc annulment committees. Although it is too early to predict a definite trend towards more extensive review, they might well mark the start of a “fourth generation” of ICSID annulment jurisprudence.

The Sempra and Enron cases

Like a number of treaty arbitrations against Argentina, Sempra v Argentina and Enron Corporation and Ponderosa Assets v Argentina concerned emergency measures introduced by the government to deal with an unprecedented economic crisis in 2001. A year later, Sempra and Enron began ICSID arbitrations contending that Argentina had breached its obligations under the US-Argentina bilateral investment treaty. Argentina’s position was that the emergency measures were permitted under Article 11 of the BIT, which provides that States are not precluded from taking “measures necessary for the maintenance of public order, the fulfilment of its obligations with respect to the maintenance or restoration of international peace or security, or the protection of its own essential security interests”.

In both cases, ICSID tribunals reasoned that Article 11 of the BIT does not define necessity and the conditions for its operation. They therefore applied customary international law in assessing whether Argentina’s invocation of the necessity defence was justified. Both tribunals reached the conclusion that Argentina failed to establish a right to invoke the “necessity” exception and held it responsible for substantive breaches of the US-Argentina BIT.

In Sempra, the ICSID tribunal viewed Article 11 as being “inseparable from the customary law standard insofar as the definition of necessity and the conditions for its operation are concerned”. After Argentina applied for annulment, the ad hoc committee held that the tribunal had made a mistake in equating the two in this way. The treaty contained stand-alone criteria that were distinct from – and potentially broader than – the customary international law position on “necessity”, it said. The committee characterised the tribunal’s mistake not as an error of law – which is not a ground for annulment – but as a complete “failure to apply the applicable law”. Such an error was serious enough to qualify as a manifest excess of the tribunal’s powers, it concluded.

The ruling is a controversial one, especially since it is not entirely clear that the Sempra tribunal did in fact apply customary international law to the exclusion of the BIT.

On the contrary, the tribunal seemed to proceed on the basis that the two regimes were equivalent when it said there was no need to “undertake a further judicial review under Article [11] given that [the] Article does not set out conditions different from customary international law in this regard.”

The ad hoc committee in Enron went down a completely different route. It held that the tribunal was entitled to equate Article 11 of the BIT and customary international law. However, it considered that the tribunal had failed to apply a number of the essential legal elements of the necessity defence under customary international law. For instance, the tribunal held that Argentina was obliged to prove that the suite of emergency measures it adopted to deal with the financial crisis was “the only way” for it to safeguard an essential security interest. The committee considered that the tribunal dealt with this issue in a “cursory” manner, noting that it had uncritically relied on the opinion of an economic expert who said that Argentina had multiple policy options in dealing with the crisis. The committee thought that the tribunal should also have assessed the viability of the alternative courses of action – and whether the measures taken were the only effective way of addressing the crisis. Its failure to grapple with such issues meant it had neglected to apply the applicable law and thereby manifestly exceeded its powers, the committee held.

Critics say the Sempra and Enron annulment committees blurred the distinction between full-fledged error of law review – which is not authorised under the Washington Convention – and the limited jurisdictional review permitted by Article 52. It is informative to contrast the decisions in these cases with the approach adopted by the ad hoc committee in one of the third generation cases, CMS Gas v Argentina, which also concerned Argentina’s response to its financial crisis. In an award similar to those of the Sempra and Enron tribunals, the CMS Gas tribunal applied the customary international law of necessity and concluded that Argentina’s actions were not excused by the necessity defence. The ad hoc committee criticised the award and the approach of the tribunal. By considering the customary international law position on necessity instead of the express language of the BIT, it said the tribunal had failed to “examine whether the conditions laid down by Article [11] were fulfilled…” The tribunal had applied Article 11 only “cryptically and defectively” the tribunal said.

Despite this finding, however, the committee refused to annul the CMS Gas award, saying that the tribunal had not manifestly exceeded its powers. There might have been a defective application of the law, it held, but the tribunal was not guilty of failing to apply the law. While the CMS committee held that an error of law, no matter how serious, could not be the basis of annulment, the Sempra and Enron committees considered that serious errors of law would likely amount to a complete failure to apply the applicable law and thereby trigger the manifest excess of powers ground for annulment.

The Helnan annulment decision

In the third of the quartet of decisions, Helnan International Hotels v Egypt, an ICSID ad hoc committee partially allowed an application by the investor for annulment. It held that the tribunal had manifestly exceeded its powers when it said that Helnan was required to exhaust local remedies before starting ICSID proceedings.

Although not as controversial as the decisions in Sempra and Enron, the Helnan annulment committee has also faced its share of criticism. The award in the case echoed the 2003 Generation Ukraine v Ukraine award, in which an investor’s claims were dismissed because it had not attempted to seek redress of a decision by a low-ranking government official in the Ukrainian courts. Following this approach, the Helnan tribunal held that the investor ought to have taken its claims to the Egyptian courts before commencing an action on the international plane. It is certainly possible to disagree with the tribunal’s approach. However, it is questionable whether its conclusion on this point was so patently unjustifiable as to amount to a “manifest excess of its powers” – especially in light of the award’s striking parallels with the one in Generation Ukraine v Ukraine.

However, the partial annulment of the award has had no impact on the overall outcome since the tribunal’s finding on this issue was not dispositive of the dispute. The tribunal had also dismissed the investor’s claims on the merits – a decision that the annulment committee did not interfere with. As a result, the tribunal’s award on the merits still stands.

Vivendi v Argentina II

In the last of the annulment decisions under consideration, Vivendi v Argentina II, Argentina sought to annul on a variety of grounds, including that the tribunal was improperly constituted and there was a serious departure from a fundamental rule of procedure. It based its application on the fact that one of the arbitrators, Gabrielle Kaufmann-Kohler, had failed to disclose that she was a member of the board of directors of UBS, one of the shareholders of Vivendi.

In its decision, the ad hoc committee severely criticised Kaufmann-Kohler for not informing the parties of her mandate for the Swiss bank and generally questioned her judgment with regard to conflict disclosures – making reference to her removal from the Yukos case by the Permanent Court of Arbitration. It said it appreciated “the argument that the Second Tribunal was no longer properly constituted after the board appointment of Professor Kaufmann-Kohler, and that there was a serious departure from a fundamental rule of procedure” which “could lead to annulment whenever justified”.

Yet, notwithstanding the prima facie existence of an annulment ground, the ad hoc committee refused to annul the award. It accepted that Kaufmann-Kohler had no actual knowledge of the connection between UBS and Vivendi until after the award was issued – meaning that “despite most serious shortcomings” her independent judgment was not impaired. Furthermore, the ad hoc committee took into account “the extraordinary length of the … case” and said that it would be “unjust to deny the Claimant the benefit of the Award”, given that it was not responsible for Kaufmann-Kohler’s actions. Although it identified serious errors in the award, the committee ultimately found that they were not grave enough to warrant annulment – as in CMS v Argentina.

A fork in the road for future annulment committees

Critics say the recent annulment decisions convert the limited jurisdictional review provided for in Article 52 of the Washington Convention into full fledged, de facto appellate proceedings – with little support from the language and underlying objectives of the Convention. If committees are going to second guess the tribunals’ interpretation of facts and application of the law, they say, the annulment process could degenerate into an exercise in academic rivalry where the members of an annulment committee strive to prove they are better arbitrators than those whose awards are under review. As Alan Redfern put it in a 1987 article, “ICSID-Losing its Appeal”, “[t]he decisions of three eminent arbitrators, appointed by or on behalf of the parties, [would then be] wiped out by another three eminent arbitrators, appointed by the President of the World Bank, in what might seem like an elaborate and expensive game of snakes and ladders.”

Given the criticism of these cases, it is small wonder that we are now hearing reiterated calls for the reform of the ICSID annulment regime. One proposal is to replace ad hoc annulment committees with a standing body of jurists (along the lines of the WTO’s appellate body) to hear challenges to ICSID awards. Such a body would issue consistent decisions, and assist in the orderly development of investment treaty jurisprudence, proponents say. But given that a number of States have recently denounced or threatened to withdraw from the Washington Convention, would an attempt to tinker with the annulment provisions at this juncture lead to demands for more comprehensive reforms – or even an overhaul of the entire ICSID architecture? Given this risk, unsurprisingly, States have little enthusiasm for reform – so the prospects of legislative change in the short term seem to be remote.

Much depends on the approach future annulment committees choose to take. There is no system of legally binding precedent under the ICSID regime and it remains to be seen whether future annulment committees will prefer the lower threshold for review in the Sempra, Enron and Helnan decisions or reinstate the widely commended principles articulated in the third generation of ICSID annulment decisions. Now that losing parties have successfully obtained the annulment of ICSID awards in a relatively large number of cases, what is clear is that others will be emboldened to seek annulment, in hope of getting a second bite at the cherry. The danger is that that this will significantly lengthen ICSID disputes and seriously undermine confidence in the efficacy of the centre’s dispute resolution regime.

This article was originally published in Volume 5, Issue 5 (October 2010) of Global Arbitration Review,