Home EJIL Analysis The Enron Annulment Decision’s exposure of Necessity’s Endemic Uncertainty: A Welcome Critique

The Enron Annulment Decision’s exposure of Necessity’s Endemic Uncertainty: A Welcome Critique

Published on October 25, 2010        Author: 

Sahib Singh is a  member of the international litigation and arbitration group at Skadden and a visiting lecturer at the University of Vienna.

As promised by my last post on the Sempra Annulment decision, this is a comment on the Enron Annulment decision of 30 July 2010. The decision is fascinating for a number of reasons, but this post shall concentrate on the Committee’s analysis of the “only means” requirement under the customary doctrine of necessity. Whilst highlighting the inherent ambiguity in the application of these words, the Committee’s inquisitorial approach may create more problems than it solves. Accordingly, this post concludes with a rough sketch of logical steps a tribunal may take in applying the “only means” requirements under custom.

I.          Background & Findings of the Committee in the Enron Annulment decision

By way of background on investor-State arbitration claims concerning Argentina, please see my last post. The Enron Annulment Committee concluded that the original award, rendered in favour of the Claimant, was to be annulled due to the Tribunal’s failure to apply the applicable customary law as represented by Article 25 of the ILC Articles on the Responsibility of State for Internationally Wrongful Acts and failure to give reasons (paras. 377-8, Enron Annulment). In particular the Committee concluded that the tribunal’s reasoning of the “only means” requirement under Article 25 was entirely insufficient: (emphasis added)        

369. The first question concerns the legal definition of the expression “only way” in Article 25(1)(a) of the ILC Articles. The Committee notes that the expression is capable of more than one possible interpretation. One potential interpretation is that it has its literal meaning, such that in the present case, the principle of necessity could be relied on by Argentina if there were genuinely no other measures that Argentina could possibly have adopted in order to address the economic crisis. As Argentina points out, there will almost inevitably be more than one way for a Government to respond to any economic crisis, and if this interpretation were correct, the principle of necessity under customary international law could rarely if ever be invoked in relation to measures taken by a Government to deal with an economic crisis. However, that would not mean that it would not be open to a Tribunal to find that this is the correct interpretation, although there are other interpretations that would be equally open to a Tribunal.

 370. For instance, another possible interpretation would be that there must be no alternative measures that the State might have taken for safeguarding the essential interest in question that did not involve a similar or graver breach of international law. Under this interpretation, if there are three possible alternative measures that a State might adopt, all of which would involve violations of the State’s obligations under international law, the State will not be prevented from invoking the principle of necessity if it adopts the measure involving the least grave violation of international law. Under this interpretation, the principle of necessity will only be precluded if there is an alternative that would not involve a breach of international law or which would involve a less grave breach of international law.

 371. A second question not addressed by the Tribunal is whether the relative effectiveness of alternative measures is to be taken into account. In adopting measures to safeguard an essential interest, a State may in practice not be in a position to know with certainty whether a given measure will prove to be effective, and reasonable minds may judge that some measures are likely to be more effective than others. For instance, suppose that there are two possible measures that a State might take in order to seek to safeguard an essential interest. One is 90 per cent probable to be 90 per cent effective to safeguard that essential interest, while the other is 50 per cent probable to be 60 per cent effective. Suppose that the former measure would (subject to the potential application of the principle of necessity) be inconsistent with obligations of the State under international law, while the latter measure would not. Would the State be precluded from invoking the principle of necessity if it adopted the former measure, on the basis that there was an alternative available? Or could the State claim that the measure taken was the “only way” that stood a very high chance of being very effective?

 372. A third question that is not specifically addressed by the Tribunal is who makes the decision whether there is a relevant alternative, and in accordance with what test? Does the Tribunal determine this at the date of its award, when the Tribunal may have the benefit of knowledge and hindsight that was not available to the State at the time that it adopted the measure in question? Or does the Tribunal determine whether, on the basis of information reasonably available at the time that the measure was adopted, a reasonable and appropriately qualified decision maker would have concluded that there was a relevant alternative open to the State? Or does customary international law recognise that reasonable minds might differ in relation to such a question, and give a “margin of appreciation” to the State in question? In that event, the relevant question for the Tribunal might be whether it was reasonably open to the State, in the circumstances as they pertained at the relevant time, to form the opinion that no relevant alternative was open.

  II.        Analysis of the Committee’s Annulment approach to “only means”

 The Committee’s analysis of the “only means” requirement under Article 25 ILC Articles raises three particular points of interest (sections B-D below). However, this post shall quickly examine the background to this requirement.

 A.           Background to the “only means” requirement under Article 25

The “only means” requirement is merely one of a number of substantive requirements under Article 25 that ensure that it shall only be rarely available. The ILC Commentaries makes only two points in regard to these words: “the plea is excluded if there are other (otherwise lawful) means available, even if they may more costly or less convenient” (para. (16), Article 25 ILC Commentaries).  The first point goes back to Judge Anzilotti’s opinion in the Oscar Chin case (p. 114), the Russian Indemnity case and the Serbian Loans case. The second, that of cost and convenience, is predominantly derived from the Gabčíkovo-Nagymaros Project case, although Ago had considered it in his 8th Report on State Responsibility (para. 14 of the 1980 Report).

Cases provide little guidance on how to treat these two words: “only means”. One may go as far to question its presence in the customary doctrine: the Serbian Loans and Russian Indemnity cases can easily be contrasted (where a very lose formulation of the requirement was considered) with the French Company of Venezuela Railroads case where the strict requirement of alternative means was not considered in upholding the defence in favour of Venezuela. 

B.           The Value of a Literal Interpretation?

The Enron Annulment Committee recognised Argentina’s argument that if a literal interpretation was to be provided to “only means”, the necessity defence “would rarely if ever be invoked … to deal with an economic crisis” (para. 369, Enron Annulment). Whilst Article 25’s application has always been envisaged as one of rare availability, this does not justify a narrow interpretation of the “only means” requirement. The value of a literal interpretation cannot be evinced to uphold the defence’s exceptionality – or vice versa. Each substantive requirement (e.g. “only means”) of Article 25 needs to be interpreted and applied with its own value.

If one was to follow a literal interpretation of the words “only means” then any alternative, irrespective of its viability and reasonableness, is available to exclude the operation of the necessity defence (see CMS Award, para. 323 onwards). Clearly, the interpretation and application of these words needs to be pragmatic but not overly inclusive (see LG&E Award, para. 257). If only for this reason, the Enron Annulment decision is notable for its willingness to confront head-on the uncertainty of Article 25; a task that most previous tribunals (including the ICJ) have shied away from.

C.           Determining what qualifies as an “alternative” means

The Enron Annulment Committee examines the uncertainty of the words “only means”. It does so by identifying factors which may determine what may constitute a relevant alternative means for the purposes of Article 25. Two specific factors are identified: (i) the comparative gravity of the breach induced by the alternative means; (ii) the comparative effectiveness of the alternative means (paras. 370-1). The Committee’s inquisitorial reasoning identifies both of these factors as possible considerations when determining whether a specific alternative means is to be considered relevant for the purposes of Article 25(1)(a) ILC Articles and therefore exclude the necessity defence.

In regard to the first of these, consideration of the comparative gravity of the breach by an alternative means, the Committee goes further than existing law. Current law enunciates the rule identified by the ILC: “the plea is excluded if there are other (otherwise lawful) means available”. The point in parenthesis is to identify as relevant those alternative means which would have been lawful. If such means are identifiable, it is only these means which could potentially exclude the necessity defence. This position is consistent with existing authority on the issue: Judge Anzilotti’s opinion in the Oscar Chin case (para. 7 of the Individual Opinion), Roberto Ago’s 8th State Responsibility Report (para. 14), and the ICJ the in Gabčíkovo-Nagymaros Project case (para. 55) and the ILC Commentaries (para. 15, Article 25 ILC Commentaries).

The Annulment Committee’s analysis on this point of gravity is worrying for two reasons. First, it presumes that, bar jus cogens norm breaches, there are identifiable hierarchical levels of breaches. This is misguided and incorrect. Second, it broadens the range of relevant alternative means that can exclude the necessity defence. Given that precedent only excludes the necessity defence for a State if an alternative legal measure is taken to safeguard its essential interests, the Committee goes further by introducing the possibility that an alternative illegal but less graver measure may exclude the necessity defence. By broadening the spectrum of possibly exclusionary alternatives the Committee’s inquisitorial reasoning, if adopted by other tribunals, would narrow the necessity defence further than its already stringent conditions.

The second factor considered by the tribunal was the relative effectiveness of the alternative measure identified. This is of great importance given that situations of necessity shall always concern grievous circumstances. The effectiveness of the measure to safeguard against such a circumstance is central to the purpose of the necessity doctrine. The most penetrating question posed by the tribunal is whether the necessity defence is precluded if the State pursues a measure which is more effective but illegal, when it possesses the capability to pursue a measure which is legal but far less effective. Formalistically, the presence of the latter should prevent the applicability of the necessity doctrine. Purposefully, the necessity doctrine is aimed at legitimating those measures which a State must take to effectively deal with a situation which threatens its essential interests or its very existence. If the doctrine does not permit the State to take those measures which would best protect it against further chaos (in this case economic chaos) and harm in a particular situation – then the doctrine serves little purpose at all. The need for the measure to adequately and effectively address the “peril” in question is at the heart of the doctrine; for necessity only ceases to operate once the situation requiring attention ceases to be one that threatens an essential interest of the State. The presumption behind this rule is that the measure taken under the necessity doctrine is effective; it does its job of safeguarding the essential interests of the State.

Given the Committee’s identification of each of the two concerns above, it should be clear that at the very least a literal and strict application of the “only means” requirement serves no purpose. Rather in determining what may or may not be a relevant alternative that may preclude the defence a reasonable, practical and empirical approach needs to be taken.

D.           Who determines what qualifies as a “relevant alternative” and when?

The Enron Annulment Committee concludes its inquisition into the above question with: “the relevant question for the Tribunal might be whether it was reasonably open to the State, in the circumstances as they pertained at the relevant time, to form the opinion that no relevant alternative was open” (para. 372, Enron Annulment). This statement is not meant as a definitive interpretation (see para. 373, Enron Annulment). It raises two particularly interesting questions.

First is the question of when the determination of the existence of a relevant alternative is made. Quite correctly, the Committee tends to the view that this determination should be examined and evaluated at the time immediately preceding and at the same time as the measures invoking necessity were initiated by the State in question. This should enable an empirical examination of the relevant evidence that the State acted upon at the relevant time. It furthermore prevents hypothesising after the event – the value of hindsight always enables the identification of options that may not have been seen at the time the measure was taken. It is quite clearly improbable that a State is able to identify all of the options available to it in dealing with an “imminent” peril. 

Second, is the question of who makes the determination at the relevant time. At the particular relevant time, it is the State that makes any determination of whether the measure is the “only means”. However, it is clearly the tribunal which determines the legality of the State’s determination. The question of who raises two particular questions: (i) if the State makes the determination at the relevant time what steps is it required to take to make this determination; (ii) when the tribunal considers the legality of State’s determination to whom does it defer, if anyone, in determining whether there were relevant alternative measures open to the State.

Turning to (i) – if it is impossible for the State to identify all of the possible alternatives open to it, then it can only undertake reasonable measures towards identifying alternatives. The Committee suggests a move towards a loose due diligence standard in this regard: has the State reasonably determined the lack of other relevant options and in doing so has it taken reasonable steps. This approach shifts the focus towards States compiling detailed and sufficient evidence and a tribunal empirically identifying the steps a State may have taken towards identifying alternatives. The question then becomes: what steps are reasonable enough in a given situation of “grave and imminent peril”? Again one finds oneself devolving into a soft standard and one of uncertainty.

Turning to (ii) – a tribunal may take two possible approaches in determining the legality of the State’s actions and its decision-making process. It may take a reasonable and commonsensical approach to identifying alternatives, or it may seek the standard that an “appropriately qualified decision maker” (para. 372, Enron Annulment) or expert may identify possible alternative measures at the relevant time. The ICJ in the Gabčíkovo-Nagymaros Project case seems to adopt the former option, whilst the original tribunal in Enron and the Annulment Committee seem to defer to the latter option. In a situation of economic necessity, deferral to an economic expert in the identification of a relevant alternative means of dealing with the financial crises is to be expected. However, not only do experts disagree, but they espouse different options at every turn (see CMS Award). Yet, the tribunal prioritised the Claimant’s expert who identified other options available to the Argentine Republic in dealing with the financial crises. This prioritisation of one expert over another, without reason, was criticised by the Annulment Committee (para. 376, Enron Annulment). This raises questions of whether an objective standard as to identifying alternative means exists under Article 25, who should adjudge such alternatives and does a tribunal applying the standard have jurisdiction to review the decision of experts who may identify alternatives.

Based on the uncertainty of the above, I post that a tribunal should take three steps in determining the existence of relevant alternative means.

(i)           The tribunal should examine the steps taken by the State, at the relevant time, to identify possible alternatives to the means its chose;

(ii)         The tribunal should then examine the reasonableness of the State’s steps in identifying alternative means, given the particular situation (i.e. imminence and level of peril faced);

(iii)       The tribunal is then able to shift the burden of demonstrating alternatives to the opposing party; to identify any particular alternative the State should have known and considered at the relevant time.

This is a rough outline based on the holes identified by the Enron Annulment Committee. It aims at removing tribunal hypotheses concerning alternative means and roots it in empirical reality. The third step is novel. Traditionally, the burden of proof for the necessity defence and the “only means” requirement lies with the State seeking to rely on it. However, if carried to its limit, this burden may devolve into the State disproving the applicability of any number of hypothetical alternative means. The purpose of steps one and two would be to enable the State to fulfil its burden reasonably and then only seek to go beyond this once the Claimant is able to identify a specific alternative means which may have been available to the State and should have been known by the State at the relevant time. This in turn also minimises and objectifies the role of experts and resorts back to the question of the evidence the State had, or should have known about, at the relevant time.

This post cannot posit a comprehensive theory answering the problems of uncertainty in the words “only means”. It has, however briefly, examined the positives to be taken out of a damning critique of literal interpretation of necessity’s requirements.

Print Friendly
Filed under: EJIL Analysis, EJIL Reports

7 Responses

  1. Matthias

    Very interesting post, as usual! Although I agree that including alternative illegal means as bars to the necessity plea certainly reduces the scope of the exception even further, I am somewhat perplexed by the statement that it is both ‘misguided and incorrect’ for the Committee to distinguish between breaches according to their gravity. This kind of ‘least-rights-infringing-measure’ test is commonly used in assessment of the proportionality of measures adopted in the pursuit of a valid public interest in human rights jurisprudence, for instance.

    So here is my question: beyond the practical and evaluative difficulties — which certainly exist — in ranking measures in terms of their potential for infringement of investor rights, what are the reasons of principle to oppose arbitrators from distinguishing alternative measures according to ‘gravity of breach’? In other terms, would the reasoning not force arbitrators to admit the necessity plea despite the existence of a far more reasonable measure simply on the ground that it is (equally) in breach of the primary obligation?

  2. Sahib Singh Sahib Singh

    Matthias, thank you for your comment. Let me begin by saying that I am far from settled on the issue that you raise: the relevance and role the factor of “gravity of breach” should play when applying the “only means” test. However, my current stance is premised on the following two points:

    First, as a matter of law, the notion of “gravity of breach” needs to be clarified. A grave breach can be determined in two ways: (a) through the character of the obligation breached (e.g. jus cogens); (b) through the intensity of the breach. There is also a third route which the ILC has taken in Article 40(1), which is that there must be a serious breach of a peremptory norm. The history of this Article and notion of a serious breach can be traced back to Ago’s 5th Report and his Article 19 – certainly it is a far more complex matter than I can do justice to in the crude classification I have given above.

    Importantly, I do not think it is clear what is meant by the tribunal when it talks of a grave breach: is it referring to the character of the obligation breached or is it referring to the intensity of action being taken in the breach of an “ordinary” norm? I have lent towards the first of these, whereas your interpretation of the judgment seems to lend towards the second.

    Second, if we were to refer to the intensity of the breach, then my reasons of principle would be as follows. The doctrine of necessity, we can all agree, is a doctrine of exception. It is to be used as a last resort and is rarely available. But specifically it may only be used:

    (a) to safeguard and essential interest against a grave and imminent peril; and

    (b) does not seriously impair an essential interest of the State or States towards which the obligation exists, or of the international community as a whole (Article 25(1) ILC Articles).

    However, if one was to introduce the “gravity of breach” requirement into the “only means” test, then a tribunal may theoretically examine the reasonableness of the measure taken by a State relying on necessity. This possibility would then allow a tribunal to examine the effects of any illegal action on the foreign investor and balance their interests with that of the State. This result would be problematic for several reasons:

    (i) The extent to which a State invoking necessity must consider the interests of the State to which it owes the international obligation is defined by Article 25(1)(b) ILC Articles. The purpose of this provision is to ensure that the interest of the invoking State outweighs all competing interests (whether community based or individuals States to whom the obligation may be owed) (para. (17), Article 25 Commentaries). The purpose of the “gravity of breach” requirement would be to evaluate whether the intensity of the breach was reasonable in impairing the rights owed to the investor or its home State. There is certainly an overlap here. If the “gravity of breach” factor was incorporated into the “only means” test, it would blur the lines between that test under Article 25(1)(a) and the test under 25(1)(b) which defines the extent to which the interest of State to whom the breached obligation was owed.

    (ii) The blurring of the lines between the two aspects of the necessity doctrine, would in my mind introduce an element of proportionality and stringent examination of effects of a breach that is currently largely absent from the doctrine of necessity. The purpose of the doctrine is to legitimise and “legalise” those acts which would otherwise be illegal due to a grave and imminent peril which may effect an essential interest. The focus of “grave and imminent peril” and “safeguarding an essential interest” requirements are to ensure that the focus of the provision is to combat these extraordinary developments and not require a proportionate balancing with the reasonableness of a measure. Where the reasonableness of the measure notion has entered international law lies primarily in the incorporation of necessity into primary norms (such as some human rights). The standards of the circumstance precluding wrongfulness doctrine, should not be conflated with standards of necessity invoked in primary rules.

    So, the above are my arguments in law and principle. However, like I said, I remain open to persuasion to another stance, since my mind is far from settled on the matter. Do let me know your thoughts.


  3. Amaka O

    I do find it suprising – from a systemic perspective – that the tribunal had at all introduced the criterion of an alternative that would involve a less grave breach of international law. It seems tantamount to having as a defence to violating an international obligation, that there is no other – lesser – obligation one could have violated. And as for the comparison with the least-right-infringing measure test used in human rights law for assessing the proportionality of State measures taken in the public interest, I somewhat doubt that that test envisages a situation where the least infringing measure concerned is not simply one that infringes less on the Convention right in issue, but rather one that violates yet another ECHR right…

  4. Sahib Singh Sahib Singh

    Amaka, I think it is important to note two points from the Committee’s decision.

    First, the Committee simply posited another possible interpretation of “only means”. It did so with the sole purpose of demonstrating that its legal application was not straightforward: and the initial tribunal should have applied a legal test for “only means” and not simply adopted an economic experts perspective of alternative means. It did not introduce “gravity” as a definitive factor – but included as one possible interpretation out of many.

    Second, I think the least infringing measure comparison in international human right law is entirely unhelpful. My understanding of this notion is that it is generally applied where rights have to be balanced and there is the introduction of an element of proportionality, so that the measure taken is one which least infringes certain human rights. For more, see my answer above. I don’t think that such balancing requirements fit within the dynamics of the necessity doctrine; nor do I think that they should be incorporated into it.


  5. Matthias

    Thank you Sahib and AmakaO for your thought provoking responses. Please excuse my late response. Let me state from the outset that I agree that the analogy from human rights law is of limited helpfulness. As you mentioned, human rights courts use proportionality tests in the context of interpretation of restrictions authorized by primary obligations. They assess whether the pursuit of a legitimate public goal – say protecting public order – could have been achieved by the adoption of measures that would restrict the right under discussion to a lesser degree (i.e. adopting restraining orders rather than imprisoning a dangerous suspect). So it is not a matter of ‘choosing between which rights to restrict’, but rather ‘choosing the least restrictive measure that would have accomplished the legitimate public goal’. It is not an issue of ‘balancing rights’, but rather of selecting policy options that infringe less on the (single, specific) right under discussion.

    When assessing whether a standard – say of ‘fair and equitable treatment’ – is met in a given policy context, the proportionality of measures adopted will certainly influence an investment arbitrators’ decision as to the existence of a breach of that primary norm. But, as you rightly pointed out, that is an issue relating to the wrongfulness of the measures adopted, and not of whether a breach, if found, could nonetheless be excused on necessity grounds.

    My original post, however, was based on what seemed to me a rather counter-intuitive conclusion of the argument being made: that necessity, as a narrow exception precluding responsibility for wrongful conduct, should not be concerned by the relative harmfulness of two equally wrongful policy options of the State invoking necessity. Under this reading, if a State facing a serious threat to its essential interests had two equally effective, but wrongful, policy options (i.e. two options that both breached fair and equitable treatment, but to differing degrees, as expressed in damages to the interests of investors), it would be under no obligation to chose the least-damaging one. As long as it had no legal options, the State was entitled to invoke necessity no matter how little concern for investor rights it displayed in choosing between wrongful (i.e. obligation breaching) options. Note that this is different from the situation described in the decision (para. 371), where the issue is whether a State should be precluded from invoking necessity, if it had a legal alternative, though a less efficient one (i.e. a measure less likely to succeed).

    Perhaps the solution to my difficulty would be for arbitrators to decide, when determining whether there was a violation of the primary norm, that the least infringing right-restrictive measure did not constitute a breach of said obligation. This would depend, of course, on the wording of the primary norm in question, and whether exceptions and restrictions were contained therein. If this were to be the case, an issue of necessity wouldn’t arise. However, I suspect, however, that this would not persuade you, for the strategy would continue blurring the distinction between primary and secondary norms.

  6. Sahib Singh Sahib Singh

    Matthias, thanks for your comment. Sorry for the delayed reply – it has been a hectic end to the week. I think you certainly raise a valid concern. I also stand by the position I stated in my above answer (specifically the second point and sub-points (i) and (ii)). However, I would like to make three points: with the third perhaps alleviating your concerns.

    (1) At the crux of your concern I think lies a deeper rooted problem: the primary rules of the BIT operate on a different legal subject plane than the secondary norm of necessity. Let me clarify. The primary norm (e.g. expropriation) operates between the investor and State. This is irrespective of the classification we give this right: some tribunals have held that the right extends as between the home State of the investor and the host State; others have held it operates between the investor and home State. The secondary norm, that of necessity, operates as between the host State and the home State of the investor. They therefore operate on different subject planes.

    If we were to state that the invocation of necessity should be concerned with investors rights, as displayed in examining the State’s choice between wrongful options, we are incorporating the rights of a party that technically should not be considered within the legal subject parameters of the necessity doctrine.

    (2) I have no objection in principle to your posited solution. But it is my understanding that the application of the substantive rights under the BIT do not really ever consider this notion of least infringing rights measure. There is a flexible treatment as to what some of the substantive rights may pertain to (e.g. FET), but not a particular consideration of whether the State could have taken another measure. Of course, investor parties posit alternative strategies as a matter of legal argument, but these have never been incorporated into the primary norms under the BIT.

    (3) I think the best, and doctrinally safest, answer lies in Article 27(b) of the ILC Articles. Ensuring that circumstances precluding wrongfulness are without prejudice as to compensation, may allow the incorporation on the “less grave” wrongful alternative being incorporated into this question of compensation. If the tribunal is provided with a less grave wrongful alternative that was viable and should have been considered by the state at the time, and this alternative would have had a less adverse monetary effect on the investor, then there should be nothing in Article 27(b) that should prevent this being considered for the question of “material loss”.

    So overall – I do sympathise with your particular concern but I think there are three good principle reasons of not involving this concern within the necessity doctrine (2 from my initial answer and point (1) from this). Nevertheless, there may be legitimate “legal space” to take into account your principled concern in Article 27(b).

  7. Matthias

    Thank you, Sahib, for taking the time to further respond to my comments. I see your position much more clearly now and I consider your arguments very compelling. On point (1), I understand that the plea of necessity operates between states, rather than between states and investors. However, as the Committee noted, the Tribunal seemed to imply that a threat to the essential interests of investors could exclude the necessity plea (Award, para. 310; Annulment Decision, para. 383). This part of the award was annulled on the ground that the the tribunal’s reasoning was unclear, but of course the Committee could not, and did not decide the point of substance, i.e. whether the essential interests of investors may preclude the appeal to necessity. I find (3) to be a very elegant solution to the problem I was concerned with: it seems to me that the deliberate choice by the host State of a more damaging policy option might be seen as a specific ground, under Article 27(b), to request compensation for losses, even if the necessity plea was successful.