Introduction
The globalised economy is witnessing a rapid increase in cross-border investments, which offer high rewards but also entail high risks. The Bilateral Investment Treaties (“BITs”) are instruments designed to mitigate these risks, lubricate transactions, and provide safeguards to the investors. While signing the BITs, the states commit to investor protection and provide legal rights to investors for securing their investments. However, a crucial, albeit challenging, task for the state is to balance these rights given to the investors with the state’s right to regulate its country.
Under the current regime of international investment law, wide range of protections are made available to the investors, which are balanced by exception clauses that allow the states to exercise their sovereign powers to regulate their territory. One such exception is the Essential Security Interest (“ESI”) clause, which allows the states to take regulatory measures that are “necessary” to protect “essential security interests” and avoid liability under the BIT.
A persistent issue, however, lies in interpreting the term “necessary” within ESI clauses. The interpretation of the ESI clauses contained in BITs has been inconsistent by different arbitral tribunals over the years. The arbitral tribunals have in some cases, drawn from customary international law’s (“CIL”) defence of “necessity” as codified in Article 25 of the International Law Commission’s (ILC) Articles on State Responsibility for interpreting the term “necessary”, however in other cases, it has independently interpreted the provisions of the treaty without relying on the CIL standard.
The defence of necessity as given in the Article 25 of the ILC articles has four elements namely (i) there must exist a grave and imminent peril; (ii) it must be for the essential interest of the state; (iii) the measures adopted by state must be the ‘only way’ to remedy the situation; and (iv) the state must not have contributed towards that situation[i].
Under the CIL, the defence of necessity only applies if a state satisfies these strict and narrow conditions. Out of these conditions, the third and fourth conditions, i.e. the concerned measure must be the only way of rectifying the situation, and the state must not have contributed towards the crisis becomes the most challenging ones to satisfy in most commercial situations and often render the defence of ESI clauses meaningless, as seen in the Argentina cases discussed below.
Against this backdrop, this piece seeks to highlight the evolving and inconsistent interpretations of the ESI clause in investment arbitration. It further argues for the adoption of a more structured approach for the interpretation of the term “necessary”, drawing from CIL but in a more flexible and fact-sensitive manner that respects both the rights of investors and the regulatory autonomy of host states.
The Evolving interpretation of ESI Clauses In BITs
It is useful to begin by examining the interpretations of ESI clauses over the years as given by various arbitral tribunals. For the sake of brevity and ease of understanding, the authors have classified the awards passed by tribunals in three broad approaches followed by the tribunals.
The first approach includes the cases of CMS v. The Republic of Argentina[ii] (“CMS”) Enron v. The Republic of Argentina[iii] (“Enron”) and Sempra v. The Republic of Argentina[iv] (“Sempra”). This phase corresponds to a very strict interpretation of the ESI Clauses and complete reliance on the CIL standard for interpreting the term “necessary”.
The second approach includes the decisions passed in the cases of LG&E v. The Republic of Argentina[v] (“LG&E”) and Continental Casualty v. The Republic of Argentina[vi] (“Continental Casualty”). This is a functional approach for the interpretation of the ESI Clauses and attempts to balance the state’s right to regulate with the investor’s right to secure its investment. This approach also includes the decisions passed by the Annulment Committee in the CMS[vii], Enron[viii] and Sempra[ix] cases.
The third approach relates to decisions passed post 2015 and includes the landmark decisions of CC/Devas v. Union of India[x] (“Devas”) and Deutsche Telecom v. Union of India[xi] (“Deutsche Telecom”), here the tribunals have resorted to a mixed interpretation swinging between a strict and diluted interpretation of the ESI clauses; however, the reasoning followed by the court in reaching these decisions has been different from the one adopted in earlier cases.
Approach -I: Strict Interpretation
The cases in the first and second approach, albeit inconsistent in interpreting the ESI Clauses, arise from near-identical facts and the common Article XI of the USA-Argentina BIT. These cases arise out of the regulatory measures, such as price control measures, currency devaluation and pesification of contracts and restriction on transfer of capital assets, undertaken in response to the severe economic crisis that hit Argentina in the 1990s.
The arbitral tribunal in all these cases was faced with a complex question of whether measures taken to protect the economy from an economic crisis could be classified as measures “necessary” for the protection of the “essential security interest”. The tribunals, despite having near-identical facts and questions of law, reached differing conclusions .
The CMS, Enron, and Sempra tribunals correlated the ESI clauses to the defence of necessity under the CIL, albeit with differing justifications. While CMS did not explain its reasoning, for corelating the two, the tribunal in Enron and Sempra explicitly stated that reliance on CIL was necessary because the BIT lacked a clear definition of the term “essential security interest”.[xii] Once they equated ESI with necessity under CIL, these tribunals evaluated Argentina’s claim based on the four parameters explained above.
The CMS tribunal held that Argentina’s economic crisis could justify invoking the ESI clause only if it led to total economic and social collapse.[xiii] The Enron[xiv] and Sempra[xv] tribunals imposed an even stricter threshold, ruling that Argentina’s crisis had to threaten the State’s very existence and independence to qualify as an ESI.
Hence, all three tribunals concluded that the crisis was not severe enough to compromise Argentina’s existence and independence, thereby not justifying the invocation of the ESI defence. In effect, these decisions reduced the ESI clause contained in the BITs to paper tigers, with no tangible effect on the rights of the parties.
Approach – II: Functional Interpretation
The LG&E and Continental Casualty tribunals drew a distinction between the CIL defence of “necessity” and ESI Clauses. It swayed away from the strict application of the CIL standard as contained in Article 25.
The Continental Casualty Tribunal distinguished between the ESI Clauses and the defence of necessity and noted that the necessity defence under CIL only precludes the wrongfulness of an act, while an ESI clause serves as a safeguard that places certain state actions outside the scope of treaty obligations altogether.[xvi] Unlike the strict standard applied by earlier tribunals, which needed total economic collapse or a threat to a State’s very existence, the LG&E and Continental tribunals held that a threat of collapse or a situation where a State’s economic foundations are under siege was sufficient to invoke an ESI clause. Hence, it allowed Argentina to claim the defence of necessity for the measures adopted by it to overcome the economic crisis.
The tribunal in LG&E and Continental Casualty believed that Argentina’s economic turmoil was so extreme that ordinary measures could not suffice, and it needed immediate and decisive action. The LG&E tribunal further emphasised that not recognising such a crisis as an essential security interest would underestimate the devastating impact of economic turmoil on citizens and governance.[xvii]
Another vital development of this phase relates to the decisions of the annulment committee in the three cases of CMS, Enron and Sempra. The annulment committee disagreed with the tribunals’ approach of equating Article XI with the CIL. It is particularly the criticism of CMS Annulment Committee that was remarkable and is relevant for our discussion here. In this case, the annulment committee, citing error in the approach of the tribunal, laid a two-step sequence of reasoning: first, the tribunal should determine whether state’s actions constituted a breach of the BIT or whether the treaty’s emergency clause excluded such a breach and second if there was a breach of terms of BIT, it should determine if the liability could be excused under the CIL.[xviii] Applying this test, it concluded that since the treaty itself had the ESI clause, it prevented the actions of the state from being classified as wrongful under the BIT in the first place. Therefore, the question of it being excused under the defence of CIL does not arise in the first place.
In 2008, the Continental Casualty Tribunal re-imagined this debate by borrowing from WTO jurisprudence and introducing a more nuanced two-part test to apply the ESI clauses: (1) the measure must “materially contribute” to protecting the State’s essential security interests, and (2) no reasonable, less treaty-infringing alternatives should be available to achieve the same objective.[xix] This approach allowed the reasonable measures taken by the state against the economic crisis to be classified as measures protected under the ESI Clauses. Hence, this approach gave a wider interpretation to the ESI clauses. This can be classified as a functional approach to interpretation that ensures the clauses are still effective and serve the necessary purpose.
Approach –III: A Mixed Bag
This approach consists of two landmark arbitral decisions one, in the case of CC/Devas v. Republic of India and second in the case of Deutsche Telekom v. Republic of India that arose from near-identical factual circumstances yet culminated in divergent outcomes. Both cases stemmed from the annulment of a satellite transponder agreement between Devas Multimedia Private Limited (“Devas”) and Antrix Corporation, a state-owned Indian enterprise. Despite the shared factual matrix and overlapping legal issues, the tribunals interpreted the Essential Security Interest (ESI) clauses differently, highlighting the inconsistency and lack of uniformity in arbitral approaches to the application of ESI provisions in investment treaties.
In the case of CC/Devas v. The Republic of India, tribunal interpreted the ESI clause in the India-Mauritius BIT by emphasizing the specific language of the treaty, which required measures to be “directed to the protection of its essential security interests” rather than “necessary for the protection of its essential security interests”, as seen in other cases discussed above. Interpreting this, the tribunal, relying on the study of UNCTAD, concluded that this distinction meant that the state was only required to demonstrate the existence of an “essential” security interest, not merely a “security interest,” and did not need to establish “necessity” for its actions since the treaty did not impose such a requirement. However, the tribunal imposed a high standard for interpreting the term “essential” by relying on its dictionary definition, which described it as “absolutely necessary; indispensable or unavoidable.”[xx] Therefore, even if it attempted to distinguish between the treaty defence and CIL, it ended up adopting a fairly strict approach rendering the ESI clauses to be meaningless.
On the contrary, the tribunal, in the case of Deutsche Telecom v. The Republic of India, while interpreting the ESI Clause in the India-Germany BIT, adopted a lower standard for the existence of the ESI. It agreed that the ESI Clause must be distinguished from the CIL defence of state of necessity, which is codified in Article 25 of the ILC Articles.[xxi] The tribunal clarified that the ESI clause requires measures to be “necessary for the protection” of essential security interests, rather than merely “related to” them. Therefore, the tribunal laid a two-fold test to interpret the term “necessary” different from the CIL standard: first, the measure must be principally targeted to protect the essential security interests at stake and second, it must be objectively required for the same taking into account whether the state had reasonable alternatives, less in conflict or more compliant with its international obligations.[xxii]
The Deutsche tribunal has departed from Continental Casualty’s test of material contribution and the least restrictive measure to examine whether the measure was principally targeted at the ESI. The refinement of the approach adopted in the case of Continental Casualty, as proposed by the tribunal in the case of Deutsche Telecom, is a welcome step[xxiii]. However, the inconsistency in interpretation of the ESI clauses continues, as this approach has not been uniformly accepted by the tribunals.
Based on the foregoing discussion, four main tests have emerged for interpreting ESI clauses: (i) the strict necessity test aligned with customary international law; (ii) the Continental Casualty test, which adopts a broader, more flexible standard; (iii) the CMS Annulment Committee’s approach, treating ESI clauses as distinct treaty-based exceptions; and (iv) the most recent test as laid in Deutsche Telecom case. However, none of these approaches has gained uniform acceptance among investment tribunals. This lack of consistency has led to significant ambiguity and legal uncertainty regarding the interpretation and application of ESI clauses.
Treaty-Based Exceptions vs. Customary Defences: Distinctions
We now move to understand the distinctions between the treaty-based exceptions like ESI clauses and defences available under customary law, like the defence of necessity contained in Article 25 of the ILC Articles. While the ESI clauses and the CIL defence of necessity share common features in that both provide the states with a justification to evade liability for taking measures that would otherwise violate treaty or international law obligations, their legal implications and requirements differ significantly.
ESI clauses are a version of Non-Precluded Measures (“NPM”) clauses which differ fundamentally from customary defences. NPM clauses exclude specific state actions from treaty protections, or in other words, it does not preclude (thereby allows) the states to take certain measures which would have otherwise lead to the violation of the treaty provisions[xxiv]. These clauses function as an ex-ante mechanism to prevent the measures from being classified as a wrongful act. In contrast, the customary defences function as ex-post justifications that may absolve a state from liability after the breach has occurred.[xxv] The necessity defence under Article 25 of ILC Articles is a CIL doctrine that applies when a state argues that breaching an international obligation is the only way to safeguard an essential interest against a grave and imminent peril.
Unlike ESI clauses, the necessity defence does not negate the breach but merely precludes the state’s responsibility for it. Moreover, under Article 27 of ARSIWA, a state that successfully invokes necessity may still be required to compensate the affected party. However, the ESI clauses prevents the measures to be classified as wrongful in the first place, thereby negating the need for the payment of any compensation.
The critical distinction between the two is that an ESI clause operates within the treaty framework and prevents a breach from occurring, whereas the necessity defence operates under general international law and merely excuses a breach while maintaining the state’s obligations. Despite this clear distinction, investment tribunals have not always treated them separately, leading to confusion in arbitral jurisprudence[xxvi].
Such NPM clauses are primary treaty rules (lex specialis), forming part of the treaty framework negotiated between states, while customary defences are secondary legal rules (lex generalis) that arise from state practice and opinio juris[xxvii].
In terms of scope, customary defences apply universally as part of international law, whereas NPM clauses are treaty-specific, varying based on the precise language negotiated by states. Importantly, equating NPM clauses with the customary necessity defence would violate the principle of effectiveness in treaty interpretation (ut res magis valeat quam pereat), as seen in WTO and ICJ jurisprudence, because customary necessity would apply regardless of whether an NPM clause exists[xxviii].
The ICJ’s Nicaragua v. USA case confirms that treaty law and CIL remain distinct, even when they appear identical in content, meaning NPM clauses do not automatically replace customary necessity unless explicitly stated or where direct conflict renders coexistence impossible. Consequently, investment treaties containing NPM clauses establish a separate, treaty-based mechanism for allocating risks between states and investors in extraordinary circumstances, functioning independently but concurrently with CIL defences[xxix]. These distinctions underscore the need for arbitral tribunals to distinguish between the treaty exceptions and defences available in the CIL and do not equate the defence of defence of necessity with the ESI clauses which shall render the latter to be non-functional and inconsequential.
Conclusion
The foregoing discussion highlights that the absence of ‘jurisprudence constante’ has rendered the ESI defence non-functional. The variations from the cases of CMS, Enron and Sampras to the recent case of Deutsche Telekom show that there is an urgent need to bring uniformity in the way the tribunals interpret the ESI Clause. The tribunals not only have ‘a duty to adopt solutions established in a consistent line of cases’, but also a ‘duty to seek to contribute to the harmonious development of investment law.’[xxx]
The necessity has failed as a legal defence, due to huge uncertainty surrounding its application in practice; therefore, the states have begun incorporating internal exceptions in investment treaties in various forms, which offer clearer, predefined regulatory space. States have been adopting interpretative shortcuts like importing self-judging clauses in the language of their BITs rather than doing justice to the intent of the parties or the rules of the Vienna Convention. This is evident from the fact that there has been a 30% increase in public health exception clauses in BITs since 2016.[xxxi] This calls for a softer and uniform application of the ESI clause by the investment tribunals, leading to more certainty and increasing the state’s confidence. This should be done by softening the standard in Article 25 of the Article on State Responsibility to make it more workable for the investment arbitration system. Despite inherent differences between the ESI Clauses and the defence of necessity contained in Article 25 of the ILC Articles, it is essential to recognize that we cannot entirely dismiss the relevance of CIL in interpreting the term “necessary”. Given the absence of any universally accepted standard for this interpretation, CIL is a vital reference point rooted in fundamental principles of international law. However, to make the defence of necessity functional for states, there is a pressing need to dilute the stringent standards traditionally associated with Article 25 while interpreting the ESI Clauses. Recent tribunal decisions, such as those in the Deutsche Telecom and Continental Casualty Case, are a welcome step that reflects an attempt to dilute these standards and allows for a more practical interpretation of the ESI clauses. However, while pursuing this flexibility, it is crucial to keep a uniform approach to interpretation. This balance will ensure an effective balance is maintained between investor protection under investment treaties and the state’s sovereign right to regulate.
[i] ILC, ‘Draft Articles on the Responsibility of States for Internationally Wrongful Acts, 2001 with Commentaries’ in Report of the International Law Commission on the work of its fifty-third session (23 April – 1 June and 2 July – 10 August 2001) 59 UN Doc. A/56/10 (2001) p. 139.
[ii] CMS Gas Transmission Co. v Argentine Republic, ICSID Case No. ARB/01/8, Award (12 May 2005) (CMS).
[iii] Enron Corporation Ponderosa Assets, L.P. v. Argentine Republic, ICSID Case No. ARB/01/3, Award (22 May 2007) (Enron).
[iv] Sempra Energy International v. Argentina, ICSID Case No. ARB/02/16, Award (28 September 2007) (Sempra).
[v] LG&E Energy Corporation and Others v Argentine Republic, ICSID Case No. ARB/04/4, Award (25 July 2007) (LG&E).
[vi] Continental Casualty Co. v Argentine Republic, ICSID Case No. ARB/03/9, Award, (5 September 2008) (Continental Casualty).
[vii] CMS Gas Transmission Company v Argentine Republic, ICSID Case No. ARB/01/8, Decision of the Ad Hoc Committee on the Application for Annulment of the Argentine Republic (25 September 2007) (CMS Annumlment Committee).
[viii] Enron Creditors Recovery Corp Ponderosa Assets, L.P. v Argentine Republic, ICSID Case No. ARB/01/3, Decision on the Application for Annulment of the Argentine Republic (30 July 2010) (Enron Annulment Committee).
[ix] Sempra Energy International v Argentina ICSID Case No. ARB/02/16, Decision on the Argentine Republic’s Application for Annulment of the Award (29 June 2010) (Sempra Annulment Committee).
[x] CC/Devas (Mauritius) Ltd., Devas Employees Mauritius Private Limited and Telecom Devas Mauritius Limited v Republic of India, PCA Case No. 2013-09, Award (25 July 2016) (CC/Devas).
[xi] Deutsche Telekom AG v Republic of India, PCA Case No. 2014-10, Award (13 December 2017) (Deutsche Telekom).
[xii] Enron, supra note 4, para 333; Sempra, supra note 5, para 376.
[xiii] CMS, supra note 3, para 355.
[xiv] Enron, supra note 4, para 306 and 307.
[xv] Sempra, supra note 5, 348 and 349.
[xvi] Continental Casualty, supra note 7, 162-167.
[xvii] August Reinisch, ‘Necessity in Investment Arbitration’ (2010) 41 Netherlands Yearbook of International Law p. 140 (August Reinisch).
[xviii] CMS Annulment Committee, supra note 8, para 134.
[xix] Continental Casualty, supra note 7, para 192-195.
[xx] CC/Devas, supra note 11, para 243.
[xxi] Deutsche Telecom, supra note 12, para 227.
[xxii] Deutsche Telecom, supra note 12, para 239.
[xxiii] Ridhi Kabra, ‘Return of the Inconsistent Application of the ‘Essential Security Interest’ Clause in Investment Treaty Arbitration: CC/Devas v India and Deutsche Telekom v India’, in Meg Kinnear and Campbell McLachlan (eds), ICSID Review – Foreign Investment Law Journal, 34(3), pp. 723 – 753.
[xxiv] Sanjay, Sujaya., Essential Security Interests in Investment Arbitration: Should ESI clauses in BITs be interpreted as per customary international law?, (2020) p. 17.
[xxv] August Reinisch, supra note 18, p. 137, 149.
[xxvi] Giovanni Zarra, ‘Orderliness and Coherence in International Investment Law and Arbitration: An Analysis Through the Lens of State of Necessity’, in Maxi Scherer (ed), Journal of International Arbitration, 2017, 34(4), pp. 653 – 678
[xxvii] William W. Burke-White and Andreas von Staden, ‘Investment Protection in Extraordinary Times: The Interpretation and Application of Non-Precluded Measures provisions in Bilateral Investment Treaties’ (2008) 48 Virginia Journal of International Law, p. 322 (William Burke).
[xxviii] William Burke, supra note 28, p. 323
[xxix] Military and Paramilitary Activities in and against Nicaragua (Nicaragua v United States) (Merits) [1986] ICJ Rep 14, paras 115-116.
[xxx] Saipem S.P.A. v. The People’s Republic of Bangladesh, ICSID Case No. ARB/05/7, Award (30 June 2009) para 90.
[xxxi] Federica Paddeu and Michael Waibel, ‘Necessity 20 Years On: The Limits of Article 25’, in Meg Kinnear and Campbell McLachlan (eds), ICSID Review – Foreign Investment Law Journal, 2022, 37(1/2), pp. 160 – 191.
Author(s)
Khyati Maurya
Student at GNLU, Gandhinagar
Saransh Sood
Student at GNLU, Gandhinagar