domingo, junio 23, 2024

Necessity in Investor-State Arbitration: The Sempra Annulment decision

Necessity in Investor-State Arbitration: The Sempra Annulment decision

Sahib Singh is a  member of the international litigation and arbitration group at Skadden and a visiting lecturer at the University of Vienna. This note was prepared before the Enron v. Argentina annulment decision became available at the beginning of August. A note on that case is forthcoming on EJIL: Talk!

On 29 June 2010, the ad hoc ICSID Annulment Committee annulled the initial award in Sempra Energy International v. Argentina, finding that the initial tribunal had exercised a manifest excess of powers. The decision is central to our understanding of necessity in international investment law, and particularly the relationship between necessity under Article XI of the Argentina-US BIT of 1991 and under customary international law. Unfortunately, the committee’s decision leaves much to be desired in terms of its interpretive methodology. The central critique of this post, is the degree of relevance the committee’s decision gives to necessity under customary international law when interpreting Article XI. It also questions the presumptive relevance of necessity under custom as an interpretive tool, when the latter can only apply if the investor does not hold substantive or procedural rights under the BIT.


The investor-state arbitration awards concerning Argentina are, for the most part, centred on the Argentine financial crisis that hit the country in late 2001. As a consequence of the crisis, Argentina undertook specific regulatory measures which liquidated the value of foreign investments (the factual matrix is far more complex, but shall not be entered into here). In the spade of investment arbitrations brought by foreign investors, Argentina has argued that it is not liable under a range of BITs due to the defence of necessity. In regards to US investors, such arguments have fallen under both customary international law and Article XI of the Argentina-US BIT. The latter reads as follows:

‘This Treaty shall not preclude the application by either Party of measures necessary for the maintenance of public order, the fulfilment of its obligations with respect to the maintenance or restoration of international peace and security, or the protection of its own essential security interests.’

Thus far six rulings have been made on the operation of necessity under Article XI and custom. Sempra Annulment is the latest. Three of the earlier decisions were against Argentina: CMS Gas Transmission (2005), Enron Corporation (2007) and Sempra Energy International (2007), through slightly varying but overarchingly common reasoning. A fourth, LG&E  (2006), found partially in favour of Argentina. Three days prior to the Sempra award being given to the claimant, an Annulment Committee strongly condemned the CMS Gas award (but did not annul); condemning the initial tribunal for its treatment of the defence of necessity. The CMS Annulment decision has marked the turning point in interpretive methodology used by tribunals. This was seen in the subsequent award in Continental Casualty (2008) and now the annulment in Sempra, which both found in favour of Argentina. However, discrepancies in reasoning still appear. Whilst significant, the Sempra annulment does contribute to such a problem.

Sempra v. Argentina

The committee in Sempra annulled the initial award due to the tribunal exercising a manifest excess of powers, in respect of its failure to apply Article XI of the BIT (para. 159, annulment). In doing so, it did not uphold annulment on the grounds of a manifest error of law or a failure to state reasons (as understood under Article 52 ICSID Convention). Given that the annulment was purely based on the application of legal rules and not on factual mischaracterisations or misapplication, it is not necessary to give a factual background. Understanding that Sempra Energy International claimed losses as a result of measures taken by Argentina during its financial crisis should suffice for the purposes of this post.

The notable findings of the committee are quoted below:

“176. This sequence of argument is illogical [necessity under custom and then under the Article XI] as the question whether a state of necessity justifies exoneration from state responsibility will become an issue only were liability is not already precluded under Article XI of the BIT. As a general rule, a treaty will take precedence over customary international law. …

188. According to Article 31(1) of VCLT, the first point of reference for interpretation of a BIT provision is the “ordinary meaning” of the words of the treaty themselves.

189. In the present case, where the BIT provides the relevant treaty language, it is necessary first and foremost to apply the provisions of the BIT. Indeed the Parties are in agreement that the BIT constitutes the applicable law. …

197. First … the Committee accepts, of course, that it may be appropriate to look at customary law as a guide to the interpretation of terms used in the BIT. It does not follow, however, that customary law (in casu, Article 25 ILC Articles) establishes a peremptory “definition of necessity and the conditions for its operations. …

199. It is apparent from this comparison [of the texts of Article 25 ILC Articles and Article XI BIT side-by-side] that Article 25 does not offer a guide to interpretation of the terms used in Article XI. The most that can be said is there are certain words of expressions are the same or similar.

200. More importantly, Article 25 is concerned with the invocation by a State Party of necessity “as a ground for precluding the wrongfulness of an act not in conformity with an international obligation of that State”. Article 25 presupposes that an act has been committed that is incompatible with the State’s international obligations and is therefore “wrongful”. Article XI, on the other hand, provides that “This Treaty shall not preclude” certain measures so that, where Article XI applies, the takes of such measures is not incomptabile with the State international obligations and is not therefore “wrongful”. Article 25 and Article XI therefore deal with quite different situations. Article 25 cannot therefore be assumed to “define necessity and the conditions for its operation” for the purposes of interpreting Article XI, still less to do so as a mandatory norm of international law. …

203. … As noted above, Article 25, deals with a situation where a State Party is in breach by a plea of necessity. Article 25 sets out the restrictive conditions in which such a plea may be admitted. Article XI of the BIT, on the other hand, expressly provides that the BIT “shall not preclude the application by either Party of measures necessary” for certain reasons or purposes.”


The decision makes several points of note. I would like to consider a few of these, in light of the previous jurisprudence dealing with the matter.

(1)   Article XI as the Primary Source of the Necessity Defence:

Previous arbitral tribunals (including Enron, CMS and Sempra) have conflated the relationship between necessity under customary international law and Article XI. Indeed, some such as the Sempra award (para. 388, award), prioritise the customary right over Article XI. The ILC correctly guards against applying the customary circumstance precluding wrongfulness, when necessity is incorporated into the primary rule (Articles on the Responsibility of States for Internationally Wrongful Acts, with Commentaries, p. 84 (para. (21) under Article 25). Such an approach not only ignores that Article XI operates as a primary rule exception to liability under the BIT, but also that the customary standard, as a secondary rule, only kicks in should (a) there be a breach of a substantive BIT provision; and (b) the Article XI exception fails.

However, the Sempra annulment decision is now the third decision/award in a row that has confirmed the primary-secondary rule distinction (following CMS Annulment, para. 129 and Continental Casualty, para. 166).  Importantly, it has  identified Article XI as the applicable legal norm of primary importance when examining the necessity defence (Sempra annulment, para. 209). Such a stance has three distinct advantages as a adjudicatory method.

First, it is an approach which correctly identifies the ruling obligation which should govern the argument. Article XI is the only rule which operates as a functioning primary rule and it is one which operates in the specific treaty environment intended to govern the dispute.

Second, it allows the correct interpretive methodology to be used. Whilst the conflation method adopted by previous tribunals did not allow a dominant or primary rule to be interpreted, this approach allows a clear, concise and logical approach to interpretation. The conflation methodology essentially jumps straight to that envisaged by Article 31(3)(c) VCLT. As a matter of treaty interpretation this is clearly the incorrect approach, since the starting point has to be Article 31(1) VCLT and its tripartite test. Identifying the operative primary rule enables the logical application of Article 31 through its step-by-step analysis.

Third, the primary-secondary rule distinction, allows a tribunal to disable or attempt to circumvent a complicated lex specialis analysis. This is a complex argument, which I shall simplify. There are two traditional types of relationships which lex specialis may regulate: (a) where a particular rule may be considered an application of a general standard in a given circumstance; or (b) where the special rule is considered as modifying, overruling or setting aside the general rule (this was implicitly used in LG&E) (para. 88, ILC Fragmentation Report). The conflation approach of the Enron, CMS and Sempra awards, treated Article 25 ILC Articles as essentially a primary rule inconsistent with Article XI BIT. However, the Sempra annulment committee ensures that Article 25 is treated a secondary norm, avoiding conflict. Although lex specialis is sometimes applied where two rules point in the same direction without express conflict (para. 93, ILC Fragmentation Report), the committee has proceeded to initially use Article 25 as a background interpretive tool to Article XI (as in Oil Platforms, paras. 43-4) (para. 197, Sempra annulment). This background interpretive influence of the customary standard is better governed within the framework of Article 31(3)(c) and the principle of ‘systemic integration’. The primary-secondary rule distinction therefore avoids a complicated question and is able to bring the question within the framework of the traditional interpretive methods of the VCLT.

(2)   Interpretation of Article XI

Having identified Article XI as the decisive and operative rule, the Sempra committee’s attempt to define its scope and application is where the decision really comes into its own, departing significantly from previous decisions. The committee notes that “Articles 25 does not offer a guide to interpretation of the terms used in Article XI.” This statement departs from tribunal’s approach in Continental Casualty which noted that it will “focus on the analysis of Art. XI and the conditions of its application, referring to the customary rule on State of Necessity (as enshrined in Art. 25 of the ILC test) only insofar as the concept there used assist in the interpretation of Art. XI itself” (para. 168). The Sempra committee reasons that customary definition can offer no guidance to interpretation due to: (a) the terms of each rule, Articles XI and 25, significantly differ (para. 199); and (b) they operate on different planes – primary and secondary – “therefore deal with quite different situations” (para. 200).

If the committee was going to make such bold statements, avoidable under its review mandate, it should have adopted a far more rigorous methodology. There are a number of problems with the committee’s approach.

(a)    The starting point of Article 31(1) is explicitly acknowledged as beginning with “the ordinary meaning” of the rule. However, the committee failed to complement its analysis of the text of Article XI this with both the “context” and “object and purpose” of the provisions and the treaty.

(b)   The failure to address these elements is significant given that: (i) these are the primary factors differentiating the application of the customary standard; (ii) some case law has presumed that the object and purpose the BIT is to apply in economic difficulties and therefore a restrictive interpretation of Article XI has followed (Enron, para. 331).

(c)    Eventual resort to Article 31(3)(c), other relevant and applicable rules of international law, may not lead to the categorical approach adopted by the Sempra committee. The two grounds stated above (in paras. 199 and 200) for Article 25’s lack of interpretive value seem to be rooted in its “relevance”. It is hard to adopt the conclusion that Article 25 offers no value whatsoever. The fact that there exists the primary-secondary rule distinction, does not preclude custom’s interpretive value, but rather goes towards the normative weight provided to the latter. Given the differences in, (a) text and (b) purpose, between the two provisions, the weight afforded is bound to be fairly low but not entirely absent. As it stands, the current approach taken by the Sempra committee nullifies the purpose of Article 31(3)(c) and essentially allows Article XI to operate by itself and without recourse to other rules of international law. More so, given that the committee does not identify other rules of international law which may assist in interpreting quite a vague provision.

The import of identifying other rules which can inform the content of Article XI, and specifically notions of “public order”, “maintenance or restoration of international peace and security” and “essential security interests”, cannot be over-emphasised. Several tribunals have noted that the initial convoluted approach is rooted in their inability to give content to the provisions of Article XI. This however, is the subject of another enquiry. For those interested, Jürgen Kuntz makes a valiant effort in his recent ICLQ article (abstract only).

I now turn to the final issue which remains unaddressed by the vast majority of arbitral decisions dealing with necessity: the ambiguity in defining the nature of investor’s right and the ability of State’s to use the customary necessity defence.

(3)   Defining the nature of the Investor’s rights and Necessity.

In her impressive recent AJIL article, Anthea Roberts tangentially hits upon the ‘old’ debate of the nature of investor’s rights under the BIT (pp. 184-5). She uses Zach Douglas’ classification of options in defining such rights:

“(1) investment treaties grant substantive and procedural rights to the treaty parties only, but investors are permitted for the sake of convenience to enforce their states’ substantive rights;

(2) investment treaties grant substantive rights to the treaty parties only, but investors are granted the procedural right to enforce their states’ substantive rights; and

(3) investment treaties grant substantive and procedural rights to investors, giving investors a procedural right to enforce their own substantive rights.”

Attempt at classification has arisen in only a few cases (concerning countermeasures and waivers). However, classification of the nature of the investor’s right has an impact on two matters regarding necessity: (a) the ability for a state to invoke the customary defence (as opposed to whether it would be successful if the substantive provisions were provided); (b) the normative weight in interpretation that tribunals can and should given to the customary norm when considering provisions such as Article XI Argentina-US BIT.

Under Robert’s scenarios (1) and (2), necessity can be invoked by the Respondent state as a circumstance precluding the wrongfulness of a BIT breach against another state. Under scenario (3) however, necessity under Article 25 cannot apply against the investor. Any breach of the BIT would be a breach against the investor. The customary defence operates to preclude wrongfulness against a breach by a State and cannot preclude wrongfulness to a third party non-state actor; the investor. The latter approach is similar to that followed in BG Group v. Argentina (para. 408) and the rational and effects of Robert’s scenario (1) were adopted in National Grid v. Argentina (para. 211). The analysis in these awards are far from explicit, but have been extrapolated by the author. Indeed, the most recent tribunal in AWG Group v. Argentina did not address the question at all, presuming the prima facie applicability of the customary rule.

Tribunals have been able to circumvent this underlying normative issue since the stringent conditions of the customary standard have not and cannot be met under investor-state arbitration (no matter the opposite and incorrect conclusion reached in LG&E, para. 257). However, Robert’s scenario (3) and its underlying rational does have a normative effect on arbitrations under the Argentina-US BIT. First, if Robert’s scenario (3) and the interpretive rational in Sempra is used then in addition to custom operating on a different normative plane (secondary, not primary), it also does not operate as the same subject plan and should therefore benefit from even less normative weight in the interpretation process of Article XI. Second, if the primary-secondary rules methodology advocated by recent tribunals is advanced, custom based necessity only kicks in once (a) a substantive breach of the BIT is shown; and (b) a defence under Article XI fails. However, scenario (3) prima facie disables the application of the customary right. Argentina would not be able to take recourse to Article 25 if an Article XI defence fails.


 In conclusion, the Sempra committee is to be credited for ingraining the primary-secondary rule distinction which clarifies the dominant and applicable rule. However, in its treatment of the residual role of Article 25 in the interpretation of Article XI, the committee is unnecessarily stringent to the exclusion of the customary standard. This is not aided by a lack of thoroughness in interpretive methodology. Finally, the operation of the necessity defence in investor-state arbitration suffers from a foundational problem, which pervades into many other areas of the regime: the lack of clarification on the nature of the investor’s rights.

The views expressed here do not necessarily reflect those of Skadden

Ver también